Wills, Wouter PJ - Efficiency and Justice in European Antitrust Enforcement PDF

JOBNAME: Wouter Wils PAGE: 1 SESS: 2 OUTPUT: Mon Jan 21 15:06:07 2008
EFFICIENCY AND JUSTICE IN EUROPEAN

ANTITRUST ENFORCEMENT
In the last few years, the public enforcement of Articles 81 and 82 EC has been
thoroughly transformed: the competition authorities of the EU Member States
have become active enforcers within the European Competition Network, the
European Commission has imposed more and higher fines than ever before,
leniency has become a major instrument of cartel detection, and some Member
States have introduced criminal penalties. The overall trend towards more and
stronger enforcement of Articles 81 and 82 EC has also rekindled discussion on
the old question of how to strike the right balance between efficient enforcement
and adequate protection of the rights of the defence. This book brings together
six essays which analyse from both a legal and an economic perspective the
powers of investigation of the European Commission and the competition
authorities of the Member States, and the corresponding procedural rights and
guarantees, the use of settlements, the theory and practice of fines and of
leniency, and the criminalisation of European antitrust enforcement.
Columns Design Ltd / Job: Wils_Soft-proofing / Division: prelims /Pg. Position: 1 / Date: 6/12
JOBNAME: Wouter Wils PAGE: 3 SESS: 6 OUTPUT: Tue Jan 22 14:26:46 2008
Efficiency and Justice in

European Antitrust
Enforcement
WOUTER PJ WILS
Columns Design Ltd / Job: Wils_Antitrust / Division: prelims /Pg. Position: 1 / Date: 22/1
Published in North America (US and Canada) by

Hart Publishing
c/o International Specialized Book Services
920 NE 58th Avenue, Suite 300
Portland, OR 97213–3786
USA
Tel: +1 503 287 3093 or toll-free: (1) 800 944 6190
Fax: +1 503 280 8832
E-mail: orders@isbs.com
Website: www.isbs.com
© Wouter PJ Wils 2008
Wouter Wils has asserted his right under the Copyright, Designs and Patents Act 1988, to be
identified as the author of this work.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, without the prior permission of Hart Publishing, or as
expressly permitted by law or under the terms agreed with the appropriate reprographic rights
organisation. Enquiries concerning reproduction which may not be covered by the above should be
addressed to Hart Publishing at the address below.
Hart Publishing, 16C Worcester Place, OX1 2JW

Telephone: +44 (0)1865 517530 Fax: +44 (0)1865 510710
E-mail: mail@hartpub.co.uk
Website: http://www.hartpub.co.uk
British Library Cataloguing in Publication Data

Data Available
ISBN: 978–1-84113–017–0
Typeset by Columns Design, Reading

Printed and bound in Great Britain by
TJ International Ltd, Padstow, Cornwall
Introduction and
Acknowledgements
In the last few years, the public enforcement of Articles 81 and 82 EC has been
thoroughly transformed. As a result of the decentralisation brought about by
Regulation 1/2003 and of the enlargement of the EU from 15 to 27 Member
States, the competition authorities of the Member States now play an important
role in the enforcement of Articles 81 and 82 EC, alongside the European
Commission in the European Competition Network (ECN).1 Meanwhile the
European Commission has to a significant extent focused its resources on the
fight against cartels, adopting recently between five and 10 major cartel decisions
per year, with total amounts of fines running into billions of euros.2 One of the
main factors contributing to this success has been the progressive refinement of
the Commission’s leniency policy. The question remains, however, whether
cartels can be effectively deterred with only financial penalties on companies, or
whether criminal penalties should be added, as has recently been done in some
Member States. The overall trend towards more and stronger enforcement of
Articles 81 and 82 EC has also rekindled discussion on the old question of how to
strike the right balance between efficient enforcement and adequate protection of
the rights of the defence.3
This book brings together six essays which analyse many of the main legal,
economic and policy issues arising from the recent growth and transformation of
the public enforcement of Articles 81 and 82 EC, and from the continuing need
to reconcile efficient enforcement with adequate protection of the rights of the
defence. The essays are all based on papers which have been published previously,
but have been updated to varying degrees so as to include recent developments as
well as the evolution of my thinking. Tables have been added to the preliminary
pages and a subject index at the end of the book.
Chapter 1 deals with the powers of investigation of the European Commission
and of the competition authorities of the Member States for the enforcement of
Articles 81 and 82 EC, and with the procedural rights and guarantees that
circumscribe or limit these powers. It focuses in particular on the question of
which law governs these matters (EC or EU law, national law, and the European
Convention on Human Rights), and by whom or how the content of this law is
1
See the enforcement statistics at http://ec.europa.eu/comm/competition/ecn/statistics.html.
2
See the statistics at http://ec.europa.eu/comm/competition/cartels/statistics/statistics.pdf.
3
See, inter alia, the Opinions of AG Geelhoed of 19 Jan 2006 and 12 Sept 2006, and the
subsequent Judgments of the ECJ of 29 June 2006 and 25 Jan 2007 in Case C–301/04 P Commission v
SGL Carbon [2006] ECR I–5915 and in Case C–411/04 P Salzgitter Mannesmann v Commission, not
yet reported.
vi INTRODUCTION AND ACKNOWLEDGEMENTS
determined (by European and national legislation, and case law of the European
Court of Justice and Court of First Instance, national courts, and the European
Court of Human Rights). This chapter is based on a paper presented at the First
Lisbon Conference on Competition Law and Economics (Lisbon, 3–4 November
2005), organised by the Portuguese Competition Authority, and published at
(2006) 29 World Competition 3 and in AM Mateus and T Moreira (eds),
Competition Law and Economics—Advances in Competition Policy and Antitrust
Enforcement (Alphen aan den Rijn, Kluwer Law International, 2007).
Chapter 2 focuses on one of the novelties under Regulation 1/2003, namely its
Article 9, which provides for formal settlement of investigations by the European
Commission into suspected infringements of Article 81 or 82 EC. The analysis
covers the origin and optimal use of this provision, the procedure for the
adoption of commitment decisions, the content and effect of such decisions, and
the scope for judicial review. This chapter is based on a paper prepared for a
public lecture at the Centre of European Law, King’s College London on 24 May
2006, and published at (2006) 29 World Competition 345.
Chapter 3 discusses, in general and from a legal and economic perspective, the
use of fines imposed on companies or other corporate entities to enforce
antitrust or competition law prohibitions such as Articles 81 and 82 EC or
sections 1 and 2 of the Sherman Act. The chapter addresses more specifically the
questions in what ways these fines contribute to competition law enforcement, on
the basis of which factors the amounts of antitrust fines should be fixed in theory,
and whether it is feasible in practice to calculate or measure such optimal fines.
An earlier version of this chapter, based on lectures at the Finnish Competition
Authority on 7 March 2003, at the University of Bologna on 28 November 2003
and at King’s College London on 21 October 2004, was published at (2006) 29
World Competition 183.
Chapter 4 focuses on the European Commission’s 2006 Guidelines on anti-
trust fines. It discusses what the purpose is of guidelines, and how foreseeable the
amount of fines should be, and analyses the method set out in the 2006
Guidelines in the light of the Commission’s past practice, the case law of the
Community Courts and the theory on optimal fines. This chapter is based on a
paper prepared for a public lecture at the Centre of European Law, King’s College
London on 15 February 2007, and published at (2007) 30 World Competition 197.
Chapter 5 discusses the theory and practice of leniency in antitrust enforce-
ment, ie the granting of immunity from penalties or the reduction of penalties
for antitrust violations in exchange for cooperation with the antitrust enforce-
ment authorities. After a description of the practice of leniency in the US and in
Europe, and of its history, the chapter analyses the positive effects and the
possible negative effects of leniency on optimal antitrust enforcement, and the
extent to which these effects can be measured. Objections of principle and
institutional problems that may constitute obstacles to the introduction of
leniency policies are discussed, as well as some further issues, namely the impact
on the effectiveness of leniency of criminal penalties on individuals, of follow-on
INTRODUCTION AND ACKNOWLEDGEMENTS vii
private damages actions, and of penalties in other jurisdictions, ‘Amnesty Plus’,

and positive financial rewards or bounties. This chapter is based on a paper
presented at the 25th Conference on New Political Economy (Saarbrücken, 12–14
October 2006), and published at (2007) 30 World Competition 64, and in M
Albert, D Schmidtchen and S Voigt (eds), Frontiers of EC Antitrust Enforcement:
The More Economic Approach, Conferences on New Political Economy 25 (Tübin-
gen, Mohr Siebeck 2007).
Finally, Chapter 6 addresses five questions concerning the criminalisation of
European antitrust enforcement: First, what do we mean by ‘criminalisation’, or
‘criminal’ enforcement (as opposed to public enforcement of a ‘civil’ or ‘admin-
istrative’ nature)? Secondly, is there a tendency in the EU Member States to
criminalise antitrust enforcement (in comparison with US antitrust enforcement
and with antitrust enforcement at the level of the EU institutions)? Thirdly, is
criminal antitrust enforcement, more specifically imprisonment, desirable (in
general, irrespective of whether it takes place at the level of the Member States or
of the EU institutions, or whether it is harmonised at EU level)? Fourthly, is it
problematic that antitrust enforcement is criminalised at the level of individual
EU Member States without parallel criminalisation at the level of the EU
institutions or without EU harmonisation? Fifthly, would it be legally possible to
criminalise antitrust enforcement at the level of the EU institutions, or to have
EU harmonisation of criminal antitrust enforcement in the Member States? This
chapter is based on papers presented at the Amsterdam Center of Law and
Economics (ACLE) Conference Remedies and Sanctions in Competition Policy
(Amsterdam, 17–18 February 2005), at the 11th European University Institute
(EUI) Competition Law and Policy Workshop Enforcement of Prohibition of
Cartels (Florence, 2–3 June 2006) and at the Forschungsinstitut für Wirtschafts-
verfassung und Wettbewerb (FIW) XXXV Brüsseler Informationstagung (Brus-
sels, 21–22 September 2006), and published at (2005) 28 World Competition 117,
at [2005]Revue Lamy de la Concurrence (August/October 2005) 139 (with French
translation), in KJ Cseres, MP Schinkel and FOW Vogelaar (eds), Criminalization
of Competition Law Enforcement: Economic and Legal Implications for the EU
Member States (Cheltenham, Edward Elgar, 2006) and in C-D Ehlermann and I
Atanasiu (eds), European Competition Law Annual 2006: Enforcement of Prohibi-
tion of Cartels (Oxford, Hart Publishing, 2007).
I am grateful to Margaret Bloom, Giuseppe Dari Mattiacci, Claus-Dieter
Ehlermann, Ferdinand Hermanns, Kirsi Leivo, Pietro Manzini, Abel Mateus,
Teresa Moreira, Maarten Pieter Schinkel, Dieter Schmidtchen, Floris Vogelaar and
Richard Whish for having invited me to the conferences and lectures which
prompted me to write the earlier versions of the essays brought together in this
book, to Filippo Amato, François Arbault, David Bailey, Margaret Bloom, André
Bouquet, Hubert de Broca, Fernando Castillo de la Torre, Fanis Christoforou, Leo
Flynn, Céline Gauer, Damien Geradin, Eric Gippini Fournier, Joe Harrington,
Erling Hjelmeng, Nicholas Khan, Bill Kovacic, Clemens Ladenburger, Xavier
Lewis, Kirti Mehta, Walter Mölls, Oke Odudu, Peter Oliver, Luc Peeperkorn,
viii INTRODUCTION AND ACKNOWLEDGEMENTS
Michele Polo, Andreas Reindl, Peter Roth, Anthony Whelan, Richard Whish and
Geert Wils for their stimulating comments on earlier versions of these essays, and
to Jose Rivas and Kluwer Law International, editor and publisher of World
Competition, as well as to all the editors and publishers of the journals and books
in which earlier versions of the essays brought together in this book were
published, for having allowed me to use (the updated versions of) these essays for
this book. All views expressed in this book are strictly personal, and should not be
construed as reflecting the opinion of the European Commission, its Legal
Service, or any of the persons mentioned above. Comments are always welcome
at Wouter.Wils@ec.europa.eu.
Wouter PJ Wils
Brussels, 30 April 2007
Contents
Introduction and Acknowledgements v

Table of Cases xv
Table of Legislation xxiii
1 Powers of Investigation and Procedural Rights and Guarantees: The

Interplay between European and National Legislation and Case Law 1
1.1 Powers of Investigation 1
1.1.1 Different Ways of Gathering Intelligence and Evidence 1
1.1.2 Cases Dealt with by the European Commission 3
1.1.2.1 Information-gathering by the Commission Itself 3
1.1.2.2 Assistance by the (Competition) Authorities of the
Member States 6
1.1.3 Cases Dealt with by a Competition Authority of a Member
State 9
1.1.3.1 Information-gathering by the National Competition
Authority Itself 9
1.1.3.2 Assistance by the Competition Authorities of Other
Member States 11
1.2 Procedural Rights and Guarantees 12
1.2.1 Procedural Rights and Guarantees Circumscribing or
Limiting Powers of Investigation 12
1.2.2 Information Collected and Used by the Commission 14
1.2.3 Information Collected and Used by the Same National
Competition Authority 19
1.2.4 Information Exchanged through the European Competition
Network 20
2 Settlements of Antitrust Investigations: Commitment Decisions

under Article 9 of Regulation 1/2003 25
2.1 Text, Origin and Optimal Use 25
2.1.1 Text 25
2.1.2 Origin 26
2.1.3 Optimal Use 29
2.2 Procedure 33
2.2.1 Initiation of Proceedings and Preliminary Assessment 33
2.2.2 Access to the File 34
x CONTENTS
2.2.3 Consultation, Adoption and Publication 36

2.3 Content and Effects 37
2.3.1 Type of Commitments 37
2.3.2 ‘No longer grounds for action by the Commission without
concluding whether or not there has been or still is an
infringement’ 37
2.3.3 ‘Binding on the undertakings’: Enforcement of
Commitment Decisions 39
2.3.4 Further Proceedings in National Courts or by National
Competition Authorities Concerning Past Behaviour 41
2.3.5 Further Proceedings in National Courts or by National
Competition Authorities Concerning Future Conduct 42
2.4 Judicial Review 44
2.5 Other Settlements 45
2.5.1 Informal Settlements 45
2.5.2 Commitment Decisions by National Competition
Authorities 45
2.5.3 Settlements Including the Payment of a Fine 46
3 Optimal Antitrust Fines: Theory and Practice 49

3.1 The Role of Fines in Antitrust Enforcement 50
3.1.1 The Three Tasks of Competition Law Enforcement 50
3.1.1.1 Clarifying the Content of the Prohibitions 50
3.1.1.2 Preventing Violations 51
3.1.1.2.1 Reducing the Opportunities to Commit
Violations 51
3.1.1.2.2 Reducing Business People’s Willingness to
Commit Violations 52
3.1.1.2.3 Altering the Balance of Expected Benefits
and Costs of Violations 52
3.1.1.3 Dealing with the Consequences of Violations 53
3.1.2 The Role of Fines 55
3.2 Optimal Fines for Single Offenders 56
3.2.1 The Basic Logic of Deterrence 56
3.2.1.1 The Deterrence Approach as Opposed to the
Internalisation Approach 56
3.2.1.2 Subjective Estimates of the Gain, the Probability of
Detection and Punishment and the Amount of the Fine 59
3.2.1.2.1 Availability Bias 59
3.2.1.2.2 Overconfidence Bias 60
CONTENTS xi
3.2.2 The Trade-off between the Level of Fines and the

Probability of Detection and Punishment 61
3.2.2.1 The Standard Economic Argument and its Limits 61
3.2.2.2 The Limits to High Fines 62
3.2.2.2.1 Inability to Pay 62
3.2.2.2.2 The Social and Economic Costs of High
Fines 63
3.2.2.2.3 Proportional Justice 64
3.2.3 Rewarding Cooperation and Efforts at Compliance 65
3.2.3.1 Cooperation with the Competition Authority’s
Investigation 66
3.2.3.2 Compliance Programmes 67
3.3 Optimal Fines for Collective Violations 68
3.3.1 Additional Possibilities for Using Fines to Prevent
Violations 68
3.3.1.1 Raising the Cost of Setting Up and Running
Cartels 69
3.3.1.2 Leniency 70
3.3.2 Additional Aspects to Consider 70
3.3.2.1 The Impact of High Fines on the Market Structure 70
3.3.2.2 Equal Treatment 71
3.4 Can the Optimal Fine be Calculated in Practice? 73
4 The European Commission’s 2006 Guidelines on Fines 77

4.1 Summary of the Guidelines 77
4.2 The Purpose of Fines 79
4.3 The Purpose of Guidelines, and the Question of Foreseeability 81
4.3.1 What are Guidelines For? 81
4.3.2 How Foreseeable should the Amount of Fines Be? 85
4.4 Basic Amount of the Fine 87
4.4.1 The New Method 87
4.4.2 Comparison with Earlier Methods 87
4.4.3 Analysis 90
4.5 Adjustments to the Basic Amount 93
4.5.1 Aggravating Circumstances 93
4.5.1.1 Repeated Infringement 94
4.5.1.2 Refusal to Cooperate With or Obstruction of the
Commission’s Investigations 100
4.5.1.3 Leader, Instigator or Coercer 101
4.5.2 Mitigating Circumstances 103
xii CONTENTS
4.5.2.1 Termination as Soon as the Commission

Intervenes 103
4.5.2.2 Negligence 104
4.5.2.3 Substantially Limited Role (Cheating) 105
4.5.2.4 Cooperation outside the Leniency Notice 107
4.5.2.5 Encouragement by Public Authorities or by
Legislation 108
4.5.3 Specific Increase for Deterrence (Large Undertakings
and/or Amount of the Illicit Gains) 108
4.5.4 Legal Maximum 110
4.5.5 Application of the Leniency Notice 111
4.5.6 Ability to Pay 111
5 Leniency: Theory and Practice 113

5.1 Introduction 113
5.1.1 Definition 113
5.1.2 Examples 113
5.1.2.1 US Department of Justice 113
5.1.2.2 The European Commission 118
5.1.2.3 Competition Authorities of the EU Member States 121
5.1.3 History 122
5.2 Contribution to Optimal Antitrust Enforcement 125
5.2.1 Optimal Antitrust Enforcement and the Imposition of
Penalties 125
5.2.2 Positive Effects of Leniency 127
5.2.2.1 Improved Collection of Evidence and Intelligence 127
5.2.2.2 Increased Difficulty of Creating and Maintaining
Cartels 131
5.2.2.3 Lower Costs of Adjudication 132
5.2.2.4 Restitution to Injured Parties 133
5.2.3 Risks of Negative Effects 134
5.2.3.1 Lowering of the Penalty Level 134
5.2.3.2 Exclusive Reliance on Leniency 136
5.2.3.3 Facilitation of the Creation and Maintenance of
Cartels 137
5.2.3.4 Negative Moral Effects 139
5.2.4 Measuring the Effects of Leniency 140
5.3 Obstacles to the Introduction of Leniency Policies 142
5.3.1 Objections of Principle 142
5.3.2 Institutional Problems 143
5.4 Some Further Issues 144
5.4.1 Leniency and Criminal Penalties on Individuals 144
CONTENTS xiii
5.4.2 Leniency and Follow-on Private Actions for Damages 147

5.4.3 Leniency in Multiple Jurisdictions 148
5.4.4 ‘Amnesty Plus’ 149
5.4.5 Positive Financial Rewards or Bounties 151
5.4.5.1 Rewards to Companies that Participated in Cartels 151
5.4.5.2 Rewards to Individual Cartel Offenders 152
5.4.5.3 Rewards to Non-offender Informants 153
6 Is Criminalisation the Answer? 155

6.1 What do We Mean by ‘Criminalisation’? 155
6.1.1 No Definition of ‘Criminal’ in EC or EU Law 155
6.1.2 Distinguishing Characteristics of Criminal Law 156
6.1.2.1 Criminal Penalties 156
6.1.2.2 Criminal Intent 157
6.1.2.3 Moral Condemnation 157
6.1.2.4 Less Strict Relationship between Penalty and
Harm 157
6.1.2.5 Criminal Powers of Investigation 158
6.1.2.6 Criminal Rights of Defence
6.1.2.7 Links between the Six Characteristics 158
6.1.3 The Wider Notion of ‘Criminal’ in the European
Convention on Human Rights 159
6.2 Is There a Tendency to Criminalise Antitrust Enforcement in the
EU Member States? 161
6.2.1 US Antitrust Enforcement 161
6.2.2 Antitrust Enforcement at the Level of the EU Institutions 164
6.2.3 The EU Framework for Antitrust Enforcement in the EU
Member States 167
6.2.3.1 Member States can Criminalise their Enforcement
of Articles 81 and 82 EC 168
6.2.3.2 Member States cannot Criminalise the
Enforcement of Only their National Competition Laws 169
6.2.4 Decriminalisation and Criminalisation of Antitrust
Enforcement in the EU Member States 173
6.2.4.1 Decriminalisation on the One Hand 173
6.2.4.2 Criminalisation on the Other Hand 175
6.3 Is Criminalisation of Antitrust Enforcement Desirable? 177
6.3.1 Five Arguments in Favour of Imprisonment 177
6.3.1.1 Effective Deterrence with Only Fines on
Companies would Require Impossibly High Fines 177
xiv CONTENTS
6.3.1.2 Fines on Companies do not Always Guarantee

Adequate Incentives for Responsible Individuals within
the Firm 181
6.3.1.3 Increased Effectiveness of Leniency and Whistle
Blowing Programmes 182
6.3.1.4 Imprisonment is a Very Effective Deterrent 183
6.3.1.5 Imprisonment Carries a Uniquely Strong Moral
Message 185
6.3.2 For Which Types of Antitrust Violations is Imprisonment
Desirable? 185
6.3.3 Are There no Equally Effective Alternatives to
Imprisonment? 186
6.3.3.1 Fines on Individuals 187
6.3.3.2 Director Disqualification 187
6.3.3.3 Private Actions for Damages 188
6.3.4 Five Conditions for Criminal Antitrust Enforcement to be
Effective not only in Theory but also in Practice 188
6.3.4.1 The Need for a Dedicated Investigator and
Prosecutor 189
6.3.4.2 The Need for Adequate Powers of Investigation 189
6.3.4.3 Judges and Juries must be Willing to Convict 191
6.3.4.4 The Need for Political and Public Support 191
6.3.4.5 Fines on Companies and Public Penalties for Other
Antitrust Violations than Hard-core Cartels Remain
Necessary 192
6.4 Is it Problematic that Antitrust Enforcement is Criminalised at the
Level of Individual EU Member States Without Parallel
Criminalisation at the Level of the EU Institutions, or Without EU
Harmonisation? 193
6.4.1 Regulation 1/2003 has been Designed to Accommodate
Criminalisation in Individual Member States 193
6.4.2 What would Be the Advantages of Criminalising the
Enforcement of EU Antitrust Law in all Member States,
through EU Harmonisation, as well as at the Level of the EU
Institutions? 195
6.5 Would it be Legally Possible to Criminalise Antitrust Enforcement
at the Level of the EU Institutions, or to Criminalise Antitrust
Enforcement in all EU Member States through EU Harmonisation? 197
6.5.1 Under the Current EC Treaty 197
6.5.2 Under the Signed but Unratified Constitutional Treaty 199
Index 203
Table of Cases
European Court of Justice
Judgments
Case 29/69 Stauder v Stadt Ulm, 12 November 1969 [1969] ECR 419:
paragraph 52 (footnote 61)
Case 41/69 ACF Chemiefarma v Commission, 15 July 1970 [1970] ECR 661:
paragraphs 209 (footnote 74), 257 (footnote 19)
Case 6/72 Europemballage and Continental Can v Commission, 21 February 1973 [1973]
ECR 215:
Case 37/79 Anne Marty v Estée Lauder, 10 July 1980 [1980] ECR 2481:
Case 136/79 National Panasonic v Commission, 26 June 1980 [1980] ECR 2056:
Case 155/79 AM&S v Commission, 18 May 1982 [1982] ECR 1575:
Joined Cases 100/80 etc. Musique Diffusion Française v Commission, 7 June 1983 [1983]
ECR 1825:
paragraphs 268 (footnote 42), 279 (footnote 70), 299 (footnote 101), 339
(footnote 176)
Case 60/81 IBM v Commission, 11 November 1981 [1981] ECR 2639:
Case 170/83 Hydrotherm Gerätebau GmbH v Compact del Dott. Ing. Mario Andreoli & C
Sas, 12 July 1985 [1985] ECR 3016:
Joined Cases 89/85 etc. Ahlström Osakeyhtiö and Others v Commission, 31 March 1993
[1993] ECR I-1575:
Case 246/86 Belasco and Others v Commission, 11 July 1989 [1989] ECR 2191:
Case 85/87 Dow Benelux v Commission, 17 October 1989 [1989] ECR 3137:
Joined Cases 133/87 etc. Nashua Corporation and Others v Commission and Council, 14
March 1990 [1990] ECR I-767:
Case 374/87 Orkem v Commission, 18 October 1989 [1989] ECR 3343:
Case 68/88 Commission v Greece, 21 September 1989 [1989] ECR 2965:
Case C-36/92 P SEP v Commission, 19 May 1994 [1994] ECR I-1932:
Case C-51/92 P Hercules v Commission, 8 July 1999 [1999] ECR I-4235:
Case C-60/92 Otto v Postbank, 10 November 1993 [1993] ECR I-5707:
Case C-199/92 P Hüls v Commission, 8 July 1999 [1999] ECR I-4383:
Columns Design Ltd / Job: Wils_Antitrust / Division: Wils_Table_Cases /Pg. Position: 1 / Date: 22/1
xvi TABLE OF CASES
Opinion 2/94 Accession to the ECHR, 28 March 1996 [1996] ECR I-1759:
Case C-137/95 P SPO and Others v Commission, 25 March 1996 [2006] ECR I-1627:
Case C-185/95 P Baustahlgewebe v Commission, 17 December 1998 [1998] ECR I-8417:
Case C-219/95 P Ferriere Nord v Commission, 17 July 1997 [1997] ECR I-4440:
Case C-231/96 Edilizia Industriale Siderurgica Srl (Edis) v Ministero delle Finanze, 15
September 1998 [1998] ECR I-4990:
Case C-309/96 Daniele Annibaldi v Sindaco del Comune di Guidonia and Presidente
Regione Lazio, 18 December 1997 [1997] ECR I-7493:
Case C-7/97 Bronner v Mediaprint, 26 November 1998 [1998] ECR I-7819:
Case C-70/97 P Kruidvat v Commission, 17 November 1998 [1998] ECR I-7213:
Case C-292/97 Kjell Karlsson and Others, 13 April 2000 [2000] ECR I-2760:
Case C-279/98 P Cascades v Commission, 16 November 2000 [2000] ECR I-9709:
Case C-297/98 P SCA Holding v Commission, 16 November 2000 [2000] ECR I-10101:
Case C-298/98 P Metsä-Serla (Finnboard) v Commission, 16 November 2000 [2000] ECR
I-10171:
Joined Cases C-238/99 P etc. Limburgse Vinyl Maatschappij (LVM) and Others v
Commission (PVC II), 15 October 2002 [2002] ECR I-8618:
Case C-354/99 Commission v Ireland, 18 October 2001 [2001] ECR I-7657:
Case C-453/99 Courage v Crehan, 20 September 2001 [2001] ECR I-6314:
paragraphs 138 (footnote 42), 139 (footnote 43), 140 (footnote 45)
Case C-94/00 Roquette Frères SA v Directeur général de la concurrence et de la répression
des fraudes, and EC Commission, 22 October 2002 [2002] ECR I-9039:
Joined Cases C-204/00 P etc. Aalborg Portland and Others v Commission, 7 January 2004
[2004] ECR I-123:
paragraphs 279 (footnote 70), 294 (footnote 91), 305 (footnotes 126 and 127),
326 (footnote 156)
Case C-253/00 Antonio Muñoz e Cia SA and Superior Fruticola SA v Frumar Ltd and
Redbridge Product Marketing Ltd, 17 September 2002 [2002] ECR I-7289
Case C-198/01 Consorzio Industriale Fiammiferi (CIF) v Autorità Garante della
Concorrenza del Mercato, 9 September 2003 [2003] ECR I-8055:
Case C-481/01 P(R) NDC Health v Commission, 11 April 2002 [2002] ECR I-3401:
Case C-60/02 Criminal proceedings against X, 7 January 2004 [2004] ECR I-651:
TABLE OF CASES xvii
Joined Cases C-65/02 P etc. ThyssenKrupp Stainless and Others v Commission, 14 July
2005 [2005] ECR I-6773:
Case C-141/02 P Commission v T-Mobile Austria, 22 February 2005 [2005] ECR I-1283:
Joined Cases C-189/02 P etc. Dansk Rørindustri and Others v Commission, 28 June 2005
[2005] ECR I-5425:
paragraphs 244 (footnote 2), 262 (footnote 29), 267 (footnote 39), 278 (footnote
69), 299 (footnote 101), 306 (footnote 128), 308 (footnote 131), 337 (footnote
173), 428 (footnote 145)
Case C-263/02 P Commission v Jégo-Quéré, 1 April 2004 [2004] ECR I-3425:
Case C-304/02 Commission v France, 12 July 2005 [2005] ECR I-6263:
Joined Cases C-387/02 etc. Criminal proceedings against SilvioBerlusconi and Others, 3
May 2005 [2005] ECR I-3565:
Case C-176/03 European Commission v Council of the European Union, 13 September
2005 [2005] ECR I-7879:
Case C-397/03 P Archer Daniels Midland v Commission, 18 May 2006 [2006] ECR
I-4429:
Case C-540/03 European Parliament v Council of the European Union, 27 June 2006
[2006] ECR I-5769:
paragraphs 56 (footnotes 67 and 70), 112 (footnote 28)
Case C-552/03 P Unilever v Commission, 28 September 2006, not yet reported in ECR:
Case C-167/04 P JCB Service v Commission, 21 September 2006, not yet reported in ECR:
Case C-289/04 P Showa Denko v Commission, 29 June 2006 [2006] ECR I-5859:
paragraphs 268 (footnote 41), 335 (footnotes 167 and 168)
Case C-295/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA, 13 July 2006
[2006] ECR I-6619:
Case C-301/04 P Commission v SGL Carbon, 29 June 2006 [2006] ECR I-5915:
paragraphs 55 (footnote 66), 59 (footnote 76), 305 (footnotes 123, 124 and 126),
399 (footnote 89), 428 (footnote 145)
Case C-308/04 P SGL Carbon v Commission, 29 June 2006 [2006] ECR I-5977:
129), 343 (footnote 181), 455 (footnote 191)
Case C-3/06 P Groupe Danone v Commission, 8 February 2007, not yet reported in ECR:
91), 294 (footnote 92), 300 (footnote 102), 301 (footnote 112), 303 (footnotes
120 and 121)
Opinions
Case 26/75 General Motors v Commission, 29 October 1975 [1975] ECR 1389
(Advocate-General Mayras):
xviii TABLE OF CASES
Case C-36/92 P SEP v Commission, 15 December 1993 [1994] ECR I-1914

(Advocate-General Jacobs):
Case C-298/98 P Finnboard v Commission, 18 May 2000 [2000] ECR I-10159
(Advoacte-General Mischo):
Case C-253/00 Antonio Muñoz y Cia SA and Superior Fruticola SA v Frumar Ltd and
Redbridge Produce Marketing Ltd, 13 December2001 [2002] ECR I-7289
(Advocate-General Geelhoed):
Joined Cases C-246/01 etc. AOK Bundesverband v Ichthyol-Gesellschaft Cordes, 22 May
2003 [2004] ECR I-2493 (Advocate-General Jacobs):
Joined Cases C-65/02 P etc. ThyssenKrupp Stainless and Others v Commission, [2005]
ECR I-6773 (Advocate-General Léger):
Case C-189/02 P Dansk Rorindustri v Commission, 8 July 2004 [2005] ECR I-5425
(Advocate-General Tizzano):
paragraphs 234 (footnote 100) , 235 (footnote 102)
Case C-289/04 P Showa Denko v Commission, 19 January 2006 [2006] ECR I-5859
(footnotes 169 and 170), 337 (footnote 171)
Case C-301/04 P Commission v SGL Carbon, 19 January 2006 [2006] ECR I-5915
Case C-308/04 P SGL Carbon v Commission, 19 January 2006 [2006] ECR I-5977
Case C-328/05 P SGL Carbon v Commission, 18 January 2007, not yet reported in ECR
(Advocate-General Mazák):
Case C-3/06 P Groupe Danone v Commission, 16 November 2006, not yet reported in
ECR (Advocate-General Poiares Maduro):
Case C-76/06 P Britannia Alloys v Commission, 1 March 2007, not yet reported in ECR
(Advocate-General Bot):
Court of First Instance

Judgments
Case T-6/89 Enichem Anic v Commission, 17 December 1991 [1991] ECR II-1695:
Case T-147/89 Société Métallurgique de Normandie v Commission, 6 April 1995 [1995]
ECR II-1061:
Case T-148/89 Tréfilunion v Commission, 6 April 1995 [1995] ECR II-1119:
Case T-151/89 Société des Treillis et Paneaux Soudés v Commission, 6 April 1995 [1995]
ECR II-1195:
TABLE OF CASES xix
Case T-24/90 Automec v Commission (Automec II), 18 September 1992 [1992] ECR
II-2250:
Case T-141/94 Thyssen Stahl v Commission, 11 March 1999 [1999] ECR II-347:
paragraph 301 (footnotes 111 and 113)
Joined Cases T-305/94 etc. Limburgse Vinyl Maatschappij and Others v Commission, 20
April 1999 [1999] ECR II-987:
Case T-327/94 SCA Holding v Commission, 14 May 1998 [1998] ECR II-1373:
Joined Cases T-25/95 etc. Cimenteries CBR and Others v Commission (Cement), 15 March
2000 [2000] ECR II-700:
Joined Cases T-213/95 and T-18/96 SCK and FNK v Commission, 22 October 1997
[1997] ECR II-1764:
Case T-87/96 Assicurazioni Generali and Unicredito v Commission, 4 March 1999 [1999]
ECR II-206:
Case T-9/97 Elf Atochem v Commission, 9 June 1997 [1997] ECR II-919:
Case T-241/97 Stork Amsterdam v Commission, 17 February 2000 [2000] ECR II-309:
Joined Cases T-45/98 etc. Krupp Thyssen Stainless and Others v Commission, 13
December 2001 [2001] ECR II-3765:
(footnote 149)
Case T-65/98 Van den Bergh Foods v Commission, 23 October 2003 [2003] ECR II-4653:
Case T-112/98 Mannesmannröhren-Werke v Commission, 20 February 2001 [2001] ECR
II-732:
Joined Cases T-191/98 etc. Atlantic Container Line and Others v Commission, 30
September 2003 [2003] ECR II-3275:
Case T-248/98 KPN v Commission, 16 November 2000 [2000] ECR I-9641:
Case T-9/99 HFB and Others v Commission, 20 March 2002 [2002] ECR II-1530:
Case T-31/99 ABB Asea Brown Boveri v Commission, 20 March 2002 [2002] ECR II-1884:
Case T-54/99 max.mobil v Commission, 30 January 2000 [2002] ECR II-313:
Case T-219/99 British Airways v Commission, 17 December 2003 [2003] ECR II-5917:
Case T-44/00 Mannesmannröhren-Werke v Commission, 8 July 2004 [2004] ECR II-2223:
Joined Cases T-67/00 etc. JFE Engineering Corp v Commission, 8 July 2004 [2004] ECR
II-2501:
Case T-213/00 CMA CGM and Others v Commission, 19 March 2003 [2003] ECR II-913:
xx TABLE OF CASES
Case T-224/00 Archer Daniels Midland v Commission, 9 July 2003 [2003] ECR II-2597:
Case T-177/01 Jégo-Quéré v Commission, 3 May 2002, [2002] ECR II-5137:
Case T-203/01 Michelin v Commission, 30 September 2003 [2003] ECR II-4071:
Joined Cases T-236/01 etc. Tokai Carbon and Others v Commission, 29 April 2004 [2004]
ECR II-1181:
Case T-241/01 SAS v Commission, 18 July 2005 [2005] ECR II-2917:
Case T-322/01 Roquette Frères v Commission, 27 September 2006, not yet reported in
ECR:
Case T-329/01 Archer Daniels Midland v Commission, 27 September 2006, not yet
reported in ECR:
142)
Case T-15/02 BASF v Commission, 15 March 2006 [2006] ECR II-497:
(footnote 133), 425 (footnote 140), 429 (footnote 149), 439 (footnote 164)
Case T-26/02 Daiichi Pharmaceutical v Commission, 15 March 2006 [2006] ECR II-713:
Case T-28/02 First Data v Commission, 17 October 2005 [2005] ECR II-4119:
Case T-33/02 Britannia Alloys & Chemicals v Commission, 29 November 2005 [2005]
ECR II-4973:
Case T-38/02 Groupe Danone v Commission, 25 October 2005 [2005] ECR II-4407:
paragraphs 270 (footnotes 44 and 47), 288 (footnote 84), 292 (footnote 89), 295
(footnote 94), 298 (footnote 99), 300 (footnote 102), 301 (footnote 112)
Case T-43/02 Jungbunzlauer v Commission, 27 September 2006, not yet reproted in ECR:
Joined Cases T-49/02 etc. Brasserie nationale and Others v Commission, 27 July 2005
[2005] ECR II-3033:
Case T-52/02 SNCZ v Commission, 29 November 2005 [2005] ECR II-5005:
Case T-59/02 Archer Daniel Midland v Commission, 27 September 2006, not yet reported
in ECR:
Joined Cases T-259/02 etc. Raiffeisen Zentralbank Österreich and Others v Commission, 14
December 2006, not yet reported in ECR:
Case T-279/02 Degussa v Commission, 5 April 2006 [2006] ECR II-897:
Joined Cases T-71/03 etc. Tokai Carbon and Others v Commission, 15 June 2005 [2005]
ECR II-10:
Case T-340/03 France Télécom v Commission, 30 January 2007, not yet reproted in ECR:
TABLE OF CASES xxi
Case T-201/04 R Microsoft v Commission, 22 December 2004 [2004] ECR II-4463:

Opinion
Case T–1/89 Rhône-Poulenc v Commission, 10 July 1991 [1991] ECR II–867 (Judge
Vesterdorf)
European Court of Human Rights

8 June 1976, Engel and Others v Netherlands, Series A No 22:
27 February 1980, Deweer v Belgium, Series A No 35:
23 June 1981, Le Compte, Van Leuven and De Meyere v Belgium, Series A No 43:
1 February 1983, Albert and Le Compte v Belgium, Series A No 58:
21 February 1984, Öztürk v Gremany, Series A No 73:
24 February 1994, Bendenoun v France, Series A No 284:
26 October 1984, De Cubber v Belgium, Series A No 86:
25 February 1997, Findlay v United Kingdom, Reports 1997-I:
pargaraph 491 (footnote 19)
10 March 2004, Senator Lines v the 15 Member States of the European Union, Application
56672/00:
3 June 2004, OOO Neste St Petersburg and Others v Russia, Applications 69042/01 etc.:
28 April 2005, Buck v Germany, Application 41604/98:
30 June 2005, Bosphorus v Ireland, Application 45036/98:
Table of Legislation
European Convention on Human Rights
Article 5: paragraph 623 (footnote 190)
Article 6: paragraphs 55, 484 to 491, 623 (footnote 190)
Charter of Fundamental Rights of the EU
Article 41: paragraph 112
EU Treaty
EC Treaty
(Articles 81 and 82 of the EC Treaty are not included, as they are discussed
throughout the whole book)
Article 10: paragraphs 42, 114
Article 83: paragraphs 255, 520, 581 (footnote 149), 619 to 624
Article 226: paragraphs 22, 40 (footnote 44), 42 (footnote 47)
Article 308: paragraphs 6??, 624
Council Regulation 1/2003 [2003] OJ L1/1
Recital 8: paragraphs 472, 532
Recital 11: paragraph 92 (footnote 13)
Recital 12: paragraphs 124, 136
Recital 13: paragraphs 80, 102, 106, 125, 141, 145, 148 (footnote 49), 160
Recital 16: paragraph 70
Recital 22: paragraphs 81, 141, 145, 148 (footnote 49)
Recital 23: paragraphs 44, 57 to 59
Recital 37: paragraph 56
Article 3: paragraphs 150, 522 to 532
Article 5: paragraphs 32, 157, 515 to 518, 604
Article 7: paragraphs 91 to 105, 107, 114, 118, 122 to 124, 161, 162
Article 9: paragraphs 76 to 160
Article 11(6): paragraph 49
Article 12: paragraphs 27, 41, 43, 67 to 75, 517, 605 to 608, 611, 613
Article 18: paragraphs 7 to 10, 26, 44, 49, 51
Article 20: paragraphs 11 to 15, 18, 19 to 22, 24, 25, 50, 57, 512
Article 21: paragraphs 16 to 18, 19 to 22, 24, 25, 36 to 39, 44, 49
Article 22: paragraphs 24, 25, 41, 42
Columns Design Ltd / Job: Wils_Soft-proofing / Division: Wils_Table_Legislation /Pg. Position: 1 / Date: 10/12
xxiv TABLE OF LEGISLATION
Article 23(1): paragraphs 8 to 10, 13, 14, 18, 23, 308, 512
Article 23(2): paragraphs 78, 87, 92, 132, 180 (footnote 35), 212, 232, 251, 256,
262, 321, 338, 339, 502 to 505
Article 23(5): paragraphs 472, 506, 507
Article 24: paragraphs 9, 10, 14, 18, 23, 79, 87, 132
Article 27: paragraphs 107, 117, 153
Commission Regulation 773/2004 [2004] OJ L123/18
Columns Design Ltd / Job: Wils_Soft-proofing / Division: Wils_Table_Legislation /Pg. Position: 2 / Date: 10/12
1
Powers of Investigation and Procedural
Rights and Guarantees: The Interplay
between European and National
Legislation and Case Law
1.1 POWERS OF INVESTIGATION
1.1.1 Different Ways of Gathering Intelligence and Evidence
1. Under Regulation 1/2003,1 the European Commission (hereafter also ‘the

Commission’) and the competition authorities of the Member States (hereafter
also ‘the national competition authorities’), together forming a network of
competition authorities (hereafter also ‘the competition authorities’ and ‘the
European Competition Network’),2 have the task of detecting and punishing
violations of Articles 81 and 82 EC.3 To fulfil this task, they need intelligence and
evidence of such violations.
2. The competition authorities may collect the necessary information (intelli-
gence and evidence) from various sources. Information could be obtained in
particular from complainants or other third parties. For certain types of antitrust
1
Council Reg 1/2003 of 16 Dec 2002 on the implementation of the rules on competition laid
down in Arts 81 and 82 of the Treaty [2003] OJ L1/1, last amended by Council Reg 1419/2006 [2006]
OJ L269/1. According to its Arts 43 and 45, this Reg replaces Council Regulation 17 [1962] JO 13/204,
[1959–62] OJ Spec Ed 87, last amended by Council Reg 1216/1999 [1999] OJ L148/5, with effect from
1 May 2004. On Reg 1/2003 generally see L Idot, Droit communautaire de la concurrence—Le nouveau
système communautaire de mise en oeuvre des articles 81 et 82 CE (Brussels, Bruylant, 2004); L
Garzaniti, J Gudofsky and J Muffat, ‘Dawn of a New Era? Powers of Investigation and Enforcement
Under Regulation 1/2003’ (2004) 72 Antitrust Law Journal 159; D Geradin (ed), Modernisation and
Enlargement: Two Major Challenges for EC Competition Law (Antwerp, Intersentia, 2004); D Cahill
and JD Cooke (eds), The Modernisation of EU Competition Law Enforcement, Reports for the XXI FIDE
Congress, Dublin 2004 (Cambridge, Cambridge University Press, 2004); WPJ Wils Principles of
European Antitrust Enforcement (Oxford, Hart Publishing, 2005); CS Kerse and N Khan, EC Antitrust
Procedure, 5th edn (London, Sweet & Maxwell, 2005); and L Ortiz Blanco (ed), EC Competition
Procedure, 2nd edn (Oxford, Oxford University Press 2006).
2
See rec 15 and Art 11 of Reg 1/2003, above n 1, and the European Commission’s Notice on
cooperation within the Network of Competition Authorities [2004] OJ C101/03.
3
The authorities also have other tasks, but those are not considered here. As to the role of private
enforcement see Ch 4, as well as nn 62 and 79 in Ch 5 of Wils, above n 1.
Columns Design Ltd / Job: Wils_Antitrust / Division: Wils_chapter1 /Pg. Position: 1 / Date: 14/1
2 INVESTIGATION AND PROCEDURAL RIGHTS
violations, in particular exclusionary practices, but also non-covert exploitative

practices, the competition authorities may be able relatively easily to obtain the
necessary intelligence and evidence through interested third parties.4 The best
information will however usually be in the hands of the undertakings that have
committed the violations and their staff. For some types of violations, in
particular for secret price cartels, the undertakings that committed the violations
and their staff may be the only ones holding the information which the
competition authorities need to detect and punish these violations.
3. Three (types of) methods could conceivably be used by the competition
authorities in order to collect intelligence and evidence of violations of Article 81
or 82 EC from the undertakings that have committed these violations (or are
suspected of having committed them)5 or from their staff.6 First, the competition
authorities could try directly to get hold of the information in the hands of the
undertakings and their staff by using direct physical force (‘dawn raids’) or covert
intrusion. Secondly, they could use compulsion in the form of threatened
sanctions for refusal to cooperate in order to make the undertakings and their
staff provide the necessary information. Thirdly, instead of using the ‘stick’ of
threatened sanctions for non-cooperation, they could also create incentives to
cooperate by using the ‘carrot’ of a promise of immunity or reduced punishment
in the event of cooperation (‘leniency’). It would appear that a combination of all
three methods is required for effective and efficient antitrust enforcement.7
4. This chapter deals with only the first and second methods of collecting
intelligence and evidence of antitrust infringements, not with leniency.8 While, in
Scott Hammond’s words, ‘leniency programs are the greatest investigative tool
ever designed to fight cartels’,9 the use of leniency does not make it possible to
dispense with the other two methods of obtaining information from antitrust
violators; rather to the contrary. Indeed, leniency can work only if cartel
4
These third parties have an incentive to help the competition authorities, so as to bring to an end
violations from which they suffer. The possibility of obtaining damages following the establishment
of the violation by the competition authorities may provide an additional incentive. See further Wils,
above n 1, Ch 5, n 62, as well as Ch 5 below, sect 5.4.5.3.
5
I will not repeat this qualification throughout the text, the legal presumption of innocence being
of course fully recognised.
6
Some of these methods may also be used to collect information from third parties, but that
aspect is not further considered here.
7
For a detailed analysis see paras 530–553 of Wils, above n 1; see also below, Ch 5, in particular
sect 5.2.2.1.
8
On leniency see Ch 5 below; Ch 3 of WPJ Wils, The Optimal Enforcement of EC Antitrust Law
(The Hague, Kluwer Law International, 2002); and Ch 5 of Wils, above n 1.
9
SD Hammond, ‘Cornerstones of an Effective Leniency Program’, paper presented before the ICN
Workshop on Leniency Programs (Sydney, 22–23 Nov 2004), available at www.usdoj.gov/atr/public/
speeches/206611.htm, at 2.
POWERS OF INVESTIGATION 3
participants perceive a risk that the competition authorities may detect the
antitrust violation without recourse to leniency.10
1.1.2 Cases Dealt with by the European Commission
5. If the European Commission deals with a case,11 it can itself collect the
necessary information or it can rely on the competition authorities of the
Member States to do so on its behalf.
1.1.2.1 Information-gathering by the Commission Itself
6. The Commission’s powers of investigation are set out in Chapter V of

Regulation 1/2003, the two main instruments being requests for information
(Article 18) and inspections (Articles 20 and 21).12
7. Under Article 18 of Regulation 1/2003, the Commission can require
undertakings and associations of undertakings to provide all necessary informa-
tion (ie either to hand over existing documents or to provide answers to
questions) specified in the request within the time-limit fixed in the request.
8. The Commission can choose either of two forms for its request: a simple
request or a decision.13 An undertaking or association of undertakings is not
obliged to respond to a simple request. Only if the undertaking or association,
intentionally or negligently, provides incorrect or misleading information can it
be penalised by the Commission under Article 23(1)(a) of Regulation 1/2003
with a fine of up to 1 per cent of its total turnover in the preceding business year.
9. On the other hand, requests made by decision are binding. Under Article
23(1)(b) of Regulation 1/2003, the Commission can impose a fine of up to 1 per
cent of their total turnover in the preceding business year on undertakings or
associations which, intentionally or negligently, supply incorrect, incomplete or
10
See below, Ch 5, in particular sect 5.2.3.2; paras 549 to 553 of Wils, above n 1; and Hammond,
above n 9, at 5, listing a (perceived) high risk of detection as one of the 3 indispensable components of
an effective leniency programme.
11
On the allocation of cases within the European Competition Network see the European
Commission’s Notice on cooperation within the Network of Competition Authorities [2004] OJ
C101/03, sect 2; D Schnichels, ‘The Network of Competition Authorities: How Will It Work in
Practice?’, in Geradin, above n 1, at 99; and Wils, above n 1, sect 1.2.5.
12
‘Inspections’ is the new term in Reg 1/2003 for what Reg 17 called ‘investigations’. Under Art 19
of Reg 1/2003, the Commission may also interview any natural or legal person who consents to be
interviewed for the purpose of collecting information relating to the subject-matter of the investiga-
tion, but this ‘power to take statements’ is not much of a power, in the absence of any obligation or
any penalties, not even for providing misleading information: see further B Vesterdorf, ‘Legal
Professional Privilege and the Privilege Against Self-Incrimination in EC Law: Recent Developments
and Current Issues’ in BE Hawk (ed), 2004 Annual Proceedings of the Fordham Corporate Law
Institute—International Antitrust Law and Policy (Huntington, NY, Juris Publishing, 2005) 701 at 727.
13
Under Art 11 of Reg 17, the Commission could use a decision only after the undertaking had
refused to answer a simple request. This limitation on the Commission’s choice of instrument has
been removed in Reg 1/2003.
misleading information or do not supply information within the required

time-limit, and under Article 24(1)(d), the Commission can impose periodic
penalty payments of up to 5 per cent of the average daily turnover in the
preceding business year per day in order to compel undertakings or associations
to supply complete and correct information which it has requested by decision.14
10. Requests for information under Article 18 of Regulation 1/2003 can be
addressed only to undertakings or associations of undertakings. Article 18(4)
provides: ‘[t]he owners of undertakings or their representatives and, in the case of
legal persons, companies or firms, or associations having no legal personality, the
persons authorised to represent them by law or by their constitution shall supply
the information requested on behalf of the undertaking or the association of
undertakings concerned. Lawyers duly authorised to act may supply the informa-
tion on behalf of their clients. The latter shall remain fully responsible if the
information supplied is incomplete, incorrect or misleading.’ The fines and
periodic penalty payments provided for in Articles 23 and 24 of Regulation
1/2003 can be imposed only on undertakings or associations of undertakings.15
11. Under Article 20 of Regulation 1/2003, the Commission may conduct ‘all
necessary inspections of undertakings and associations of undertakings’. The
officials and other persons authorised by the Commission to conduct the
inspection have the power to enter any premises of the undertaking or associa-
tion of undertakings, to examine all business-related records and to take or
obtain copies or extracts. Article 20(2)(d) allows the Commission inspectors ‘to
seal any business premises and books or records for the period and to the extent
necessary for the inspection’.16 Under Article 20(2)(e), they are also empowered
‘to ask any representative or member of staff of the undertaking or association of
undertakings for explanations on facts or documents relating to the subject-
matter and purpose of the inspection and to record the answers’.17
12. Article 20(5) of Regulation 1/2003 provides that ‘officials of as well as
those authorised or appointed by the competition authority of the Member State
in whose territory the inspection is to be conducted shall, at the request of that
14
The maximum amounts of these penalties are new compared to those under Reg 17, where the
maximum fine was €5,000 and the maximum amount for periodic penalty payments €1,000 per day.
15
See below n 31 on the notion of ‘undertaking’.
16
This latter power was not provided for in Reg 17.
17
Under Reg 17, the Commission officials could only seek explanations relating to the books and
records under examination: see Judgment of the ECJ of 26 June 1980 in Case 136/79 National
Panasonic v Commission [1980] ECR 2056, para 15, and the Order of the CFI of 9 June 1997 in Case
T–9/97 Elf Atochem v Commission [1997] ECR II–919, para 23. It should immediately be pointed out,
however, that Reg 1/2003 does not provide for any sanction to be imposed on the members of staff for
failing to provide answers or for providing incorrect, incomplete or misleading answers. Penalties can
only be imposed on the undertaking that fails to correct within a time-limit set by the Commission an
incorrect, incomplete or misleading answer given by a member of staff: see paras 13 and 14 below, and
further Vesterdorf, above n 12, at 727–9.
authority or of the Commission, actively assist the officials and other accompa-
nying persons authorised by the Commission. To this end, they shall enjoy the
[same] powers’.
13. The Commission can again choose either of two forms of inspection: on
the basis of a simple authorisation or on the basis of a decision. If the inspection
is based on a simple authorisation, the undertaking or association of undertak-
ings is not obliged to submit to it. However, under Article 23(1)(c) and (d) of
Regulation 1/2003, the Commission can impose on the undertaking or associa-
tion a fine of up to 1 per cent of the total turnover in the preceding business year
where, intentionally or negligently, it produces the required business-related
records in incomplete form or where, in response to a question asked under
Article 20(2)(e), it gives an incorrect or misleading answer or fails to correct
within a time-limit set by the Commission an incorrect, incomplete or mislead-
ing answer given by a member of staff.18
14. If the inspection is based on a decision, the undertaking or association is
required to submit to it. In addition to the penalties which can be imposed in
case of an inspection based on a simple authorisation, the Commission can also,
under Article 23(1)(c) and (d) of Regulation 1/2003, impose fines of the same
amount where the undertaking or association refuses to submit to the inspection,
or where, in response to a question asked under Article 20(2)(e), it fails or refuses
to provide a complete answer on facts relating to the subject-matter and purpose
of the inspection. Under Article 24(1)(e) of Regulation 1/2003, the Commission
can also impose periodic penalty payments of up to 5 per cent of the average
daily turnover in the preceding business year per day in order to compel
undertakings or associations to submit to an inspection ordered by decision.19
15. Moreover, Article 20(6) of Regulation 1/2003 provides that, where the
Commission officials conducting the inspection find that an undertaking
opposes an inspection ordered by decision, the Member State concerned shall
afford them the necessary assistance of the police or of an equivalent enforce-
ment authority, so as to enable them to conduct their inspection.
16. Under Article 21 of Regulation 1/2003, the Commission also has the
power to order by decision an inspection of ‘any other premises …, including the
homes of directors, managers and other members of staff of the undertakings
and associations of undertakings concerned’ if ‘a reasonable suspicion exists that
books or other records related to the business and to the subject-matter of the
inspection, which may be relevant to prove a serious violation of Article 81 or
Article 82 of the Treaty, are being kept in [those premises]’.20
17. Like in the case of an inspection at business premises, the officials and
other persons authorised by the Commission to conduct the inspection have the
18
No sanctions can be imposed by the Commission on the member of staff: see above n 17.
19
See above n 14.
20
This power is new compared to Reg 17; see in this context App 41604/98, Buck v Germany,
ECtHR judgment.
power to examine all business-related records and to take or obtain copies or

extracts, but they have no power to put seals on premises or documents or to ask
for explanations.21
18. Article 20(6) of Regulation 1/2003, which provides for assistance by the
national authorities, including the police, in case of opposition, also applies to
inspections under Article 21 of Regulation 1/2003,22 but the powers of the
Commission to impose fines or periodic penalty payments under Articles 23 and
24 of Regulation 1/2003 are not applicable.23
1.1.2.2 Assistance by the (Competition) Authorities of the Member States
19. As already indicated above,24 when the Commission conducts an inspection

either of business premises under Article 20 or of private homes under Article 21
of Regulation 1/2003, its inspectors can receive the ‘active assistance’ of inspectors
designated by the competition authority of the Member State in whose territory
the inspection is conducted. Article 20(5) of Regulation 1/2003 provides that the
Commission can request such assistance, or—interestingly—the national compe-
tition authority has the right to provide such assistance at its own request. The
assisting national inspectors have the same powers as the Commission inspectors,
as set out in Articles 20 and 21 of Regulation 1/2003.
20. As also already indicated above,25 Article 20(6) of Regulation 1/2003
provides that, where the Commission officials conducting the inspection find
that an undertaking opposes an inspection ordered by decision, the Member
State concerned shall afford them the necessary assistance of the police or of an
equivalent enforcement authority, so as to enable them to conduct their inspec-
tion.
21. Apart from the ability of the Commission itself to impose fines or
periodic penalty payments to compel undertakings to submit to an inspection,26
the Commission thus relies on the authorities of the Member State in whose
territory the inspection takes place to overcome opposition to its inspections. The
Commission inspectors cannot themselves use any force.
22. In the application of the principles of effectiveness and of equivalence—
general principles of Community law, developed in the case law of the Court of
Justice and applicable to all instances where Community law entrusts Member
States with a role in the enforcement of Community law27—the assistance which
21
See above, paras 11 and 12.
22
The provisions of Art 20(5) also apply: see above, paras 15 and 12.
23
24
Above n 22 and para 12.
25
Above, paras 15 and 18.
26
See above, para 14.
27
See, inter alia, Judgments of the ECJ of 21 Sept 1989 in Case 68/88 Commission v Greece [1989]
ECR 2965, paras 23–25, of 15 Sept 1998 in Case C–231/96 Edis [1998] ECR I–4990, paras 34 and
36–37, and of 18 Oct 2001 in Case C–354/99 Commission v Ireland [2001] ECR I–7657, para 46; see
the Member States provide to overcome opposition to Commission inspections

must be effective, and at least equivalent to what the Member State would
provide for in comparable situations of enforcement of its own national law. If a
Member State were to fail to provide the necessary assistance to Commission
inspections, the Commission could bring an action against that Member State
before the Court of Justice under Article 226 EC.28
23. Ultimately the effectiveness of the fines and periodic penalty payments
which the Commission can impose under Articles 23 and 24 of Regulation
1/2003 to compel undertakings to submit to inspections or to answer requests for
information, and to punish the provision of incorrect, incomplete or misleading
information,29 also depends on assistance by the Member States. Indeed, if the
company were to refuse to pay the fine or penalty payment, the Commission
would make use of Article 256 EC, which provides that a national authority
designated for this purpose by the Member State in the territory of which the
Commission decision imposing the financial penalty is to be executed shall
append an order for its enforcement to the Commission decision, allowing that
decision to be enforced in accordance with the rules of civil procedure in force in
that Member State.
24. Apart from obtaining assistance from the Member States for the inspec-
tions which it conducts itself, the Commission can also request the competition
authorities of the Member States to conduct inspections on its behalf. Article
22(2) of Regulation 1/2003 provides as follows:
At the request of the Commission, the competition authorities of the Member States
shall undertake the inspections which the Commission considers to be necessary under
Article 20(1) or which it has ordered by decision pursuant to Article 20(4). The officials
of the competition authority of the Member States who are responsible for conducting
these inspections as well as those authorised or appointed by them shall exercise their
powers in accordance with their national law.
25. The powers of the national inspectors conducting inspections under Article
22(2) of Regulation 1/2003 are thus not those set out in Articles 20 and 21, but
those provided for in the national law of the Member State concerned. However,
it follows again from the principles of effectiveness and equivalence30 that these
powers must be effective, and at least equivalent to what the Member State would
provide for in comparable situations of enforcement of its own national law.
further, for a detailed and very interesting discussion of the application of the principles of
equivalence and effectiveness in the application of Arts 81 and 82 EC, P Oliver, ‘Le règlement 1/2003
et les principes d’efficacité et d’équivalence’ [2005] Cahiers de droit européen 343.
28
And if the problem were to persist after the ECJ had found against the Member State in the
action under Art 226 EC, the Commission could bring a further action under Art 228 EC, in which
the ECJ could impose a fine and/or periodic penalty payment on the Member State: see generally
Judgment of the ECJ of 12 July 2005 in Case C–304/02 Commission v France [2005] ECR I–6263.
29
See above, paras 8, 9, 13 and 14.
30
See above, para 22.
26. Another power for the Commission is provided for in Article 18(6) of
Regulation 1/2003: ‘[a]t the request of the Commission the governments and
competition authorities of the Member States shall provide the Commission with
all the necessary information to carry out the duties assigned to it by this
Regulation’.
27. Finally, Article 12(1) of Regulation 1/2003 provides that ‘[f]or the purpose
of applying Articles 81 and 82 of the Treaty the Commission and the competition
authorities of the Member States shall have the power to provide one another
with and use in evidence any matter of fact or of law, including confidential
information’.31 Even without a request from the Commission, the national
competition authorities could thus provide the Commission with information
which the Commission could subsequently use both as intelligence and as
evidence for the enforcement of Articles 81 and 82 EC.
31
As explained below at paras 69–73, Art 12(3) of Reg 1/2003 limits the ability to use exhanged
information in evidence to impose sanctions on natural persons. This limitation is however of very
limited practical importance where the Commission is the authority receiving the information, as the
Commission can impose sanctions only on undertakings. Neither the EC Treaty nor Reg 1/2003
contains any definition of the term ‘undertaking’. The ECJ has held that ‘in competition law, the term
“undertaking” must be understood as designating an economic unit … even if in law that economic
unit consists of several persons, natural or legal’: Judgment of 12 July 1985 in Case 170/83
Hydrotherm v Compact del Dott Ing Mario Andreoli [1985] ECR 3016 at para 11. In practice the
Commission does not simply impose fines on undertakings thus defined as economic units. For
reasons of enforceability, it rather addresses its fining decisions to entities possessing legal personality,
typically companies, to which the violation committed by the undertaking is imputed: Opinion of
Judge Vesterdorf, acting as AG, in Case T–1/89 Rhône-Poulenc v Commission [1991] ECR II–869 at
916, Judgment of the CFI of 20 Apr 1999 in Joined Cases T–305/94 etc Limburgse Vinyl Maatschappij
and Others v Commission [1999] ECR II–945 at para 978. As to the choice of the person or persons to
whom a violation is to be imputed, the general rule formulated by the Community Courts is that
‘when … a violation is found to have been committed, it is necessary to identify the natural or legal
person who was responsible for the operation of the undertaking at the time when the violation was
committed, so that it can answer for it’: (Judgment of the CFI of 17 Dec 1991 in Case T–6/89 Enichem
Anic v Commission [1991] ECR II–1695 at para 236, and Judgment of the ECJ of 16 Nov 2000 in Case
C–279/98 P Cascades v Commission [2000] ECR I–9709 at para 78. If the undertaking found to have
committed a violation were found to consist of an unincorporated business (a single trader, with or
without employees, who had not incorporated his or her business, or a professional exercising his or
her profession alone and unincorporated, or several natural persons operating a single business
without any employment relationship between them and without any form of legal person), the
Commission would necessarily have to address its fining decision to the natural person or persons
operating the business. However, in the more than 100 decisions in which the Commission has
imposed fines under Regs 17 and 1/2003 hitherto, this situation has never occurred: hitherto, all fines
have been imposed on companies or other legal persons. Indeed, the undertakings found to have
committed violations of Art 81 or 82 EC invariably either coincided with a single company or other
legal person which was the obvious addressee of the fining decision, or consisted of a group of
companies in which case fines were imposed on one or more of the companies in the group. In
Pre-Insulated Pipes, one of the undertakings was controlled and managed by a natural person, Dr W
Henss, but the Commission appears to have made an effort (for reasons unexplained) to avoid
imposing the fine on him, identifying instead a collection of companies which it held jointly and
severally liable for the fine; see recs 157–160 and Art 3(d) of the Decision of 21 Oct 1998,
Pre-Insulated Pipe Cartel [1999] OJ L24/1, and para 105 of the judgment of the CFI of 20 Mar 2002 in
Case T–9/99 HFB and Others v Commission [2002] ECR II–1530, in which the Court found that the
European Commission had ‘intentionally’ not established Dr Henss’s liability.
1.1.3 Cases Dealt with by a Competition Authority of a Member State
28. If a competition authority of a Member State deals with a case, it can either
itself collect the necessary information or it may receive assistance from compe-
tition authorities of other Member States.
1.1.3.1 Information-gathering by the National Competition Authority Itself
29. The powers of investigation of a competiion authority of a Member State are

in principle those provided for in the national law of that Member State.
Regulation 1/2003 has not harmonised the powers of investigation of the
national competition authorities.32
30. In many Member States, the powers of the national competition authori-
ties for the enforcement of Articles 81 and 82 EC are similar to the Commission’s
powers of investigation under Regulation 1/2003, including requests for informa-
tion and inspections.33
31. In some Member States, the powers of investigation of the national
competition authorities appear to be stronger than the Commission’s powers
under Regulation 1/2003.
32. In some Member States, these stronger powers of investigation reflect
stronger sanctions for violations of Article 81 or 82 EC or related offences.34
Whereas Regulation 1/2003 allows the Commission to impose fines only on
undertakings and associations of undertakings for violations of Article 81 or 82
EC, Article 5 allows national authorities to impose ‘any other penalty provided
for in their national law’, including in particular prison sanctions for directors or
managers responsible for their companies’ violation of Article 81 or 82 EC.35
Some Member States may even be under a Community law obligation to
criminalise their enforcement of Articles 81 and 82 EC. Indeed, it follows from
the principle of equivalence that, if a Member State provides, for instance,
imprisonment sanctions to enforce a cartel prohibition under national law, it
must also provide for such sanctions to enforce the cartel prohibition in Article
81 EC.36
33. In those Member States where such criminal sanctions are provided for, or
where criminal sanctions can be imposed not directly for violations of Article 81
32
See however below, paras 37–40, as to the consequences flowing from the principles of
equivalence and effectiveness.
33
See above, paras 6–18.
34
The stronger sanctions mean that the offences are considered as more serious, which justifies
more means being deployed to detect and punish such offences.
35
On the need for prison sanctions to deter effectively the most serious violations of Art 81 EC see
below, Ch 6, in particular sect 6.3.
36
See above,para 22, and sect 6.2.3.2 below.
or 82 EC but for a related offence,37 it can be expected that correspondingly

stronger criminal powers of investigation are also provided for, including for
instance powers to carry out directed surveillance and to use covert human
intelligence sources.38
34. Even in the absence of stronger sanctions for violations of Articles 81 and
82 EC or related offences, the national laws of several Member States give their
national competition authorities stronger powers to investigate violations of
Articles 81 and 82 EC than Regulation 1/2003 gives to the Commission, for
instance in that not only undertakings and associations of undertakings but also
individual members of staff can be required to produce specified documents and
information, and that failure to cooperate with investigations is a criminal
offence, with directors, managers or other officers of companies or other bodies
corporate being liable to punishment if they have consented to or connived at an
offence or an offence results from neglect on their part.39
35. On the other hand, in some Member States the national competition
authorities’ powers of investigation appear to be weaker than those of the
Commission under Regulation 1/2003.
36. For instance, in at least one Member State there has been discussion as to
whether the national competition authority, when conducting inspections at
business premises, has the power to examine and take copies not only of business
records on paper but also of electronically stored information. In a few Member
States, the national competition authorities do not have a power comparable to
the Commission’s power under Article 21 of Regulation 1/2003 to conduct
inspections at private homes, but can only inspect business premises.
37. If the national competition authorities concerned lack these powers not
only for the enforcement of Articles 81 and 82 EC but also for the enforcement of
their national competition laws, and if the laws of the Member States concerned
do not allow for the inspection of electronically stored information or for
searches at private homes in relation to other offences under national law of a
similar nature and importance to violations of Articles 81 and 82 EC, the absence
of such powers is not problematic under the principle of equivalence.40
38. It can, however, be questioned under the principle of effectiveness.41 Even
if Regulation 1/2003 provides for parallel competences of the Commission and
the competition authorities of the Member States, and any ineffectiveness of
37
On the relationship between the cartel offence in s 188 of the Enterprise Act 2002 in the UK and
Reg 1/2003, see below, paras 527–532.
38
See for instance in the UK, the powers for investigating criminal cartels under the Enterprise
Act 2002, described in OFT Guidance Powers for Investigating Criminal Cartels (London, OFT, Jan
2004); see also OFT Codes of practice Covert surveillance in cartel investigations and Covert human
intelligence sources in cartel investigations (London, OFT, Aug 2004), all available at www.oft.gov.uk.
39
See for instance in the UK, the powers of investigation of the OFT under the Competition Act
1998, described in OFT Guideline 404 Powers of Investigation (London, OFT, Dec 2004), available at
www.oft.gov.uk.
40
See above, para 22.
41
Ibid.
enforcement in any Member State due to weak powers of investigation could thus
be parried by increased enforcement by the Commission in that Member State, it
is clear from the whole of Regulation 1/2003 that its purpose is fully to involve all
national competition authorities in the enforcement of Articles 81 and 82 EC.42 It
can thus be argued that Regulation 1/2003 and the principle of effectiveness
impose on Member States the obligation to provide their competition authorities
with effective powers of investigation.43
39. It appears rather obvious that powers of investigation which are limited to
paper documents and do not cover electronically stored information cannot be
effective in the twenty-first century. The Commission’s power to conduct inspec-
tions at private homes is based on the following reasoning in recital 26 of
Regulation 1/2003: ‘[e]xperience has shown that there are cases where business
records are kept in the homes of directors or other people working for an
undertaking. In order to safeguard the effectiveness of inspections, therefore,
officials and other persons authorised by the Commission should be empowered
to enter any premises where business premises may be kept, including private
homes’. It is not apparent why this need for powers to search private homes so as
to safeguard the effectiveness of inspections would be any different in the case of
investigations by national competition authorities.
40. It can thus be argued that all Member States are under an obligation to
provide their national competition authorities with the power, when conducting
inspections at business premises, to examine and copy not only business records
on paper but also all relevant electronically stored information, as well as with a
power to conduct inspections at private homes.44
1.1.3.2 Assistance by the Competition Authorities of Other Member States
41. Article 22(1) of Regulation 1/2003 provides as follows:

The competition authority of a Member State may in its own territory carry out any
inspection or other fact-finding measure under its national law on behalf and for the
account of the competition authority of another Member State in order to establish
42
See also the Joint Statement of the Council and the Commission on the functioning of the
network of competition authorities, Council Doc 15435/02 ADD 1 of 10 Dec 2002.
43
See also Art 35(1), first sentence, of Reg 1/2003: ‘[t]he Member States shall designate the
competition authority or authorities responsible for the application of Articles 81 and 82 of the Treaty
in such a way that the provisions of this regulation are effectively complied with’.
44
It does not follow that, in those Member States that have not (yet) provided for a power to
inspect private homes in their legislation, the competition authorities could today claim to have such
power by virtue of (the primacy of) Community law. This would be contrary to the principle of legal
certainty, which is a general principle of Community law: see Judgments of the ECJ of 7 Jan 2004 in
Case C–60/02 X [2004] ECR I–651 at para 61 and of 3 May 2005 in Joined Cases C–387/02 etc
Berlusconi [2005] ECR I–3565 at para 74. But the Commission could bring an action against the
Member State before the ECJ under Art 226 EC, and if necessary subsequently under Art 228 EC, to
make the Member State enact the necessary legislation.
whether there has been an infringement of Article 81 or Article 82 of the Treaty. Any
exchange and use of the information collected shall be carried out in accordance with
Article 12.45
42. This provision creates a possibility for national competition authorities to
use their own powers of investigation to help each other, not an obligation to do
so, unlike Article 22(2) of Regulation 1/2003 which obliges national competition
authorities to undertake inspections on behalf of the Commission at the latter’s
request.46 However, it can be argued that the obligation of loyalty of the Member
States towards the Community under Article 10 EC, read together with Regula-
tion 1/2003, would not be respected by a Member State whose competition
authority unreasonably refused reasonable requests for assistance from other
national competition authorities.47
43. Even if this possibility is not expressly mentioned in Regulation 1/2003,
nothing prevents the Commission from using its powers of investigation on
behalf of a national competition authority. The transmission of the information
thus collected and its subsequent use by the national competition authority
would again be governed by Article 12 of Regulation 1/2003.48
1.2 PROCEDUR AL RIG HTS AND GUAR ANTEES
1.2.1 Procedural Rights and Guarantees Circumscribing or Limiting Powers

of Investigation
44. A number of procedural rights and guarantees circumscribe or limit the

competition authorities’ (use of) powers of investigation. To give just a few
examples: Article 18 of Regulation 1/2003 stipulates that the Commission can
make requests for information only ‘[i]n order to carry out the duties assigned to
it by this Regulation’, and that the request or decision ‘shall stipulate the legal
basis and purpose of the request’. The Court of Justice has clarified that this
implies that the request must identify, ‘with reasonable precision’, the suspected
infringement of Article 81 or 82 EC, and that it can be made only if ‘the
Commission could reasonably suppose, at the time of the request, that the
document [or other information requested] would help it to determine whether
45
On Art 12 of Reg 1/2003, see paras 67–75 below.
46
See above, para 24 above; see also above, paras 12, 15, 19, 20 and 26 on the obligations for
national (competition) authorities to assist the Commission under Arts 18(6) and 20(5) and (6) of
Reg 1/2003.
47
Respect for this obligation could be ensured by actions before the ECJ against the uncoopera-
tive Member State either by the Commission under Art 226 EC or by other Member States under Art
227 EC, and, if necessary, subsequently by the Commission under Art 228 EC.
48
On Art 12 of Reg 1/2003 see sect 1.2.4 below.
PROCEDURAL RIGHTS AND GUARANTEES 13
the alleged infringement had taken place’.49 According to the Orkem judgment of
the Court of Justice,50 as well as recital 23 of Regulation 1/2003, the Commission
cannot use its powers under Article 18 to compel undertakings to admit that they
have committed an infringement of Article 81 or 82 EC, but companies are
obliged to answer factual questions and to provide documents, even if this
information may be used to establish against them the existence of an infringe-
ment. As to inspections in private homes, Article 21(3) of Regulation 1/2003
provides that a Commission decision ordering such inspection ‘cannot be
executed without prior authorisation from the national judicial authority of the
Member State concerned’.
45. These procedural rights and guarantees appear to be based to a large
extent on considerations of principle.51 For instance, in Orkem, the Court of
Justice based its holding that the Commission cannot use requests for informa-
tion to compel undertakings to admit that they have committed an infringement
of Articles 81 and 82 EC on the need to safeguard the rights of the defence.52
Similarly, the Court of Justice based its ruling in AM&S that the Commission
cannot use its powers of investigation to take or to compel the production of
certain lawyer–client communications (legal professional privilege) on the need
to respect the rights of the defence.53 Considerations of principle, such as the
need to safeguard the rights of the defence, thus limit the available use of
investigative powers, possibly leading to less effective or less efficient antitrust
enforcement, as these considerations of principle are given precedence over the
objective of effective and efficient antitrust enforcement.54
46. Apart from considerations of principle, procedural rights and guarantees
may also be based on utilitarian or instrumental considerations, in that they are
49
Opinion of AG Jacobs of 15 Dec 1993 in Case C–36/92 P SEP v Commission [1994] ECR I–1914
(explicitly endorsed by the ECJ in its judgment of 19 May 1994 in the same case at paras 30 and 21.
50
Judgment of 18 Oct 1989 in Case 374/87 Orkem v Commission [1989] ECR 3343, paras 35–40.
51
I am using here the term ‘principle’, as opposed to ‘policy’, in the sense given by Ronald
Dworkin, who in his book Taking Rights Seriously (Princeton, NJ, Harvard University Press, 1978),
defines ‘policy’ as the kind of standard ‘that sets out a goal to be reached, generally an improvement in
some economic, political, or social feature of the community’ and ‘principle’ as a standard ‘that is to
be observed, not because it will advance or secure an economic, political, or social situation deemed
desirable, but because it is a requirement of justice or fairness or some other dimension of morality’.
52
Above n 50, paras 32–34.
53
Judgment of 18 May 1982 in Case 155/79 AM&S v Commission [1982] ECR 1575, in particular
paras 18, 20, 21 and 23; for a thorough analysis of this judgment, and more generally of legal
professional privilege, see E Gippini-Fournier, ‘Legal Professional Privilege in Competition Proceed-
ings Before the European Commission: Beyond the Cursory Glance’ in BE Hawk (ed), 2004 Annual
Proceedings of the Fordham Corporate Law Institute—International Antitrust Law and Policy (Hunting-
ton, NY, Juris Publishing, 2005) 587.
54
In the language of Ronald Dworkin’s Taking Rights Seriously, above n 51, individual rights thus
‘trump’ utilitarian concerns. As argued by RH Pildes, ‘Why Rights Are Not Trumps: Social Meanings,
Expressive Harms, and Constitutionalism’ (1998) 27 Journal of Legal Studies 725 at 729, the actual
practice of courts’ protection of fundamental rights shows that rights are however ‘not general trumps
against appeals to the common good or anything else; instead, they are better understood as
channelling the kinds of reasons government can invoke when it acts in certain arenas’.
considered to enhance the effectiveness and efficiency of antitrust enforcement.

For instance, the privilege against self-incrimination could at least in part be
explained by concerns relating to the reliability of the information collected.55 It
is also often claimed that legal professional privilege would lead to increased
compliance with the law.56
47. The remainder of this chapter does not attempt to list all the existing
procedural rights and guarantees circumscribing or limiting the powers of
investigation of the Commission and the national competition authorities,57 but
to describe generally which law governs these matters (EC or EU law, national
law, or the European Convention on Human Rights), and by whom or how the
content of this law is determined (by European and national legislation, and case
law of the European Court of Justice and Court of First Instance, national courts,
and the European Court of Human Rights). A distinction is made between three
situations: first, that where information is collected and used by the Commission;
secondly, that where information is collected and used by the same national
competition authority; and, thirdly, that where information is exchanged through
the European Competition Network58 to be used by a different competition
authority from the one that collected it.
1.2.2 Information Collected and Used by the Commission
48. As to information collected and used by the Commission, the applicable

procedural rights and guarantees are essentially those provided for in Regulation
1/2003, in Commission Regulation 773/2004, and in the case law of the Commu-
nity Courts.
49. A number of procedural rights and guarantees are provided for in
Regulation 1/2003. For instance, as already mentioned above, Article 18 stipulates
that the Commission can make requests for information only ‘[i]n order to carry
out the duties assigned to it by this Regulation’ and that the request or decision
‘shall stipulate the legal basis and purpose of the request’. As to inspections in
private homes, Article 21(3) provides that a Commission decision ordering such
55
See Wils, above n 1, paras 545–548, and Vesterdorf, above n 12, at 703.
56
This claim is however highly questionable: see L Kaplow and S Shavell, ‘Legal Advice About
Information to Present in Litigation: Its Effects and Social Desirability’ (1989) 102 Harvard Law
Review 565; L Kaplow and S Shavell, ‘Legal Advice About Acts Already Committed’ (1990) 10
International Review of Law and Economics 149; S Shavell, ‘Legal Advice’ in P Newman (ed), New
Palgrave Dictionary of Economics and the Law (Basingstoke, Palgrave Macmillan, 1998), ii, 516–520; E
Gippini-Fournier, above n 53, atb596–606; and I Baum, ‘The Corporate Lawyer Client Privilege: A
Comparative Law Analysis from Law & Economics Perspective’ [2006] German Working Papers in Law
and Economics, Paper 13.
57
See further K Dekeyser and C Gauer, ‘The New Enforcement System for Articles 81 and 82 EC
and the Rights of Defence’ in BE Hawk (ed), 2004 Annual Proceedings of the Fordham Corporate Law
Institute—International Antitrust Law and Policy (Huntington, NY, Juris Publishing, 2005) 549–85;
and the works referred to in n 1 above.
58
See above, nn 2 and 11, and para 1.
inspection ‘cannot be executed without prior authorisation from the national

judicial authority of the Member State concerned’.
50. Some further procedural rights and guarantees are set out in Regulation
773/2004, the Commission’s implementing regulation based on Article 33 of
Regulation 1/2003.59 For instance, Article 4(3) of Regulation 773/2004 provides
that:
In cases where [pursuant to Article 20(2)(e) of Regulation 1/2003] a member of staff of
an undertaking who is not or was not authorised by the undertaking or by the
association of undertakings to provide explanations on behalf of the undertaking or
association of undertakings has been asked for explanations, the Commission shall set a
time-limit within which the undertaking or the association of undertakings may
communicate to the Commission any rectification, amendment or supplement to the
explanations given by such member of staff. The rectification, amendment or supple-
ment shall be added to the explanations as recorded.
51. The case law of the Community Courts is clearly very important in deter-
mining the procedural rights and guarantees applicable to information collected
and used by the Commission. First, the Community Courts interpret, and thus
clarify the full meaning of, the regulatory provisions. For instance, the provisions
of Article 18 of Regulation 1/2003, which stipulate that the Commission can
make requests for information only ‘[i]n order to carry out the duties assigned to
it by this Regulation’ and that the request or decision ‘shall stipulate the legal
basis and purpose of the request’, have been interpreted by the Court of Justice as
implying that the request must identify, ‘with reasonable precision’, the suspected
infringement of Article 81 or 82 EC, and that it can be made only if ‘the
Commission could reasonably suppose, at the time of the request, that the
document [or other information requested] would help it to determine whether
the alleged infringement had taken place’.60
52. Secondly, the Community Courts find procedural rights and guarantees
on the basis of general principles of Community law. According to the case law of
the Court of Justice, fundamental rights are part of the general principles of
Community law, and the Court of Justice and the Court of First Instance thus
ensure respect for these rights on the basis of Article 220 EC.61 Examples already
mentioned above are the Orkem rule that the Commission cannot use requests
for information to compel undertakings to admit that they have committed an
infringement of Articles 81 and 82 EC, and the AM&S rule that the Commission
certain lawyer–client communications (legal professional privilege).62
59
Commission Reg 773/2004 of 7 April 2004 relating to the conduct of proceedings by the
Commission pursuant to Arts 81 and 82 of the EC Treaty [2004] OJ L123/18.
60
See above n 49.
61
Judgment of 12 Nov 1969 in Case 29/69 Stauder v Stadt Ulm [1969] ECR 419, para 7.
62
See above nn 50 and 53 and paras 41 and 45.
53. In defining and applying fundamental rights, the Community Courts

draw from the constitutional traditions common to the Member States and from
international treaties on which Member States have collaborated and to which
they are signatories. In that respect the Community Courts have accorded
particular significance to the European Convention on Human Rights.63
54. The core of this case law is now enshrined in Article 6(2) EU:
The Union shall respect fundamental rights, as guaranteed by the European Conven-
tion for the Protection of Human Rights and Fundamental Freedoms signed in Rome
on 4 November 1950 and as they result from the constitutional traditions common to
the Member States, as general principles of Community law.
55. The Court of Justice has also made it clear in its case law that, in thus
ensuring respect for the fundamental rights laid down in the European Conven-
tion on Human Rights, the Community Courts must take into account the case
law of the European Court of Human Rights.64 For instance, in Hüls the Court of
Justice referred to two judgments of the European Court of Human Rights
concerning the notion of ‘criminal’ within the meaning of the European Conven-
tion on Human Rights (ECHR) to conclude that the presumption of innocence
as set out in Article 6(2) ECHR applies to procedures relating to infringements of
Articles 81 and 82 EC that may result in the imposition of fines or periodic
penalty payments.65 In PVC II, concerning the right not to give evidence against
oneself, the Court of Justice confirmed that further developments in the case law
63
Convention for the Protection of Human Rights and Fundamental Freedoms, available at
www.echr.coe.int/Eng/basicTexts.htm; see eg Judgments of the ECJ of 18 Dec 1997 in Case C–309/96
Annibaldi [1997] ECR I–7493 and of 17 Dec 1998 in Case C–185/95 P Baustahlgewebe [1998] ECR
I–8417 and Judgment of the CFI of 20 Feb 2001 in Case T–112/98 Mannesmannröhren-Werke [2001]
ECR II–732.
64
Whereas the ECHR and the case law of the ECtHR thus play an important role in determining
the procedural rights and guarantees applicable to information collected and used by the Commis-
sion, via the case law of the Community Courts, the possibilities for the undertakings concerned to
bring a claim of alleged violation of their fundamental rights directly before the ECHR appear
extremely limited. As the EC or the EU is not a signatory to the ECHR (and could not become so
under the current Treaties; see Opinion 2/94 (Accession to the ECHR) of the ECJ of 28 Mar 1996,
[1996] ECR I–1759; Art I-9 of the signed but unratified Treaty establishing a Constitution for Europe
[2004] OJ C310/1, would make accession of the EU to the ECHR possible and indeed mandatory), it
is not possible to bring an action against the Community or its institutions before the ECtHR. In its
judgment of 30 June 2005 in App 45036/98 Bosphorus v Ireland, the ECtHR has clarified that acts of
Community institutions, to the extent that they are implemented through Member States’ acts (which
is almost always the case; see also above, paras 20–23), are not per se exempted from scrutiny by the
Court when it considers applications against Member States concerning such implementing acts. The
ECtHR, considering that the Community protects fundamental rights in a manner equivalent to the
system of the ECHR, accords a presumption of ECHR compatibility to Community acts and their
implementation by Member States; that presumption can however be rebutted if, in the circum-
stances of the case, it appears that the fundamental rights protection at the Community level were
manifestly deficient; see further S Douglas-Scott, ‘A Tale of Two Courts: Luxembourg, Strasbourg and
the Growing European Human Rights Acquis’ (2006) 43 CML Rev 629.
65
Judgment of 8 July 1999 in Case C–199/92 P Hüls v Commission [1999] ECR I–4383, paras
149–150; see however the Admissibility Decision of the ECtHR of 3 June 2004 in Apps 69042/01 etc
OOO Neste St Petersburg and Others v Russia.
of the European Court of Human Rights since the Orkem judgment must be
taken into account by the Community Courts.66
56. Whereas the European Convention on Human Rights thus plays an
important role in the case law, some judgments of the Community Courts also
mention the Charter of Fundamental Rights of the European Union.67 This
document was ‘solemnly proclaimed’ by the European Parliament, the Council
and the Commission on 7 December 2000.68 The signed but unratified Treaty
establishing a Constitution for Europe comprises this Charter (slightly modified)
and would thus make it legally binding.69 In the absence of such Treaty incorpo-
ration, the legal status of the Charter remains somewhat unclear. Recital 37 of
Regulation 1/2003, however, provides that ‘[t]his Regulation respects the funda-
mental rights and observes the principles recognised in particular by the Charter
of Fundamental Rights of the European Union. Accordingly, this Regulation
should be interpreted and applied with respect to those rights and principles’.70
57. With regard to some procedural rights and guarantees, Regulation 1/2003
codifies the case law of the Community Courts. For instance, Article 20(8), which
defines the role of national courts in granting authorisation for assistance by
national authorities in case of opposition to an inspection ordered by the
Commission, restates the Roquette judgment of the Court of Justice.71 The
second sentence of recital 23 of Regulation 1/2003 restates the Orkem case law
concerning self-incrimination.72
66
Judgment of 15 Oct 2002 in Joined Cases C–238/99 P etc Limburgse Vinyl Maatschappij (LVM)
and Others v Commission [2002] ECR I–8616, para 274. It does not necessarily follow however that the
Orkem judgment would no longer be good law. Indeed, as I have argued in detail elsewhere (Wils,
above n 1, sects 5.1.2 and 5.2.2), it is not at all clear that the ECtHR would grant a wider scope of
protection under the privilege against self-incrimination to legal persons in proceedings such as those
under Reg 1/2003, to the extent that these proceedings can only lead to the imposition of fines on
legal persons, than the protection granted under Orkem. It should also be pointed out that in several
jurisdictions only natural persons, not legal persons, can invoke the privilege against self-
incrimination: see the Judgment of the US Sup Ct in Braswell v United States 487 US 99 (1988) and
the Judgment of the German Const Ct of 26 Feb 1997, 1 BvR 2172/96, and the Opinion of AG
Geelhoed of 19 Jan 2006 in Case C–301/04 P Commission v SGL Carbon [2006] ECR I–5915, paras
62–67, and the Judgment of the ECJ of 29 June 2006 in the same case, paras 43 and 49.
67
See for instance the Judgment of the ECJ of 27 June 2006 in Case C–540/03 European
Parliament v Council of the European Union [2006] ECR I–5769, para 38, concerning the validity of
Dir 2003/86/EC on the right to family reunification. The CFI relied to some extent on the Charter in
its Judgment of 30 Jan 2000 in Case T–54/99 max.mobil v Commission [2002] ECR II–313, para 48,
and in its Judgment of 3 May 2002 in Case T–177/01 Jégo-Quéré v Commission [2002] ECR II–5137,
paras 45–51, but both these judgments were annulled on appeal: by the ECJ; see Judgments of 22 Feb
2005 in Case C–141/02 P Commission v T-Mobile Austria [2005] ECR I–1283 and of 1 Apr 2004 in
Case C–263/02 P Commission v Jégo-Quéré [2004] ECR I–3425.
68
[2000] OJ C364/1.
69
[2004] OJ C310/1; see further L Ortega, ‘Fundamental Rights in the European Constitution’
(2005) 11 European Public Law 363.
70
Cf Judgment of the ECJ of 27 June 2006 in Case C–540/03 European Parliament v Council of the
European Union [2006] ECR I–5769, para 38.
71
Judgment of 22 Oct 2002 in Case C–94/00 Roquette [2002] ECR I–9011.
72
See above n 50 and para 44.
58. Such legislative codification of case law raises the question to what extent
the Community Courts, in later cases where they may be asked by litigants to go
further in granting procedural rights and guarantees than in their earlier case law,
should defer to the legislative codification or to the views expressed in the
legislative process. In my view the answer should be different depending on
whether the proposed extension of procedural rights is based on considerations
of principle or rather on utilitarian or instrumental considerations.73
59. If, for instance, the Community Courts were to come to the view that
fundamental rights of the defence required a broader right not to give evidence
against oneself than recognised under the current Orkem case law,74 the fact that
recital 23 of Regulation 1/2003 reflects the current case law should not prevent
the Community Courts from reaching such a conclusion. Indeed, in the hierarchy
of norms general principles of Community law take precedence over legislative
acts, and it is the Courts’ task to find, interpret and apply these principles.
60. The situation is different, however, when proposed extensions of proce-
dural rights or guarantees are based on utilitarian or instrumental considerations
relating to the effectiveness and efficiency of antitrust enforcement. A topical
example is the question whether the ruling in AM&S that the Commission
certain lawyer–client communications (legal professional privilege)75 should be
extended to cover not only independent lawyers, as is the current state of the law,
but also in-house legal counsel.76 As already mentioned above,77 the Court of
Justice based its ruling in AM&S on the need to respect the rights of the defence.
I personally cannot see how one could consider that the ability to consult in
confidence an independent lawyer would not be sufficient to guarantee the rights
of the defence, thus creating a need to extend legal professional privilege to
in-house counsel. Indeed, there is a wide choice of independent lawyers under-
takings could turn to, and those undertakings which can afford to have in-house
counsel can undoubtedly also afford to pay an independent lawyer. In fact the
extension of legal professional privilege to in-house lawyers may lead to less
protection of the fundamental rights of the defence, in that it may lead large
undertakings to use only in-house counsel, thus reducing the availability of
independent lawyers to the detriment of smaller companies that cannot afford
in-house counsel. Not surprisingly then, a large number of the arguments used
by advocates of extending the AM&S ruling to in-house counsel are not based on
considerations of principle, but rather on utilitarian or instrumental considera-
tions relating to the effectiveness and efficiency of antitrust enforcement. It is in
73
74
This is unlikely to happen, for good reasons: see Opinion of AG Geelhoed of 19 Jan 2006 in
Case C–301/04 P Commission v SGL Carbon [2006] ECR I–5915, paras 62–67, and the Judgment of
the ECJ of 29 June 2006 in the same case, at paras 43 and 49; and above n 66.
75
76
See Cases T–125/03 and T–253/03 Akzo Nobel Chemicals v Commission, pending.
77
particular often argued that legal professional privilege for in-house counsel
would lead to better compliance with Articles 81 and 82 EC.78 These arguments
were also brought up in the legislative process leading to the adoption of
Regulation 1/2003. Indeed, the Economic and Monetary Committee of the
European Parliament initially adopted an amendment providing for the exten-
sion of legal professional privilege to in-house counsel,79 but this amendment,
which was strongly opposed by the Commission, was subsequently rejected in a
plenary session of the Parliament by 404 votes to 69.80 The Council also declined
to include any such extension in Regulation 1/2003. The European Parliament,
Council and Commission thus rejected the arguments that the extension of legal
professional privilege to in-house counsel would be beneficial for the enforce-
ment of Articles 81 and 82 EC. Given that the assessment of what is beneficial for
effective and efficient antitrust enforcement belongs to these three institutions
rather than the Community Courts, this should in my view lead the Community
Courts to dismiss any arguments relating to the effectiveness and efficiency of
antitrust enforcement which may be raised before them in support of claims to
extend legal professional privilege to in-house counsel.81
1.2.3 Information Collected and Used by the Same National Competition

Authority
61. For information collected and used by the same national competition
authority, the applicable procedural rights and guarantees are primarily those
provided for in the legislation of the Member State concerned and the case law of
the national courts of that Member State.
62. In those Member States where criminal sanctions are provided for viola-
tions of Article 81 or 82 EC,82 correspondingly stronger procedural rights and
guarantees will exist.
63. Given that all Member States are signatories to the European Convention
on Human Rights, the procedural rights and guarantees resulting from its
provisions, as interpreted by the European Court of Human Rights, are applica-
ble in all Member States.83
64. Moreover, according to the case law of the Court of Justice, ‘the require-
ments flowing from the protection of fundamental rights in the Community legal
78
See above n 56.
79
Amendment No 10 contained in the Evans Report, EP Session document of 21 June 2001, PE
296.005.
80
Plenary Session of 6 Sept 2001, PE 308.749, at 35.
81
See also Gippini-Fournier, above n 53, at 606.
82
See above, para 32.
83
It is possible for the national courts of a Member State to interpret the ECHRs more generously
than the ECtHR. The interpretation given by the ECtHR functions as a minimum, given that the
undertakings or individuals concerned can bring an action before it after exhaustion of domestic
remedies.
order are also binding on the Member States when they implement Community
rules’.84 Given that national competition authorities when investigating violations
of Articles 81 and 82 EC are implementing Community rules, it follows that they
must respect all procedural rights and guarantees which, according to the case
law of the Court of Justice, flow from the protection of fundamental rights in the
Community legal order.85
65. For example, the AM&S rule that powers of investigation cannot be used
to take or to compel the production of certain communications with independ-
ent lawyers, which the Court of Justice deduced from the need to protect the
rights of the defence,86 thus applies not only to the European Commission but
also to the national competition authorities when investigating violations of
Article 81 or 82 EC.
66. National law could of course provide for wider procedural rights and
guarantees than those flowing from the protection of fundamental rights in the
Community legal order. For instance, in four of the 25 Member States,87 legal
professional privilege covers not only independent lawyers but also in-house
counsel.
1.2.4 Information Exchanged through the European Competition Network
67. When information is collected by one competition authority (the Commis-

sion or a national competition authority) and exchanged through the network to
be used by another competition authority, regard should be had to Article 12(1)
and (3) of Regulation 1/2003.
68. As already mentioned above, Article 12(1) of Regulation 1/2003 reads as
follows:
For the purpose of applying Articles 81 and 82 of the Treaty the Commission and the
competition authorities of the Member States shall have the power to provide one
another with and use in evidence any matter of fact or of law, including confidential
information.
69. This rule is however limited by the following exception in Article 12(3) of
Regulation 1/2003:
84
Judgment of 13 Apr 2000 in Case C–292/97 Karlsson and Others [2000] ECR I–2760, para 37.
85
This conclusion is not contradicted by the ECJ Judgment of 10 Nov 1993 in Case C–60/92 Otto
v Postbank [1993] ECR I–5707, in which the Court held that Community law does not require that
the Orkem rule (see above n 50 and para 44) be applied by a national court in private litigation based
on Art 81 or 82 EC. As is clear from paras 14 and 15 of Otto v Postbank, the Court came to this
conclusion not on the ground that the requirements flowing from the protection of fundamental
rights in the Community legal order would not apply in national proceedings (on the contrary, it
reaffirmed their applicability), but on the ground that the necessary procedural guarantees in civil
litigation are not the same as those in administrative proceedings.
86
See above, para 45.
87
Belgium, Greece, Ireland and the UK.
Information exchanged pursuant to paragraph 1 can only be used in evidence to

impose sanctions on natural persons where:
—the law of the transmitting authority foresees sanctions of a similar kind in relation
to an infringement of Article 81 or Article 82 of the Treaty or, in the absence thereof,
—the information has been collected in a way which respects the same level of
protection of the rights of defence of natural persons as provided for under the national
rules of the receiving authority. However, in this case, the information exchanged
cannot be used by the receiving authority to impose custodial sanctions.88
70. Some additional clarification is provided for in the first, third and fourth
sentences of recital 16 of Regulation 1/2003:
Nothwithstanding any national provision to the contrary, the exchange of information
and the use of such information in evidence should be allowed between the members of
the network even where the information is confidential. … When the information
exchanged is used by the receiving authority to impose sanctions on undertakings,
there should be no limit to the use of the information than the obligation to use it for
the purpose for which it was collected given the fact that the sanctions imposed on
undertakings are of the same type in all systems. The rights of defence enjoyed by
undertakings in the various systems can be considered as sufficiently equivalent.
71. Finally, the three last sentences of paragraph 27 of the Commission Notice
on cooperation within the Network of Competition Authorities89 explain further:
Article 12 of [Regulation 1/2003] takes precedence over any contrary law of a Member
State. The question whether information was gathered in a legal manner by the
transmitting authority is governed on the basis of the law applicable to this authority.
When transmitting information the transmitting authority may inform the receiving
authority whether the gathering of the information was contested or could still be
contested.
72. When information is collected by one competition authority (the transmit-
ting authority) and exchanged through the European Competition Network to be
used by another competition authority (the receiving authority), this informa-
tion can thus always be used as intelligence by the receiving authority. As to its
use in evidence, a distinction is made between the use to impose sanctions on
natural persons and the use to impose sanctions on legal persons. The receiving
authority can use the information in evidence to impose custodial sanctions only
if the law of the transmitting authority provides for such sanctions in relation to
violations of Article 81 or 82 EC. As regards other kinds of sanctions against
natural persons (such as fines or director disqualification), the receiving author-
ity can use the exhanged information in evidence if either the law of the
transmitting authority provides for the same kind of sanctions or the informa-
tion has been collected in a way which respects the same level of protection of the
88
See also the last sentence of Art 22(1) of Reg 1/2003, quoted above in para 41.
89
Above n 2.
rights of the defence of natural persons as provided for under the national rules
of the receiving authority. As regards sanctions on legal persons, the receiving
authority can always use the exchanged information in evidence if it has been
lawfully collected according to the law applicable to the transmitting authority,
even if the receiving authority could not itself have collected it or could not have
used it if it had collected it, in accordance with its own national law.
73. This means for instance that the Office of Fair Trading, which itself under
United Kingdom law could not use its powers of investigation to take or compel
undertakings to give certain correspondence with in-house legal counsel, could
receive and use in evidence such information if collected by the European
Commission or the German or French competition authorities, given that in EC
law and in the national laws of Germany, France, and indeed most other Member
States, legal professional privilege does not cover in-house lawyers.90
74. This ability of a receiving authority to use in evidence exchanged informa-
tion which it could not itself have collected, or could not have used if so collected,
has been criticised by Marcos Araujo as an ‘unacceptable’ ‘circumvention’ of
procedural rights and guarantees ‘that may erode fundamental rights’.91
75. I would consider such fears to be wholly unwarranted. First, the problem
concerns only legal persons, not natural persons.92 Secondly, as also explained
above,93 all members of the European Competition Network must respect the
procedural rights and guarantees flowing from the European Convention on
Human Rights, as interpreted by the European Court of Human Rights, as well as
all procedural rights and guarantees which, according to the case law of the Court
of Justice, flow from the protection of fundamental rights in the Community
legal order.94 The ‘eroded fundamental rights’ can thus only be rights of legal
persons which are recognised neither in the case law of the European Court of
Human Rights nor in the fundamental rights case law of the Court of Justice.
Indeed, it is difficult to find any example other than the already mentioned
example of legal privilege for in-house counsel.95 As argued above, the arguments
for extending legal professional privilege to in-house lawyers appear quite weak.96
90
See above, paras 41, 6- and 66; OFT Guideline 404, above n 39, para 6.3 (‘Whilst UK privilege
rules would apply to cases being investigated in the UK by the OFT on its own behalf, the OFT could
be sent the communications of in-house lawyers … by an NCA from another Member State where the
communication of such lawyers are not privileged. Under those circumstances, the OFT may use the
documentation received from the other NCA in its investigation.’); and Vesterdorf, above n 12, at
721–3.
91
M Araujo, ‘The Respect of Fundamental Rights Within the European Network of Competition
Authorities’ in BE Hawk (ed), 2004 Annual Proceedings of the Fordham Corporate Law Institute—
International Antitrust Law and Policy (Huntington, NY, Juris Publishing, 2005) 511, at 528 and 530.
92
93
See above, paras 52–55 and 63–65.
94
Including eg the Orkem and AM&S case law: see above, paras 44, 45 and 65.
95
See above, para 73. In earlier publications, I have mentioned an example concerning the
privilege against self-incrimination (see, inter alia, Wils, above n 1, para 207); see however the
judgment of the German Const Ct, of 26 Feb 1997 referred to above in n 66.
96
Finally, if an ‘erosion of fundamental rights’ were ever seen to be happening, one

could safely expect the Community Courts and the courts of the Member States
with a lower level of protection to react by increasing procedural rights and
guarantees.97
97
See Wils, above n 1, para 208, n 282, and Vesterdorf, above n 12, at 722–3 (‘Regulation No
1/2003 therefore sets the stage for a comprehensive interpenetration of the national procedural laws.
It must be stressed that it will not necessarily level down the national laws: nothing in Regulation No
1/2003 prevents a Member State from increasing the current level of protection granted in its
jurisdiction for the collection of evidence’).
2
Settlements of Antitrust Investigations:
Commitment Decisions under Article 9
of Regulation No 1/2003
2.1 TEXT, OR IG IN AND OPTIMAL USE
2.1.1 Text
76. Regulation 1/2003 is the main implementing regulation governing the

enforcement of Articles 81 and 82 EC.1 It sets out in particular the powers of the
European Commission for the purpose of applying Articles 81 and 82 EC.
77. Article 9, which is part of Chapter III, ‘Commission decisions’, reads as
follows:
Article 9
Commitments
1. Where the Commission intends to adopt a decision requiring that an infringement
be brought to an end and the undertakings concerned offer commitments to meet
the concerns expressed to them by the Commission in its preliminary assessment,
the Commission may by decision make those commitments binding on the
undertakings. Such a decision may be adopted for a specified period and shall
conclude that there are no longer grounds for action by the Commission.
2. The Commission may, upon request or on its own initiative, reopen the proceed-
ings:
1
OJ L269/1. According to its Arts 43 and 45, this reg replaces Council Reg 17 [1962] JO 13/204,
[1959–62] OJ Spec Ed 87, last amended by Council Reg 1216/1999 [1999] OJ L148/5, with effect from
1 May 2004. On Reg 1/2003 generally see L Idot, Droit communautaire de la concurrence—Le nouveau
système communautaire de mise en oeuvre des articles 81 et 82 CE (Brussels, Bruylant, 2004); D Geradin
(ed), Modernisation and Enlargement: Two Major Challenges for EC Competition Law (Antwerp,
Intersentia, 2004); D Cahill and JD Cooke (eds), The Modernisation of EU Competition Law
Enforcement, Reports for the XXI FIDE Congress, Dublin 2004 (Cambridge, Cambridge University
Press, 2004); WPJ Wils, Principles of European Antitrust Enforcement (Oxford, Hart Publishing, 2005);
CS Kerse and N Khan, EC Antitrust Procedure, 5th edn (London, Sweet & Maxwell, 2005); and L Ortiz
Blanco (ed), EC Competition Procedure, 2nd edn (Oxford, Oxford University Press, 2006).
26 SETTLEMENTS OF ANTITRUST INVESTIGATIONS
(a) where there has been a material change in any of the facts on which the decision was
based;
(b) where the undertakings concerned act contrary to their commitments;
(c) where the decision was based on incomplete, incorrect or misleading information
provided by the parties.
78. Article 23(2)(c) of Regulation 1/2003 provides that the Commission ‘may by
decision impose fines [not exceeding 10 per cent of the total turnover in the
preceding business year] on undertakings … where, intentionally or negligently
… they fail to comply with a commitment made binding by a decision pursuant
to Article 9’.
79. Article 24(1)(c) provides that the Commission ‘may, by decision, impose
on undertakings … periodic penalty payments not exceeding 5 % of the average
daily turnover in the preceding business year per day and calculated from the
date appointed by the decision, in order to compel them … to comply with a
commitment made binding by a decision pursuant to Article 9’.
80. Recital 13 of Regulation 1/2003—the recital directly corresponding to
Article 9—reads as follows:
Where, in the course of proceedings which might lead to an agreement or practice
being prohibited, undertakings offer the Commission commitments such as to meet its
concerns, the Commission should be able to adopt decisions which make those
commitments binding on the undertakings concerned. Commitment decisions should
find that there are no longer grounds for action by the Commission without concluding
whether or not there has been or still is an infringement. Commitment decisions are
without prejudice to the powers of the competition authorities and courts of the
Member States to make such a finding and decide upon the case. Commitment
decisions are not appropriate in cases where the Commission intends to impose a fine.
81. The last sentence of recital 22—the recital corresponding to Article 16,
which deals with the impact of Commission decisions on national courts and
competition authorities—repeats:
Commitment decisions adopted by the Commission do not affect the power of the
courts and the competition authorities of the Member States to apply Articles 81 and 82
of the Treaty.
2.1.2 Origin
82. Regulation 17,2 which governed the enforcement of Articles 81 and 82 EC
from 1962 until its replacement as from 1 May 2004 by Regulation 1/2003, did
not contain a provision similar to Article 9 of Regulation 1/2003. Settlements of
investigations by the Commission into suspected infringements of Article 81 or
82 EC did however occur.
2
See above n 1.
TEXT, ORIGIN AND OPTIMAL USE 27
83. As regards Article 82 EC, the general enforcement system under Regula-
tion 17 was not very different from the current system under Regulation 1/2003.
The Commission would normally start an investigation only if it suspected an
infringement, and with a view to adopting a decision finding such infringement,
ordering its termination and/or imposing a fine. In a number of cases, however,
including some well-known ones, such as the IBM case,3 the Commission did not
proceed up to the adoption of a decision, but rather closed the case after the
undertaking concerned had offered an agreed remedy.
84. Even if Regulation 17 was silent in this respect, there did not appear to be
any legal obstacle to such informal settlements. Indeed, the Commission had
under Regulation 17—as it still has today under Regulation 1/2003—a broad
discretion as to which infringements it chooses to pursue, depending on its
enforcement priorities.4 This discretion logically also entails the discretion to
close a case under investigation where the Commission considers that the interest
in further pursuing the case has disappeared as a result of an agreed remedy
having been offered by the undertaking concerned.
85. Under the silence of Regulation 17, such informal settlements had only
those legal consequences that flow from general principles of Community law. In
application of the principle of protection of legitimate expectations,5 the infor-
mal settlement would normally prevent the Commission from reopening the
case, except on the basis of a material change in the facts, or if the undertaking
did not respect its commitments, or if the undertaking had misled the Commis-
sion during the settlement talks.6 This broadly corresponds to what is now
explicitly provided for in Article 9(2) of Regulation 1/2003.
86. The practice of informal settlements under Regulation 17 was however
perceived to have at least two defects. First, apart from (using the threat of)
reopening the case, the Commission did not have any means to ensure that the
undertaking respected the commitments it had offered and on the basis of which
the case had been closed. Secondly, the informal settlement practice lacked
transparency towards directly interested third parties (who might benefit from
the commitments and whose vigilance might be of help in monitoring the
respect of the commitments) as well as towards the wider business and legal
community (who might learn from the precedent). Settlements were usually
3
Commission’s XIVth Report on Competition Policy (Brussles, EC Commission, 1984), pts 94–95.
For a list of further examples see R Whish, Competition Law, 5th edn (London, Lexis Nexis
Butterworth, 2003) at 210.
4
See Judgment of the CFI of 18 Sept 1992 in Case T–24/90 Automec II [1992] ECR II–2250, and
Commission Notice on the handling of complaints by the Commission under Arts 81 and 82 of the
EC Treaty [2004] OJ C101/65.
5
See KPE Lasok and T Millett, Judicial Control in the EU: Procedures and Principles (Richmond,
Richmond Law & Tax, 2004) at 353–64.
6
The exact scope of the legitimate expectations would of course depend on the specific assurances
given or not given by the Commission, in particular the wording of the letter or other document
informing the undertaking of the closure of the case; compare with Judgment of the CFI of 17 Feb
2000 in Case T–241/97 Stork Amsterdam [2000] ECR II–309.
described in the Commission’s annual Report on Competition Policy, but often

without much detail. Only in one case, SWIFT, were the commitments published
in the Official Journal.7
87. The introduction of Article 9 into Regulation 1/2003 addresses these
defects. First, Articles 23(2)(c) and 24(1)(c) of Regulation 1/2003 allow the
Commission to impose fines and periodic penalty payments in the event of
failure to comply with commitments made binding pursuant to Article 9.8
Secondly, decisions pursuant to Article 9, like all other decisions provided for in
Chapter III of Regulation 1/2003, must be published in accordance with Article
30.9
88. As regards Article 81 EC, the general enforcement system under Regula-
tion 17 was quite different from the current system. Regulation 17 provided for
an opportunity for undertakings to notify their agreements to the Commission
so as to obtain a negative clearance or exemption decision, and the Commission
was in principle under an obligation to act upon such notifications.10 In cases
where the Commission had started an investigation with a view to finding an
infringement, ordering its termination and/or imposing fines, but where the
undertakings concerned had offered an agreed remedy, the settlement could thus
be formalised through an exemption decision. Article 8(1) of Regulation 17
explicitly provided for the ability to attach conditions and obligations to exemp-
tion decisions, and Article 15(2)(a) allowed the Commission to impose fines in
case of failure o respect such obligations.
89. It is interesting to note that, nothwithstanding this availability of a formal
instrument to close cases, the Court of Justice also accepted the legality of the
Commission’s practice of closing cases informally by way of so-called ‘comfort
letters’.11
90. Finally, Article 9 of Regulation 1/2003 is also inspired by the US example
of the consent decrees or consent orders through which investigations by the
Department of Justice or the Federal Trade Commission can be closed.12
7
[1997] OJ C335/3.
8
See below, para 132.
9
10
Judgment of the CFI of 22 Oct 1997 in Joined Cases T–213/95 and T–18/96 SCK and FNK v
Commission [1997] ECR II–1746; see further WPJ Wils, The Optimal Enforcement of EC Antitrust Law
(Kluwer Law International, 2002) at 111–13.
11
Judgment of the ECJ of 10 July 1980 in Case 37/79 Marty v Lauder [1980] ECR 2481.
12
See DP Ducor, ‘Settlement of Competition Conduct Violations at the United States Antitrust
Agencies and at the European Commission—Some Observations’ and JR Atwood, ‘Observations on
Negotiating Government Antitrust Settlements in the United States’, both in BE Hawk (ed), Annual
Proceedings of the Fordham Corporate Law Institute 2005 (Huntington, NY, Juris, 2006), and Wils,
above n 10, at 160–1.
2.1.3 Optimal Use
91. As is apparent from the first part of the text of Article 9, commitment
decisions are conceived as substitutes for infringement decisions under Article 7
of Regulation 1/2003.
92. An infringement decision under Article 7 of Regulation 1/2003 normally
contains a finding of an infringement of Article 81 or 82 EC and an order to
bring the infringement to an end, which may include any behavioural or
structural remedies which are proportionate to the infringement found and
necessary to bring it effectively to an end.13 In the same decision the Commission
may also, on the basis of Article 23(2)(a) of Regulation 1/2003, impose fines.
93. Such infringement decisions can thus contribute in six ways to the
enforcement of Articles 81 and 82 EC:14
(1) the finding of the infringement can contribute to clarifying the content of
the prohibitions laid down in Article 81 or 82 EC, thus allowing better
respect for these prohibitions in the future;
(2) the termination order helps bringing the infringement to an end;
(3) especially if accompanied by a fine, the finding of the infringement acts as a
public censure, which—apart from contributing to the deterrence and
punishment effects mentioned under (4) and (5) below—tends to confirm
those who already respect the prohibitions of Articles 81 and 82 EC in their
law-abiding attitude;
(4) the imposition of fines, if sufficiently high, will contribute to deterring both
the undertakings concerned and others from committing future infringe-
ments;
(5) to the extent that the fines equal the gain obtained through the infringement,
their imposition leads to disgorgement of the illicit gains; to the extent that
the fines exceed this gain, they may be valued as punishment;
(6) the finding of the infringement may facilitate follow-on private actions for
damages, leading to compensation for the victims of the infringement.
94. Whereas infringement decisions can thus contribute much to the enforce-
ment of Articles 81 and 82 EC, these results are usually obtained only at the end
of relatively long and costly procedures. Indeed, the Commission can only adopt
an infringement decision after having fully heard the undertakings concerned,
and the undertakings can bring an application for annulment of the decision
before the Court of First Instance, as well as a further appeal, limited to points of
law, before the Court of Justice.
13
According to the last sentence of Art 7(1) and the second sentence of rec 11 of Reg 1/2003, an
infringement decision can also consist only of a finding of a past infringement, if the Commission has
a legitimate interest in making such a finding.
14
For a systematic discussion of the objectives of the enforcement of Arts 81 and 82 EC and the
different ways in which these objectives can be reached see sect 3.1.1 below.
95. For instance, in the Irish Ice Cream case, the Commission in March 1998,
after an administrative procedure which had taken several years, adopted a
decision finding that the practice of Unilever’s Irish subsidiary in providing its
distributors with freezer cabinets for free but under the condition that they could
only be used to store Unilever’s ice cream products, constituted, in the circum-
stances of the Irish market at the time, an infringement of Articles 81 and 82 EC,
and ordered Unilever to put an end to this practice. Unilever brought an
application for annulment of the decision before the Court of First Instance and
obtained from its President an order staying the execution of the Commission’s
decision while the case was pending. The order to terminate the infringement
thus became effective only after the Court of First Instance rejected the applica-
tion for annulment in October 2003.15
96. To the extent that the case requires a clarification of the law—as was
probably the case in Irish Ice Cream—it may very well be unavoidable to go
through these relatively long and costly procedures, as an authoritative clarifica-
tion of the law can only be obtained by bringing the case before the courts.
97. In other cases, however, where the main concern is to bring a current
infringement to an end, it would appear to be in the public interest to obtain this
result in a quicker way, so that the infringement can no longer cause its effects
and that the resources which both the authorities and the companies concerned
would spend on longer proceedings are saved. Commitment decisions as pro-
vided for in Article 9 of Regulation 1/2003 are the instrument which makes this
possible.
98. Optimally, commitment decisions should thus be used instead of
infringement decisions (only) in those cases where the benefit in terms of an
earlier termination of the infringement and the saving of the cost of longer
proceedings outweigh the benefit of the other contributions to the enforcement
of Articles 81 and 82 EC which infringement decisions could make, in terms of
clarification of the law, public censure, deterrence, disgorgement of illicit gains
and punishment, and facilitation of follow-on actions for compensation.
99. Having thus established the optimal use of commitment decisions, the
next question is whether one can assume that settlement decisions will be used
when optimal, and only in those cases, or whether there are any discernible risks
of insufficient or excessive use.
100. Given that both the Commission and the undertakings concerned
directly benefit from the cost savings resulting from the avoidance of the longer
infringement proceedings, there would not appear to be much risk of insufficient
use of commitment decisions. On the other hand, one could identify three
possible sources of risk of excessive use of commitment decisions.
15
Unilever subsequently brought an appeal against the judgment of the CFI, which was rejected
by the ECJ by Order of 28 Sept 2006: Case C–552/03 P Unilever v Commission, not yet reported, but
Unilever did not seek a further suspension of the execution of the termination order while its appeal
before the ECJ was pending.
101. The first source of risk of more than optimal use of commitment
decisions relates to the fact that most of the benefits of infringement decisions for
the enforcement of Articles 81 and 82 EC, namely the public censure, deterrence,
disgorgement of illicit gains and punishment, and facilitation of follow-on
actions for compensation, are strictly negative from the perspective of the
undertakings that are the addressees of the infringement decisions.16 The under-
takings concerned will thus have a systematic bias in favour of commitment
decisions rather than infringement decisions.
102. In order to ensure that the benefits of infringement decisions in terms of
clarification of the law, public censure, deterrence, disgorgement of illicit gains
and punishment and facilitation of follow-on actions for compensation are
adequately weighed in the choice between commitment decisions and infringe-
ment decisions, and that commitment decisions are thus not used more often
than optimal, it is essential that the undertakings concerned are not given any
right to a commitment decision. The Commission should instead have very
broad discretion to choose between commitment decisions and infringement
decisions, allowing it to weigh up all the benefits and disadvantages of the two
courses of action. The text of Article 9 of Regulation 1/2003 supports such a
discretion, in that it states that the Commission ‘may’ adopt a commitment
decision where the undertakings concerned offer commitments. The last sentence
of recital 13 of Regulation 1/2003, according to which ‘[c]ommitment decisions
are not appropriate in cases where the Commission intends to impose a fine’, also
reflects the concern for the benefits of infringement decisions to be weighed by
the Commission before it agrees to close a case by way of a commitment decision.
103. The second risk of undesirable use of commitment decisions relates to
the abolition of the notification system. As already mentioned above, Regulation
17 provided for the ability of undertakings to notify their agreements to the
Commission so as to obtain a negative clearance or exemption decision, to which
conditions and obligations could be attached. This notification system was
abolished by Regulation 1/2003. One of the reasons for this reform was that the
notification system distorted the Commission’s enforcement priorities, making it
spend its resources on notifications, which tended to involve only minor infringe-
ments, if any, instead of detecting and pursuing the most serious infringements
of Articles 81 and 82 EC.17 For the same reason, it would be inopportune if the
mechanism of settlement decisions could be used to make the Commission
spend its resources on cases which involved only minor, if any, infringements.
The text of Article 9 of Regulation 1/2003 reflects this concern, in that it restricts
the use of commitment decisions to cases ‘[w]here the Commission intends to
adopt a decision requiring that an infringement be brought to an end’, and only
on the basis of ‘concerns expressed … by the Commission in its preliminary
16
Only the clarification of the law may also be beneficial for the undertakings concerned.
17
For a detailed analysis of this and the other reasons for abolishing the notification system see
Wils, above n 1, at 3–21.
assessment’. It is thus ensured that commitment decisions are only used in cases
which involve a sufficiently serious (suspected) infringement, justifying in prin-
ciple the adoption of an infringement decision by the Commission.
104. The third risk of excessive use of commitment decisions relates to the
possible temptation for competition authorities, or their staff, to try to obtain
desired results beyond the scope of their legal powers. With regard to the use of
consent decrees in the USA, Douglas Melamed observed in 1995 that these
‘enable government officials to address issues that are politically or economically
important but legally ambiguous’, and that ‘some consent decree remedies appear
to go beyond what the government could realistically anticipate as a remedy in a
contested case’.18 In a system governed by the rule of law, it is important that
public authorities do not act beyond their legal powers, however useful that
action may otherwise also appear. Regulation 1/2003, which is based on Article 83
EC, only grants powers to the Commission to implement the prohibitions laid
down in Articles 81 and 82 EC. Commitment decisions should thus only be used
for commitments that are proportionate and necessary to bring effectively to an
end a suspected infringement of Article 81 or 82 EC, ie the type of remedies
which the Commission would be able to impose if it proceeded to adopt an
infringement decision.
105. In the case of infringement decisions, the (frequently used) possibility of
bringing an application for annulment of the decision before the Court of First
Instance guarantees that no remedies are imposed that go beyond what is
proportional and necessary to bring the infringement of Article 81 or 82 EC
effectively to an end. In the case of commitment decisions, this control mecha-
nism is in practice removed. Indeed, the saving of the cost and delay of judicial
review is precisely one of the benefits justifying the use of commitment decisions.
The requirement that the undertakings concerned must consent to the adoption
of the commitment decision, which can only include commitments offered by
the undertakings, acts instead as guarantee that no disproportionate or unneces-
sary remedies are imposed. This guarantee is however an imperfect one, to the
extent that, faced with the alternative of long and costly infringement proceed-
ings, in particular risk-averse undertakings may at the margin still be prepared to
offer commitments that go beyond what is proportional and necessary to enforce
Articles 81 and 82 EC. The risk of this happening could be reduced to some
extent by ensuring that the undertakings’ consent is informed consent, by giving
the undertakings the right of access to the Commission’s file,19 but some risk will
remain. It thus remains important for the Commission to exercise some self-
restraint in this respect. In the USA, Douglas Melamed suggested in 1995 that ‘the
antitrust agencies could announce and adhere to a policy of seeking consent
decrees only if they have decided to litigate in the event that the defendant balks,
18
AD Melamed, ‘Antitrust: The New Regulation’ (Fall 1995) 10 Antitrust 13, at 13–14.
19
PROCEDURE 33
and to seek by consent decree only remedies that they believe they would obtain
if the case were litigated. Second, the agencies should consider a policy of refusing
to enter into a consent decree if the complaint alleges conduct that, if proven,
would not fairly certainly be illegal; in other words, the agencies could endeavor
to make new law or advance new theories only by litigation’.20
2.2 PROCEDURE
2.2.1 Initiation of Proceedings and Preliminary Assessment
106. According to the text of Article 9 of Regulation 1/2003, commitment

decisions require that ‘the Commission intends to adopt a decision requiring that
an infringement be brought to an end’, and that the undertakings offer commit-
ments ‘to meet the concerns expressed to them by the Commission in its
preliminary assessment’. Article 2(1) of Regulation 773/2004 clarifies that this
requires that proceedings have been initiated.21 The last sentence of recital 13 of
Regulation 1/2003, which states that ‘[c]ommitment decisions are not appropri-
ate in cases where the Commission intends to impose a fine’, further indicates
that Article 9 proceedings are entered into at a stage where the Commission has
already made up its mind about whether or not the infringement warrants the
imposition of a fine. All this indicates that the ‘preliminary assessment’ follows an
investigation of the case, sufficiently serious to have allowed the Commission to
take a preliminary position about the existence of an infringement and about the
likely imposition of fines, and sufficiently detailed to serve as a benchmark
against which to evaluate any commitment proposals.
107. In cases where the Commission has already sent a statement of objec-
tions, as it is required to do under Article 27(1) of Regulation 1/2003 before
taking an infringement decision pursuant to Article 7, this statement of objec-
tions could function as preliminary assessment. This is however not the only
possibility. Indeed, a preliminary assessment could be adopted and communi-
cated to the undertakings concerned in a different form, shorter and less detailed
than a statement of objections. The preliminary assessment in any event consti-
tutes a formal act of the Commission,22 which must be adopted either by the
20
Melamed, above n 18, at 14–15.
21
Commission Reg 773/2004 of 7 Apr 2004 relating to the conduct proceedings by the Commis-
sion pursuant to Arts 81 and 82 of the EC Treaty [2004] OJ L123/18.
22
The fact that the preliminary assessment constitutes an act of the Commission does however
not imply that it can be subject to an action for annulment before the CFI. Indeed, as the ECJ has held
in its judgment of 11 Nov 1981 in Case 60/81 IBM v Commission [1981] ECR 2639, concerning the
possibility of bringing an action for annulment against a statement of objections, ‘it is clear from the
case-law that in principle an act is open to review only if it is a measure definitively laying down the
position of the Commission … on the conclusion of that procedure, and not a provisional measure to
pave the way for a final decision’ (point 10 of the judgment).
Commission (College) itself or, on the basis of an empowerment, by one or more

of its Members in the name of the Commission and under the Commission’s
responsibility.23
2.2.2 Access to the File
108. As has been pointed out recently by Richard Whish, whether there is a right
of access to the file in Article 9 proceedings constitutes one of the ‘unanswered
questions’ concerning commitment decisions.24
109. Article 15(1) of Regulation 773/2004 provides that, ‘[i]f so requested, the
Commission shall grant access to the file to the parties to whom it has addressed
a statement of objections’. There is thus clearly a right of access to the file in those
Article 9 proceedings in which the preliminary assessment takes the form of a
statement of objections. But what if the preliminary assessment does not take this
form?
110. The Commission’s recent Notice on access to the file does not answer the
question. Indeed, according to its first paragraph, it deals with access to the file
only in proceedings leading to decisions on the basis of Articles 7, 8, 23 and 24(2)
of Regulation 1/2003, and ‘does not cover the possibility of the provision of
documents in the context of other proceedings’.25
111. According to consistent case law of the Community Courts, ‘[a]ccess to
the file is … one of the procedural guarantees intended to protect the rights of
defence and to ensure, in particular that the right to be heard … can be exercised
effectively’,26 and ‘observance of the rights of defence constitutes a fundamental
principle of Community law … and must therefore be observed prior to the
adoption of any decision likely to have an adverse effect on the undertakings
concerned’.27
23
The Commission has currently empowered its Member in charge of Competition matters to
determine and issue the preliminary assessment within the meaning of Art 9 of Reg 1/2003. As with
all other empowerments, its exercise requires the prior approval of the Commission’s Legal Service.
The same Member is also empowered to determine and issue statements of objections, but the
exercise of this latter empowerment further requires the agreement of the President of the Commis-
sion.
24
R Whish, ‘Commitment Decisions under Article 9 of the EC Modernisation Regulation: Some
Unanswered Questions’ in M Johansson, N Wahl and U Bernitz (eds), Liber Amicorum in Honour of
Sven Norberg—A European for All Seasons (Brussels, Bruylant, 2007).
25
Commission Notice on the rules for access to the Commission file in cases pursuant to Arts 81
and 82 EC, Arts 53, 54 and 57 of the EEA Agreement and Council Reg 139/2004 [2005] OJ C325/7.
26
Judgment of the CFI of 15 Mar 2000 in Joined Cases T–25/95 etc Cimenteries CBR and Others v
Commission [2000] ECR II–508, para 142.
27
Judgment of the CFI of 4 Mar 1999 in Case T–87/96 Assicurazioni Generali and Unicredito v
Commission [1999] ECR II–206, para 88.
PROCEDURE 35
112. Article 41 of the Charter of Fundamental Rights of the European

Union,28 entitled ‘Right to good administration’, reaffirms the right of access to
the file in the following terms:
Every person has the right to have his or her affairs handled impartially, fairly and
within a reasonable time by the institutions and bodies of the Union.
This right includes:
(a) the right of every person to be heard, before any individual measure which would
affect him or her adversely is taken;
(b)the right of every person to have access to his or her file, while respecting the
legitimate interests of confidentiality and of professional and business secrecy;
….
113. Commitment decisions pursuant to Article 9 of Regulation 1/2003 clearly
have an adverse effect on the undertakings concerned, as they make the commit-
ments binding on them, and expose them to the risk of fines and periodic penalty
payments in case of failure to comply with the commitments.
114. The fact that the undertakings concerned assent to the commitment
decision by offering the commitments does not alter this. The voluntary element
only relates to the choice between an infringement decision pursuant to Article 7
of Regulation 1/2003 and a commitment decision pursuant to Article 9. Indeed,
as is clear from the text of Article 9, commitments are only offered ‘[w]here the
Commission intends to adopt a decision requiring that an infringement be
brought to an end’. The undertaking is thus the object of proceedings which have
been initiated against it, and only has a degree of choice as to two possible final
decisions, both of which have an adverse effect.29
28
This document was ‘solemnly proclaimed’ by the European Parliament, the Council and the
Commission on 7 Dec 2000 [2000] OJ C364/1. The signed but unratified Treaty establishing a
Constitution for Europe comprises this Charter (slightly modified) and would thus make it legally
binding [2004] OJ C310/1; see further L Ortega, ‘Fundamental Rights in the European Constitution’
(2005) 11 European Public Law 363. Without such Treaty incorporation, the legal status of the Charter
remains somewhat unclear. recital 37 of Reg 1/2003, however, provides that ‘[t]his Regulation respects
the fundamental rights and observes the principles recognised in particular by the Charter of
Fundamental Rights of the European Union. Accordingly, this Regulation should be interpreted and
applied with respect to those rights and principles’; cf Judgment of the ECJustice of 27 June 2006 in
Case C–540/03 European Parliament v Council of the European Union [2006] ECR I–5769, para 38.
29
Cf Judgment of the ECJ of 31 Mar 1993 in Joined Cases 89/85 etc Ahlström Osakeyhtiö and
Others v Commission [1993] ECR I–1625, para 181, and with Judgment of the CFI of 15 Mar 2006 in
Case T–15/02 BASF v Commission [2006] ECR II–497, para 58. The choice between two possible
decisions is of course only available to the extent that the Commission is willing to close the case
through a settlement decision. As argued above (in para 102), there can be no right to a commitment
decision, the Commission having a very broad discretion to choose between commitment decisions
and infringement decisions, allowing it to weigh all the benefits and disadvantages of the two courses
of action for the optimal enforcement of Arts 81 and 82 EC.
115. As argued above,30 the recognition of a right of access to the file would
also appear useful in that it tends to reduce the risk of imposition of commit-
ments that go beyond the Commission’s powers to enforce the prohibitions laid
down in Articles 81 and 82 EC.
116. The recognition of a right of access to the file does of course not preclude
the ability of for undertakings to waive the exercise of their right in individual
cases.
2.2.3 Consultation, Adoption and Publication
117. Article 27(4) of Regulation 1/2003 provides that ‘[w]here the Commission
intends to adopt a decision pursuant to Article 9 …, it shall publish a concise
summary of the case and the main content of the commitments or of the
proposed course of action. Interested third parties may submit their observations
within a time limit which is fixed by the Commission in its publication and
which may not be less than one month’.
118. Third parties that are opposed to the case being closed by a commitment
decision, or that would at least want stronger commitments, could strengthen
their procedural position by lodging a complaint within the meaning of Article
7(1) of Regulation 1/2003. This requires that they ‘can show a legitimate interest’
within the meaning of Article 7(2) and that they use Form C annexed to
Regulation 773/2004.31 If the Commission is not convinced by the complaint,
Articles 7 and 8 of Regulation 773/2004 require it to inform the complainant of
its reasons, give the complainant access to the documents on which it bases its
assessment (expunged of business secrets and other confidential information),32
and, eventually, take a reasoned decision rejecting the complaint.
119. Article 14(1) of Regulation 1/2003 provides that the Commission shall
consult the Advisory Committee on Restrictive Practices and Dominant Posi-
tions, which is composed of representatives of the competition authorities of the
Member States, prior to the taking of a commitment decision.
120. Commitment decisions are adopted by the Commission (College) itself.
121. Finally, Article 30 of Regulation 1/2003 provides that the Commission
shall publish commitment decisions taken pursuant to Article 9. The publication
shall state the names of the parties and the main content of the decision, having
regard to the legitimate interest of undertakings in the protection of their
business secrets.
30
Above para 105.
31
See Art 5 of Reg 773/2004, and Commission Notice on the handling of complaints, above n 4.
32
In accordance with Art 8(2) of Reg 773/2004, the complainant may (only) use the documents to
which it has thus had access for the purposes of judicial or administrative proceedings for the
application of Arts 81 and 82 EC.
CONTENT AND EFFECTS 37
2.3 CONTENT AND EFFECTS
2.3.1 Type of Commitments
122. As already argued above,33 the type of commitments that could be the
object of commitment decisions pursuant to Article 9 of Regulation 1/2003
corresponds to the type of remedies which the Commission could have imposed
in an infringement decision pursuant to Article 7 if it had continued its
infringement proceedings and found the infringement of Article 81 or 82 EC
which it suspected.
123. This follows from the text of Article 9, according to which commitment
decisions can only be taken ‘[w]here the Commission intends to adopt a decision
requiring that an infringement be brought to an end’ and the commitments are
offered by the undertakings concerned ‘to meet the concerns expressed to them
by the Commission in its preliminary assessment’, and from the fact that
Regulation 1/2003 has its legal basis in Article 83 EC which concerns (only) the
enforcement of Articles 81 and 82 EC.
124. Article 7(1) of Regulation 1/2003 specifies that the Commission may
impose in infringement decisions ‘any behavioural or structural remedies which
are proportionate to the infringement committed and necessary to bring the
infringement effectively to an end. Structural remedies can only be imposed
either where there is no equally effective behavioural remedy or where any
equally effective behavioural remedy would be more burdensome for the under-
taking concerned than the structural remedy’. According to the last sentence of
recital 12 of Regulation 1/2003, ‘[c]hanges to the structure of an undertaking as it
existed before the infringement was committed would only be proportionate
where there is a substantial risk of a lasting or repeated infringement that derives
from the very structure of the undertaking’.
2.3.2 ‘No longer grounds for action by the Commission without concluding
whether or not there has been or still is an infringement’
125. According to the second sentence of Article 9(1) of Regulation 1/2003,

commitment decisions ‘shall conclude that there are no longer grounds for action
by the Commission’. According to the second sentence of recital 13, ‘[c]ommit-
ment decisions should find that there are no longer grounds for action by the
Commission without concluding whether or not there has been or still is an
infringement’.
126. Commitment decisions thus leave open the questions whether or not the
past conduct of the undertakings concerned (the agreement or practice without
33
Above, para 104.
the commitments) constituted a violation of Articles 81 and 82 EC, and whether

or not the future conduct of the undertakings concerned (assuming that they
comply with the commitments) constitutes a violation of Articles 81 and 82 EC.
127. The logic behind this is that, even if the Commission must have had
serious doubts about the compatibility of the undertakings’ past conduct with
Articles 81 and 82 EC (otherwise it could not have ‘intend[ed] to adopt a decision
requiring that an infringement be brought to an end’, as the text of Article 9 of
Regulation 1/2003 requires), it has not gone through the full proceedings
(statement of objections, hearing, etc) required reliably to make a finding of an
infringement, and the undertakings’ offer of commitments as provided for in
Article 9 does not constitute a recognition of the existence of a past infringement.
As the Commission has not gone through the full proceedings which would have
enabled it to make a reliable finding of a past infringement, it is equally unable to
make a reliable finding that compliance with the commitments makes the
agreement or practice compatible with Articles 81 and 82 EC.
128. The absence of a finding that compliance with the commitments makes
the agreement or practice compatible with Articles 81 and 82 EC does not
prevent the Commission from concluding that ‘there are no longer grounds for
action by [it]’, because the Commission is under no obligation to act against all
infringements of Articles 81 and 82 EC. Indeed, as already mentioned above,34
the Commission has a broad discretion as to which infringements it chooses to
pursue.
129. The adoption of a commitment decision does therefore not imply that
the Commission must have considered that compliance with the commitments
makes the agreement or practice compatible with Articles 81 and 82 EC.35 It
rather implies that the Commission must have considered that either there would
no longer be an infringement, or action against the remaining infringement
would not have fitted in its enforcement priorities, while saving the resources
which it would have had to spend to decide which of the two was the case.
130. Article 9(2) of Regulation 1/2003 specifies that the Commission ‘may,
upon request or on its own initiative, reopen the proceedings (a) where there has
been a material change in any of the facts on which the decision was based; (b)
where the undertakings concerned act contrary to their commitments; or (c)
where the decision was based on incomplete, incorrect or misleading information
provided by the parties’. Following such a reopening, the proceedings could then
again be closed either by an infringement decision, finding an infringement of
Article 81 or 82 EC, imposing remedies for its termination, and imposing a fine
34
Above, para 84.
35
Cf J Temple Lang, ‘Commitment Decisions and Settlements with Antitrust Authorities and
Private Parties under EU Antitrust Law’ in Hawk (ed), above n 12, and CJ Cook, ‘Commitment
Decisions: the Law and Practice Under Article 9’ (2006) 29 World Competition 211 at 228.
for this infringement (and/or for failure to comply with the commitments), or by
another commitment decision, or simply by the withdrawal of the first commit-
ment decision.
2.3.3 ‘Binding on the undertakings’: Enforcement of Commitment Decisions
131. According to the text of Article 9(1) of Regulation 1/2003, commitment

decisions make the commitments ‘binding on the undertakings’ that have offered
them.
132. Regulation 1/2003 explicitly provides for three possible actions the
Commission could take to enforce commitment decisions:
— Article 24(1)(c) allows the Commission to impose periodic penalty pay-
ments of up to 5 per cent of the average daily turnover in the preceding
business year per day to compel the undertakings to comply with the
commitments;
— Article 23(2)(c) empowers the Commission to impose fines on the under-
takings concerned, up to 10 per cent of their turnover in the preceding
business year, in case of intentional or negligent non-compliance; and
— Article 9(2)(b) allows the Commission to reopen proceedings where the
undertakings act contrary to their commitments, possibly leading to the
adoption of an infringement decision, finding an infringement of Article 81
or 82 EC, imposing remedies for its termination, and imposing fines for this
infringement.
133. Another unanswered question concerning commitment decisions is
whether they can also be enforced by private parties before national courts.36 If
the undertakings that are the addressees of a commitment decision fail to comply
with the commitments, could a private third party that claims to be harmed by
this failure obtain in a national court an injunction ordering the first undertak-
ings to comply with the commitments, or damages to compensate for the harm
caused to it by the failure to comply?
134. There is certainly no obstacle in Community law to national law provid-
ing for the possibility of such actions. Less obvious is the question whether
Community law requires that such actions can be brought.
135. In those Member States where national law has empowered the national
competition authority to adopt commitment decisions for the enforcement of
Articles 81 and 82 EC or for the enforcement of the equivalent provisions of
national competition law, similar to the commitment decisions provided for in
Article 9 of Regulation 1/2003,37 and where national law has provided for the
36
See J Davies and M Das, ‘Private Enforcement of Commission Commitment Decisions: a Steep
Climb not a Gentle Stroll’ in Hawk (ed), above n 12 and Whish, above n 24.
37
As the national law of many Member States indeed does: see below, para 158.
possibility of private enforcement of these commitment decisions, alongside the

possibility for the national competition authority to enforce its commitment
decision through fines, periodic penalty payments and reopening of proceedings
comparable to the enforcement possibilities provided to the Commission for the
enforcement of Article 9 decisions under Regulation 1/2003, it probably follows
from the principle of equivalence, a general principle of Community law which
requires equal treatment of comparable situations of Community and of purely
national origin,38 that Community law also requires a possibility for private
enforcement of the Commission’s commitment decisions. But what if national
law does not provide for private enforcement of comparable commitment
decisions adopted by national competition authorities?
136. In its initial proposal for what has become Regulation 1/2003,39 the
Commission had explicitly envisaged the possibility of private enforcement of
settlement decisions. Indeed, recital 12 of the proposed regulation stated that the
Commission should be able to adopt decisions making commitments binding ‘so
that the commitments can be relied upon by third parties before national courts
and failure to comply with them can be punished by the imposition of fines and
periodic penalty payments’. This wording was however removed by the Council,
leaving Regulation 1/2003 silent on the subject.
137. In its Notice on cooperation with national courts,40 the Commission has
affirmed that ‘[a]part from the application of Articles 81 and 82 EC, national
courts are also competent to apply acts adopted by EU institutions in accordance
with the EC Treaty or in accordance with the measures adopted to give the Treaty
effect. National courts may thus have to enforce Commission decisions [footnote:
E.g. a national court may be asked to enforce a Commission decision taken
pursuant to Articles 7 to 10, 23 and 24 of [Regulation 1/2003]]’. In a memoran-
dum posted on the Commission’s website,41 it is stated even more firmly that
‘national courts must enforce the commitments by any means provided for by
national law, including the adoption of interim measures’.
138. It is certainly correct that the silence of Regulation 1/2003 does not rule
out that Community law may require the possibility of private enforcement of
commitment decisions. Indeed, according to the case law of the Court of Justice,
‘the Treaty has created its own legal order, which is integrated into the legal
systems of the Member States and which their courts are bound to apply. The
subjects of that legal order are not only the Member States but also their
nationals. Just as it imposes burdens on individuals, Community law is also
intended to give rise to rights which become part of their legal assets. Those
38
See generally P Oliver, ‘Le règlement 1/2003 et les principes d’efficacité et d’équivalence’ [2005]
Cahiers de droit européen 343, and para 22 above.
39
COM(2000)582 final of 27 Sept 2000; see further Wils, above n 10, at 161.
40
Commission Notice on the co-operation between the Commission and the courts of the EU
Member States in the application of Arts 81 and 82 EC [2004] OJ C101/54, para 7 and n 15.
41
MEMO/04/217 dated 17 Sept 2004.
rights arise not only where they are expressly granted by the Treaty but also by
virtue of obligations which the Treaty imposes in a clearly defined manner both
on individuals and on the Member States and the Community institutions’.42
139. The question appears to be in essence whether the nature and the
function of commitment decisions, as provided for in Regulation 1/2003, are
such that commitment decisions are to be considered as creating individual
rights for third parties, and that it must be considered that the full effectiveness of
Articles 81 and 82 EC, or more specifically of Article 9 of Regulation 1/2003,
requires private enforcement of commitment decisions alongside the enforce-
ment by the Commission explicitly provided for in Regulation 1/2003.43
140. It is not obvious that the answer to this question must be positive.
Indeed, given that a commitment decision leaves open whether or not there has
been or still is an infringement,44 failure to comply with a commitment decision
does not necessarily imply that Article 81 or 82 EC is being infringed. In any
event, whether or not the commitment decision is being complied with, third
parties can always bring an action for damages or for injunctive relief in case of
infringement of Article 81 or 82 EC.45 The possibilities of enforcement by the
Commission itself, through fines, periodic penalty payments and reopening of
proceedings, would appear quite effective to ensure compliance with commit-
ment decisions. The fact that the Council did not follow the Commission’s initial
proposal to refer to the possibility of private enforcement in the recitals of
Regulation 1/2003 tends to confirm this appreciation. Finally, as has been pointed
out by John Davies and Manish Das, given that ‘the Commission will already be
seized of the facts and familiar with the underlying objectives of its commitment
decision in any given instance’, the Commission may indeed be ‘the most
appropriate enforcer’ of ‘what are, after all, its own decisions’.46
2.3.4 Further Proceedings in National Courts or by National Competition

Authorities Concerning Past Behaviour
141. Recitals 13 and 22 of Regulation 1/2003 clearly state that commitment

decisions do not conclude ‘whether or not there has been … an infringement.
Commitment decisions are without prejudice to the powers of the competition
authorities and courts of the Member States to make such a finding and decide
upon the case’.
42
Judgment of the ECJ of 20 Sept 2001 in Case C–453/99 Courage and Crehan [2001] ECR
I–6314, para 19.
43
Cf the Judgment of the ECJ of 17 Sept 2002 and the Opinion of AG Geelhoed in Case C–253/00
Muñoz [2002] ECR I–7289, and Courage v Crehan, above n 42.
44
45
See Courage v Crehan, above n 42; Opinion of AG Jacobs of 22 May 2003 in Joined Cases
C–246/01 etc AOK Bundesverband [2004] ECR I–2493, para 104; and below, para 146.
46
Davies and Das, above n 36.
142. The fact that the Commission has closed its proceedings through a
commitment decision does therefore not in any way prevent interested private
parties from bringing or continuing private actions in national courts to obtain
damages on the basis that the past conduct of the addressees of the commitment
decision (before the commitments were complied with) constituted an infringe-
ment of Article 81 or 82 EC. Of course, as the infringement decision does not
conclude that there has been an infringement, the existence of the infringement
remains to be proven. That the Commission had initially ‘intend[ed] to adopt a
decision requiring that an infringement be brought to an end’, as the text of
Article 9 of Regulation 1/2003 requires, and must thus have had serious doubts
about the compatibility of the undertakings’ past conduct with Articles 81 and 82
EC, is however a fact which the national courts may take into account in their
assessment.47
143. Neither a finding by a national court that there has been no infringement
nor a finding that there has been an infringement in any way undermines the full
effect of (the operative part of) the Commission’s commitment decision. There
can thus be no question of the decision of the national court ‘running counter’ to
the Commission’s decision within the meaning of Article 16(1) of Regulation
1/2003. For the same reason, neither the one finding nor the other would go
against Article 10 EC.
144. Similarly, the fact that the Commission has closed its proceedings
through a commitment decision does not in any way prevent a national compe-
tition authority from finding that the past conduct of the addressees of the
commitment decision (before the commitments were complied with) constituted
an infringement of Article 81 or 82 EC, and possibly also imposing a fine for this
infringement. The finding of an infringement and the imposition of a fine by a
national competition authority would not ‘run counter’ to the Commission’s
decision within the meaning of Article 16(2) of Regulation 1/2003, nor go against
Article 10 EC, because it would not in any way undermine the full effect of (the
operative part of) the commitment decision.
2.3.5 Further Proceedings in National Courts or by National Competition

Authorities Concerning Future Conduct
145. Recitals 13 and 22 of Regulation 1/2003 state no less clearly that commit-
ment decisions do not conclude ‘whether or not there … still is an infringement.
Commitment decisions are without prejudice to the powers of the competition
authorities and courts of the Member States to make such a finding and decide
upon the case’.
47
Cf the Order of the CFI of 17 Oct 2005 in Case T–28/02 First Data v Commission [2005] ECR
II–4119, para 50.
146. The fact that the Commission has closed its proceedings through a
commitment decision does not therefore in any way prevent interested private
parties from bringing private actions in national courts seeking injunctive relief
on the basis that the conduct of the addressees of the commitment decision,
although in full compliance with the commitments, constitutes an infringement
of Article 81 or 82 EC. Indeed, as explained above,48 the adoption of a commit-
ment decision does not imply that the Commission must have considered that
compliance with the commitments makes the agreement or practice compatible
with Articles 81 and 82 EC, but merely that the Commission must have
considered that either there would no longer be an infringement, or action
against the remaining infringement would not have fitted into its enforcement
priorities, while saving the resources which it would have had to spend to decide
which of the two it was.
147. On the basis of its own full analysis, a national court can thus find that
there is still an infringement, and grant injunctive relief. Such a decision would
not ‘run counter’ to the Commission’s decision within the meaning of Article
16(1) of Regulation 1/2003, nor go against Article 10 EC, unless the injunction
would make it impossible for the undertaking concerned to comply (also) with
the commitments made binding by the Commission’s decision.
148. Similarly, there is no legal obstacle to a national competition authority
finding, on the basis of its own full analysis, that there is still an infringement of
Article 81 or 82 EC, and imposing further remedies to bring this infringement to
an end, or indeed accepting further commitments to that effect. For the same
reasons as in the case of national courts, such action by a national competition
authority would not ‘run counter’ to the Commission’s decision within the
meaning of Article 16(2) of Regulation 1/2003, nor go against Article 10 EC, as
long as the additional remedies do not make it impossible to comply with the
commitments made binding by the Commission’s decision.49
149. In practice, it would seem unlikely that national authorities which are
closely involved in the adoption of the commitment decision by the Commission,
in particular through the Advisory Committee,50 would often want to take such
action.51
48
Above, paras 126–128.
49
As regards Art 11(6) of Regulation 1/2003, which provides that the initiation of proceedings by
the Commission relieves national authorities of their competence to apply Arts 81 and 82 EC, it must
be considered that this relieving of competence is transitory and comes to an end with the adoption
of the commitment decision which closes the Commission’s proceedings. Any opposite reading of Art
11(6) would be incompatible with the clear statement in recs 13 and 22 that national competition
authorities can make a finding that there still is an infringement.
50
See above, para 119.
51
See the example of the Spanish competition authority closing its parallel investigation in the
Coca-Cola case after the Commission had accepted commitments, decsribed by H Brokelman, ‘ECN
Enforcement Practice in Selected Member States: Spain and Portugal’, paper presented at the 14th
Annual Brussels IBC Conference on ‘Advanced EC Competition Law’, 17–18 Nov 2005.
150. In the light of the second sentence of Article 3(2) of Regulation 1/2003,
national competition authorities (like national courts) may of course also inter-
vene to enforce on their territory national laws which prohibit or sanction
unilateral conduct engaged in by undertakings and which are stricter than Article
82 EC.
2.4 JUDICIAL RE V IEW
151. The undertakings concerned cannot challenge before the Court of First
Instance the Commission’s preliminary assessment, because, like a statement of
objections, it constitutes merely a preliminary measure (for either a later
infringement decision or a later commitments decision).52 They cannot challenge
either any rejection by the Commission of an offer of commitments, as again this
rejection constitutes merely a preliminary measure (for a later infringement
decision).53
152. There can be no doubt that the undertakings concerned can bring an
application for annulment of the commitment decision, given that they are the
addressees of it and that it adversely affects their legal position, in that it makes
the commitments binding on them. There would however not appear to be many
grounds on which one could imagine such a challenge being brought in practice
and having a chance of success. One possible ground would be the inclusion by
the Commission in its decision of commitments that go beyond what the
undertakings have offered.
153. Member States could always bring an application for annulment of the
commitment decision before the Court of First Instance, and other interested
third parties could do so if they were directly and individually concerned by the
decision. Participation in the administrative proceedings, by submitting observa-
tions in response to the publication pursuant to Article 27(4) of Regulation
1/2003, may distinguish a potential applicant individually for this purpose.54
However, the grounds on which successful challenges could be brought appear to
be relatively limited here as well. One possibility would be procedural grounds,
for instance inadequate consultation of the Advisory Committee or inadequate
publication pursuant to Article 27(4) of Regulation 1/2003. Given that a commit-
ment decision leaves open the question whether or not there still is an infringe-
ment of Article 81 or 82 EC, the existence of such an infringement could not be a
ground for annulment.
52
See above n 22.
53
Cf Judgment of the ECJ of 14 Mar 1990 in Joined Cases 133/87 and 150/87, Nashua Corporation
and Others v Commission [1990] ECR 767 concerning the rejection of an offer in an anti-dumping
case.
54
Cf Judgment of the ECJ of 17 Nov 1998 in Case C–70/97 P Kruidvat v Commission [1998] ECR
I–7213.
OTHER SETTLEMENTS 45
2.5 OTHER SETTLEMENTS
2.5.1 Informal Settlements
154. The existence of Article 9 of Regulation 1/2003 does not prevent the
Commission from otherwise informally closing cases. Indeed, given that the
Commission has a wide discretion whether or not to pursue specific cases of
suspected infringements of Articles 81 and 82, depending on its enforcement
priorities, it also has the freedom to close an investigation it has started when the
undertakings concerned have altered their conduct or have committed them-
selves to do so, with the result that the case no longer appears to have any
priority.
155. As already mentioned above, under Regulation 17, which provided for a
notification system and formal exemption decisions, the Court of Justice also
accepted the legality of the Commission’s practice of closing cases informally by
way of ‘comfort letters’.55
156. Of course informal settlements have the disadvantage that the Commis-
sion has no means of enforcing compliance with any commitments, apart from
reopening proceedings or threatening to do so.56
2.5.2 Commitment Decisions by National Competition Authorities
157. Article 5 of Regulation 1/2003 states that ‘[t]he competition authorities of

the Member States shall have the power to apply Articles 81 and 82 of the Treaty
in individual cases. For this purpose, acting on their own initiative or on a
complaint, they may take the following decisions: —requiring that an infringe-
ment be brought to an end, …, —accepting commitments, …’.
158. On the basis of this provision and following the example of Article 9 of
Regulation 1/2003, most Member States have in their national law empowered
their national competition authorities to adopt commitment decisions compara-
ble to those which the Commission can adopt.57 As long as the national
competition authorities have the power to adopt infringement decisions, and that
the effectiveness of the enforcement of Articles 81 and 82 EC by the national
55
Above n 11.
56
57
See for instance for the UK s 31D(1) of the Competition Act 1988; see further OFT, ‘Enforce-
ment, Incorporating the Office of Fair Trading’s guidance as to the circumstances in which it may be
appropriate to accept commitments’ (London, OFT, Dec 2004), available at www.oft.gov.uk/, and
Judgment (Observations on procedure and binding commitment) of the Competition Appeals
Tribunal of 29 Nov 2004 in Case 1026/2/3/04 Wanadoo v Ofcom [2004] CAT 20.
authority is thus ensured, it would however not appear that Community law
requires that the national competition authority can also adopt commitment
decisions.58
2.5.3 Settlements Including the Payment of a Fine
159. Contrary to the situation in the USA, where antitrust investigations can be
settled by consent decrees including payment of fines, and indeed even impris-
onment, settlements of Commission investigations pursuant to Article 9 of
Regulation 1/2003 cannot include the payment of a fine.
160. The last sentence of recital 13 of Regulation 1/2003 clearly states that
commitment decisions ‘are not appropriate in cases where the Commission
intends to impose a fine’. This is related to the fact (recalled just before in the
same recital) that commitment decisions leave open the question whether an
infringement of Article 81 or 82 EC has been committed. Indeed, in those cases
where the infringement is of such a nature that a fine is indicated, it would
normally also be indicated to make a finding of an infringement, for reasons of
public censure, and to facilitate follow-on private actions for compensation.59
161. It would however not appear impossible to construct a system of
settlements involving fines on the basis of Article 7 of Regulation 1/2003. Under
an earlier version of its Leniency Notice,60 the European Commission also
granted a reduction of the fine (typically of 10 per cent) where, after receiving a
statement of objections, the undertaking informed the Commission that it did
not substantially contest the facts on which the European Commission based its
allegations.61 The European Commission is currently considering introducing a
new settlement policy further along this line, offering a fine reduction in
exchange for not merely non-contestation of the facts, but rather a recognition of
the infringement and a commitment to pay the fine.62
162. Provided that the fine reduction is not disproportionately large, such a
mechanism would not appear incompatible with fundamental rights.63 The
proceedings would still be closed by decision pursuant to Article 7 of Regulation
58
Cf Wils, above n 38.
59
60
Commission Notice on the non-imposition or reduction of fines in cartel cases [1996] OJ
C207/4. This notice was replaced in 2002; see below, para 362 and, n 26.
61
The legality of the 1996 Leniency Notice has been confirmed by the ECJ; see, inter alia,
Judgment of 16 Nov 2000 in Case C–298/98 P Metsä-Serla (Finnboard) v Commission [2000] ECR
I–10171, paras 56–58.
62
See speech by Commissioner N. Kroes, ‘Reinforcing the fight against cartels and developing
private antitrust damage actions: two tools for a more competitive Europe’ (Commission/IBA Joint
Conference on EC Competitiona Policy, 8th March 2007), available at http://ec.europa.eu/
commission_barroso/kroes/speeches_en.html ; see also below, paragraphs 377 and 409 to 412.
63
Compare with Judgment of the European Court of Human Rights of 27 Feb 1980, Deweer v
Belgium, A/35.
OTHER SETTLEMENTS 47
1/2003, but the decision would require less reasoning, and it would be much less
likely to be challenged before the Court of First Instance.
3
Optimal Antitrust Fines: Theory and
Practice
163. This chapter discusses the use of fines imposed on companies or other
corporate entities (hereafter also ‘corporate fines’) to enforce antitrust or compe-
tition law prohibitions such as Articles 81 and 82 of the EC Treaty or sections 1
and 2 of the Sherman Act. I will particularly focus on the enforcement of Articles
81 and 82 EC,1 not only because these prohibitions are the ones I am most
familiar with, but also because corporate fines play a particularly important role
in EU antitrust enforcement.2
164. The chapter addresses more specifically the questions in what ways
corporate fines contribute to competition law enforcement, on the basis of which
factors the amount of antitrust fines should be fixed in theory, and whether it is
feasible in practice to calculate or measure such optimal fines.
165. Throughout most of the chapter, I assume that corporate fines are the
only sanctions used to deter antitrust violations, ie that they are not combined
with fines on individuals, imprisonment or other individual sanctions, nor with
private damages. This assumption is still broadly realistic for the existing situa-
tion in the EU.3 The chapter does not address the question whether it would be
desirable to combine corporate fines with such other sanctions,4 nor whether or
how the presence of these other sanctions should affect the fixing of the amount
of fines on companies.
1
Art 81 EC prohibits agreements or concerted practices which restrict competition and lack
redeeming virtue, whereas Art 82 prohibits abuse of a dominant position: see generally R Whish,
Competition Law, 5th edn (London, Lexis Nexis Butterworth, 2003).
2
See RH Pate, ‘Antitrust in a Transatlantic Context—From the Cicada’s Perspective’, address to the
‘Antitrust in a Transatlantic Context’ Conference (Brussels, 7 June 2004), available at www.usdoj.gov/
atr/public/speeches/203973.htm, at 3: ‘[i]t is a fact I do not like to advertise that, during Commis-
sioner Monti’s tenure, the EC has surpassed the United States in terms of monetary penalties levied
against cartelists’; see also immediately below as to the limited importance of other penalties or
sanctions (fines on individuals, imprisonment, private damages) in the enforcement of EU competi-
tion law.
3
On individual sanctions see D Cahill (ed) and JD Cooke (General Rapporteur), The Modernisa-
tion of EU Competition Law Enforcement in the EU—FIDE 2004 National Reports (Cambridge,
Cambridge University Press, 2004); on private actions for damages, see Ashurst (a firm), Study on the
Conditions of Claims for Damages in Case of Infringement of EC Competition Rules (Aug 2004),
available at http://europa.eu.int/comm/competition/antitrust/others/private_enforcement/
index_en.html#damages.
4
On the question whether individual sanctions should be added see Ch 6 below; on the question
whether private actions for damages should be encouraged in Europe see WPJ Wils, ‘Should Private
50 OPTIMAL ANTITRUST FINES: THEORY AND PRACTICE
166. I also assume throughout the chapter that fines are imposed in only one
jurisdiction. The article does therefore not deal with any questions related to the
simultaneous or successive imposition of fines in different jurisdictions.5
3.1 THE ROLE OF FINES IN ANTIT R UST ENFORCEMENT
3.1.1 The Three Tasks of Competition Law Enforcement
167. Considered broadly,6 the enforcement of antitrust prohibitions such as

Articles 81 and 82 EC could be said to entail three tasks: (1) clarifying the content
of the prohibitions, (2) preventing violations of these prohibitions, and (3)
dealing with the consequences when violations have nevertheless happened.
3.1.1.1 Clarifying the Content of the Prohibitions
168. The first task of antitrust enforcement is to clarify the content of the
antitrust prohibitions. Indeed, Articles 81 and 82 EC, like sections 1 and 2 of the
Sherman Act, contain only generally worded standards.7 Their content is clarified
through judgments or decisions in individual cases,8 as well as through guidelines
issued by the competition authorities.9 This also allows the content of the
Antitrust Enforcement Be Encouraged in Europe?’ (2003) 26 World Competition 473, reprinted in

WPJ Wils, Principles of European Antitrust Enforcement (Oxford, Hart Publishing, 2005) 111–27.
5
On the relevance of fines imposed in the US for the imposition of fines in the EU see the
Judgment of the ECJ of 29 June 2006 in Case C–308/04 P SGL Carbon v Commission [2006] ECR
I–5977, paras 26–39. On the imposition of fines by several competition authorities or courts within
the EU see WPJ Wils, ‘The Principle of Ne Bis in Idem in EC Antitrust Enforcement: A Legal and
Economic Analysis’ (2003) 26 World Competition 131, and ch 1, sects 1.2.5 and 1.2.6, and 3 of Wils,
Principles, above n 4.
6
In this broad sense, ‘enforcement’ of Arts 81 and 82 EC covers the whole of the ‘implementation’
of these provisions, the latter being the term used in the title of Council Reg 1/2003 of 16 Dec 2002 on
the implementation of the rules on competition laid down in Arts 81 and 82 of the Treaty [2003] OJ
L1/1, last amended by Council Reg 1419/2006 [2006] OJ L269/1, the basic reg laying down the rules
on the enforcement of Arts 81 and 82 EC.
7
On the economics of rules versus standards see generally I Ehrlich and RA Posner, ‘An Economic
Analysis of Legal Rulemaking’ (1974) 3 Journal of Legal Studies 257 and L Kaplow, ‘Rules versus
Standards’ (1992) 42 Duke Law Journal 557.
8
For examples of such judgments and decisions clarifying the content of Arts 82 and 81 EC see:
Judgment of the ECJ of 26 Nov 1998 in Case C–7/97 Bronner v Mediaprint [1998] ECR I–7817, and
Dec 98/531/EC of the European Commission of 11 Mar 1998 in Case IV/34.073 Van den Bergh Foods
[1998] OJ L246/1 and Judgment of the CFI of 23 Oct 2003 in Case T–65/98 Van den Bergh Foods v
Commission, [2003] ECR II–4653; see also Art 10 of Reg 1/2003, above n 6, and European
Commission Notice on informal guidance relating to novel questions concerning Arts 81 and 82 of
the EC Treaty that arise in individual cases (guidance letters) [2004] OJ C101/78.
9
See for instance the European Commission’s Guidelines on vertical restraints [2000] OJ C291/1,
and Guidelines on the applicability of Art 81 EC to horizontal co-operation agreements [2001] OJ
C3/2.
Columns Design Ltd / Job: Wils_Antitrust / Division: Wils_Chapter3 /Pg. Position: 2 / Date: 22/1
THE ROLE OF FINES IN ANTITRUST ENFORCEMENT 51
prohibitions to evolve over time, following the development of economic

theory,10 and the fluctuations of political or societal preferences as to the bounds
of legitimate private and public power.11
3.1.1.2 Preventing Violations
169. The second and central task of antitrust enforcement is to prevent viola-
tions of the antitrust prohibitions. To see the different ways in which one could
try to prevent violations from occurring, one could start by asking under what
conditions companies, or the individuals taking the decision within the com-
panies, are likely to commit antitrust violations.12 The first condition is that they
need an opportunity to commit a violation. Secondly, they need to be willing to
commit a violation. Corporate managers are not necessarily just maximisers of
profits for themselves and their principals. They may feel a moral responsibility
to live within the law whether or not they are likely to be caught, and this
normative commitment could trump their interest calculus.13 Indeed, psycho-
logical research suggests that normative commitment is generally an important
factor explaining compliance with the law.14 Thirdly, the companies or individual
decision-makers need incentives to commit violations, in that the expected
benefits to them of the violations exceed the expected costs. One could thus try to
prevent antitrust violations by influencing any of these three conditions.
3.1.1.2.1 Reducing the Opportunities to Commit Violations
170. First, one could try to reduce the opportunities to commit violations. This
can be done through anticipatory (ex ante) intervention. Under the EC Merger
Regulation,15 the prohibition of concentrations which significantly impede effec-
tive competition is enforced through a system of mandatory prior notification and
authorisation.16 Most jurisdictions have similar ex ante enforcement mechanisms
10
A good example is the evolution of economic thought on predatory pricing: see United States v
AMR Corporation, 335 F 3d 1109 (10th Cir, 2003).
11
See G Amato, Antitrust and the Bounds of Power (Oxford, Hart Publishing, 1997).
12
See WPJ Wils, The Optimal Enforcement of EC Antitrust Law (the Hague, Kluwer Law
International, 2002), sect 8.2.
13
CD Stone, ‘Sentencing the Corporation’ (1991) 71 Boston University Law Review 383 at 389.
14
See TR Tyler, Why People Obey the Law (Princeton, NJ, Yale University Press, 1990).
15
Council Reg 139/2004 of 20 Jan 2004 on the control of concentrations between undertakings
(the EC Merger Reg) [2004] OJ L24/1.
16
Legally speaking, the EC Merger Reg does not contain a prohibition of concentrations which
significantly impede effective competition. It rather contains an obligation to notify proposed
concentrations and a prohibition to put them into effect before the notification, during the
Commission procedure and thereafter if the procedure results in a decision finding that the
concentration will significantly impede effective competition. These obligations are the reflection of
the choice of an enforcement system based on mandatory prior notification and advance clearance, to
enforce the underlying substantive prohibition of concentrations which significantly impede effective
competition.
for mergers.17 Before 1 May 2004, a (half-hearted) system of prior notification also
existed for Article 81 EC, but this proved unworkable and has therefore been
abolished by Regulation 1/2003.18 Injunctions ordering impending or current
violations to be stopped are also a form of anticipatory intervention.19
3.1.1.2.2 Reducing Business People’s Willingness to Commit Violations
171. Secondly, one could try to prevent antitrust violations by reducing business
people’s willingness to commit violations through a strengthening of their
normative commitment to the antitrust rules.20 Competition authorities could
try to do this through competition advocacy. The imposition of punishment for
violations also plays an important role. Indeed, the public punishment of those
who violate the antitrust prohibitions not only has a deterrent effect, in that it
helps create a credible threat of punishment for those who would be willing to
commit violations on the basis of a profit calculation,21 but also has moral effects,
in that it sends a message to the spontaneously law-abiding, reinforcing their
moral commitment to the rules.22
3.1.1.2.3 Altering the Balance of Expected Benefits and Costs of Violations
172. Thirdly, one can try to prevent antitrust violations by altering the balance
of expected benefits and expected costs of violations. This can be done by
prosecuting and punishing violations, and creating a credible threat of penalties
which weighs sufficiently in the balance of expected costs and benefits to deter
calculating companies from committing antitrust violations. Apart from such
deterrence, other methods could also be used to alter the balance of expected
costs and benefits. Setting up and maintaining a successful cartel requires effort.
The cartel members have to select and coordinate their behaviour on mutually
17
On the economics of ex ante versus ex post enforcement see Wils, above n 12, ch 5 and sect 6.2.2.
18
See ch 1 of Wils, Principles, above n 4. Prohibitions such as Arts 81 and 82 are by their nature
unsuited for ex ante enforcement, because of the very high number of agreements and business
decisions the legality of which would have to be checked, and because of the ease of concealing
violations: see ch 5 and sect 6.2.2 of Wils, above n 12.
19
See below n 28.
20
See generally KG Dau-Schmidt, ‘An Economic Analysis of the Criminal Law as a Preference-
Shaping Policy’ [1990] Duke Law Journal 1; CR Sunstein, ‘On the Expressive Function of the Law’
(1996) 144 University of Pennsylvania Law Review 2021, DM Kahan, ‘Social Influence, Social Meaning,
and Deterrence’ (1997) 83 Virginia Law Review 349, NK Katyal, ‘Deterrence’s Difficulty’ (1997) 95
Michigan Law Review 2385, GE Lynch, ‘The Role of Criminal Law in Policing Corporate Misconduct’
(1997) 60 Law and Contemporary Problems 23, DM Kahan, ‘Social Meaning and the Economic
Analysis of Crime’ (1998) 27 Journal of Legal Studies 609, and KG Dau-Schmidt, ‘Preference Shaping
by the Law’ in P Newman (ed), The New Palgrave Dictionary of Economics and the Law (London,
Macmillan, 1998) 84.
21
22
See J Adenaes, ‘The Moral or Educative Influence of Criminal Law’ (1971) 27 Journal of Social
Issues 17, and ‘General Prevention Revisited: Research and Policy Implications’ (1975) 66 Journal of
Criminal Law & Criminology 338, at 341–3, Lynch, above n 20, at 46–7, and the other literature
referred to in n 20 above.
consistent, collusive strategies, allowing the cartel participants as a group to

increase profits, and providing for a fair distribution of profits between them.
They also need to develop mechanisms to discourage cheating, involving moni-
toring, rewards and punishments. In a dynamic economy, successful cartels may
have to develop an organisational structure that allows them to solve these
problems continuously.23 Antitrust enforcement can increase the cost of setting
up and running cartels in different ways. One very important measure is to make
cartel agreements legally unenforceable, as Article 81(2) EC does. As will be
explained further in this chapter,24 the cost of setting up and running cartels can
also be increased through leniency policies and through the use of other
aggravating or attenuating circumstances affecting the amount of the penalty
imposed on those violations which are detected and punished. Finally, the
possible use of injunctions to stop continuation of violations which are detected
when still in progress also has an impact on the expected benefits of violations.25
3.1.1.3 Dealing with the Consequences of Violations
173. Attempts to prevent violations of antitrust prohibitions such as Articles 81

and 82 EC are unlikely ever to lead to the total elimination of all violations. The
reasons for this are both economic and psychological.
174. The economic reason is that the enforcement measures to prevent
violations are normally not without cost. The detection, prosecution and punish-
ment of violations has a significant administrative cost, which includes both the
cost borne by the public sector (cost of competition authorities, prosecutors and
courts) and that borne by the businesses and individuals concerned (cost of
lawyers and experts, management time). Apart from administrative costs, the
pursuit of deterrence could also have undesirable side-effects. For instance, errors
or the risk of errors in the imposition of sanctions could lead to lawful and
economically desirable conduct being deterred. Given the existence of these costs,
it is unlikely to be optimal to pursue full prevention of antitrust violations.
Depending on the one hand on the cost of achieving different degrees of
prevention, and on the other hand on how much value society attaches to the
avoidance of antitrust violations, the optimum will be to pursue a certain degree
of prevention, which in all likelihood will be less than 100 per cent.26
23
MC Levenstein and VY Suslow, ‘What Determines Cartel Success?’ (2006) 44 Journal of
Economic Literature 43 at 44–5; see also JE Harrington, ‘How Do Cartels Operate?’ (June 2006),
available at www.econ.jhu.edu/People/Harrington/CartelOperations-6.28.06.pdf.
24
See paras 223–228 below; see also Ch 5.
25
See below n 28.
26
See generally, with regard to deterrence, GJ Stigler, ‘The Optimum Enforcement of Laws’ (1970)
78 Journal of Political Economy 526 and KG Elzinga and W Breit, The Antitrust Penalties: A Study in
Law and Economics (Princeton, NJ, Yale University Press, 1976) at 9–15.
175. The psychological reason, which is further explained below,27 is that, in

the probability estimates they make, people tend to rely disproportionately on
those incidents which can easily be brought to mind. This implies that, even if the
competition authorities managed at some point in time to detect so many
violations and to impose such high fines that all companies will be deterred from
committing new violations, this situation will not last, as over time the memory
of those successful prosecutions will fade, and violations will thus again be
committed.
176. If it is thus unrealistic to achieve total prevention of all violations of the
antitrust prohibitions, there can be a further task for antitrust enforcement in
dealing with the consequences of violations that have taken place.
177. In cases where the anti-competitive harm caused by the violation is still
unfolding, it may be possible to limit or mitigate the harm through the use of
injunctions.28
178. Even if the harm has fully unfolded, there could still be a task for
antitrust enforcement in the pursuit of corrective justice. Two (partly overlap-
ping) aspects could be distinguished. First, corrective justice could be pursued
through disgorgement, by taking from the antitrust violator any benefits from the
violation.29 Secondly, corrective justice could be pursued through compensation,
by making the party which wrongfully committed the violation compensate
other parties who innocently suffered the consequences of the violation.30
179. If one takes the view that the sole goal of antitrust is to enhance total
economic efficiency,31 the pursuit of corrective justice is by assumption a waste of
resources. However, if one considers that antitrust is, at least in part, concerned
with avoiding wealth transfers from consumers to firms with market power, as I
do and I believe most people do,32 the pursuit of corrective justice through
compensation can be a legitimate task for antitrust enforcement. The pursuit of
this goal, like the pursuit of the goal of preventing antitrust violations, is of
course again not without cost.33 How much compensation is worth the cost of
obtaining it depends on the one hand on how high the cost is, and on the other
hand on how much value society attaches to the pursuit of corrective justice
through compensation in this area.34
27
Below, paras 195–196.
28
As indicated above, paras 170 and 172, injunctions can also be an instrument of ex ante
prevention, and their possible use may also effect the expected benefits of contemplated infringe-
ments.
29
See US Federal Trade Commission, Policy Statement on Monetary Equitable Remedies in
Competition Cases (25 July 2003), available at www.ftc.gov/os/2003/07/disgorgementfrn.htm.
30
See K Roach and MJ Trebilcock, ‘Private Enforcement of Competition Laws’ (1996) 34 Osgoode
Hall Law Journal 461 at 496.
31
32
33
34
In Wils, ‘Should Private Antitrust Enforcement Be Encouraged in Europe?’, above n 4, I argued
at 486–8 that, because there does not appear to be a clear social need, and because truly achieving
3.1.2 The Role of Fines
180. With regard to the first task of clarifying the content of the antitrust
prohibitions, fines do not appear to play a significant role.35
181. Fines are however an important instrument in the prevention of viola-
tions. As already indicated,36 the imposition of fines on companies that are found
to have breached the antitrust prohibitions could in three ways contribute to
preventing such violations. First, it may have a deterrent effect by creating a
credible threat of being prosecuted and fined which weighs sufficiently in the
balance of expected costs and benefits to deter calculating companies from
committing antitrust violations. Secondly, it may at the same time have a moral
effect, in that it sends a message to the spontaneously law-abiding, reinforcing
their moral commitment to the antitrust prohibitions. Thirdly, through leniency
policies and through the use of other aggravating or attenuating circumstances
affecting the amount of the fine imposed, the cost of setting up and running
cartels can be raised. These different effects will be analysed in more detail
below.37
182. Finally, whereas fines will normally have disgorgement of the unjust
enrichment as one of their effects,38 the proceeds of fines normally go into the
public budget rather than to the victims of the antitrust violations, and fines
could thus at most be said to contribute to the pursuit of corrective justice
through compensation in an abstract and indirect way.39
corrective justice in the antitrust context is in practice a very difficult task, an increased focus of EC
antitrust enforcement on compensation may be difficult to justify.
35
Indeed, in those cases where a competition authority took a decision or brought a prosecution
so as to clarify that a certain behaviour, the illegality of which was previously not clear, violated the
antitrust prohibitions, the imposition of a fine would legally not be possible. In the case of fines
imposed by the European Commission for violation of Art 81 or 82 EC, the ‘intent or negligence’
required by Art 23(2) of Reg 1/2003, above n 6, would be lacking; see also Judgment of the CFI of 30
Sept 2003 in Joined Cases T–191/98 etc Atlantic Container Line and Others v Commission [2003] ECR
II–3275, paras 1611–1633, and the European Commission Decision of 24 June 2004 in Case
COMP/38.549 Belgian Architects’ Association [2005] OJ L4/10, para 137. Requiring intent or negli-
gence also makes perfect economic sense. Fining companies for conduct they cannot avoid at
reasonable cost will have no deterrent effect, while it may generate wasteful expense: see R Posner, The
Economics of Justice (Cambridge, Mass, Harvard University Press, 1983) at 225; R Posner, ‘An
Economic Theory of the Criminal Law’ (1985) 85 Columbia Law Review 1193 at 1221–3; and Wils,
above n 12, at 30.
36
37
Corporate fines are of course not unique in having these effects, which may also, and maybe
even better, be obtained through or in combination with other types of penalties: see above n 4.
38
Indeed, as explained below, in para 184, in order to deter, fines must exceed the expected gain
from the violation, multiplied by the inverse of the probability that a fine will be imposed. Fines thus
set will therefore normally exceed the unjust enrichment, and will thus automatically have disgorge-
ment as one of their effects.
39
See however Decs of the European Commission of 21 Oct 1998, Pre-Insulated Pipe Cartel
[1999] OJ L24/1, rec and of 30 Oct 2002, Nintendo [2003] OJ L255/33, recs 440 and 441, granting
reductions of the fines imposed on ABB and Nintendo respectively for violations of Art 81 EC in
recognition of the fact that ABB had paid ‘substantial compensation to Powerpipe and its previous
owner’ and that Nintendo had ‘offered substantial financial compensation to third parties identified
3.2 OPTIMAL FINES FOR SING LE OFFENDERS
183. In order to examine on the basis of which factors the amount of fines
should be fixed, it may be useful to concentrate first on the relatively simple case
of fines for antitrust violations committed by a single offender, such as abuses of
a dominant position within the meaning of Article 82 EC committed by a single
dominant company. The case of fines for collective violations, such as price-
fixing or market-sharing cartels, will be examined afterwards.
3.2.1 The Basic Logic of Deterrence
184. The idea of deterrence is to create a credible threat of penalties which

weighs sufficiently in the balance of expected costs and benefits to deter calculat-
ing companies from committing antitrust violations. Deterrence through the use
of fines will work if, and only if, from the perspective of the company contem-
plating whether or not to commit a violation the expected fine exceeds the
expected gain from the violation. The expected fine equals the nominal amount
of the fine discounted by the probability that a fine is effectively imposed. Certain
types of violations are more easily detectable than others. Some companies may
also be better at avoiding apprehension than others, possibly because they are
more experienced.40 The probability of actually being fined obviously also
depends on the competition authorities’ enforcement priorities and their avail-
able resources. If, for instance, the probability of detection and punishment is one
in five, the expected fine is only one fifth of the nominal amount. In order to
deter, the nominal amount of the fine must then be at least five times larger than
the expected gain. The minimum fine for deterrence to work thus equals the
expected gain from the violation multiplied by the inverse of the probability of a
fine being effectively imposed.41
3.2.1.1 The Deterrence Approach as Opposed to the Internalisation Approach
185. Under the deterrence approach set out just above, the optimal fine should
exceed the expected gain from the violation multiplied by the inverse of the
in the Statement of Objections as having suffered financial harm as a result of [Nintendo]’s violation’.
The European Commission has refused to grant such reductions in other cases, and the CFI has
recently confirmed that addressees of fining decisions have no right to obtain such reductions: see
paras 349–355 of the judgment of 27 Sept 2006 in Case T–59/02 Archer Daniels Midland v
Commission, not yet reported.
40
See LA Bebchuk and L Kaplow, ‘Optimal Sanctions and Differences in Individuals’ Likelihood
of Avoiding Detection’ (1993) 13 International Review of Law and Economics 217.
41
Interest should also be added to the fine to cover the delay between the point in time at which
the gain was (expected to be) obtained and the point in time at which the fine is paid.
OPTIMAL FINES FOR SINGLE OFFENDERS 57
probability of a fine being effectively imposed, so as to eliminate all violations.42

A different approach has been advocated by Gary Becker and by William
Landes.43 Under their internalisation approach, the optimal fine equals the net
harm caused to persons other than the offender, again multiplied by the inverse
of the probability of a fine being effectively imposed. The optimal fine thus set
makes the offender internalise all the costs and benefits of the violation, thus
leading the offender to commit the ‘efficient violations’ the total benefits of which
exceed the total costs, while deterring ‘inefficient violations’ the total costs of
which exceed the total benefits.
186. The difference between the deterrence approach and the internalisation
approach to antitrust fines mainly reflects, I believe, a difference in views about
the primary goal of the antitrust prohibitions.
187. The internalisation approach to antitrust fines fits the Chicago School
view that the primary goal of the antitrust prohibitions is to maximise total
economic welfare or total economic efficiency, ie the sum of the economic
welfare of both buyers and sellers (consumers and producers).44 If one adheres to
this view about the primary purpose of antitrust, one would logically also
embrace the internalisation approach to antitrust fines, as it ensures that the
welfare of antitrust violators and that of their victims are given equal weight.
188. The internalisation approach to antitrust fines appears however unsuit-
able if one has a different or broader view about the primary purpose of
antitrust, as I do and I believe most people do. If one takes the view that the
primary goal of the antitrust prohibitions is to prevent extractions of consumers’
wealth by firms with market power, ie to prevent wealth transfers from consum-
ers to producers,45 one would naturally opt for the deterrence approach to
antitrust fines, as it aims at deterring all antitrust violations, irrespective of
whether the offender’s gain exceeds the harm caused to consumers. If one
42
See however above, paras 173–175 as to the economic and psychological reasons why deterrence
is unlikely ever to be complete.
43
GS Becker, ‘Crime and Punishment: An Economic Approach’ (1968) 76 Journal of Political
Economy 169 and WM Landes, ‘Optimal Sanctions for Antitrust Violations’ (1983) 50 The University
of Chicago Law Review 652.
44
As has been pointed out by RH Lande, ‘Chicago’s False Foundation: Wealth Transfers (Not Just
Efficiency) Should Guide Antitrust’ (1989) 58 Antitrust Law Journal 631 at 638, this view has been
somewhat deceptively presented in terms of maximising ‘consumer welfare’, a term which rather
corresponds to the view, mentioned just below, that the primary goal of antitrust is to prevent wealth
transfers from consumers to producers. For a recent example of this confusing use of the term
‘consumer welfare’ see KN Hylton, Antitrust Law—Economic Theory and Common Law Practice
(Cambridge, Cambridge University Press, 2003) at 43–4.
45
See RH Lande, ‘Wealth Transfers as the Original and Primary Concern of Antitrust: The
Efficiency Interpretation Challenged’ (1982) 34 Hastings Law Journal 67, and ‘Chicago’s False
Foundation’, above n 44; see also the European Commission’s Guidelines on the application of Art
81(3) of the [EC] Treaty [2004] OJ C101/97, at para 43: ‘[n]egative effects on consumers in one
geographical market or product market cannot normally be balanced against and compensated by
positive effects for consumers in another unrelated geographic market or product market’, such
compensation only being possible ‘provided that the group of consumers affected by the restriction
and benefitting from the efficiency gains are substantially the same’.
considers that the primary goal of the antitrust prohibitions is more broadly to
protect consumer choice or consumer sovereignty,46 or to preserve competitive
structure and open market values,47 the deterrence approach to fines would
similarly be indicated.
approach to antitrust fines also reflects a difference in views as to the degree of
moral condemnation to be attached to antitrust violations. Borrowing terms
used by Robert Cooter and by John Coffee,48 the internalisation approach merely
seeks to ‘price’ antitrust violations, whereas the deterrence approach seeks to
‘sanction’ or ‘prohibit’ such violations. Under the deterrence approach, the
antitrust fine is ‘a sanction for doing what is forbidden’, whereas it is ‘the price of
doing what is permitted’ under the internalisation approach.
approach to antitrust fines not only reflects philosophical differences as to the
primary goal of the antitrust prohibitions and the degree of moral condemnation
to be attached to antitrust violations. It also makes a practical difference for the
administrative cost and the effectiveness of antitrust enforcement.
191. First, the internalisation approach appears more difficult to apply in
practice.49 Indeed, to determine the optimal fine in a concrete case, one has to
quantify the harm to parties other than the offender, which includes not only the
monopoly transfer but also the portion of the deadweight loss borne by consum-
ers,50 as well as the probability of a fine being imposed. These factors are very
hard to quantify in practice,51 which means that the imposition of fines will be
administratively costly and/or that errors will be made which will weaken
deterrence. On the other hand, the deterrence approach also requires some
46
See RH Lande, ‘Consumer Choice as the Ultimate Goal of Antitrust’ (2001) 62 University of
Pittsburgh Law Review 503 and NW Averitt and RH Lande, ‘Consumer Sovereignty: A Unified Theory
of Antitrust and Consumer Protection Law’ (1997) 65 Antitrust Law Journal 713.
47
See EM Fox, ‘What is Harm to Competition? Exclusionary Practices and Anticompetitive Effect’
(2002) 70 Antitrust Law Journal 371; see also Judgment of the ECJ of 21 Feb 1973 in Case 6/72
Europemballage and Continental Can v Commission [1973] ECR 215, para 26: ‘[Article 82 EC] is not
only aimed at practices which may cause damage to consumers directly, but also at those which are
detrimental to them through their impact on an effective competition structure’; Order of the
President of the ECJ of 11 Apr 2002 in Case C–481/01 P(R) NDC Health v Commission [2002] ECR
I–3401, para 84: ‘the reasoning … cannot be accepted without reservation, in so far as it could be
understood as excluding protection of the interests of competing undertakings from the aim pursued
by Article 82 EC, even though such interests cannot be separated from the maintenance of an effective
competition structure’; and Judgment of the CFI of 17 Dec 2003 in Case T–219/99 British Airways v
Commission [2003] ECR II–5917, para 311: ‘Article 82 is aimed at penalising even an objective
detriment to the structure of competition itself ’.
48
R Cooter, ‘Prices and Sanctions’ (1984) 84 Columbia Law Review 1523 and JC Coffee,
‘Paradigms Lost: The Blurring of the Criminal and Civil Law Models—And What Can Be Done
About It’ (1992) 101 Yale Law Journal 1875.
49
See also E David, Les sanctions des pratiques anticoncurrentielles en droit comparé, Doctoral
Dissertation (Université Robert Schuman, Strasbourg III, 10 Dec 2004), para 1732.
50
See Landes, above n 43, and Hylton, above n 44, at 44–5.
51
quantification of the expected gain and of the probability of a fine being

imposed, but the need for precision is much lower, because it is only required
that the fine exceed the expected gain discounted by the probability of a fine
being imposed.52
192. Secondly, the deterrence approach also appears more effective because
the imposition of fines under this approach is likely to have stronger moral
effects. As explained above,53 the imposition of punishment for antitrust viola-
tions has not only a deterrent effect, in that it helps to create a credible threat of
punishment for those who would be willing to commit violations on the basis of
a profit calculation, but also has a moral effect, in that it sends a message to the
spontaneously law-abiding, reinforcing their moral commitment to the rules.
This moral effect is weakened, or even eliminated, under the internalisation
approach, because the antitrust fine is not presented as ‘a sanction for doing what
is forbidden’, but rather as the ‘price for doing what is permitted’.54
193. For all these reasons, the deterrence approach appears to me the better
one, and I will thus primarily focus on this approach in the remainder of this
chapter.55
3.2.1.2 Subjective Estimates of the Gain, the Probability of Detection and

Punishment and the Amount of the Fine
194. For deterrence to work, it is necessary that, from the perspective of the
company (or the individual decision-maker deciding for the company) contem-
plating a possible antitrust violation, the expected fine exceed the expected gain.
What thus counts is the potential offender’s subjective estimate of the gain, of the
probability of detection and punishment, and of the amount of the fine in case of
detection and punishment. These subjective estimates are not necessarily accu-
rate. Indeed, cognitive psychology suggests two reasons why this may not be the
case.
3.2.1.2.1 Availability Bias
195. To predict accurately the probability of detection and punishment, and the
amount of the fine, potential offenders would have to consider all existing
information about the probability of detection and the amount of fines in
general, and adjust the statistical probability thus derived with any particularised
information about their specific situation. In practice, people tend to rely
52
See however paras 203–212 below as the limits on very high fines.
53
54
See para 189 above.
55
See however below, nn 61 and 83 and para 239, where the internalisation approach is also
considered.
disproportionately on those incidents which can easily be brought to mind,

because they are recent, happened close to them, or were well publicised.56
196. This phenomenon has several implications for the use of fines to deter
antitrust violations. First, it confirms the importance for competition authorities
to give maximum publicity to the cases in which they successfully discover and
punish violations, in particular the cases with the highest fines.57 Secondly, it
implies that it will never be possible to achieve complete deterrence. Indeed, even
if the competition authorities managed at some point in time to detect so many
violations and to impose such high fines that all companies were deterred from
committing new violations, this situation would not last, as over time the
memory of those successful prosecutions would fade, and violations would thus
again be committed.58 Thirdly, for the same reason, it also implies that the fact
that violations are still discovered does not necessarily justify the conclusion that
the level of the fines imposed in the past was insufficient and should thus be
raised.
3.2.1.2.2 Over-confidence Bias
197. It appears that people consistently tend to overestimate the probability of

good things happening to them, and underestimate the probability of bad things
happening to them, in particular when the event in question is perceived to be
controllable.59
198. Applied to deterrence through antitrust fines, this suggests that com-
panies and individual decision-makers within them will tend to overestimate the
gain from the antitrust violation and to underestimate the probability of being
caught.60 Ex post estimates of the gain obtained from the violation will thus tend
to be systematically below the ex ante expectations of potential offenders, and ex
post estimates of the probability of a fine being imposed will systematically tend
to exceed the probability as perceived ex ante by potential offenders. If fines are
56
See RB Korobkin and TS Ulen, ‘Law and Behavioral Science: Removing the Rationality
Assumption from Law and Economics’ (2000) 88 California Law Review 1051 at 1085–90 and A
Tversky and D Kahneman, ‘Availability: A Heuristic for Judging Frequency and Probability’ in D
Kahneman, P Slovic and A Tversky (eds), Judgment under Uncertainty: Heuristics and Biases
(Cambridge, Cambridge University Press, 1982) at 163.
57
Such publicity is also desirable because of the moral effects of punishment: see above, paras 169
and 171; see however below, paras 208–212 as to the requirements of proportional justice.
58
59
See ND Weinstein, ‘Optimistic Biases About Personal Risks’ (1989) 246 Science 1232 and
Korobkin and Ulen, above n 56, at 1091–5.
60
See also ML Denger, ‘Too Much or Too Little’, American Bar Association, Section of Antitrust
Law, Antitrust Remedies Forum, available at www.abanet.org/antitrust/remedies, at 5: ‘most cartel
participants don’t think they will get caught’.
calculated on the basis of ex post estimates of gain and probability, they will thus
systematically tend to be too low to deter, unless an upward adjustment is made
to counter this effect.61
3.2.2 The Trade-off between the Level of Fines and the Probability of
Detection and Punishment
3.2.2.1 The Standard Economic Argument and its Limits
199. Given that deterrence requires that, from the perspective of the company
contemplating whether or not to commit a violation, the expected fine exceed the
expected gain from the violation, and given that the expected fine equals the
nominal amount of the fine discounted by the probability that a fine is effectively
imposed, deterrence can be achieved through different combinations of the level
of fines and of the probability of detection and punishment. The same result
could be achieved either with very high fines and a low probability of such fines
being imposed, or with lower fines and a high probability of detection and
punishment.
200. Given that the detection and punishment of antitrust violations have a
significant administrative cost, including not only the costs borne by the compe-
tition authorities and courts but also those borne by the companies concerned
(cost of lawyers and other experts, as well as lost management time), basic
economic reasoning would plead for a strategy of very high fines which are only
rarely imposed.62
201. There are however two (sets of) reasons why this strategy may not be
optimal, or could at least not be carried too far. The first possible reason relates to
the availability bias discussed above.63 If companies or individual decision-
makers are biased in their estimates of the probability of detection and punish-
ment and of the amount of the fine, in that they rely disproportionately on those
incidents which can easily be brought to mind because they are recent, happened
close to them, or were well publicised, a strategy of very high fines which are
rarely imposed may on the one hand make sense, in that the very high amounts
of the fines imposed may give them more publicity and thus make them more
memorable. On the other hand, a sufficient frequency of punishment would
seem more efficient, assuming that if a company (or the individual decision-
maker within the company) knows or knows of a company which has been fined
61
See below, para 240 . It should be noted that the same problem arises under the internalisation
approach: see above, paras 187–193.
62
See Becker, above n 43.
63
for a comparable antitrust violation, this information is likely to be available and

thus to cause an overestimate of the probability of being fined.64
202. The second set of reasons why a strategy of very high fines combined
with a low probability of detection and punishment may not be optimal, or could
at least not be carried too far, is that very high fines may be problematic in several
respects, in particular because they risk exceeding the ceiling of companies’
ability to pay, because high fines, even if they do not exceed ability to pay, may
have substantial social and economic costs, and because they may be unaccept-
able from the perspective of proportional justice.65
3.2.2.2 The Limits to High Fines
3.2.2.2.1 Inability to Pay
203. For several types of antitrust violations, it would not appear abnormal for
the company that is contemplating committing the violation to expect as gain
from the violation extra profits of 5 per cent of turnover in the products
concerned over five years. In this example, if only one out of 10 such violations is
detected and punished, the fine required for deterrence would have to exceed 250
per cent of the annual turnover in the products concerned. If only one out of 20
violations is punished, the fine will have to exceed 500 per cent of annual
turnover.
204. Even if the company had retained the extra profits until the fine was
imposed, these profits would pay for only one tenth or one twentieth respectively
of the fine. It is more likely that these profits would not have been retained but
rather would have been paid out in taxes, dividends, salaries and wages.66 Even
liquidating the assets of the company concerned will often be unlikely to generate
64
Korobkin and Ulen, above n 56, at 1089.
65
Apart from these 3 main problems, which are discussed further immediately below, risk-
bearing cost also argues against increasing fines and lowering probability of punishment: see AM
Polinsky and S Shavell, ‘The Optimal Tradeoff between the Probability and Magnitude of Fines’
(1979) 69 American Economic Review 881, L Kaplow, ‘The Optimal Probability and Magnitude of
Fines for Acts that Definitely Are Undesirable’ (1992) 12 International Review of Law and Economics 3,
and also MK Block and JG Sidak, ‘The Cost of Antitrust Deterrence: Why Not Hang a Price Fixer Now
and Then?’ (1990) 68 Georgetown Law Journal 1131. The question could also be raised whether a
strategy of very high fines and low probability of punishment would not pose problems for marginal
deterrence, in that antitrust violators, once they are committing the violation, could no longer be
deterred from making the violation worse, by expanding its scope, duration or intensity, because there
would be no possibility left for threatening them with an ever higher fine if they did so. However, this
concern may not be of much practical importance, as the expansion of the scope, duration or
intensity would increase the risk of detection and punishment, all the more so if one (reasonably)
assumed that competition authorities would, in selecting which violations to prosecute, give priority
to the most serious ones: see generally S Shavell, ‘A Note on Marginal Deterrence’ (1992) 12
International Review of Law and Economics 345.
66
GJ Werden and MJ Simon, ‘Why price fixers should go to prison’ [1987] The Antitrust Bulletin
917 at 928 n 35, referring to empirical estimates that unions are able to capture most of the monopoly
profits earned by US manufacturing firms.
enough revenues to pay the fine. Indeed, many companies’ annual turnover
exceeds their assets. A fine of 250 or 500 per cent of annual turnover in the
products concerned by the violation could thus only be paid out of the assets of
large, diversified companies or companies with very high asset-to-sales ratios.67
205. What would thus really happen if one were to adopt a policy of detecting
and punishing only one out of 10 or one out of 20 violations and impose
correspondingly high fines? If such high fines were really imposed, many of the
companies concerned would be forced into bankruptcy. This would, however,
entail undesirable social costs, because, in the absence of perfect markets, it
would hurt not only managers and shareholders, on whom the bankruptcy may
be considered to have a desirable deterrent effect, but also all other stake-holders
in the firm: employees, suppliers, customers, creditors and tax authorities.68
206. Because of the social costs of bankruptcy, the more likely outcome is that
the high fines will not be imposed, but rather lower ones, corresponding to the
firms’ ability to pay. Indeed, both the European Commission and the US
Department of Justice take into account ability to pay when setting the amount
of fines.69 But this means that the fines really imposed thus end up being lower
than those required to deter.70
3.2.2.2.2 The Social and Economic Costs of High Fines
207. Even if high fines stay below the level of inability to pay, their imposition is
likely to be costly. High fines may have undesirable side-effects. In the absence of
perfect markets, high fines imposed on companies will have an incidence on all
the stake-holders in the firm. Bondholders and other creditors will suffer a
diminution in the value of their securities. Employees may suffer from cost-
cutting campaigns induced by the need to pay the fine. Tax receipts will be
reduced. Finally, consumers may end up suffering. Indeed, if the firm competes in
67
Ibid at 928–9.
68
RH Kraakman, ‘Corporate Liability Strategies and the Costs of Legal Controls’ (1984) 93 Yale
Law Journal 857 at 882.
69
The European Commission’s guidelines on the method of setting fines mentions the undertak-
ing’s ‘inability to pay in a specific social and economic context’ as a factor to be taken into account in
setting the amount of the fine: see Guidelines on the method of setting fines imposed pursuant to Art
23(2)(a) of Reg 1/2003 [2006] OJ C210/2; see further below, Ch 4, in particular paras 341–343.
70
It should be noted that even if the social costs of bankruptcy were disregarded and companies
were penalised into bankruptcy, effective deterrence would not be at all guaranteed. Indeed, the firm’s
inability to pay the fine means that its shareholders can externalise part of the fine risk, as their
liability is limited to their shareholding. The managers’ exposure is similarly limited by the ease with
which they can find alternative employment. If the probability of detection and punishment is
sufficiently low, shareholders and managers may thus decide to run the risk of their firms being fined
into bankruptcy.
a product market characterised by imperfect competition (as will often be the

case), the fine may be partly recovered from consumers in the form of higher
prices.71
3.2.2.2.3 Proportional Justice
208. A strategy of very high fines combined with a low probability of detection
and punishment risks being unacceptable in the light of the principle of propor-
tionality of penalties. This principle has been restated recently in Article 49(3) of
the Charter of Fundamental Rights of the European Union, which provides that
‘the severity of penalties must not be disproportionate to the criminal offence’.72
209. The principle of proportionality of penalties reflects the retributive view
of punishment. Indeed, the utilitarian conception of punishment, which justifies
fines being set at the level required for optimal deterrence at the lowest cost,
competes for the allegiance of the legal system with the retributive view of
punishment. Under the latter view, punishment is not justified by its future
consequence of deterring harmful conduct, but rather on the ground that it is
morally fitting that a person who does wrong should suffer in proportion to his
wrongdoing.73 The law generally reflects a mixture of both conceptions. This is
visible in the Court of Justice’s affirmation that the fines imposed by the
European Commission for violations of Articles 81 and 82 EC ‘have as their
object to punish illegal conduct as well as to prevent it being repeated’.74 The
retributive view imposes a constraint on the pursuit of optimal deterrence in the
form of the principle of proportionality of penalties.75
71
JC Coffee, ‘”No Soul to Damn: No Body to Kick”: an Unscandalized Inquiry into the Problem of
Corporate Punishment’ (1981) 79 Michigan Law Review 387 at 401–2.
72
[2000] OJ C364/1 at 20; see above, para 56. The Charter has since been incorporated as Part II
in the (signed but not ratified) Treaty establishing a Constitution for Europe [2004] OJ C310/41. Art
II-109(3) of this Treaty is the (identical) equivalent of Art 49(3) of the Charter. See also D van Zyl
Smit and A Ashworth, ‘Disproportionate Sentences as Human Rights Violations’ (2004) 67 MLR 541.
73
J Rawls, ‘Two Concepts of Rules’ (1955) 64 Philosophical Review 3 at 4–5.
74
Judgment of 15 July 1970 in Case 41/69 ACF Chemiefarma [1970] ECR 661, para 173 (author’s
translation of the French ‘ont pour but de réprimer des comportements illicites aussi bien que d’en
prévenir le renouvellement’). Similarly, the Commission has stated that ‘the purpose of fines is twofold:
to impose a pecuniary sanction on the undertaking for the violation and prevent a repetition of the
offence, and to make the prohibition in the Treaty more effective’: Thirteenth Report on Competition
Policy 1983 (Brussels, EC Commission, 1984) at para 62.
75
See further HLA Hart, Punishment and Responsibility—Essays in the Philosophy of Law (Oxford,
Clarendon Press, 1968), J Adenaes, ‘The Morality of Deterrence’ (1970) 37 University of Chicago Law
Review 649. JS Parker, ‘Criminal Sentencing Policy for Organizations: The Unifying Approach of
Optimal Penalties’ (1989) 26 American Criminal Law Review 513 at 563–6, defends the view that there
is no real conflict between deterrence and proportionality, because the probability of detection would
reflect a second dimension of harm directly related to the offender’s culpability. By choosing an
offence with a lower probability of detection, the offender would deserve a higher penalty. Parker’s
argument appears unconvincing to the extent that the low probability of detection results from a
policy choice, justified by the saving of administrative costs, to adopt an enforcement strategy of high
penalties and low probability of punishment.
210. Respect for the principle of proportionality of penalties may in fact also
make sense within a utilitarian conception of optimal law enforcement. Indeed,
as explained above,76 the imposition of penalties on antitrust violators not only
has a deterrent effect, in that it helps create a credible threat of punishment for
those who would be willing to commit violations on the basis of a profit
calculation, but also has a moral effect, in that it sends a message to the
spontaneously law-abiding, reinforcing their moral commitment to the rules.
The strength of this moral effect is likely to depend on whether the punishment is
perceived as just, and in particular not disproportionate.77
211. If for instance only one antitrust violation out of 10 or only one out of
20 is detected and punished, the deterrent fine has to exceed the expected gain
from the violation by a factor of 10 or 20 respectively. Such a large multiplication,
justified not solely by the need to achieve effective deterrence but also by a desire
to save administrative costs, risks being unacceptable from a proportional justice
perspective.78
212. The maximum of twice the gross gain as provided for in the US under
the Criminal Fines Improvement Act79 may reflect the limit of what multiplica-
tion is considered acceptable from a proportional justice perspective. In the EU,
Regulation 1/2003 provides that fines imposed by the European Commission
cannot exceed 10 per cent of the total (consolidated) turnover of the company
concerned in the preceding business year.80 This ceiling appears to reflect more
generally concerns with very high fines, not just from the perspective of propor-
tional justice but also as to the risk of inability to pay, and the social and
economic costs of high fines.
3.2.3 Rewarding Cooperation and Efforts at Compliance
213. As explained above, antitrust fines should in principle exceed the expected
gain from the violation multiplied by the inverse of the probability of a fine being
effectively imposed, but the amount may in some cases have to be limited to a
lower level because of the economic and social costs of very high fines and
because of the requirements of proportional justice. The question I now turn to is
whether the amount of the fine should also be affected by the company’s
76
At paras 169 and 171.
77
On the importance of the perception of punishment being just see Tyler, above n 14. See further
also CR Sunstein, D Schkade and D Kahneman, ‘Do People Want Optimal Deterrence?’ (2000) 29
Journal of Legal Studies 237, providing empirical evidence that people reject law-enforcement policies
that increase or decrease penalties because of the probability of detection.
78
See also E David, ‘La détermination du montant des amendes sanctionnant les infractions
complexes: régime commun ou régime particulier?’ (2000) 36 Revue Trimestrielle de Droit européen
511 at 526–7, and J Waldfogel, ‘Criminal Sentences as Endogenous Taxes: Are They “Just” or
“Efficient”?’ (1993) 36 Journal of Law and Economics 139.
79
18 USC § 3571(d) (1984) (as amended in 1987).
80
Art 23(2), second subpara, of Reg 1/2003, above n 6; see further below, paras 338–339 , .
cooperation with the competition authority’s investigation or by the company’s

efforts at complying with the antitrust prohibitions.
3.2.3.1 Cooperation with the Competition Authority’s Investigation
214. When the competition authority investigates an antitrust violation, the

extent to which the company concerned cooperates with the investigation can
make a big difference. If the company does not cooperate at all, the competition
authority will have to use its compulsory powers of investigation to collect the
evidence of the violation, and will have to go through the entire administrative
and judicial proceedings to establish the violation, bring it to an end if it is still
continuing, and impose the fine. At the other extreme, if the company cooperates
fully, it will at the outset of the investigation terminate the violation, provide the
competition authority with all available evidence and admit the violation, thus
dispensing with the need to go through the entire administrative and judicial
proceedings.81
215. Cooperation with the competition authority’s investigation can thus in
two ways be advantageous for antitrust enforcement: first, it reduces the admin-
istrative cost as well as the duration of the investigation and prosecution.
Secondly, if the violation was still continuing at the outset of the investigation,
the cooperation brings the violation to an end earlier.
216. Both of these effects justify a lowering of the fine. Indeed, the saving of
administrative cost will allow the competition authority to redeploy the resources
saved to other investigations, thus increasing the overall probability of detection
and punishment, and allowing a corresponding lowering of the general level of
fines. The shortening of the delay between the time of the violation and the time
of the imposition of the fine reduces the amount of the fine required for
deterrence, because of the lesser need for interest to cover the delay.82 The shorter
duration of the violation itself correspondingly reduces the gain from the
violation, and thus also lowers the required fine.83
81
On the law and economics of the collection of intelligence and evidence of antitrust violations
see WPJ Wils, ‘Self-incrimination in EC Antitrust Enforcement: A Legal and Economic Analysis’
(2003) 26 World Competition 567 and ch 5 of Principles, above n 4; see also Ch 5 below.
82
See above n 41.
83
Under the internalisation approach (see paras 185–193), a similar result is achieved through a
different reasoning. Under that approach, the fine should in principle equal the net harm to persons
other than the offender. One of the components of this harm is the competition authority’s and the
courts’ costs of investigating and punishing the violation. A lowering of these costs because of
cooperation thus automatically lowers the optimal fine. To the extent that the cooperation also limits
the duration of the violation, the optimal fine will also correspondingly go down, as the harm to
consumers will be lower.
217. Lowering the fine because of cooperation is of course only justified if

and to the extent that the cooperation has the beneficial effects described just
above.84
3.2.3.2 Compliance Programmes
218. Should antitrust fines also be reduced to reward companies that put in
place or have put in place compliance programmes? In the US, a well-designed
compliance programme may, in some circumstances, help the company qualify
for sentence mitigation under the sentencing guidelines, so long as the employees
who committed the violation were not ‘high-level personnel’ of the company.85
In the EU, ‘the [European] Commission considers that it is not appropriate to
take the existence of a compliance programme into account as an attenuating
circumstance for a cartel infringement, whether committed before or after the
introduction of such a programme’.86
219. If they reflect a genuine commitment to antitrust compliance at the
highest levels within the company, and are well-designed, compliance pro-
grammes can no doubt be very useful both to prevent antitrust violations and to
detect such violations as early as possible.
220. It is not obvious, however, that this should translate into reduced fines. If
fines are set at the level required for deterrence, namely above the expected gain
from the violation multiplied by the inverse of the probability that a fine will be
imposed, companies will already have all the necessary incentives to prevent the
commitment of antitrust violations.
221. Moreover, as Stephen Calkins has pointed out,
A company has a wide array of ways to increase its compliance with various laws. It can
emphasize the quality of its people, by hiring honest employees, encouraging them to
lead healthy lives, and taking care of them in times of need. It can create good
incentives, by tying compensation to long-term results, by refraining from exerting
undue pressure, and by paying supra-competitive wages employees will not want to risk
84
See Judgment of the ECJ of 16 Nov 2000 in Case C–-297/98 P SCA Holding v Commission
[2000] ECR I–10101, para 36, confirming the Judgment of the CFI of 14 May 1998 in Case T–327/94:
‘[t]he Court of First Instance correctly held in that regard, in paragraph 156 of the contested
judgment, that a reduction in the fine on grounds of cooperation during the administrative
procedure is justified only if the conduct of the undertaking in question enabled the Commission to
establish the existence of an infringement more easily and, where relevant, to bring it to an end’; see
also JM Connor, Global Price Fixing: Our Customers are the Enemy (Boston, Mass, Kluwer Academic
Publishers, 2001), at 387–9, on the risk of unnecessarily generous concessions.
85
US Sentencing Commission, Guidelines Manual (Nov 2005), available at www.ussc.gov/
2005guid/TABCON05.htm, §§ 8B2.1 and 8C2.5; and WJ Kolasky, ‘Antitrust Compliance Programs:
The Government Perspective’, address to the Corporate Compliance 2002 Conference (San Francisco,
12 July 2002), available at www.usdoj.gov/atr/public/speeches/11534.pdf.
86
Dec of 3 Dec 2003 in Case COMP/E-23/38.359 Electrical and mechanical carbon and graphite
products; see also Judgments of the CFI of 20 Mar 2002 in Case T–31/99 ABB Asea Brown Boveri v
Commission [2002] ECR II–1884, para 221, and of 9 July 2003 in Case T–224/00 Archer Daniels
Midland v Commission [2003] ECR II–2597, paras 280–281.
losing. It can teach and remind. It can monitor and audit. And it can threaten with
whatever draconian consequences are in its powers …. Some companies will be better
at one approach, some at another, most at some mix; but it would be surprising were
the same approach right for all. Accordingly, it would seem self-evident that govern-
ment should set out penalties for violating the law and leave it to firms to determine
how best to respond to those penalties.87
3.3 OPTIMAL FINES FOR COLLECTIVE VIOLATIONS
222. Most of the above analysis concerning fines for single offenders applies
equally to fines for collective antitrust violations such as price-fixing or market-
sharing cartels. The case of collective violations is however different in some
important respects.
3.3.1 Additional Possibilities for Using Fines to Prevent Violations
223. As explained above, in the case of a single offender, the necessary condition
for deterrence to work is that, from the perspective of the company contemplat-
ing whether or not to commit a violation, the expected fine (being the nominal
amount of the fine discounted by the probability that it is effectively imposed)
must exceed the expected gain from the violation. How should this be transposed
to the case of collective infringements such as price-fixing or market-sharing
cartels? Should this condition be fulfilled for each of the offenders? Or for the
group of all the offenders taken together? It appears that neither of the two
conditions is necessary, but each is sufficient. If for each member of the
contemplated cartel the expected fine exceeds the expected gain, the cartel will
obviously be deterred. But given that the success of the cartel is likely to require
the participation of most, if not all, of the members, it may be sufficient for
deterrence to work that for some of them the expected fine exceed the expected
gain. On the other hand, if for the group of prospective cartel members taken
together the sum of expected fines exceeds the expected gain, the cartel will also
be deterred, as it will be impossible to distribute the fine risk and the expected
gain between the different cartel members in a way which makes participation
profitable for all of them. But again, deterrence may even work in some situations
where collectively the excepted gain exceeds the sum of expected fines, if for a
sufficient number of participants the expected fine exceeds the expected gain and
the cartel members do not manage to organise the necessary transfers between
them so as to make participation profitable for all prospective participants.
Deterrence is thus conceptually more complicated in the case of collective
87
S Calkins, ‘Corporate Compliance and the Antitrust Agencies’ Bi-Modal Penalties’ (1997) 60
Law and Contemporary Problems 127 at 147.
OPTIMAL FINES FOR COLLECTIVE VIOLATIONS 69
violations than in the case of single offenders, but may already be effective at
lower overall fine levels, if the added complication of the internal dynamics of the
group of violators is well exploited.
3.3.1.1 Raising the Cost of Setting Up and Running Cartels
224. Setting up and running a successful cartel takes effort and requires the
participation of the different cartel members in various respects. A number of
tasks have to be performed: the agreed price or other factors must be determined;
the joint profit must be allocated through quotas or otherwise; and measures
have to be taken to prevent cheating, by developing a sufficiently strong sense of
solidarity and mutual trust and/or by monitoring and punishing deviations. For
all these tasks to be carried out, someone has to take the initiative and someone
has to do the job: someone has to convene the meeting, someone has to make the
price or quota calculations, someone has to do the monitoring and someone has
to do the punishing.88
225. The setting up and running of cartels can be made more difficult (and
hence more costly, and cartels thus less profitable, which means they can be
deterred at lower overall fine levels) by threatening higher fines for those cartel
members who play active roles in setting up and running them, while offering the
perspective of reduced fines for those cartel members who are exclusively passive.
This makes the setting up and functioning of cartels more difficult, because,
faced with the prospect of higher fines, there will be fewer volunteers to play
active roles. Those who nevertheless agree to do so are likely to want compensa-
tion in the form of a larger part of the gain, which is likely to be difficult to agree
on, and to weaken the sense of solidarity and mutual trust within the group.89
226. Fines are indeed used in this way in the different antitrust jurisdictions.
In the EU, for instance, it is well-established in the case law of the Court of Justice
and the Court of First Instance and in the practice of the European Commission
that the role of ringleader or instigator, or the fact of having taken retaliatory
measures against recalcitrant or deviating cartel members, is taken into account
in determining the amount of the fine as aggravating circumstances, regularly
resulting in increases of 50 per cent for ringleaders, whereas an exclusively passive
role is treated as an attenuating circumstance.90
88
To some extent these tasks can be outsourced to specialist service providers. Hence the need for
the activity of such service providers to be covered by the antitrust prohibitions and to be subject to
the threat of sufficiently deterrent fines; see eg the role of AC Treuhand in the organic peroxides
cartel: European Commission Dec of 10 Dec 2003 in Case COMP/E-2/37.857 Organic Peroxides; an
appeal against this decision is pending before the CFI as Case T–99/04.
89
See NK Katyal, ‘Conspiracy Theory’ (2003) 112 Yale Law Journal 1307 at 1341–6 and 1363–7.
90
See sects 4.5.1.3 and 4.5.2.3 below.
3.3.1.2 Leniency
227. The group dynamics of collective violations can be further exploited

through a well-designed leniency policy. Following the example of the US
Antitrust Division Corporate Leniency Policy,91 the European Commission’s
leniency notice92 promises immunity from fines to the cartel member that is the
first to provide the Commission with evidence allowing the Commission to start
a formal investigation or to prove the violation of Article 81 EC, for which the
Commission has not yet got sufficient evidence, and that continues to cooperate
fully with the Commission throughout the whole procedure leading to the
imposition of fines on the other cartel members. The Commission’s leniency
notice also promises fine reductions to those other cartel members who start
cooperating later during the procedure and provide additional valuable evidence,
the percentage of the fine reduction depending on whether they are the first to do
so (reduction of 30 to 50 per cent), the second (20 to 30 per cent) or subsequent
(up to 20 per cent).
228. As in the case of violations committed by a single offender, cooperation
with the competition authority’s investigation is advantageous for antitrust
enforcement, in that it reduces the administrative cost as well as the duration of
the investigation and punishment, and may bring the violation to an earlier end,
if the violation is still continuing. Leniency moreover makes the setting up and
running of cartels more difficult, in that it fractures trust within the cartel, thus
raising the need for costly monitoring and reducing the cartel members’ willing-
ness to share with each other information which could later be used to obtain
leniency.93
3.3.2 Additional Aspects to Consider
3.3.2.1 The Impact of High Fines on the Market Structure
229. Concerning fines for single offenders, it has already been pointed out
above94 that fines which exceed the company’s ability to pay would lead to
bankruptcy, and that, even below the level of inability to pay, the imposition of
high fines may entail substantial social and economic costs. A dimension is added
to this possible problem in the case of collective violations for which fines are
simultaneously imposed on all or most of the companies competing in a given
market.
91
US Department of Justice, Corporate Leniency Policy (10 Aug 1993); see further sect 5.1.2.1
below.
92
Commission notice on immunity from fines and reduction of fines in cartel cases [2006] OJ
C298/17; see further sect 5.1.2.2 below.
93
See Katyal, above n 89, at 1346–55 and 1382–4, and Ch 5 below, in particular paras 393–408.
94
OPTIMAL FINES FOR COLLECTIVE VIOLATIONS 71
230. Imagine by way of illustration that a cartel which has been functioning
very profitably for many years is discovered in some local market, on which five
companies with comparable market shares are active, four of which are not active
in other markets and have limited financial reserves, whereas the fifth is a
subsidiary of a large, financially strong, international group. All five companies
have been equally active in the cartel, and none has cooperated with the
competition authority’s investigation. If one were to focus exclusively on deter-
rence, one would advocate in such a case the imposition of high fines on all five
cartel participants. The risk, however, is that this could lead to the elimination of
the four smaller competitors and the domination of the market by the fifth,
stronger company.
231. In a variation on the same example, it could also be that the fifth,
stronger company is the one which disclosed the existence of the cartel to the
competition authority, and for that reason qualifies for immunity from fines.95
232. To avoid a deterioration in the market structure as a result of the
imposition of the fines, one should thus, in all cases where high fines are imposed
and where there is a significant difference in the ability to pay of the various
cartel members, differentiate the amount of the fines imposed on the different
companies so as to reflect their respective ability to pay. This can be done either
by reducing the fines imposed on the companies with the lower ability to pay, or
by imposing generally lower fines but then increasing them for the companies
with a higher ability to pay. In EC fining practice, such differentation is indeed
made, through the capping of fines at the statutory ceiling of 10 per cent of
overall consolidated turnover of the firms concerned, and also through the use of
so-called ‘multipliers’ to increase the fines imposed on companies whose size far
exceeds that of the other cartel members.96
3.3.2.2 Equal Treatment
233. According to well-established case law of the Court of Justice and Court of
First Instance, the European Commission, when imposing fines, ‘is not entitled to
disregard the principle of equal treatment, a general principle of Community law
95
This was roughly the situation with the Luxembourg beer cartel: see European Commission
Dec of 5 Dec 2001 in Case COMP/37.800/F3 Luxembourg Brewers [2002] OJ L253/21, and judgment
of the CFI of 27 July 2005 in Joined Cases T–49/02 etc Brasserie nationale and Others v Commission
[2005] ECR II–3033; see also para 336 below, referring to the Opinion of AG Geelhoed of 19 Jan 2006
in Case C–289/04 P Showa Denko v Commission (Graphite electrodes) [2006] ECR I–5859, para 61 and
n 16.
96
On the 10% ceiling see para 212 above and paras 338 and 339 below; on the ‘multipliers’, see
paras 334–336 below. In the Luxembourg Brewers case, above n 95, only low fines were imposed
(without much explanation). Obviously such low fines pose a problem for deterrence. The good
solution would have been to send a few managers of the Luxembourg brewing companies to prison,
which would have allowed strong deterrence without causing harm to the structure of the Luxem-
bourg beer market, but this option was not available, since prison sanctions are currently not
provided for in EC law: see above, para 165 and Ch 6 below.
which is infringed … where comparable situations are treated differently or

different situations are treated in the same way, unless such difference is objec-
tively justified’.97 With respect to collective violations, it is thus necessary to
examine ‘the relative gravity of the participation of each [company]’.98
234. This requirement of equal justice is a variation on the requirement of
proportional justice discussed above.99 Indeed, equal treatment is the ‘”relative”
aspect of the test of proportionality’.100 As already argued above, respect for the
principle of proportionality, including the principle of equal treatment, not only
reflects the retributive view of punishment, but also makes sense within a
utilitarian conception of optimal law enforcement, in that the strength of part of
the effects of punishment is likely to depend on whether the punishment is
perceived as just, including non-discriminatory.101
235. The principle of equal treatment could be considered not only in the case
of collective violations, but also in the case of violations committed by a single
offender, by comparison between several violations. In practice, however, the
issue of respect for the principle of equal treatment arises primarily in cases of
collective violations, as the treatment of the different members of the same cartel
provides an obvious base for comparison.102
236. Whereas the ‘absolute’ aspect of proportional justice discussed above
may constitute a real constraint in the determination of the amount of fines, in
that it excludes fines which exceed by a high multiple the gain derived from the
violation or the harm caused by it, to compensate for a low probability of
detection and punishment,103 my impression is that the need to respect the
principle of equal treatment when determining the amounts of the fines for a
collective violation does not usually modify the outcome which one would
anyway reach when modulating the relative amounts of the fines in function of
the more or less active role played by the different cartel members, with a view to
raising the cost of setting up and running cartels,104 and in function of the
relative ability to pay of the different companies, with a view to avoiding a
detrimental impact on the market structure.105
97
Judgment of the CFI of 13 Dec 2001 in Joined Cases T–45/98 and T–47/98 Krupp Thyssen
Stainless and Acciai Speciali Terni v Commission [2001] ECR II–3765, para 237.
98
Judgment of the ECJ of 8 July 1999 in Case C–51/92 P Hercules Chemicals v Commission [1999]
ECR I–4235, para 110.
99
100
See Opinion of AG Tizzano of 8 July 2004 in Case C–189/02 P Dansk Rorindustri v Commission
[2005] ECR I–5425, paras 107–109.
101
102
See also the Opinion of AG Tizzano, above n 100, para 108.
103
104
105
See above, paras 229–232. As regards leniency, the granting of immunity and reductions of
fines strictly on the basis of the timing of the cooperation (immunity only for the first; diminishing
reductions for the second, third and following) is not contrary to the principle of equal treatment, as
all cartel members have equal chances to be the first depending on their own initiative, and as this
practice is in any event objectively justified in order to create a race to cooperate and thus fracture the
CAN THE OPTIMAL FINE BE CALCULATED IN PRACTICE? 73
3.4 CAN THE OPTIMAL FINE BE CALCULATED IN PR ACTICE?
237. Would it be possible in practice for a competition authority or a court, with

the help of the best experts trained in econometrics, to measure or estimate
reliably the optimal fine for a given antitrust violation on the basis of the theory
explained above?
238. Consider first an offence committed by a single offender, for instance the
abuses of a dominant position committed by Microsoft in that it refused to
supply interoperability information and allow its use for the purpose of develop-
ing and distributing work group server operating system products, and made the
availability of the Windows Client PC Operating System conditional on the
simultaneous acquisition of Windows Media Player.106 According to the theory,
in this relatively simple case, where there is no issue of rewarding cooperation
with the investigation or of limiting the amount of the fine because of inability to
pay, the fine should be set so as to exceed the gain which Microsoft expected, at
the time it decided to commit the violation, to obtain from the violation,
multiplied by the inverse of what Microsoft at that time expected to be the
probability that a fine would be imposed. If Microsoft had itself quantified its
estimates at the time and these calculations had been discovered in the investiga-
tion, these could be used. Otherwise it does not appear possible to measure this
econometrically.
239. The task would not be any easier if one adopted the internalisation
approach rather than the deterrence approach.107 Indeed, under the internalisa-
tion approach, the optimal fine would be equal to the net harm caused by the
violation to all persons other than Microsoft (or rather Microsoft’s subjective
estimate of this at the time it decided to commit the infringement), again
multiplied by the inverse of what Microsoft, at the time it decided to commit the
violation, expected to be the probability that a fine would be imposed. The main
component of this net harm is the consumer welfare forgone as a result of the
innovation which was blocked by Microsoft’s violation. It would not seem less
difficult to measure this econometrically.
240. One might argue that it would be easier to calculate the optimal fine for
other types of violations, for example price cartels.108 Unfortunately this does not
appear to be the case. It may be true that for a price cartel it is not impossible
trust within the cartel. The competition authority would however violate the principle of equal
treatment if it somehow interfered with the race to be the first, thus altering the time ranking from
what it would have been if based only on the cooperating companies’ own initiative: see Judgment of
the CFI of 13 Dec 2001 in Joined Cases T–45/98 and T–47/98 Krupp Thyssen Stainless and Acciai
Speciali Terni v Commission [2001] ECR II–3765, paras 237 to 249; see also Ch 5 below, in particular
paras 429 and 437.
106
See Dec of the European Commission of 24 Mar 2004 in Case COMP/C-3/37.792 Microsoft,
and Order of the President of the CFI of 22 Dec 2004 in Case T–201/04 R Microsoft v Commission
[2004] ECR II–4463.
107
See above, paras 185 –192.
108
See Hylton, above n 44, at 47.
(though no doubt difficult and thus costly) to provide reliable estimates of the
harm caused by the cartel, or the gain for all the cartel members taken together,
but this is not the theoretically correct measure. Indeed, neither the actual harm
nor the actual gain is the relevant measure, but the subjectively expected harm or
gain, discounted by the subjectively expected probability that a fine would be
imposed. Moreover, the fine thus calculated would only be the correct measure if
one could assume that the collectivity of all cartel members could be assimilated
to a single decision-maker with a single, undivided self-interest, which is mani-
festly not a realistic assumption. As explained above, cartels are the result of a
dynamic interaction between different self-interested cartel members, and there
is, even at the level of theory, no single formula of what fines for the different
cartel members are optimal to deter the cartel.109 Regard should be had to the
economics and psychology of setting up and running cartels, and fines should
correspondingly be modulated, in particular so as to discourage companies from
playing an active role in the setting up and running of the cartel.110 There is again
no way to measure all this econometrically.
241. Even if it thus appears unfeasible in practice to measure econometrically
the theoretically optimal fine for a given antitrust violation, the theory on
optimal fines certainly remains useful as general guidance for the practice of
fixing the amount of antitrust fines.
242. In some cases, it may happen that at least some of the factors which
determine the theoretically optimal fine, or some reasonably close proxy, can in
fact be measured. When such information is available, it should of course be
taken into account in fixing the amount of the fine. One should however beware
of the risk of giving excessive weight to whatever factor happens to be measur-
able, while ignoring other equally relevant, but less measurable factors. A clear
example of this risk is the one decision in the past 40 years in which the European
Commission was able to measure exactly the profit derived from the antitrust
violation, and then impose this exact figure as the fine, thus disregarding
completely the need to take into account the less than certain probability of
detection and punishment.111
243. One should in my view also avoid any system in which the fixing of the
amount of the fine is based upon the quantification of the gain obtained by the
offender or the harm caused by the antitrust violation. In such a system, the
burden of proof will always be on the competition authority or prosecutor.112
109
110
111
Commission Dec of 25 Mar 1992 in Case IV/30.717-A Eurocheque: Helsinki Agreement [1992]
OJ L95/5.
112
In the EU, according to well-established case law, the European Commission is not obliged to
prove any actual impact on the market in order to be able to impose the necessary fines for antitrust
violations such as price cartels which are anti-competitive by object, but if it chooses to refer to such
effects in setting the amount of the fine, it must prove what it claims: see judgments of the CFI of 15
Mar 2000 in Joined Cases T–25/95 etc Cimenteries CBR and Others v Commission [2000] ECR
CAN THE OPTIMAL FINE BE CALCULATED IN PRACTICE? 75
Given the inherent difficulty in measuring these quantities, and thus also in
proving them to the requisite legal standard, and given also the informational
disadvantage which the authority or prosecutor is likely to have in this regard as
compared to the defendant companies, what can be proven is likely to be
systematically below reality, with fines which are too low and hence under-
deterrence as a result.113 Even assuming that account will also be taken of the
probability of detection and punishment, fixing the amount of the fine on the
basis of ex post values of the gain or harm and probability of punishment is also
likely to lead to under-deterrence because, as explained above, prospective
antitrust violators are likely to overestimate the gain and underestimate the
probability of detection and punishment, as well as the harm.114
II–1570, paras 4862–3; of 19 Mar 2003 in Case T–213/00 CMA CGM and Others v Commission [2003]
ECR II–913, para 280; and of 9 July 2003 in Case T–224/00 Archer Daniels Midland v Commission
[2003] ECR II–2597, paras 148–171. Similarly, the Commission is not required, in order to determine
fines, to establish that the infringement brought an unlawful advantage for the undertakings
concerned, but if it chooses to refer to such gain in setting the amount of the fine, it must prove what
it claims: see Judgments in Joined Cases T–25/95 etc, paras 4881–2; and Case T–213/00, paras 340–3.
113
See also J Davidow, ‘Recent US Antitrust Developments of International Relevance’ (2004) 27
World Competition 407 at 412, commenting at the increase in the US of the maximum antitrust fine
from $10 million to $100 million by the Antitrust Criminal Penalty Enhancement and Reform Act of
2003: ‘[a]t first glance, it is difficult to see why the enforcers need higher fines, since they have
imposed fines of up to $500 million already, by using a “double the damage” approach already set out
in US law. Fines such as those, however, were negotiated. Litigation to prove such large figures would
be expensive and uncertain of being sustained, so the enforcers wish to obtain authority to set a high
fine based on discretion or generalised standards.’
114
4
The European Commission’s 2006
Guidelines on Fines
4.1 SUMMAR Y OF THE GUIDELINES
244. On 1 September 2006, the European Commission published new Guide-

lines on the method it will use when setting fines to be imposed on undertakings
for violations of the prohibitions on restrictive agreements and abuse of a
dominant position laid down in Articles 81 and 82 of the EC Treaty and Articles
53 and 54 of the EEA Agreement (‘the 2006 Guidelines’, ‘the Guidelines’ or ‘the
new Guidelines’).1 These new guidelines replace earlier guidelines dating from
1998 (‘the 1998 Guidelines’).2
245. The Guidelines set out a two-step methodology for the setting of fines.3
First, the Commission will determine a basic amount for each undertaking.
Secondly, it may adjust that basic amount upwards or downwards.
246. The basic amount will be set as a percentage of the value (before tax) of
the undertaking’s sales of goods or services to which the infringement directly or
1
Guidelines on the method of setting fines imposed pursuant to Art 23(2)(a) of Reg 1/2003
[2006] OJ C210/2; see H de Broca, ‘The Commission Revises its Guidelines for Setting Fines in
Antitrust Cases’ [Autumn 2006] Competition Policy Newsletter 1. The Guidelines also deal with fines
imposed on associations of undertakings, but these are not analysed here. On the notion of
‘undertaking’ see below, paras 502–504; WPJ Wils, ‘The Undertaking as Subject of EC Competition
Law and the Imputation of Infringements to Natural and Legal Persons’ (2000) 25 European Law
Review 99; ch 7 of WPJ Wils, The Optimal Enforcement of EC Antitrust Law (The Hague, Kluwer Law
International, 2002); and A Montesa and A Givaja, ‘When Parents Pay for their Children’s Wrongs:
Attribution of Liability for EC Antitrust Infringements in Parent–Subsidiary Scenarios’ (2006) 29
World Competition 555.
2
Guidelines on the method of setting fines imposed pursuant to Art 15(2) of Reg 17 and Art
65(5) of the ECSC Treaty [1998] OJ C9/3. On these earlier guidelines see WPJ Wils, ‘The Commis-
sion’s New Method for Calculating Fines in Antitrust Cases’ (1998) 23 European Law Review 252, and
chr 4 of Wils, The Optimal Enforcement, above n 1. According to point 38 of the 2006 Guidelines, the
latter apply in all cases where a statement of objections is notified after 1 Sept 2006; on the
transitional question see Judgment of the ECJ of 28 June 2005 in Joined Cases C–189/02 P etc Dansk
Rorindustri and Others v Commission (Pre-insulated pipes) [2005] ECR I–5425, paras 198–233;
Judgment of the ECJ of 8 Feb 2007 in Case C–3/06 P Danone v Commission (Belgian beer), not yet
reported, paras 87–92; and Judgment of the CFI of 27 Sept 2006 in Case T–329/01 Archer Daniels
Midland v Commission (Sodium gluconate), not yet reported, paras 374–381; H de Broca, ‘Les
nouvelles Lignes directrices de la Commission en matière d’amendes’ [2006] 4 Concurrences 37, pts
40–47; and E Barbier de la Serre, ‘Les lignes directrices de 2006 pour le calcul des amendes’ [2006] 9
Revue Lamy de la concurrence 14, pts 53–55.
3
Guidelines, above n 1, pts 9–11.
78 THE COMMISSION’S 2006 GUIDELINES ON FINES
indirectly relates in the relevant geographic area within the EEA.4 The Commis-
sion will normally take the sales made by the undertaking during the last full
business year of its participation in the infringement.5
247. The percentage to be applied to the value of sales will depend on the
gravity of the infringement.6 The assessment of gravity will be made on a
case-by-case basis for all types of infringement, taking into account all the
relevant circumstances of the case. As a general rule, the percentage applied on
account of gravity will be up to 30 per cent. To determine the percentage within
this scale, the Commission will have regard to a number of factors, such as the
nature of the infringement, the combined market share of all the undertakings
concerned, the geographic scope of the infringement and whether or not the
infringement has been implemented. For hard-core cartels (horizontal price-
fixing, market-sharing and output-limitation agreements) the percentage will
generally be set at the higher end of the scale.7
248. To take into account the duration of the infringement, the amount thus
determined will be multiplied by the number of years of participation in the
infringement.8 In addition, irrespective of the duration of the undertaking’s
participation in the infringement, the Commission will, for hard-core cartels,
include in the basic amount a sum of between 15 and 25 per cent of the value of
sales. The Commission may also apply such an additional amount in the case of
other infringements.9
249. Once the basic amount of the fine is determined, the Commission may
make adjustments upwards or downwards for a number of reasons. First, the
basic amount may be increased or decreased on account of aggravating or
extenuating circumstances. The Guidelines contain a non-exhaustive list of
aggravating circumstances (recidivism, obstruction of investigations, and role of
leader or instigator of the infringement) and of extenuating circumstances
(including negligence, and authorisation or encouragement by public authorities
4
Ibid, pts 12, 13 and 17. In the case of world-wide market-sharing cartels or other infringements
with a wider geographic scope than the EEA and involving several undertakings, the Commission
may determine the share of the sales of each undertaking in this wider geographic market and apply
this share to the aggregate sales within the EEA, so as better to reflect the relative weight of each
undertaking in the infringement: ibid, pt 18. On this last point see de Broca, above n 1, at 3, and
Judgments of the CFI of 29 Apr 2004 in Joined Cases T–236/01 etc Tokai Carbon and Others v
Commission (Graphite electrodes) [2004] ECR II–1181, paras 196–204, and of 15 June 2005 in Joined
Cases T–71/03 etc Tokai Carbon and Others v Commission (Specialty graphite) [2005] ECR II–10, paras
180–189.
5
Guidelines, above n 1, pt 13.
6
Ibid, pt 19.
7
Ibid, pts 20–23.
8
Ibid, pt 24. Periods of less than 6 months will be counted as half a year; periods of between 6
months and one year will be counted as one year.
9
Ibid, pt 25. In the press release and memorandum issued when the Commission adopted the
2006 Guidelines, this additional amount independent of duration is called an ‘entry fee’: IP/06/857
and MEMO/06/256 of 28 June 2006.
THE PURPOSE OF FINES 79
or by legislation).10 Whereas the increase or decrease on account of other

aggravating or mitigating circumstances is not specified, it is stated that the
increase on account of recidivism will be by up to 100 per cent of the basic
amount for each earlier finding of the same or a similar infringement.11
250. Secondly, the Commission may apply a specific increase for deterrence in
the case of undertakings which have a particularly large turnover beyond the
sales of goods or services to which the infringement relates and/or to ensure that
the fine exceeds the amount of gains improperly made as a result of the
infringement where it is possible to estimate that amount.12
251. Thirdly, whenever the amount thus set exceeds the statutory maximum
of 10 per cent of the total turnover in the preceding business year of the
undertaking concerned, as laid down in Article 23(2) of Regulation 1/2003,13 it
will be capped at that ceiling.14
252. Fourthly, the Commission will apply its Leniency Notice, under which
immunity from fines or reduction of fines is granted in cartel cases to undertak-
ings that have cooperated with the Commission by voluntarily providing intelli-
gence and/or evidence of the infringement, in accordance with the criteria set out
in that notice.15
253. Fifthly, in exceptional cases, the Commission may reduce the fine on
account of the undertaking’s inability to pay in a specific social and economic
context, on the basis of objective evidence that imposition of the normal fine
would irretrievably jeopardise the economic viability of the undertaking con-
cerned and cause its assets to lose all their value.16
254. Finally, the particularities of a given case or the need to achieve deter-
rence in a particular case may justify departing from the methodology set out in
the Guidelines.17
4.2 THE PUR POSE OF FINES
255. Article 83(1) of the EC Treaty instructs the Council to lay down ‘the
appropriate regulations and directives to give effect to the principles set out in
Articles 81 and 82’. Article 83(2)(a) adds that these regulations or directives shall
be designed in particular ‘to ensure compliance with the prohibitions laid down
in Article 81(1) and in Article 82 by making provision for fines’.
10
Guidelines, above n 1, pts 28 and 29.
11
Ibid, pt 28.
12
Ibid, pts 30 and 31.
13
down in Arts 81 and 82 of the Treaty [2003] OJ L1/1, as last amended by Council Reg 1419/2006 of 25
Sept 2006 [2006] OJ L269/1.
14
15
Ibid, pt 34; see further below, para 340 and Ch 5.
16
Ibid, pt 35.
17
Ibid, pt 37. The Commission may, in certain cases, impose a symbolic fine: ibid, pt 36.
256. The Council has done so through Regulation 1/2003, Article 23(2) of
which empowers the Commission to impose fines on undertakings that, either
intentionally or negligently, infringe Article 81 or Article 82 of the EC Treaty.18
257. The Court of Justice has held that such fines ‘have as their objective to
suppress illegal conduct as well as to prevent it being repeated’.19 Similarly, the
Commission has stated that ‘the purpose of the fines is twofold: to impose a
pecuniary sanction on the undertaking for the infringement and prevent a
repetition of the offence, and to make the prohibition in the Treaty more
effective’.20
258. As argued in more detail above,21 the imposition of fines on undertak-
ings that are found to have breached the antitrust prohibitions can in several
ways contribute to the enforcement of these prohibitions. First, the imposition of
fines may have a deterrent effect, in that it creates a credible threat of prosecution
and punishment which weighs sufficiently in the balance of expected costs and
benefits to deter calculating undertakings from committing antitrust violations.22
259. Secondly, in the case of collective violations such as cartels, the differen-
tiation of penalties according to the role played by the different conspirators can
have the effect of raising the cost of setting up and running cartels. Setting up and
maintaining a successful cartel requires effort. The cartel members have to select
18
See above n 13. Reg 1/2003 replaced as of 1 May 2004 the earlier Council Reg 17 of 6 Feb 1962
[1962] JO 13/204, [1959–62] OJ Spec Ed 87, Art 15(2) of which had basically the same content. Art 5
of Reg 1/2003 also provides for fines imposed by the competition authorities of the Member States for
violations of Art 81 or 82 EC: see chs 1 and 4 of WPJ Wils, Principles of European Antitrust
Enforcement (Oxford, Hart Publishing, 2005).
19
Judgment of 15 July 1970 in Case 41/69 Chemiefarma v Commission [1970] ECR 661, para 173
(my translation of the French ‘ont pour objet de réprimer des comportements illicites aussi bien que d’en
prévenir le renouvellement’).
20
EC Commission, Thirteenth Report on Competition Policy 1983 (Brussels, EC Commission,
1983) para 62.
21
At sect 3.1.
22
’Deterrence’ is here understood in the usual sense of trying to eliminate all violations. Another
strand of the economic literature, following GS Becker, ‘Crime and Punishment: An Economic
Approach’ (1968) 76 Journal of Political Economy 169 and WM Landes, ‘Optimal Sanctions for
Antitrust Violations’ (1983) 50 University of Chicago Law Review 652, and including more recently KN
Hylton, Antitrust Law—Economic Theory and Common Law Evolution (Cambridge, Cambridge
University Press, 2003) ch 2, and C Veljanovski, ‘Cartel Fines in Europe: Law, Practice and Deterrence’,
forthcoming in (2007) 30 World Competition 65, uses the term ‘deterrence’ to denote something else,
which I would rather call ‘internalisation’. Whereas under the deterrence approach, the optimal fine
should exceed the expected gain from the violation multiplied by the inverse of the probability of a
fine being effectively imposed, so as to eliminate all violations, under the internalisation approach, the
optimal fine equals the net harm caused to persons other than the offender, again multiplied by the
inverse of the probability of a fine being effectively imposed. The optimal fine thus set makes the
offender internalise all the costs and benefits of the violation, thereby leading the offender to commit
the ‘efficient violations’ the total benefits of which exceed the total costs while deterring ‘inefficient
violations’ the total costs of which exceed the total benefits. See sect 3.2.1.1 above for the reasons why
I believe that the deterrence approach is the better one as far as antitrust fines are concerned; see also
in that sense the study of the French Competition Council: Conseil de la Concurrence, Rapport
annuel 2005, Etude thématique ‘Sanctions, injonctions, engagements, transaction et clémence: les
instruments de la mise en oeuvre du droit de la concurrence’ (Paris, 2006), available at www.conseil-
concurrence.fr/doc/etude_thematique05.pdf, at 103–5.
THE PURPOSE OF GUIDELINES 81
and coordinate their behaviour according to mutually consistent, collusive strat-

egies, allowing the cartel participants as a group to increase profits, and providing
for a fair distribution of profits between them. They also need to develop
mechanisms to discourage cheating, involving monitoring, rewards and punish-
ments. In a dynamic economy, successful cartels may have to develop an
organisational structure that allows them to solve these problems continuously.23
A well-designed fining policy can make these tasks more difficult by imposing
higher penalties for cartel members that play active roles in setting up and
running the cartel, and lower penalties for cartel members that remain exclusively
passive, and by offering deviators the opportunity of applying for leniency.24
260. Thirdly, the public punishment of those who violate the antitrust prohi-
bitions may also have a moral effect, by sending a message to the spontaneously
law-abiding, reinforcing their moral commitment to the antitrust prohibitions.25
Indeed, corporate managers are not necessarily just maximisers of profits for
themselves and their principals. They may feel a moral responsibility to live
within the law whether or not they are likely to be caught, and this normative
commitment could trump their interest calculus.26 Psychological research sug-
gests that normative commitment is generally an important factor explaining
compliance with the law.27
261. Finally, the imposition of penalties may contribute to corrective justice,
in the form of disgorgement of unjust enrichment.28
4.3 THE PUR POSE OF GUIDELINES, AND THE QUESTION OF FORESEEABILIT Y
4.3.1 What are Guidelines For?
262. The fines imposed by the Commission on undertakings that have infringed
Article 81 or 82 EC have as their sole legal basis Article 23(2) of Regulation
23
MC Levenstein and VY Suslow, ‘What Determines Cartel Success?’ (2006) 44 Journal of
Economic Literature 43 at 44–5; see also JE Harrington, ‘How Do Cartels Operate?’ (June 2006),
available at www.econ.jhu.edu/People/Harrington/CartelOperations-6.28.06.pdf.
24
See above, paras 223–228,309–313, and Ch 5, paras 324–327.
25
Criminal Law & Criminology 338 at 341–3; KG Dau-Schmidt, ‘An Economic Analysis of the Criminal
Law as a Preference-shaping Policy’ [1990] Duke law Journal 1; CR Sunstein, ‘On the Expressive
Function of the Law’ (1996) 144 University of Pennsylvania Law Review 2021; DM Kahan, ‘Social
Influence, Social Meaning, and Deterrence’ (1997) 83 Virginia Law Review 349; NK Katyal, ‘Deter-
rence’s Difficulty’ (1997) 95 Michigan Law Review 2385; GE Lynch, ‘The Role of Criminal Law in
Policing Corporate Misconduct’ (1997) 60 Law and Contemporary Problems 23; DM Kahan, ‘Social
Meaning and the Economic Analysis of Crime’ (1998) 27 Journal of Legal Studies 609; and KG
Dau-Schmidt, ‘Preference Shaping by the Law’ in P Newman (ed), The New Palgrave Dictionary of
Economics and the Law (London, Macmillan, 1998) 84.
26
27
28
See above, paras 178 and 182, and below, paras 413 and 414.
1/2003, which only specifies that, for each undertaking, ‘the fine shall not exceed
10 % of its total turnover in the preceding business year’, and that, ‘in fixing the
amount of the fine, regard shall be had both to the gravity and to the duration of
the infringement’.29
263. Regulation 1/2003 does not impose any obligation on the Commission
to publish further guidance on the amount or method of setting fines.30
264. Indeed, for three decades, from the imposition of the first fines in 196931
to the publication of the 1998 Guidelines,32 the Commission imposed fines
without having published any guidance. In its decisions imposing fines the
Commission always listed a number of factors which it had taken into account in
fixing the amount of the fine, but never explained how these factors had brought
it to reach the precise figure of the fine imposed.33 Some information on the
method which the Commission used could be found in judgments of the
Community Courts, as the Commission sometimes revealed details of its calcu-
lation during the appeal procedure, or was obliged by the Courts to do so.34
Sometimes additional information was also provided in the press release or at the
press conference which, at least in major cases, followed the adoption of the
decision by the Commission.35
265. Some addressees of fining decisions claimed before the Community
Courts that this situation amounted to a breach of the requirement, laid down in
Article 253 EC, for Commission decisions to state the reasons on which they are
based. Even if this argument was rejected, the Court of First Instance nevertheless
observed in 1995 in three judgments concerning the Welded Steel Mesh cartel that
it would be ‘desirable’ for the Commission to make more transparent its method
for setting fines.36
29
See Judgment of the ECJ of 28 June 2005 in Joined Cases C–189/02 P etc Dansk Rorindustri and
Others v Commission (Pre-insulated pipes) [2005] ECR I–5425, paras 211–214, and Judgment of the
ECJ of 8 Feb 2007 in Case C–3/06 P Danone v Commission (Belgian beer), not yet reported, paras
24–28. The absence of further precision as to the amount or method of calculation of the fines is not
contrary to the principle that penalties must be defined by law, as contained in Art 7 of the European
Convention on Human Rights (ECHR); see Judgment of the CFI of 5 Apr 2006 in Case T–279/02
Degussa v Commission (Methionine) [2006] ECR II–897, paras 34–98, appeal pending as Case
C–266/06 P Degussa v Commission (Methionine), and Judgment of the CFI of 27 Sept 2006 in Case
T–43/02 Jungbunzlauer v Commission (Citric acid), not yet reported, paras 69–92.
30
By way of comparison, in the UK s 38 of the Competition Act 1998 requires the Office of Fair
Trading to publish guidance as to the appropriate amount of fines: see OFT’s Guidance as to the
Appropriate Amount of a Penalty (London, OFT, Dec 2004), para 1.7.
31
Commission Decs of 16 July 1969 in Quinine [1969] OJ L192/5 and of 24 July 1969 Dyestuffs
[1969] OJ L195/11.
32
See above, text to n 2.
33
For a more detailed discussion of the situation before the 1998 Guidelines see Wils, above n 1,
chs 2 and 4.
34
See for instance Judgment of 7 June 1983 in Joined Cases 100–103/80 Musique Diffusion
française v Commission [1983] ECR 1825, para 103.
35
For instance, press release IP/94/1108 of 30 Nov 1994 concerning the Cement cartel explained
that ‘calculation is normally based on the Community turnover in the products concerned’.
36
Judgments of 6 April 1995 in Cases T–148/89 Tréfilunion v Commission (Welded Steel Mesh)
[1995] ECR II–1119 para 142; T–147/89 Société Métallurgique de Normandie v Commission (Welded
266. Those judgments led to the publication by the Commission of its 1998
Guidelines.37 According to their first paragraph, to which the 2006 Guidelines
also refer,38 the principles outlined in those guidelines ‘should ensure the trans-
parency and impartiality of the Commission’s decisions, in the eyes of the
undertakings and of the Court of Justice alike, while upholding the discretion
which the Commission is granted under the relevant legislation to set fines
within the limit of 10 % of overall turnover. This discretion must, however,
follow a coherent and non-discriminatory policy which is consistent with the
objectives pursued in penalizing infringements of the competition rules’.
267. The Court of Justice has held that, ‘in adopting [guidelines] and
announcing by publishing them that they will henceforth apply to the cases to
which they relate, [the Commission] imposes a limit on the exercise of its
discretion and cannot depart from those rules under the pain of being found,
where appropriate, to be in breach of the general principles of law, such as equal
treatment or the protection of legitimate expectations’.39 While the Guidelines
‘may not be regarded as rules of law which [the Commission] is always bound to
observe, they nevertheless form rules of practice from which [the Commission]
may not depart in an individual case without giving reasons that are compatible
with the principle of equal treatment’.40
268. According to settled case law, the EC Treaty and Regulation 1/2003 leave
the Commission ‘a particularly wide discretion as regards the choice of factors to
be taken into account for the purposes of determining the amount of the fines,
…, without the need to refer to a binding or exhaustive list of criteria which must
be taken into account’.41 A degree of discretion is indeed inevitable, given the
diversity of cases and the wide range of factors relevant for setting optimal
fines.42
Steel Mesh) [1995] ECR II–1061 and T–151/89 Société des Treillis et Paneaux Soudés v Commission
(Welded Steel Mesh) [1995] ECR II–1195. In the Judgment of 14 May 1998 in Case T–309/94 KPN v
Commission (Cartonboard) [1998] ECR II–1007, paras 77–78, the CFI appeared to go further,
implying an obligation for the Commission fully to set out its calculation method in its decisions, but
this was reversed on appeal by the ECJ in the Judgment of 16 Nov 2000 in Case C–248/98 P KPN v
Commission (Cartonboard) [2000] ECR I–9641, paras 44–48.
37
See above n 2.
38
2006 Guidelines, above n 1, pt 3.
39
Judgment of the ECJ of 28 June 2005 in Joined Cases C–189/02 P etc Dansk Rorindustri and
Others v Commission (Pre-insulated pipes) [2005] ECR I–5425, para 211.
40
Judgment of the ECJ of 18 May 2006 in Case C–397/03 P Archer Daniels Midland v Commission
(Amino acids) [2006] ECR I–4429, para 91.
41
Judgment of the ECJ of 29 June 2006 in Case C–289/04 P Showa Denko v Commission (Graphite
electrodes) [2006] ECR I–5859, para 36, and Judgment of the ECJ of 8 Feb 2007 in Case C–3/06 P
Danone v Commission (Belgian beer), not yet reported, para 37.
42
See Judgment of the ECJ of 7 June 1983 in Joined Cases 100–103/80 Musique Diffusion
Française [1983] ECR 1825, paras 105 and 106; Ch 3 above; and Aristotle, Politics, Bk III, Ch 17:
‘[s]ome things can, and others cannot, be comprehended under the law, and this is the origin of the
vexed question whether the best law or the best man should rule. For matters of detail about which
men deliberate cannot be included in legislation.’
269. The Commission’s discretion is however not unlimited. According to

equally settled case law of the Community Courts, the Commission, when
imposing fines, ‘is not entitled to disregard the principle of equal treatment, a
general principle of Community law which is infringed … where comparable
situations are treated differently or different situations are treated in the same
way, unless such difference is objectively justified’.43
270. The Guidelines can be useful in ensuring respect for the principle of
equal treatment. As the Court of First Instance has stated, ‘while the method of
calculating the amount of fines contained in the Guidelines is, admittedly, not the
only permissible method, it is capable of ensuring a coherent decision-making
practice in relation to the imposition of fines, which in turn guarantees equality
of treatment for undertakings which are penalised for infringements of the rules
of competition law’.44 The Guidelines make it easier for the Commission itself to
follow a consistent fining policy, and to resist pressure for unjustified special
treatment in individual cases. It makes it easier for the addressees of fining
decisions to understand why the fine was set at the level it was, thus possibly
reducing the number of appeals.45 The fact that fines are seen to be imposed in a
consistent way may also reinforce the moral effects of their imposition.46 Finally,
it facilitates the task of the Community Courts in verifying the respect of the
principle of equal treatment by the Commission.47
271. However, because of the diversity of cases and the wide range of factors
relevant to setting optimal fines,48 and because the principle of equal treatment
also requires different cases to be treated differently,49 a degree of discretion has
to be retained, and guidelines cannot provide full foreseeability of the exact
43
Judgment of the CFI of 13 Dec 2001 in Joined Cases T–45/98 and T–47/98 Krupp Thyssen
Stainless and Acciai Speciali Terni v Commission (Alloy surcharge) [2001] ECR II–3765, para 237; see
further Judgment of the ECJ of 31 Mar 1993 in Joined Cases C–89/85 etc Ahlström Osakeyhtiö and
Others v Commission (Wood Pulp) [1993] ECR I–1629, paras 196–197; Judgment of the ECJ of 21 Sept
2006 in Case C–167/04 P JCB Service v Commission, not yet reported, para 205; and Opinion of AG
Bot of 1 Mar 2007 in Case C–76/06 P Britannia Alloys v Commission (Zinc phosphate), not yet
reported, paras 99–100 and 176–180; see also below, nn 89 and 119, the text to n 101, and sect 2.3.2.2
above.
44
Judgment of the CFI of 25 Oct 2005 in Case T–38/02 Groupe Danone v Commission (Belgian
Beer) [2005] ECR II–4407, para 523.
45
See JRM Killick, ‘Viewpoint: The 2006 Fining Guidelines: Two Steps Forward but One Step
Back?’ (Nov 2006), available at http://eccp.esapience.org/index.php?&id=60&action=907, at 9, and
the Welded Steel Mesh judgments, above n 36. In the same logic, it may lead more undertakings to
accept settlements if, as is currently being studied, the Commission were to introduce a settlement
mechanism for antitrust cases involving fines: see above, paras 159–162 above and below, paras 376
and 377.
46
See above, para 260, and Tyler, above n 27 above.
47
See Judgment in Case T–38/02, above n 44.
48
See above, text to nn 41–42.
49
amount or level of fines. As the Court of First Instance has stated, ‘[t]he objective
of the Guidelines is therefore transparency and impartiality, and not the foresee-
ability of the level of the fines’.50
4.3.2 How Foreseeable Should the Amount of Fines Be?
272. There appears to be a wide divergence of opinion as to how foreseeable or

predictable fines should be. It is often argued, or implied, that the more
predictable fines are, the better.51 On the other hand, enforcement officials,
authorities and courts have often warned against the risk of too high a degree of
foreseeability.52
273. As already mentioned, the diversity of cases and the wide range of factors
relevant for setting optimal fines make it impossible to draft guidelines which
lead simultaneously to appropriate amounts of fines and to full predictability of
these amounts in all cases.53 Attempts to achieve full foreseeability would thus
inevitably lead to under-deterrence in some instances and/or disproportionately
high fines in other instances.54
274. Even if this problem were not to arise, the relationship between the
foreseeability of fines and their deterrent and other enforcement effects would
not be entirely straightforward.55 At a first level of analysis, it would appear that
the logic of deterrence through fines is to induce undertakings to refrain from
committing antitrust violations by altering the potential offender’s balance of
expected cost and benefit, and it should thus be fully foreseeable that fines would
50
Judgment of the CFI of 15 March 2006 in Case T–15/02 BASF v Commission (Vitamins) [2006]
ECR II–497, para 250; see also Judgment of the CFI of 27 Sept 2006 in Case T–43/02 Jungbunzlauer v
Commission (Citric acid), not yet reported, para 84.
51
See for instance the arguments of Degussa, as reported in paras 34 ff of the Judgment of the CFI
of 5 Apr 2006 in Case T–279/02 Degussa v Commission (Methionine), not yet reported, appeal pending
as Case C–266/06 P Degussa v Commission (Methionine), and those of Archer Daniels Midland, as
reported in para 49 of the Judgment of the CFI of 27 Sept 2006 in Case T–329/01 Archer Daniel
Midland (Sodium gluconate), not yet reported; and Killick, above n 45.
52
See for instance L Gyselen, ‘The Commission’s Fining Policy in Competition Cases—“Questo è
il catalogo’’’ in PJ Slot and A McDonnell (eds), Procedure and Enforcement in E.C. and U.S.
Competition Law (London, Sweet & Maxwell, 1993) 63 at 64; House of Lords, Select Committee on
the European Communities, Enforcement of Community Competition Rules, Session 1993–94, 1st
Report, at 23, para 62; Opinion of Judge Vesterdorf acting as AG in Case T–1/89 Rhône-Poulenc v
Commission (Polypropylene) [1991] ECR II–869 at 1027; Judgment of the CFI of 5 Apr 2006 in Case
T–279/02 Degussa v Commission (Methionine), not yet reported, para 83, appeal pending as Case
C–266/06 P Degussa v Commission (Methionine), and Judgment of the CFI of 27 Sept 2006 in Case
T–43/02 Jungbunzlauer v Commission (Citric acid), not yet reported, para 84.
53
See above, text to nn 41–42 and 48–50.
54
See also G Des Rosiers, ‘Hear No Evil, See No Evil, Speak No Evil? The Effects of Guidelines
Sentencing on the Behavior of Corporations and their Insiders’ (1997) 18 Research in Law and
Economics 65 at 70 and CS Diver, ‘The Optimal Precision of Administrative Rules’ (1983) 93 Yale Law
Journal 65 at 78; on the requirement to respect the principle of proportionality see above, paras
208–212 and 233–236.
55
On the different enforcement effects of fines see above, paras 258–261.
have this effect.56 There are however reasons to qualify this conclusion. First,
deterrence does not require that the expected fine (discounted for the probability
of detection and punishment) equals the expected illicit gain, but rather that the
expected fine exceeds the expected gain by a sufficient safety margin.57 Secondly,
because the marginal hurtfulness of fines increases with their amount, especially
when bankruptcy or severe financial distress becomes a possibility, undertakings
(or the individuals taking the decisions inside them) tend to be risk averters. This
implies that a less determinate structure for setting fines generates more deter-
rence.58 Thirdly, to the extent that fines may not be high enough fully to deter,59
precision as to the amount of fines to be expected may lead some undertakings
that would otherwise have been law-abiding without making any calculations to
make a cost-benefit calculation and to conclude that they have an interest in
committing antitrust violations.60 More generally, excessive precision as to the
amount of the fines is likely to weaken the moral effects of the imposition of
fines.61 Fourthly, in the case of cartels, differentiation of penalties depending on
the role played by each of the conspirators has the effect of raising the cost of
creating and maintaining cartels.62 Uncertainty as to the precise amount of the
fines increases this effect, as it becomes more difficult for the conspirators to
come to an agreement on who should bear what risks and for what reward.63
56
See Wils, above n 1, at 73.
57
See Opinion of AG Geelhoed of 19 Jan 2006 in Case C–289/04 P Showa Denko v Commission
(Graphite electrodes), not yet reported, para 58; Judgment of the CFI of 27 Sept 2006 in Case T–329/01
Archer Daniels Midland v Commission (Sodium gluconate), not yet reported, para 49; see also above n
22, concering the distinction between the deterrence approach and the internalisation approach to
fines; and Wils, above n 1, at 22–24.
58
JC Coffee, ‘Corporate Crime and Punishment: A Non-Chicago View of the Economics of
Criminal Sanctions’ (1980) 17 American Criminal Law Review 419 at 430–1, who also points out that
the opposite is to true for sentences of imprisonment.
59
On the question to what extent deterrence can be achieved by imposing only fines on
undertakings see below, Ch 6, and Wils, above n 1, ch 8.
60
See Des Rosiers, above n 54, at 72; CS Diver, above n 54, at 78; F Arbault, ‘La politique de la
Commission en matière d’amendes antitrust: récents développements, perspectives d’avenir’ (Sum-
mer 2003) Competition Policy Newletter 1 at 6.
61
See C Harding, ‘Effectiveness of Enforcement and Legal Protection’, paper presented at the EUI
2006 EU Competition Law and Policy Workshop, forthcoming in C-D Ehlermann and I Atanasiu,
European Competition Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart Publish-
ing, 2007), available at www.iue.it/RSCAS/Research/Competition/2006(papers).shtml, at 4–5; Wils.
‘Optimal Antitrust Fines: Theory and Practice’, Ch 3 above, n 00, at 193; and above, text to nn 25–27.
62
See above, para 259, and below, paras 309–313 and 324–327.
63
NK Katyal, ‘Conspiracy Theory’ (2003) 112 Yale Law Journal 1307 at 1342–3.
BASIC AMOUNT OF THE FINE 87
4.4 BASIC AMOUNT OF THE FINE
4.4.1 The New Method
275. As described above,64 the Guidelines contain a two-step methodology for

the setting of fines. The first step consists in the determination of a basic amount
for each undertaking. The basic amount is set as a percentage of the value (before
tax) of the undertaking’s sales of goods or services to which the infringement
directly or indirectly relates in the relevant geographical area within the EEA.65
The Commission will normally take the sales made by the undertaking during
the last full business year of its participation in the infringement. The percentage
to be applied to the value of sales will depend on the gravity of the infringement
and on its duration. The assessment of gravity will be made on a case-by-case
basis for all types of infringement, taking into account all the relevant circum-
stances of the case. As a general rule, the percentage applied on account of gravity
will be up to 30 per cent. To determine the percentage within this scale, the
Commission will have regard to a number of factors, such as the nature of the
infringement, the combined market share of all the undertakings concerned, the
geographical scope of the infringement and whether or not the infringement has
been implemented. For hard-core cartels (horizontal price-fixing, market-sharing
and output-limitation agreements) the percentage will generally be set at the
higher end of the scale. To take into account the duration of the infringement, the
amount thus determined will be multiplied by the number of years of participa-
tion in the infringement. In addition, irrespective of the duration of the under-
taking’s participation in the infringement, the Commission will, for hard-core
cartels, include in the basic amount a sum of between
15 and 25 per cent of the value of sales. The Commission may also apply such an
additional amount in the case of other infringements.
4.4.2 Comparison with Earlier Methods
276. Although the Commission had not published any guidance before 1998,66
it had in fact progressively developed a method of calculating fines as a percent-
age (varying between two and nine, depending on the gravity and the duration of
the infringement) of the annual turnover in the product and geographical market
concerned by the infringement.67
64
65
See above n 4.
66
67
See Wils, above n 1, at 31–4 and 70–2.
277. Under the 1998 Guidelines,68 the determination of the basic amount of
the fine required as a first step that the infringement be classified as either minor,
serious, or very serious, depending on its nature, its actual impact on the market,
where this could be measured, and the size of the relevant geographical market.
‘Likely fines’ were between €1,000 and €1 million for minor infringements,
between €1 and €20 million for serious infringements, and ‘above 20 million €’
for very serious infringements. Within these ranges, the basic amount of the fine
was set, taking into account ‘the nature of the infringement’ and the need to
ensure ‘a sufficiently deterrent effect’. For infringements involving several under-
takings, such as cartels, there was a possibility of differentiating between these
undertakings to reflect ‘considerable disparity between the sizes of the undertak-
ings’. Except for infringements of ‘short duration (in general, less than one year)’,
the basic amounts thus determined were then increased on account of duration,
up to 50 per cent in the case of ‘medium duration (in general, one to five years)’,
and ‘up to 10 % per year’ in the case of ‘long duration (in general, more than five
years)’.
278. The 1998 Guidelines were contested before the Community Courts with
the argument that they violated the principle of equal treatment in that they were
not based on the turnover in the affected market. The Court of Justice has
strongly rejected this argument, stating that:
since the calculation method recommended by the Guidelines envisages that numerous
factors will be taken into account in assessing the gravity of the infringement for the
purpose of determining the amount of the fine, including in particular the profits
secured by the infringement or the need to ensure the deterrent effect of the fines, it
seems to correspond better with the principles laid down by [Regulation 1/2003] as
interpreted by the Court of Justice, notably in its judgment in Musique Diffusion
française and Others v Commission, than what was alleged to be the Commission’s
earlier practice, referred to by the applicants, in which the relevant turnover played a
predominant and relatively mechanical role.69
279. Indeed, in Musique Diffusion française, the Court of Justice held that:
in assessing the gravity of an infringement regard must be had to a large number of
factors, the nature and importance of which vary according to the type of infringement
in question and the particular circumstances of the case. These factors may, depending
on the circumstances, include the volume and value of the goods in respect of which
the infringement was committed and the size and economic power of the undertaking
and, consequently, the influence which the undertaking was able to exert on the market.
It follows that, on the one hand, it is permissible, for the purpose of fixing the fine, to
68
Above n 2.
69
Others v Commission (Pre-insulated Pipes) [2005] ECR I–5425, para 260; see also Judgment of 18 May
2006 in Case C–397/03 P Archer Daniels Midland v Commission (Amino acids), not yet reported, para
101: ‘Community law contains no principle that the penalty be proportionate to the undertaking’s
size on the product market in respect of which the infringement was committed’.
have regard both to the total turnover of the undertaking, which gives an indication,
albeit approximate and imperfect, of the size of the undertaking and of its economic
power, and to the proportion of that turnover accounted for by the goods in respect of
which the infringement was committed, which gives an indication of the scale of the
infringement. On the other hand, it follows that it is important not to confer on one or
the other of these figures an importance disproportionate in relation to the other
factors.70
280. In practice, the Commission made extensive and increasing use of the
ability under the 1998 Guidelines to differentiate between the different undertak-
ings involved in the same cartel, because of significant disparity in their size, so as
to make the starting amounts of the fines for each undertaking roughly propor-
tional to its respective turnover in the affected market. This practice of ‘group-
ings’ (undertakings being grouped into categories of comparable turnover) has
been accepted by the Courts.71 It also appears that, at least to some extent, the
Commission in fact took into account the size of the product market in the
determination of the basic amount of the fine on account of gravity.72
281. It would thus appear that, leaving aside the way in which duration is
taken into account, the new method for determining the basic amount of the fine
is not very different from the way in which the Commission has come to apply
the 1998 Guidelines in practice.73
282. It is also interesting to compare the new method for determining the
basic amount of fines with the method set out in the Guidelines for the Setting of
Fines published by the Netherlands Competition Authority on 21 December
2001.74 In their introduction, these Dutch guidelines state that ‘the [1998]
Guidelines drawn up by the European Commission cannot be taken as the point
70
Judgment of the ECJ of 7 June 1983 in Joined Cases 100–103/80 Musique Diffusion française v
Commission [1983] ECR 1825, paras 120–121. In its Judgment of 7 Jan 2004 in Joined Cases C–204/00
P etc Aalborg Portland and Others v Commission (Cement) [2004] ECR I–123, para 91, the ECJ has
stated more positively that ‘objective factors such as … the extent of the market affected … must be
taken into account’ in setting the fine. ‘The analysis must also take into account the relative
importance and market share of the undertakings responsible’.
71
See, inter alia, Judgments of the CFI of 19 Mar 2003 in Case T–213/00 CMA CGM and Others v
Commission (FETTCSA) [2003] ECR II–913, para 385, and of 29 Apr 2004 in Joined Cases T–236/01
etc Tokai Carbon and Others v Commission (Graphite electrodes) [2004] ECR II–1181, para 217, and
Judgment of the ECJ of 29 June 2006 in Case C–308/04 P SGL Carbon v Commission (Graphite
electrodes), not yet reported, paras 54–56.
72
See for instance Commission Decision of 21 Nov 2001 in Vitamins [2003] OJ L6/1, para 675:
‘[t]he Commission, moreover, for the purposes of determining the starting amounts of the fines,
takes into consideration the size of each of the different vitamin markets’; see also Judgments of the
CFI of 15 Mar 2006 in Case T–15/02 BASF v Commission (Vitamins) [2006] ECR II–497, para 134,
and of 27 Sept 2006 in Case T–322/01 Roquette Frères v Commission, not yet reported, paras 147–150.
73
This impression of continuity is reinforced by pt 26 of the 2006 Guidelines, above n 1, which
states that, ‘where the value of sales by undertakings participating in the infringement is similar but
not identical, the Commission may set for each of them an identical basic amount. Moreover, in
determining the basic amount of the fine, the Commission will use rounded figures’. See further
Barbier de la Serre, above n 2, pts 1–23.
74
Available at www.nmanet.nl/Images/14_10518_tcm16–75169.pdf, and analysed in M van Oers
and B van der Meulen, ‘The Netherlands Competition Authority and its Policy on Fines and Leniency’
of departure without adaptation. The European Commission uses categories of

infringements, in accordance with the [1998] Guidelines, to which fixed fines
apply. A disadvantage of a system of fixed fines is that small undertakings are
affected relatively more harshly than larger undertakings.’ Instead the Dutch
guidelines provide that the basic amount of the fine is calculated as a percentage
of the turnover of the undertaking in the affected market and throughout the
entire duration of the infringement. The percentage depends on the seriousness
of the infringement, with a value of between 15 and 30 per cent for very serious
infringements. This appears in effect very similar to the Commission’s new
method in the 2006 Guidelines.75
4.4.3 Analysis
283. The 2006 Guidelines themselves explain the choice of the new method for
determining the basic amount of the fine as follows: ‘[t]he combination of the
value of sales to which the infringement relates and of the duration of the
infringement is regarded as providing an appropriate proxy to reflect the eco-
nomic importance of the infringement as well as the relative weight of each
undertaking in the infringement’.76
284. As I have argued in more detail above,77 deterrence, at least in the case of
antitrust violations committed by a single undertaking,78 requires that, from the
(2003) 26 World Competition 25; see also Opinion of AG Tizzano of 8 July 2004 in Case C–189/02 P
Dansk Rorindustri v Commission (Pre-insulated Pipes) [2005] ECR I–5425, n 59.
75
There are 3 relatively minor differences between the Commission’s new method and the Dutch
method for determining the basic amount: (1) the Commission’s new method no longer bothers with
classifying infringements as either minor, serious or very serious; as has been pointed out by de Broca,
above n 1, at 1, this classification appears to be a largely unnecessary step in practice; (2) whereas the
Netherlands Competition Authority takes as a starting point the relevant turnover throughout the
entire duration of the infringement, the Commission will normally take the sales in the last full
business year of the undertaking’s participation in the infringement, and multiply this figure by the
number of years of the infringement: see further below, para 289; (3) at least for hard-core cartels, the
Commission will also add an amount of between 15 and 25% of annual turnover, irrespective of the
duration of the infringement: see further below, para 290.
76
Above n 1, pt 6.
77
Ch 3, in particular para 184.
78
In the case of collective violations such as cartels, the situation is more complicated: if for each
member of the contemplated cartel the expected fine exceeds the expected gain, the cartel will
obviously be deterred. But given that the success of the cartel is likely to require the participation of
most, if not all, of the members, it may be sufficient for deterrence to work that for some of them the
expected fine exceeds the expected gain. On the other hand, if for the group of prospective cartel
members taken together the sum of expected fines exceeds the expected gain, the cartel will also be
deterred, as it will be impossible to distribute the fine risk and the expected gain between the different
cartel members in a way which makes participation profitable for all of them. But again, deterrence
may even work in some situations where collectively the excepted gain exceeds the sum of expected
fines, if for a sufficient number of participants the expected fine exceeds the expected gain and the
cartel members do not manage to organise the necessary transfers between them so as to make
participation profitable for all prospective participants. Deterrence is thus conceptually more compli-
cated in the case of collective violations than in the case of single offenders, but may already be
perspective of the undertaking contemplating whether or not to commit a

violation, the expected fine, discounted by the probability that a fine is effectively
imposed, exceeds the expected gain from the violation. For deterrence to work,
fines should thus in principle be set at a level exceeding the expected gain from
the violation multiplied by the inverse of the probability of a fine being effectively
imposed.
285. It would not appear possible in practice for a competition authority or a
court, even with the help of the best experts trained in econometrics, to measure
or estimate reliably this theoretically correct minimum fine for a given antitrust
violation.79 The first problem is that the potential offender’s subjective ex ante
estimates of the gain and the probability of detection and punishment do not
necessarily correspond to what is objectively measurable ex post.80 As for the
probability of detection and punishment, its ex post estimation would anyway
appear extremely difficult, if not impossible, because by definition only the
detected antitrust violations are known to the authorities, not the undetected
ones. As for the gain from the violation, its ex post estimation may in principle be
possible, but quantification of the gain in each case is administratively costly and
likely to lead to under-deterrence. Indeed, the burden of proof will always be on
the competition authority, and, given the inherent difficulty in quantifying the
gain and thus also in proving it to the requisite legal standard, and given also the
informational disadvantage which the authority is likely to have in this regard as
compared to the defendant undertakings, what can be proven is likely to remain
systematically below reality.81
286. If it would thus not be a good idea to try to estimate the theoretically
correct minimum fine for each given antitrust violation, the insight that, for
deterrence to work, fines should in principle be set at a level exceeding the
expected gain from the violation multiplied by the inverse of the probability of a
fine being effectively imposed, may still be helpful in finding a workable method
of setting fines. The expected gain from the violation can safely be assumed to be
positively correlated with the undertaking’s turnover in the affected market
throughout the duration of the infringement. The turnover is of course not the
only relevant factor. In the paradigmatic case of a price cartel, the illicit gain of
the cartel as a whole will depend not only on the size of the affected market and
the duration of the cartel, but also on how much the cartel has been able to raise
the price, and the extent to which this price rise has depressed demand for the
effective at lower overall fine levels, if the added complication of the internal dynamics of the group of
violators is well exploited: see further Ch 3 above, in particular para 223.
79
See also Opinion of AG Geelhoed of 19 Jan 2006 in Case C–289/04 P Showa Denko v
Commission (Graphite electrodes) [2006] ECR I–5859, paras 57–60.
80
See Ch 3 above, in particular paras 194–198 and 237–243.
81
Ibid, in particular para 243.
cartel members’ products or services. Moreover, not only the gain but also the
probability of detection and punishment should be taken into account in setting
the fine.
287. The Guidelines allow for these other factors to be taken into account
within the wide range of ‘up to 30%’ for the proportion taken of the affected
turnover.82 The Guidelines provide that, ‘in order to decide whether the propor-
tion of the value of sales to be considered in a given case should be at the lower
end or at the higher end of that scale, the Commission will have regard to a
number of factors, such as the nature of the infringement, the combined market
share of all the undertakings concerned, the geographic scope of the infringe-
ment and whether or not the infringement has been implemented’. For hard-core
cartels, ‘which are usually secret’, the percentage ‘will generally be set at the
higher end of the scale’.83
288. The ‘nature of the infringement’ is obviously relevant for assessing both
its profitability and its likelihood of detection. The combined market share of all
the undertakings concerned and the geographical scope of the infringement are
both relevant for the ability of, for instance, a price cartel to raise prices without
losing too much custom to outsiders. Whether the infringement has been
implemented will obviously affect its profitability.84 Finally, it also makes perfect
sense to set higher fines for secret infringements, as their probability of detection
and punishment is lower.
289. As already pointed out above,85 the Commission’s 2006 Guidelines,
unlike the guidelines of the Netherlands Competition Authority, do not calculate
the basic amount of the fine as a percentage of the value of the relevant sales
82
In this respect the new method for determining the basic amount of the fine in the 2006
Guidelines is very different from, and a huge improvement over, the method which the Commission
used before 1998, under which a narrow range of 2 to 9% was applied to the relevant turnover during
one business year only (thus not multiplied by the number of years of duration of the infringement),
thereby not allowing for duration and other relevant factors other than the size of the affected market
to be taken duly into account: see above, para 276, and Wils, above n 1, at 32–4.
83
Above n 1, pts 21–23.
84
In the 1998 Guidelines, one of the criteria for assessing the gravity of the infringement and
determining the basic amount of the fine was the ‘actual impact [of the infringement] on the market,
where this can be measured’: see above, para 277. Basing the setting of the fine on an assessment of its
actual impact on the market is not a good idea, for the same reason as basing it on an assessment of
the actual illicit gain is not a good idea: see above, para 285. Indeed, given that such an assessment is
inherently difficult, that the burden of proof will always be on the competition authority, and that the
authority will be at an informational disadvantage compared to the defendant undertaking, such a
system is administratively expensive and bound to lead to assessments below reality, and thus to
under-deterrence: see above, paras 237–243. The CFI has however held that the Commission can
presume, in cartel cases, that the agreement had an actual impact on the market if it can establish that
the agreement has been implemented: see Judgment of the CFI of 25 Oct 2005 in Case T–38/02
Groupe Danone v Commission (Belgian Beer) [2005] ECR II–4407, para 148 and Judgment of the CFI
of 14 Dec 2006 in Joined Cases T–259/02 etc Raiffeisen Zentralbank Österreich v Commission (Austrian
banks—‘Lombard Club’), not yet reported, paras 283–288. This then leads in practice to the same
result as now under the 2006 Guidelines, where the stated criterion is ‘whether or not the
infringement has been implemented’.
85
Above n 75.
ADJUSTMENTS TO THE BASIC AMOUNT 93
throughout the duration of the infringement, but rather as a percentage of the

value of sales in one year (’normally’ the last full business year of the undertak-
ing’s participation in the infringement) multiplied by the number of years of
participation in the infringement. This approach makes the determination of the
basic amount easier for the Commission, reduces the scope for manipulation by
the defendant undertakings (which are of course in this matter at an informa-
tional advantage compared to the Commission), and for unproductive litigation
before the Court of First Instance. The Guidelines state that the Commission will
‘normally’ take the last full business year of the undertaking’s participation in the
infringement, as it cannot be ruled out that in some special cases the figure for
that year could be particularly unrepresentative.
290. Finally, the Guidelines also provide that, irrespective of the duration of
the undertaking’s participation in the infringement, the Commission will, for
hard-core cartels, include in the basic amount a sum of between 15 and 25 per
cent of the value of sales, and may also do so for other infringements.86 As the
Guidelines themselves explain, the rationale is ‘to deter companies from even
entering into illegal practices’.87
4.5 ADJUST MENTS TO THE BASIC AMOUNT
291. Once the basic amount of the fine has been determined, the second step in
the methodology set out in the Guidelines allows for the Commission to take into
account ‘circumstances that result in an increase or decrease in the basic amount
of the fine’, ‘on the basis of an overall assessment which takes account of all the
relevant circumstances’.88
4.5.1 Aggravating Circumstances
292. The basic amount may be increased where the Commission finds that there
are aggravating circumstances, of which the Guidelines give three examples.89
86
87
Above n 1, pt 7.
88
89
Ibid, pt 28. In its Judgment of 29 June 2006 in Case C–308/04 P SGL Carbon v Commission
(Graphite electrodes), not yet reported, para 71, the ECJ held that, ‘whereas the basic amount of the
fine is set according to the infringement, its gravity is determined by reference to numerous factors, in
respect of which the Commission has a wide discretion. To take into account aggravating circum-
stances when setting the fine is consistent with the Commission’s task of ensuring compliance with
the competition rules.’ According to settled case law, ‘the mere fact that the Commission has found in
its previous decisions that certain factors did not constitute an aggravating circumstance for the
purpose of determining the amount of the fine does not mean that it is obliged to do so also in a
4.5.1.1 Repeated Infringement
293. The first example of an aggravating circumstance listed in the Guidelines is

‘where an undertaking continues or repeats the same or a similar infringement
after the Commission or a national competition authority has made a finding
that the undertaking infringed Article 81 or 82; the basic amount will be
increased by up to 100% for each such infringement established’.90
294. The Court of Justice has held that, in setting the amount of the fine,
‘[t]he analysis must … take into consideration … any repeated infringement’,91
and that the circumstance of repeated infringement is ‘not only a relevant factor
but also a particularly important factor and a very significant indication of the
gravity of the infringement for the purpose of assessing the amount of the fine in
the context of effective deterrence’.92
295. The 1998 Guidelines also listed ‘repeated infringement of the same type
by the same undertaking(s)’ as an aggravating circumstance. Out of the 74 cases
in which the Commission applied the 1998 Guidelines up to the end of 2006, 17
cases involved a finding of repeated infringement by at least one undertaking,
making a total of 28 findings of repeat offending.93 Almost invariably, such
finding led to an increase by 50 per cent of the basic amount of the fine.94
subsequent decision’: see, inter alia, Judgment of the CFI of 25 Oct 2005 in Case T–38/02 Groupe
Danone v Commission (Belgian Beer) [2005] ECR II–4407, para 57.
90
Ibid, pt 28, first indent.
91
Judgment of the ECJ of 7 Jan 2004 in Joined Cases C–204/00 P etc Aalborg Portland and Others
v Commission (Cement) [2004] ECR I–123, para 91, and Judgment of the ECJ of 8 Feb 2007 in Case
C–3/06 P Danone v Commission (Belgian Beer), not yet reported, para 26. The CFI had already held
earlier in its Judgment of 17 Dec 1991 in Case T–6/89 Enichem Anic v Commission (Polypropylene)
[1991] ECR II–1707, para 295, that ‘the fact that the Commission in the past already found an
undertaking guilty of infringing the competition rules and penalized it for that infringement may be
treated as an aggravating factor as against that undertaking but the absence of any previous
infringement is a normal circumstance which the Commission does not have to take into account as
a mitigating factor’.
92
Judgment of the ECJ of 8 Feb 2007 in Case C–3/06 P Danone v Commission (Belgian Beer), not
yet reported, para 47.
93
Above n 2, sect 2, first indent. Also before the 1998 Guidelines, the Commission took repeated
infringement into account as an aggravating factor in several decisions: see Opinion of AG Poiares
Maduro of 16 Nov 2006 in Case C–3/06 P Groupe Danone v Commission (Belgian Beer), not yet
reported, para 25. In a number of cases under the 1998 Guidelines, the repeat offenders had been
condemned already several times and/or the same group of undertakings had been involved both in
the earlier and in the new infringement; see for instance the Synthetic rubber case (Commission
Decision of 29 November 2006, IP/06/1647), in which Eni, Shell and Bayer received a fine increase of
50 % because all three had been earlier condemned by the Commission for cartels in the polypropyl-
ene, PVC and citric acid sectors; see also notes 98 and 106 below.
94
In the Belgian beer case, the increase for Danone was only of 40 %; see Judgment of the CFI of
25 Oct 2005 in Case T-38/02 Groupe Danone v Commission (Belgian beer) [2005] ECR II–4407, paras
312–313. In British Sugar, an increase of 75% was imposed jointly for the aggravating circumstances
of British Sugar having been the instigator of the infringement, and having acted contrary to a
compliance programme for the introduction of which it had received a fine reduction for an earlier
infringement: see Commission Dec of 14 Oct 1998, [1999] OJ L76/1, paras 206–210, and below, n
119.
296. As has been pointed out by Christopher Harding and Alun Gibbs, these
figures indicate an ‘awesome level of recidivism on the part of major companies
who appear as usual suspects in the world of business cartels. In short, this
suggests a confirmed culture of business delinquency’.95
297. The 2006 Guidelines appear to differ in three respects from (the practice
under) the 1998 Guidelines.96 First, not only are earlier findings by the European
Commission of similar infringements of Article 81 or 81 EC now taken into
account, but also such findings by the competition authorities of the EU Member
States. This is a logical evolution, following the establishment by Regulation
1/2003 of the European Competition Network (ECN), in which the European
Commission and the competition authorities of the EU Member States share the
task of investigating and prosecuting infringements of Articles 81 and 82 EC.97
Secondly, whereas the practice under the 1998 Guidelines was to increase the
basic amount by 50 per cent on account of repeated infringements, the 2006
Guidelines provide for an increase ‘by up to 100%’. Thirdly, this increase applies
‘for each [earlier] infringement established’.98
298. As regards the justification for imposing higher fines in case of repeated
infringement, the Court of First Instance has held that:
it must be recalled to mind that, for the purpose of determining the amount of the fine,
the Commission must ensure that its action has the necessary deterrent effect ….
Recidivism is a circumstance which justifies a significant increase in the basic amount
of the fine. Recidivism constitutes proof that the sanction previously imposed was not
sufficiently deterrent.99
299. It would indeed appear that the fact that an undertaking continues or
repeats an infringement not very long after it has been fined for a similar
infringement shows that the first fine has not been sufficiently deterrent. One
95
C Harding and A Gibbs, ‘Why Go to Court in Europe? An Analysis of Cartel Appeals
1995–2004’ (2005) 30 European Law Review 349 at 369. Further statistics for cartels discovered
throughout the world in the period 1990–2005, with rankings of top cartel recidivists (mostly
European), can be found in JM Connor and CG Helmers, ‘Statistics on Modern Private International
Cartels, 1990–2005’, Purdue University Dept. of Agricultural Economics Working Paper # 06–11 (Nov
2006), available at www.agecon.purdue.edu/working_papers/workingpaper.connor.11.10.06.pdf, at
23–5.
96
See also de Broca, above n 1, at 4–5.
97
See above n 13; Wils, Principles, above n 18, ch 1; and the ECN’s website at http://ec.europa.eu/
comm/competition/ecn/index_en.html.
98
In the Synthetic rubber case, see above n 93, application of the 2006 instead of the 1998
Guidelines could thus have led to an increase in the basic amount of the fines for Eni, Shell and Bayer
by up to 300% instead of by 50%.
99
Judgment of the CFI of 30 Sept 2003 in Case T–203/01 Michelin v Commission [2003] ECR
II–4071, para 293; see also Judgment of the CFI of 25 Oct 2005 in Case T–38/02 Groupe Danone v
Commission (Belgian Beer) [2005] ECR II–4407, para 348 (using in the English version the weaker
term ‘evidence’ instead of ‘proof’; this appears to be a translation error, the original French version of
both judgments using the term ‘preuve’). In the original French version of both judgments, the terms
corresponding to ‘a significant increase’ are ’une augmentation considérable’ (a considerable increase),
which appears stronger; see also above, text to n 92.
may however wonder whether the solution to this problem should be to raise the
second fine specifically for the repeat offender, or rather to increase the fine level
generally for all offenders. If the first fine was insufficient to deter the specific
undertaking, could one not assume a general problem of insufficient deterrence
for all undertakings?100 Indeed, according to settled case law of the Court of
Justice, the Commission can at any time raise the level of fines so as to reinforce
their deterrent effect when it finds that practices that have long been established
as being unlawful are still relatively frequent.101
300. There are several reasons why a finding of repeated infringement may
justify an increased fine specifically for the repeat offender rather than (or in
addition to) a general increase in the level of fines. First, as has been pointed out
by the Community Courts, the fact of the repeated infringement may reveal ‘a
characteristic specific to the perpetrator of the infringement, namely its propen-
sity to commit such infringements’.102 Indeed, some undertakings may have a
higher propensity to commit antitrust violations than others, because they, or
rather the individuals taking the decisions inside them, have a weaker moral
commitment to respect the law, or are more willing to take legal risks, or differ in
some other relevant aspect from the average undertaking.103 To deter such
undertakings from committing infringements (not only repeated ones, but also
first infringements), higher fines would be required for them than for other
undertakings. At the time of a first infringement, it may however not be possible
for a competition authority or a court to observe these specific characteristics
with an acceptable degree of accuracy or at a reasonable cost, allowing a
differentiation in fines between the undertakings with these specific characteris-
tics and the others. It may not be possible either, in particular for reasons of
proportional justice, to set the level of fines for all undertakings at the higher
level which would be required to deter the undertakings with a higher propensity
100
One could even argue that the problem of insufficient deterrence could be greater for the other
undertakings, as their subjective estimate of the probability that they will be caught and punished
may be lower than that of the undertaking which has already been fined; see the discussion on
subjective estimates and availability bias above in paras 194–198, and DA Dana, ‘Rethinking the
Puzzle of Escalating Penalties for Repeat Offenders’ (2001) 110 Yale Law Journal 733 at 743–53 and
759–72.
101
Judgment of the ECJ of 7 June 1983 in Joined Cases 100–103/80 Musique Diffusion française v
Commission [1983] ECR 1825, paras 108–109, and Judgment of the ECJ of 28 June 2005 in Joined
Cases C–189/02 P etc Dansk Rorindustri and Others v Commission (Pre-insulated Pipes) [2005] ECR
I–5425, para 169.
102
Beer) [2005] ECR II–4407, para 349, confirmed upon appeal in Judgment of the ECJ of 8 Feb 2007 in
Case C–3/06 P Danone v Commission (Belgian Beer), not yet reported, para 39.
103
On the role of an undertaking’s culture, with a case study of Archer Daniels Midland, see JM
Conley and WM O’Barr, ‘Crime and Custom in Corporate Society: A Cultural Perspective on
Corporate Misconduct’ (1997) 60 Law and Contemporary Problems 5; see also J Sonnenfeld and PR
Lawrence, ‘Why Do Companies Succumb to Price Fixing?’ [July–Aug 1978] Harvard Business Review
145; and Wils, above n 21, at 193–4.
to commit infringements.104 The best that can then be done in practice is to wait
until the undertakings with a higher propensity to commit infringements have
committed repeated infringements, and then to raise the fines specifically for
them.105 Secondly, higher fines for repeat offenders may be needed to compensate
for a lower probability of detection, on the assumption that undertakings learn
from a first investigation and prosecution by a competition authority how they
can better hide their infringements or better organise their defence.106 Thirdly,
higher fines for repeat offenders may be needed to compensate for decreasing
social sanctions, in that the reputational, commercial or social damage from
being fined for an antitrust violation may be lower the second or subsequent time
than the first time.107 Fourthly, in a situation where the fines imposed (dis-
counted for the probability of detection and punishment) are in reality below the
illicit gains from the antitrust violations,108 but where the amounts of the fines
are not sufficiently foreseeable for undertakings fully to understand this,109 it may
take a first imposition of a fine for an undertaking fully to understand that
violations are profitable. Higher fines for repeat offenders may then be required
to reinforce deterrence for these undertakings. Finally, higher fines for repeat
offenders may be required for moral reasons. Indeed, the visible fact of repeat
offending risks weakening the moral commitment to the law of the spontane-
ously law-abiding, and the imposition of higher fines counters this effect, by
expressing increased moral condemnation.110
301. As regards the conditions under which higher fines can be imposed on
account of repeated infringement, the Court of First Instance has held that
‘recidivism, as understood in a number of national legal systems, implies that a
person committed fresh infringements after having been penalised for similar
104
On the principle of proportionality as a constraint on the pursuit of optimal deterrence, as well
as other limits to high fines, see above, paras 203–212, and below, paras 338 and 339.
105
See also RA Posner, Economic Analysis of Law, 3rd edn (Boston, Mass, Little, Brown &
Company, 1986) at 214; AM Polinsky and DL Rubinfeld, ‘A Model of Optimal Fines for Repeat
Offenders’ (1991) 46 Journal of Public Economics 291; and Dana, above n 100, at 755.
106
See Wils, above n 1, at 36–7, and Polinsky and Rubinfeld, above n 105, at 303; see also, on the
importance of learning by cartels, Levenstein and Suslow, above n 23, and Harrington, above n 23,
and, on the role of the cartel defendant bar, Harding and Gibbs, above n 95, 368–9; see however also
above n 100 on the countervailing considerations relating to possible increased attention by the
investigating authorities and availability bias. See also Katyal, above n 63, at 1381–2, arguing for
increased penalties in cases where the same group of cartelists commit repeated infringements
together, with the argument that this circumstances suggest a tightly knit conspiracy in which
defection is more difficult: see above n 93 for the example of Eni, Shell and Bayer.
107
See Dana, above n 100, at 772–6; cf also TJ Miceli and C Bucci, ‘A Simple Theory of Increasing
Penalties for Repeat Offenders’ (2005) 1 Review of Law and Economics No 1, Art 5.
108
For evidence suggesting that this is the case in Europe today, or at least was before the 2006
Guidelines see Connor and Helmers, above n 95; see also above n 59.
109
110
See Dana, above n 100, 776–83, and text to nn 25–27.
infringements’.111 It has later added that, ‘bearing in mind the objective it

pursues, the concept of repeated infringement does not necessarily imply that a
fine has been imposed in the past, but merely that a finding of infringement has
been made in the past’.112 The notion of repeated infringement thus does not
cover contemporaneous infringements.113 As regards the identity of the person
having committed the previous and the new infringements, given that the
prohibitions of Articles 81 and 82 EC apply to ‘undertakings’ with an economic
meaning, a finding of repeated infringement can be based on the imposition of
an earlier fine on a sister company, if both companies are directly or indirectly
owned by the same parent company and thus form the same undertaking.114
302. The aggravating circumstance of repeated infringement is taken into
account only if a ‘similar infringement’ has been repeated. In one case, French
Beer, the Commission considered that a price-fixing cartel is not ‘similar’ to a
market-sharing cartel.115 It is not obvious that such a narrow approach is
necessary. As already mentioned, the Court of First Instance has held that
‘recidivism, as understood in a number of national legal systems, implies that a
person committed fresh infringements after having been penalised for similar
infringements’.116 In the comparison with national legal systems, with their
broad criminal laws, all hard-core violations of Article 81 or 82 EC may qualify as
‘similar infringements’. Moreover, many national legal systems know both the
notions of specific recidivism (repetition of the same or a similar offence) and
general recidivism (committing an offence after having been penalised for some
other, not necessarily similar, offence).117 One could also consider the issue in the
111
Judgments of the CFI of 11 Mar 1999 in Case T–141/94 Thyssen Stahl v Commission (Steel
Beams) [1999] ECR II–347, para 617, and of 30 Sept 2003 in Case T–203/01 Michelin v Commission
[2003] ECR II–4071, para 284.
112
Beer) [2005] ECR II–4407, para 363, confirmed upon appeal in Judgment of the ECJ of 8 Feb 2007 in
Case C–3/06 P Danone v Commission (Belgian Beer), not yet reported, para 41.
113
Judgments of the CFI of 11 Mar 1999 in Case T–141/94 Thyssen Stahl v Commission (Steel
Beams) [1999] ECR II–347, para 618; see also Case T–53/03 BPB v Commission (Plasterboard),
pending before the CFI; see also the case of SGL Carbon, which received from the European
Commission a 33% discount (called ‘volume discount’ by SB Völcker, ‘Developments in EC Compe-
tition Law in 2003: An Overview’ (2004) 41 CML Rev 1027 at 1045) from its fine for participation in
the Electrical and mechanical carbon and graphite products cartel (Commission Dec of 3 Dec 2003,
paras 358–360) because of its bad financial situation and the fact that it had already been fined for its
participation in two other contemporaneous cartels.
114
See Judgment of the CFI of 30 Sept 2003 in Case T–203/01 Michelin v Commission [2003] ECR
II–4071, para 290, and the literature on the notion of ‘undertaking’ referred to above in n 1.
115
Commission Dec of 29 Sept 2004 in French Beer [2005] OJ L184/57, para 91. On the other
hand, the fact that the infringements took place in different markets or sectors appears irrelevant: see
ibid.
116
Above, text to n 111.
117
See for instance French and Hungarian criminal law; alternatively, the similarity of the crimes
may be a factor to be weighed up by the sentencing judge in determining by how much the penalty is
increased on account of recidivism: see for instance Ch 9 Sect 4 of the Swedish Penal Code: ‘[i]n
determining the appropriate punishment, the court … shall … take reasonable account of whether
the accused has previously been guilty of crime. In this connection, special consideration shall be
light of the justification which the Community Courts have given for increasing
fines for repeat offenders: that the fact of the repeated infringement may reveal ‘a
characteristic specific to the perpetrator of the infringement, namely its propen-
sity to commit such infringements’.118 Does it make sense to distinguish between
a propensity to take part in market-sharing cartels and a propensity to take part
in price-fixing cartels? Or are both part of a single propensity to commit
hard-core antitrust violations?119
303. Given that the Guidelines do not contain a maximum period of time
between the starting date of the new infringement, or the date of the new
decision, and the date at which the earlier infringement was penalised or found,
concerns have been expressed that this would allow a system of ‘everlasting’ or
‘perpetual’ repeated infringement.120 In its recent judgment in Belgian Beer, the
Court of Justice held that ‘the finding and the appraisal of the specific character-
istics of a repeated infringement come within the Commission’s discretion and
that the Commission cannot be bound by any limitation period when making
such a finding. … repeated infringement is an important factor which the
Commission must appraise, since the purpose of taking repeated infringement
into account is to induce undertakings which have demonstrated a tendency
towards infringing the competition rules to change their conduct. The Commis-
sion may, therefore, in each individual case, take into consideration the indicia
given to the extent of any previous criminality, to the time that has been elapsed between the crimes,
and to whether the previous and the new criminality are similar in nature or whether in both cases
they are of an especially serious character.’
118
Above, text to n 102.
119
In British Sugar, the Commission increased by 75% the basic amount of the fine for an Art 81
infringement by British Sugar jointly for the aggravating circumstances of British Sugar having been
the instigator of the infringement, and having acted contrary to a compliance programme for the
introduction of which it had received a fine reduction for an earlier Art 82 infringement. The
Commission noted that the compliance programme covered both Arts 81 and 82 EC, but also pointed
out that ‘it has to be recalled that both [Art 81] and [Art 82] serve the common aim laid down in
point (g) of Article 3 of the Treaty, of establishing a system ensuring that competition in the internal
market is not distorted’: Commission Dec of 14 Oct 1998 [1999] OJ L76/1, para 209. Having learned
from the British Sugar experience, the Commission today no longer takes the adoption of a
compliance programme into account as a mitigating circumstance in setting the amount of fines. This
change of policy has been approved by the CFI, which has held that ‘the mere fact that in certain of its
previous decisions the Commission took the implementation of a compliance programme into
consideration as a mitigating factor does not mean that it is obliged to act in the same manner in any
given case …, especially where the infringement in question is … a clear infringement of [Art 81 or 82
EC]. The Commission is therefore not required to take a circumstance such as that into account as a
mitigating factor, provided that it adheres to the principle of equality of treatment, which requires
that it should not assess the matter differently for any undertaking addressed by the same decision’:
Judgment of the CFI of 9 July 2003 in Case T–224/00 Archer Daniels Midland v Commission (Amino
acids) [2003] ECR II–2597, paras 280–281; see also below n 138.
120
See the argument of Danone, as summarised in Judgment of the ECJ of 8 Feb 2007 in Case
C–3/06 P Danone v Commission (Belgian Beer), not yet reported, para 33, and K Nordlander, ‘The
Commission’s Policy on Recidivism: Legal Certainty for Repeat Offenders?’ (2005) 2 Competition Law
Review 55 at 66–8.
which confirm such a tendency, including, for example, the time that has elapsed
between the infringements in question’.121
4.5.1.2 Refusal to Cooperate with or Obstruction of the Commission’s

Investigations
304. The second example of an aggravating circumstance listed in the Guidelines

is ‘refusal to cooperate with or obstruction of the Commission in carrying out its
investigations’.122
305. According to settled case law, Regulation 1/2003 places the undertaking
being investigated under ‘an obligation to cooperate actively, which means that it
must make available to the Commission all information relating to the subject-
matter of the investigation’.123 ‘That obligation to cooperate means that the
undertaking may not evade requests for production of documents on the ground
that by complying with them it would be required to give evidence against
itself’.124 The Commission is also ‘entitled to compel an undertaking, if necessary
by adopting a decision, to provide all necessary information concerning such
facts as may be known to it’.125 However, because the Commission must ensure
that the rights of the defence are not impaired during preliminary inquiry
procedures, it ‘may not compel an undertaking to provide it with answers which
might involve an admission on its part of the existence of an infringement which
it is incumbent upon the Commission to prove’.126 ‘Equally, respect for the rights
of the defence requires that the undertaking concerned must have been afforded
the opportunity, during the administrative procedure, to make known its views
on the truth and relevance of the facts and circumstances alleged and on the
documents used by the Commission to support its claim that there has been an
infringement of [Article 81 or 82] of the Treaty’.127
306. In this light, the Community Courts have approved the use of refusal to
cooperate with or obstruction of the Commission’s investigations as an aggravat-
ing circumstance, stating that ‘the conduct of the undertaking during the
administrative procedure may be one of the factors to be taken into account
when fixing the fine’, but that ‘the aggravating circumstance consisting in the
121
Judgment of the ECJ of 8 Feb 2007 in Case C–3/06 P Danone v Commission (Belgian Beer), not
yet reported, paras 38 and 39.
122
Guidelines, above n 1, pt 28, second indent. Sect 2, second indent, of the 1998 Guidelines,
above n 2, similarly listed ‘refusal to cooperate with or attempts to obstruct the Commission in
carrying out its investigations’ as an aggravating circumstance.
123
Judgment of the ECJ of 29 June 2006 in Case C–301/04 P Commission v SGL Carbon (Graphite
electrodes), not yet reported, para 40; on the Commission’s powers of investigation and the rights of
the defence see generally Ch 1 above.
124
Judgment in Case C–301/04 P, above n 123, para 48.
125
Ibid, para 41.
126
v Commission (Cement) [2004] ECR I–123, paras 63 and 65, and Judgment in Case C–301/04 P, above
n 123, para 42.
127
Judgment in Joined Cases C–204/00 P etc, above n 126, para 66.
refusal to cooperate with or attempts to obstruct the Commission in carrying out

its investigations cannot be applied when the undertaking concerned is merely
exercising its rights of defence’.128
307. An example can be found in Graphite Electrodes, where the basic amount
of the fine for SGL was increased by 85 per cent because of three aggravating
circumstances, including ‘its attempts to obstruct the Commission proceedings
by giving warnings to other companies of the forthcoming investigation’.129
308. For certain types of refusal to cooperate, the Commission also has the
power to impose separate, procedural fines under Article 23(1) of Regulation
1/2003.130 The Commission is however not obliged to make use of this power,
and the fact that this power exists is not an obstacle to taking the refusal to
cooperate or the obstruction into account as an aggravating factor in the setting
of the fine for the substantive infringement.131
4.5.1.3 Leader, Instigator or Coercer
309. The third example of an aggravating circumstance listed in the Guidelines

is ‘role of leader in, or instigator of, the infringement; the Commission will also
pay particular attention to any steps taken to coerce other undertakings to
participate in the infringement and/or any retaliatory measures taken against
other undertakings with a view to enforcing the practices constituting the
infringement’.132
310. According to settled case law:
where an infringement has been committed by a number of undertakings, it is
necessary, in determining the amount of the fines, to establish their respective roles in
128
Others v Commission (Pre-insulated Pipes) [2005] ECR I–5425, paras 351 and 353.
129
Commission Dec of 18 July 2001 in Graphite electrodes [2002] OJ L100/1, para 160, and
Judgment of the ECJ of 29 June 2006 in Case C–308/04 P SGL v Commission (Graphite Electrodes), not
yet reported, paras 69 and 70. For further examples see D Geradin and D Henry, ‘The EC Fining
Policy for Violations of Competition Law: An Empirical Review of the Commission Decisional
Practice and the Community Courts’ Judgments’ [2005] European Competition Law Journal 401 at
446–7.
130
Above n 13; see above, paras 8 and 14.
131
See Judgment of the CFI of 20 Mar 2002 in Case T–9/99 HFB and Others v Commission
(Pre-insulated Pipes) [2002] ECR II–1498, paras 474–564; Judgment of the ECJ of 28 June 2005 in
Joined Cases C–189/02 P etc Dansk Rorindustri and Others v Commission (Pre-insulated Pipes) [2005]
ECR I–5425, paras 348–362; and Commission Dec of 30 Nov 2005 in Industrial Bags, not yet
published, para 792. If however the Commission were to impose a fine under Art 23(1) of Reg 1/2003
for certain facts, as well as to use the same facts as an aggravating circumstance in setting the amount
of the fine for the substantive infringement under Art 23(2), this may be questioned under the
principle of ne bis in idem; on the latter principle see generally WPJ Wils, ‘The Principle of Ne Bis in
Idem in EC Antitrust Enforcement: A Legal and Economic Analysis’ (2003) 26 World Competition 131.
132
Guidelines, above n 1, pt 28, third indent. Sect 2, third and fourth indents, of the 1998
Guidelines, above n 2, similarly listed ‘role of leader in, or instigator of the infringement’ and
‘retaliatory measures against other undertakings with a view to enforcing practices which constitute
an infringement’ as aggravating circumstances.
the infringement throughout the duration of their participation in it …. It follows, in

particular, that the role of ‘ringleader’ played by one or more undertakings in a cartel
must be taken into account for the purposes of calculating the amount of the fine, in so
far as the undertakings which played such a role must therefore bear special responsi-
bility in comparison with other undertakings.133
311. Under the 1998 Guidelines, the Commission has in 21 cases decided before
the end of 2006 increased the basic amount of the fine for one or two undertak-
ings, because they had acted as keader or instigator in the cartel, or coerced others
to participate, by percentages ranging from 20 to (more often) 50 per cent.134
312. As already mentioned above,135 in the case of collective violations such as
cartels, the differentiation of penalties according to the role played by the
different conspirators is an effective instrument for raising the cost of setting up
and running cartels. Indeed, setting up and running a successful cartel takes
effort and requires the participation of the different cartel members in various
respects. A number of tasks have to be performed: the agreed price or other
factors must be determined; the joint profit must be allocated through quotas or
otherwise; and measures have to be taken to prevent cheating, by developing a
sufficiently strong sense of solidarity and mutual trust and/or by monitoring and
punishing deviations. For all these tasks to be carried out, someone has to take
the initiative and someone has to do the job: to convene the meeting, to make the
price or quota calculations, to do the monitoring and the punishing.136
313. The setting up and running of cartels can be made more difficult (and
hence more costly, and cartels thus less profitable, which means they can be
deterred at lower overall fine levels) by threatening higher fines for those cartel
members that play active roles in setting up and running them, while offering the
perspective of reduced fines for those cartel members that are exclusively passive.
This makes the setting up and functioning of cartels more difficult because, faced
with the prospect of higher fines, there will be fewer volunteers to play active
roles. Those who nevertheless agree to do so are likely to want compensation in
133
Judgment of the CFI of 29 Apr 2004 in Joined Cases T–236/01 etc Tokai Carbon and Others v
Commission (Graphite Electrodes) [2004] ECR II–1181, para 301. On the notions of ‘leader’ and
‘instigator’ see further Judgment of the CFI of 15 Mar 2006 in Case T–15/02 BASF v Commission
(Vitamins) [2006] ECR II–497, paras 315–354.
134
See also the case of SGL in Graphite Electrodes, above, text to n 129, where ‘SGL’s role as one of
the ringleaders and instigators of the cartel’ was one of the 3 aggravating circumstances justifying
together an increase of 85% of the basic amount of the fine. For further examples see Geradin and
Henry, above n 129, 443–5. See also Opinion of AG Mazák of 18 Jan 2007 in Case C–328/05 P SGL
Carbon v Commission (Specialty graphite), not yet reported, para 60.
135
Above, para 259.
136
To some extent these tasks can be outsourced to specialist service providers. Hence the need for
the activity of such service providers to be covered by the antitrust prohibitions and to be subject to
the threat of sufficiently deterrent fines; see eg the role of AC-Treuhand in the Organic Peroxides
cartel, Commission Dec of 10 December 2003; an appeal against this decision is pending before the
CFI as Case T–99/04 AC-Treuhand v Commission (Organic peroxides).
the form of a larger part of the gain, which is likely to be difficult to agree on, and
to weaken the sense of solidarity and mutual trust within the group.137
4.5.2 Mitigating Circumstances
314. The basic amount of the fine may be reduced where the Commission finds
that there are mitigating circumstances, of which the Guidelines give five exam-
ples.138
4.5.2.1 Termination as Soon as the Commission Intervenes
315. The first example of a mitigating circumstance listed in the Guidelines is

‘where the undertaking concerned provides evidence that it terminated the
infringement as soon as the Commission intervened: this will not apply to secret
agreements or practices (in particular, cartels)’.139
316. The 1998 Guidelines listed as an extenuating circumstance ‘termination
of the infringement as soon as the Commission intervenes (in particular when
the Commission carries out checks)’.140 The main difference is that the 2006
Guidelines exclude the application of this extenuating circumstance to cartels or
other secret practices.141
317. The justification for this mitigating circumstance in the case of clear-cut
infringements is indeed far from obvious. As the Court of First Instance has
pointed out: ‘termination of an infringement only after the Commission has
intervened should not be rewarded in the same way as an independent initiative
of the offending party, and merely constitutes an appropriate and normal
reaction to that intervention’; ‘if termination of an infringement as soon as the
Commission intervenes were to be recognised as an attenuating circumstance,
that would unduly impair the effectiveness of Article 81(1) EC by weakening
both the penalty and its deterrent effect’; ‘the classification of the continuation of
an infringement after the Commission intervenes as an aggravating circumstance
… rightly constitutes an incentive to terminate the infringement, but … does not
reduce the penalty or its deterrent effect’.142
137
See NK Katyal, ‘Conspiracy Theory’ (2003) 112 Yale Law Journal 1307 at 1341–6 and 1363–7.
On the question whether ringleaders should also be excluded from the benefit of leniency see below,
paras 425 and 430, nn 140 and 150.
138
Guidelines, above n 1, pt 29. Implementation of compliance programmes is not taken into
account as a mitigating circumstance: see above n 119, and sect 3.2.3.2 above.
139
Guidelines, above n 1, pt 29, first indent.
140
1998 Guidelines, above n 2, sect 3, third indent.
141
According to de Broca, above n 1, at 5, this corresponds to what was already the Commission’s
practice under the 1998 Guidelines; see also Barbier de la Serre, above n 2, pt 34.
142
Judgment of 27 Sept 2006 in Case T–329/01 Archer Daniels Midland v Commission (Sodium
gluconate), not yet reported, paras 278–279, appeal pending as Case C–510/06 P Archer Daniels
Midland v Commission (Sodium gluconate); see also Judgments of the CFI of 8 July 2004 in Case
318. One could add that, under the 2006 Guidelines, an incentive to terminate
the agreement already flows from the fact that the basic amount of the fine is
proportional to the duration of the agreement.143
4.5.2.2 Negligence
319. The second example of a mitigating circumstance listed in the Guidelines is

‘where the undertaking provides evidence that the infringement has been com-
mitted as a result of negligence’.144
320. The 1998 Guidelines similarly listed as an extenuating circumstance
‘infringements committed as a result of negligence or unintentionally’.145
321. Article 23(2), first subparagraph, of Regulation 1/2003 lays down, as a
condition that must be fulfilled to enable the Commission to impose fines, that
the infringement must have been committed intentionally or negligently.146 As
the Court of Justice has pointed out, this initial condition is distinct from the
determination of the amount of the fine, which depends on the gravity and
duration of the infringement.147
322. Before the 1998 Guidelines, the Commission had in several decisions
taken intent into account as an aggravating circumstance in the calculation of the
fine.148 There are at least two reasons for imposing higher fines for intentional
infringements than for negligent ones. First, those who deliberately infringe the
law generally have more opportunity to hide their tracks. They thus face a lower
T–44/00 Mannesmannröhren-Werke v Commission (Seamless steel tubes) [2004] ECR II–2223, paras
280–284, and of 15 June 2005 in Joined Cases T–71/03 etc Tokai Carbon and Others v Commission
(Specialty graphite) [2005] ECR II–10, para 292.
143
See above, paras 248 and 276. One could counterargue that there is still scope for an extra
incentive because, in the setting of the basic amount of the fine, periods of less than 6 months are
counted as half a year, and periods of between 6 months and one year as one year; see above n 8.
However, a further problem of possibly perverse incentives may arise to the extent that termination as
soon as the Commission intervenes is recognised as a mitigating circumstance, whereas termination
before such intervention is not; see J-F Bellis, ‘La détermination des amendes pour infraction au droit
communautaire de la concurrence—bilan de cinq années d’application des lignes directrices de 1998’
[2003] Cahiers de droit européen 373 at 385–6, and Geradin and Henry, above n 129, at 451.
144
Guidelines, above n 1, pt 29, second indent.
145
1998 Guidelines, above n 2, sect 3, fifth indent.
146
Requiring intent or negligence obviously makes sense, because it determines whether the
conduct can be deterred. Fining undertakings for conduct they cannot avoid at reasonable cost will
have no deterrent effect, while it may generate wasteful expense: see R Posner, The Economics of Justice
(Boston, Mass, Harvard University Press, 1983) at 225; R Posner, ‘An Economic Theory of the
Criminal Law’ (1985) 85 Columbia Law Review 1193 at 1221–3; and Wils, above n 1, at 30.
147
Order of the ECJ of 25 Mar 1996 in Case C–137/95 P SPO and Others v Commission (Dutch
building and construction industry) [1996] ECR I–1627, paras 53–54, and Judgment of the ECJ of 17
July 1997 in Case C–219/95 P Ferriere Nord v Commission (Welded Steel Mesh) [1997] ECR I–4440,
paras 32–33.
148
Eg, Commission Dec of 19 Dec 1990 in Soda-ash—Solvay [1991] OJ L152/21 and Commission
Dec of 10 Jan 1996 in ADALAT [1996] OJ L210/1.
probability of detection and should therefore be fined more heavily for deter-
rence to occur.149 Secondly, higher fines for intentional offenders may be required
for moral reasons. Indeed, as in the case of repeat offending, the visible fact of
intentional offending risks weakening the moral commitment to the law of the
spontaneously law-abiding, and the imposition of higher fines counters this
effect, by expressing increased moral condemnation.150
323. Following the 1998 Guidelines, the Commission no longer takes intent
into account as an aggravating circumstance but instead recognises negligence as
an extenuating circumstance. This change reflects the fact that intentional
infringements are in fact far more common than negligent ones,151 given the
wide interpretation of the former concept. Indeed, according to settled case law,
‘it is not necessary for an undertaking to have been aware that it was infringing
[Article 81 EC] for an infringement to be regarded as having been committed
intentionally; it is sufficient that it could not have been unaware that the
contested conduct had as its object the restriction of competition’.152
4.5.2.3 Substantially Limited Role (Cheating)
324. The third example of a mitigating circumstance listed in the Guidelines is

‘where the undertaking provides evidence that its involvement in the infringe-
ment is substantially limited and thus demonstrates that, during the period in
which it was a party to the offending agreement, it actually avoided applying it by
adopting competitive conduct in the market: the mere fact that an undertaking
participated in an infringement for a shorter duration than others will not be
regarded as a mitigating circumstance since this will already be reflected in the
basic amount’.153
149
See LA Bebchuk and L Kaplow, ‘Optimal Sanctions and Differences in Individuals’ Likelihood
of Avoiding Detection’ (1993) 13 International Review of Law and Economics 217 at 223; Posner, ‘An
Economic Theory’, above n 146 at 1221–2; and Wils, above n 1, at 36–7.
150
151
Indeed, the extenuating circumstance of negligence appears to have been found in none of the
74 decisions taken under the 1998 Guidelines before the end of 2006; see however Geradin and Henry,
above n 129, at 452–3 and Judgment of the CFI of 27 July 2005 in Joined Cases T–49/02 etc Brasserie
nationale and Others v Commission (Luxembourg Beer) [2005] ECR II–3033, paras 191–194, on the
application in 3 decisions of the extenuating circumstance of ‘existence of a reasonable doubt on the
part of the undertaking as to whether the restrictive conduct does indeed constitute an infringement’,
a circumstance which was listed in the 1998 Guidelines but no longer figures in the 2006 Guidelines.
152
Judgment of the ECJ of 11 July 1989 in Case 246/86 Belasco and Others v Commission [1989]
ECR 2191, para 41. Intention can thus be derived from the awareness either that the behaviour is
forbidden by Community law or of the behaviour’s anti-competitive object. The Community Courts
have never provided a definition of negligence. In his Opinion of 29 Oct 1975 in Case 26/75 General
Motors v Commission [1975] ECR 1389, AG Mayras wrote that ‘the concept of negligence must be
applied where the author of the infringement, although acting without any intention to perform an
unlawful act, has not foreseen the consequences of his action in circumstances where a person who is
normally informed and sufficiently attentive could not have failed to foresee them’.
153
Guidelines, above n 1, pt 29, third indent.
325. The 1998 Guidelines similarly listed as an extenuating circumstance

‘non-implementation in practice of the offending agreements or practices’.154
The more elaborate wording in the new Guidelines appears to be inspired by the
Judgment of the Court of First Instance in Daiichi Pharmaceutical v Commission,
in which the Court stated that, in order to be granted a reduction in its fine on
account of non-implementation in practice of the offending agreement, an
undertaking must show that, ‘during the period in which [it] was party to the
offending agreements, it actually avoided implementing them by adopting com-
petitive conduct on the market or, at the very least, .. . it clearly and substantially
breached the obligations relating to the implementation of the cartel to the point
of disrupting its very operation’.155
326. This mitigating circumstance should be understood against the back-
ground of the settled case law according to which:
it is sufficient for the Commission to show that the undertaking concerned participated
in meetings at which anti-competitive agreements were concluded, without manifestly
opposing them, to prove to the requisite standard that the undertaking participated in
the cartel. … Nor is the fact that an undertaking does not act on the outcome of a
meeting having an anti-competitive purpose such as to relieve it of responsibility for
the fact of its participation in a cartel, unless it has publicly distanced itself from what
was agreed in the meeting …. Neither is the fact that an undertaking has not taken part
in all aspects of an anti-competitive scheme or that it played only a minor role in the
aspects in which it did participate material to the establishment of the existence of an
infringement on its part. Those factors must be taken into consideration only when the
gravity of the infringement is assessed and if and when it comes to determining the
fine.156
327. Granting a reduction of the fine to cartel members that cheated on their
cartel commitments obviously makes sense, because it creates an incentive for
cheating, thus undermining the effectiveness of cartels. It is however important
not to be too generous on this account, so as to maintain adequate incentives for
cheaters to make use of the ability to denounce the cartel to the Commission and
seek immunity under the Leniency Notice, thereby allowing the cartel to be
discovered and penalised.157
154
1998 Guidelines, above n 2, sect 3, second indent.
155
Judgment of the CFI of 15 Mar 2006 in Case T–26/02 Daiichi Pharmaceutical v Commission
(Vitamins) [2006] ECR II–713, para 113.
156
v Commission (Cement) [2004] ECR I–123, paras 81–86.
157
Commission Notice on Immunity from fines and reduction of fines in cartel cases [2006] OJ
C298/17; see further below, para 340 and Ch 5.
4.5.2.4 Cooperation Outside the Leniency Notice
328. The fourth example of a mitigating circumstance listed in the Guidelines is

‘where the undertaking concerned has effectively cooperated with the Commis-
sion outside the scope of the Leniency Notice and beyond its legal obligation to
do so’.158
329. According to its point 1, the Leniency Notice:
sets out the framework for rewarding cooperation in the Commission investigation by
undertakings which are or have been party to secret cartels affecting the Community.
Cartels are agreements and/or concerted practices between two or more competitors
aimed at coordinating their competitive behaviour on the market and/or influencing
the relevant parameters of competition through practices such as the fixing of purchase
or selling prices or other trading conditions, the allocation of production or sales
quotas, the sharing of markets including bid-rigging, restrictions of imports or exports
and/or anti-competitive actions against other competitors.159
330. For types of infringements of Articles 81 and 82 EC other than secret
cartels, cooperation with the Commission can thus be rewarded under this
fourth mitigating circumstance.160 Only cooperation ‘beyond [the undertaking’s]
legal obligation to do so’ can be rewarded by a reduction in the fine. Indeed, as
already mentioned above, undertakings being investigated have certain obliga-
tions of active cooperation, and failure to respect these obligations is taken into
account as an aggravating circumstance leading to an increase in the fine.161
Finally, the cooperation must have been ‘effective’, which means that it must have
enabled the Commission more easily to establish the existence of the infringe-
ment, and, where relevant, to bring it to an end.162
158
Guidelines, above n 1, pt 29, fourth indent. This mitigating circumstance also figured in sect 3,
sixth indent, of the 1998 Guidelines.
159
Above n 157.
160
See however also Commission Dec of 20 Oct 2005 in Case COMP/B-2/38.281 Italian Raw
Tobacco, paras 385–398.
161
See above, paras 304–308, and Judgment of the CFI of 30 Jan 2007 in Case T–340/03 France
Télécom v Commission, not yet reported, para 276: ‘according to established case-law, cooperation in
the investigation which does not go beyond that which undertakings are already obliged to provide
under [Regulation No 1/2003] does not warrant a reduction in the fine’.
162
See Judgment of the ECJ of 16 Nov 2000 in Case C–297/98 P SCA Holding v Commission
(Cartonboard) [2000] ECR I–10101, para 36, confirming the Judgment of the CFI of 14 May 1998 in
Case T–327/94: ‘[t]he Court of First Instance correctly held in that regard, in paragraph 156 of the
contested judgment, that a reduction in the fine on grounds of cooperation during the administrative
procedure is justified only if the conduct of the undertaking in question enabled the Commission to
establish the existence of an infringement more easily and, where relevant, to bring it to an end’; see
further sect 3.2.3.1 above, and Ch 5 below.
4.5.2.5 Encouragement by Public Authorities or by Legislation
331. The fifth and last example of a mitigating circumstance listed in the
Guidelines is ‘where the anti-competitive conduct of the undertaking has been
authorized or encouraged by public authorities or by legislation’.163
332. This mitigating circumstance must be interpreted in the light of the
settled case law according to which ‘where undertakings are required by national
legislation to engage in anti-competitive conduct, they cannot … be held
accountable for infringement of Articles 81 EC and 82 EC’; however, ‘if a national
law merely encourages, or makes it easier for undertakings to engage in anti-
competitive conduct, those undertakings remain subject to Articles 81 EC and 82
EC and may incur penalties …. It must none the less be made clear that, although
such a situation cannot lead to acceptance of practices which are likely further to
exarcerbate the adverse effects on competition, it nevertheless means that when
the level of the penalty is set the conduct of the undertakings concerned may be
assessed in the light of the national legal framework, which is a mitigating
factor’.164
333. An example of this mitigating circumstance can be found in the French
Beef case, where the Commission reduced the basic amount of the fine by 30 per
cent on account of ‘the forceful intervention of the French Minister for Agricul-
ture in favour of the conclusion of such an agreement. This was reflected in his
speeches in parliament and his press release the day the agreement was con-
cluded.’165
4.5.3 Specific Increase for Deterrence (Large Undertakings and/or Amount of

the Illicit Gains)
334. Under the heading ‘Specific increase for deterrence’, the Guidelines state
that the Commission ‘will pay particular attention to the need to ensure that fines
have a sufficient deterrent effect; to that end, it may increase the fine to be
imposed on undertakings which have a particularly large turnover beyond the
sales of goods or services to which the infringement relates. The Commission will
163
Guidelines, above n 1, pt 29, fifth indent. The accompanying n 1 adds: ‘[t]his is without
prejudice to any action that may be taken against the Member State concerned’.
164
Judgment of the ECJ of 9 Sept 2003 in Case C–198/01 CIF [2003] ECR I–8055, paras 51 and
56–57; the Court also held that national competition authorities have the power and duty to disapply
any national law that requires undertakings to engage in anti-competitive conduct, with the result
that, from that time onwards, the undertakings ‘can no longer claim that they are obliged by that law
to act in breach of the Community competition rules. Their future conduct is therefore liable to be
penalised’: ibid, para 55.
165
Commission Dec of 2 Apr 2003 in Case COMP/C.38.279/F3 French Beef [2003] OJ L209/12,
para 176.
also take into account the need to increase the fine in order to exceed the amount
of gains improperly made as a result of the infringement where it is possible to
estimate that amount’.166
335. As regards increasing the fine for undertakings which have a particularly
large turnover beyond the sales of goods or services to which the infringement
relates (known as ‘deterrence multiplier’), the Court of Justice has recently
confirmed in Showa Denko v Commission its settled case law according to which
‘the size of the undertaking concerned is one of the factors which may be taken
into account for the purpose of calculating the fine and, therefore, setting the
“deterrence multiplier’’’.167 The Court of Justice also approved the finding by the
Court of First Instance that ‘owing to its “enormous” worldwide turnover by
comparison with the turnovers of the other members of the cartel, the applicant
could more readily raise the necessary funds to pay its fine, which, if the fine was
to have a sufficiently deterrent effect, justified the application of a multiplier’.168
336. Advocate-General Geelhoed has given two further explanations of why a
‘deterrence multiplier’ for very large undertakings appears justified. First, ‘in the
event of collective infringement like a cartel as opposed to an infringement by a
single offender, the Commission must also consider the subsequent effects of the
fines and take into account the size of a given company. The example given by the
Commission is a cartel consisting of one big player and several small players. The
big player cooperated with the Commission and received immunity under the
Leniency Notice. In such a case very high fines could have put the smaller players
out of business, in which case the Commission’s intervention would have
resulted in a monopoly.’169 Secondly, ‘as regards the argument that larger com-
panies do not care less about small amounts than small companies, that assumes,
in the first place, perfect information and perfect rationality. That is hard for an
individual calculating the risk involved in an offence and it is hard for the
undertakings involved in a cartel too. Furthermore, it cannot be denied that there
is a difference between small and large companies, in the sense that a small fine
may escape the attention of the board of a parent company of the group, while a
huge amount will not. A huge fine is likely to attract the attention of the board
and therefore is likely to stimulate compliance with the competition rules with
regard to future behaviour.’170
337. As regards the possible increase in the fine in order to exceed the amount
of the illicit gains, the theoretical relevance of the amount of the illicit gain for
the setting of the optimal fine, as well as the practical impossibility in most cases
166
Guidelines, above n 1, pts 30 and 31.
167
Judgment of the ECJ of 29 June 2006 in Case C–289/04 P [2006] ECR I–5859, para 29.
168
Ibid, para 18.
169
Opinion of AG Geelhoed of 19 Jan 2006 in ibid, para 61 and n 16; see also above, paras
229–232.
170
Ibid, para 62.
of estimating this amount with sufficient precision and at a reasonable adminis-

trative cost, has already been discussed above.171 This should however not rule
out that, in those rather unlikely circumstances in which the Commission has a
reliable estimate of the illicit gains available, it could take this amount into
account, at least to ensure that the fine exceeds it. Indeed, a fine below the illicit
gains could certainly never be a sufficient deterrent, even under the unrealistic
assumption that all infringements are detected and penalised.172 According to
settled case law, ‘the profit which the undertakings were able to derive from their
practices is one of the factors to be considered in assessing the gravity of the
infringement and … taking that factor into account is designed to ensure that the
fine is deterrent’.173 Ensuring that the fine exceeds the illicit gain also contributes
to corrective justice, in the form of disgorgement of unjust enrichment.174
4.5.4 Legal Maximum
338. The Guidelines recall that the ‘final amount of the fine shall not, in any
event, exceed 10 % of the total turnover in the preceding business year of the
undertaking … , as laid down in Article 23(2) of Regulation No 1/2003’.175
339. The Court of Justice has held that ‘this limit seeks to prevent fines from
being disproportionate in relation to the size of the undertaking and since only
the total turnover can effectively give an approximate indication of that size, the
aforementioned percentage must … be understood as referring to the total
turnover’.176
171
See above, paras 284–288; see also Opinion of AG Geelhoed in Showa Denko, above n 167,
paras 57–60.
172
See also above, para 242, discussing Commission Dec of 25 Mar 1992 in Case IV/30.717-A
Eurocheque: Helsinki Agreement [1992] OJ L95/5.
173
Others v Commission (Pre-insulated Pipes) [2005] ECR I–5425, para 292.
174
175
Guidelines, above n 1, pt 32; see above n 13 and para 262
176
Judgment of 7 June 1983 in Joined Cases 100–103/80 Musique Diffusion française v Commission
[1983] ECR 1825, para 119; see also Judgments of the CFI of 15 June 2005 in Joined Cases T–71/03 etc
Tokai Carbon and Others v Commission (Specialty graphite) [2005] ECR II–10, paras 388–390, and of
29 Nov 2005 in Case T–33/02 Britannia Alloys & Chemicals v Commission (Zinc phosphate) [2005]
ECR II–4973, paras 33–52; see further above, paras 202–212 (on the requirements of proportional
justice and other reasons for avoiding excessively high fines) and Wils, above n 1, at 44 and 159 (on
the origin of the 10% limit in the ECSC Treaty, and the suggestion made by the OECD Competition
Law and Policy Division in 1999 to raise the 10% ceiling); see also I Bos and MP Schinkel, ‘On the
Scope for the European Commission’s 2006 Fining Guidelines under the Legal Fine Maximum’
(2006) 2 Journal of Competition Law and Economics 673.
4.5.5 Application of the Leniency Notice
340. The Guidelines recall that, after the amount of the fine has been calculated
and possibly capped at the ceiling of 10 per cent of total turnover, the Commis-
sion will apply its Leniency Notice.177 As has been pointed out by the Court of
First Instance, the Commission’s practice of applying the Leniency Notice after
the application of the 10 per cent ceiling ‘ensures that the Leniency Notice is fully
effective: if the basic amount was significantly in excess of the 10% limit before
the application of the Leniency Notice and that limit could not be applied
immediately, the incentive for the undertaking concerned to cooperate with the
Commission would be much less, since the final fine would be reduced to 10% in
any event, with or without the undertaking’s cooperation’.178
4.5.6 Ability to Pay
341. Finally, the Guidelines state that, ‘in exceptional cases, the Commission
may, upon request, take account of the undertaking’s inability to pay in a specific
social and economic context. It will not base any reduction granted for this
reason in the fine on the mere finding of an adverse or loss-making financial
situation. A reduction could be granted solely on the basis of objective evidence
that the imposition of the fine as provided for in these Guidelines would
irretrievably jeopardise the economic viability of the undertaking concerned and
cause its assets to lose all their value’.179
342. This possibility was also more briefly mentioned in the 1998 Guidelines,
according to which, ‘depending on the circumstances, account should be taken …
of certain objective factors such as [the undertakings’] real ability to pay in a
specific social context, and the fines should be adjusted accordingly’.180
343. The Court of Justice has recently recalled that, ‘according to settled
case-law, the Commission is not required, when determining the amount of the
fine, to take into account the poor financial situation of an undertaking, since
recognition of such an obligation would be tantamount to giving unjustified
competitive advantages to undertakings least well adapted to the market consi-
tions’.181 ‘That case-law is in no way called into question by [the 1998 Guide-
lines], which [state] that an undertaking’s real ability to pay must be taken into
177
Commission Notice on Immunity from fines and reduction of fines in cartel cases [2006] OJ
C298/17; for a detailed discussion of the law and economics of leniency see Ch 5 below.
178
Judgment of the CFI of 29 Nov 2005 in Case T–52/02 SNCZ v Commission (Zinc phosphate)
[2005] ECR II–5005, para 41.
179
180
1998 Guidelines, above n 2, sect 5(b).
181
Judgment of the Court of Justice of 29 June 2006 in Case C–308/04 P SGL Carbon v
Commission (Graphite electrodes) [2006] ECR I–5977, para 105. See also Opinion of AG Mazák of 18
Jan 2007 in Case C–328/05 P SGL Carbon v Commission (Specialty graphite), not yet reported, para 93:
account. That ability can be relevant only in a ‘specific social context’, namely the
consequences which payment of a fine could have, in particular, by leading to an
increase in unemployment or deterioration in the economic sectors upstream or
downstream of the undertaking concerned’.182
‘the fact that a measure taken by a Community authority results in the insolvency or liquidation of a
particular undertaking is not precluded as such by Community law’.
182
Ibid, para 106; see further Killick, above n 45, at 6, and A Stephan, ‘The Bankruptcy Wildcard in
Cartel Cases’, CCP Working Paper 06–5 (Mar 2006).
5
Leniency: Theory and Practice
5.1 INT RODUCTION
5.1.1 Definition
344. This chapter deals with leniency in antitrust enforcement, defined here as
the granting of immunity from penalties or the reduction of penalties for
antitrust violations in exchange for cooperation with the antitrust enforcement
authorities.
345. The cooperation could consist in the provision of intelligence and/or
evidence of the antitrust violations, and/or in the recognition of the violation and
acceptance of the reduced penalty.1 It could possibly also involve acceptance of
remedial or compensatory measures.2
346. The penalties that are waived or reduced could be any penalties that can
be imposed or sought by the antitrust enforcement authorities in the jurisdiction
concerned: fines on companies, fines on individuals, director disqualification
and/or imprisonment.3
5.1.2 Examples
5.1.2.1 US Department of Justice
347. The US Department of Justice has a Corporate Leniency Policy (also

known as corporate amnesty or corporate immunity policy), under which it
1
Typically the cooperation and the penalties that are waved or reduced relate to the same antitrust
violation; see however below, paras 456–463, as to penalty reductions for one violation in exchange
for cooperation concerning another violation.
2
See paras 413 and 414 below; this ch does not deal with settlements of antitrust investigations in
cases where no penalties are envisaged but only remedial or compensatory measures; on such
settlements see Ch 1 above.
3
On the choice of penalties for antitrust offences see below, paras 443–450 and Ch 6.
114 LENIENCY: THEORY AND PRACTICE
accords immunity to corporations reporting their illegal antitrust activity at an

early stage.4 Immunity means in this case not charging such a firm criminally for
the activity being reported.5
348. Immunity will always be granted if, at the time the corporation comes
forward to report the illegal activity, the US Department of Justice has not
received information about the illegal activity from any other source. A number
of other conditions must also be met, in particular that the corporation, upon its
discovery of the illegal activity being reported, took prompt and effective action
to terminate its part in the activity; that the corporation reports the wrongdoing
with candour and completeness and provides full, continuing and complete
cooperation with the US Department of Justice throughout the investigation;
that, where possible, the corporation makes restitution to injured parties; and
that the corporation did not coerce another party to participate in the illegal
activity and clearly was not the leader in, or originator of, the activity.6
349. If a corporation thus qualifies for immunity, all its directors, officers and
employees who admit their involvement as part of the corporate confession will
receive immunity, in the form of not being charged criminally with the illegal
activity, if they admit their wrongdoing with candour and completeness and
continue to assist the US Department of Justice throughout the investigation.7
350. If the US Department of Justice has already received information about
the illegal activity, and possibly already started an investigation, immunity may
still be granted to a corporation that is the first to come forward and qualify for
immunity, provided that, at the time the corporation comes in, the US Depart-
ment of Justice does not yet have evidence against the company that is likely to
result in a sustainable conviction. The same conditions as to prompt termination,
full cooperation and restitution have to be met. Moreover, immunity will be
granted only if the US Department of Justice determines that granting immunity
would not be unfair to others. In applying this last condition, the primary
considerations are how early the corporation comes forward and whether it
coerced another party to participate in the illegal activity or clearly was the leader
in, or originator of, the activity.8
4
US Department of Justice, Corporate Leniency Policy (10 Aug 1993), available at
www.usdoj.gov/atr/public/guidelines/0091.pdf.
5
Ibid, at 1.
6
Ibid, at 1–2.
7
Ibid, at 4. Where a corporation is granted immunity after the US Department of Justice has
already received information about the illegal activity, under the conditions described below (para
350), the directors, officers and employees who come forward with the corporation will be considered
for immunity from criminal prosecution on the same basis as if they had approached the US
Department of Justice individually, under the conditions of the Leniency Policy for Individuals: see
below, para 352.
8
Ibid, at 2–3.
INTRODUCTION 115
351. As a result of a relatively recent statutory modification,9 companies that

have been accepted by the US Department of Justice in its immunity programme
and that are found by the court in follow-on actions for damages to have
provided satisfactory cooperation with the private claimants are liable for only
single instead of treble damages, and the normal rule of joint and several liability
for co-conspirators does not apply to the immunity applicant.10
352. The US Department of Justice also has a Leniency Policy for Individuals,
which applies to individuals who approach the Division on their own behalf, not
as part of a corporate confession.11 Leniency again means here immunity from
criminal prosecution.12 Immunity will be granted to an individual reporting
illegal antitrust activity before an investigation has begun if, at the time the
individual comes forward, the US Department of Justice has not received
information about the illegal activity from any other source; the individual
reports the wrongdoing with candour and completeness and provides full,
continuing and complete cooperation with the US Department of Justice
throughout the investigation; and the individual did not coerce another party to
participate in the illegal activity and clearly was not the leader in, or instigator of,
the activity.13
353. The Corporate Leniency Policy and the Leniency Policy for Individuals
apply only to (hard-core) cartels. Indeed, although the text of the criminal
provisions in the US Sherman Act covers all types of restrictive agreements as
well as unilateral monopolistic behaviour, by long tradition the US Department
of Justice limits criminal prosecution to hard-core violations of section 1 of the
Sherman Act such as price fixing, bid rigging, market division, or customer
allocation schemes among horizontal competitors.14
354. As already mentioned, beneficiaries of the Corporate Leniency Policy
and the Leniency Policy for Individuals are not charged criminally. As regards the
9
Antitrust Criminal Penalty Enhancement and Reform Act of 2004, Pub L 108–237, §§ 201–214,
118 Stat 661 (22 June 2004), 15 USC § 1 note (2004 Act).
10
Ibid, § 213.
11
US Department of Justice, Leniency Policy for Individuals (10 Aug 1994), available at
www.usdoj.gov/atr/public/guidelines/0092.pdf; see also above n 7.
12
Ibid, at 1.
13
Ibid, at 1–2.
14
SW Waller, ‘The Incoherence of Punishment in Antitrust’ (2003) 78 Chicago-Kent Law Review
207 at 216, referring in n 48 to US v Dunham Concrete Prods, Crim No 1842 (ED La 1969) as the last
criminal monopolisation indictment discovered by that author; and US Antitrust Modernization
Commission, Report and Recommendations (Washington, DC, Antitrust Modernization Commission,
Apr 2007), at 296–7 and 303. As the US Department of Justice has no powers to impose or to seek the
imposition of civil or administrative fines, it only seeks prospective, injunctive relief against other
violations of s 1 and all violations of s 2 of the Sherman Act. The Federal Trade Commission has
relatively recently in competition cases applied its power, which it had traditionally used in consumer
protection cases, to seek disgorgement of unlawful profits. Otherwise deterrence and punishment are
left to private treble-damages actions; see again Waller, above, at 210, 218 and 230–2; S Calkins,
‘Corporate Compliance and the Antitrust Angencies’ Bi-Modal Penalties’ (1997) 60 Law and Contem-
porary Problems 127 at 136–9 and 150; and the US Antitrust Modernization Commission report,
above, at 285–91.
other corporations and individuals that are charged, the US Department of

Justice may enter into plea agreements or plea/cooperation agreements.15
355. The US Department of Justice has discretion whether or not it enters
into a plea agreement. One of the minimal requirements is that the defendant is
willing to assume responsibility for its conduct, ie admits guilt.16 Under a plea
agreement, an agreed joint recommendation is normally made to the court as to
the penalty to be imposed (amount of the corporate fine, or length of the prison
term and/or amount of the individual fine). The agreement is binding on the
defendant and on the US Government, but not on the court. It appears however
that courts normally follow such agreements.17 The defendant also agrees to
waive the right to appeal the sentence imposed by the court, if that sentence
follows the joint recommendation.
356. The plea agreement is made in the light of the US Sentencing Guidelines,
which give the courts advice on sentencing.18 The Sentencing Guidelines deter-
mine guideline ranges for prison terms, individual fines and corporate fines,
depending on the volume of commerce affected by the antitrust offence and on
the defendant’s culpability.19
357. Under Sections 3E1.1 and 8C2.5 of the Sentencing Guidelines, recogni-
tion and acceptance of responsibility for the offence, possibly including the
voluntary payment of restitution, reduces the defendant’s culpability.
358. Section 1B1.8 of the Sentencing Guidelines provides that where a
defendant agrees to cooperate with the government by providing information
concerning unlawful activities of others, and as part of that cooperation agree-
ment the government agrees that self-incriminating information provided pursu-
ant to the agreement will not be used against the defendant, then such
information shall not be used in determining the applicable guideline range,
unless the information was already known to the government prior to entering
15
See US Department of Justice, Antitrust Division, Grand Jury Manual, Ch 9, Plea Agreements
(Nov 1991), available at www.usdoj.gov/atr/public/guidelines/207144.pdf.
16
Ibid, at 7. In limited circumstances, where employees have died or are not subject to the
corporation’s control, a corporate defendant’s agreement to accept as true the Government’s version
of the offence may be enough.
17
J Ratliff, ‘Plea Bargaining in EC Anti-Cartel Enforcement—A System Change’, paper presented
at the 11th EUI Competition Law and Policy Workshop (Florence, 2–3 June 2006), available at
www.iue.it/RSCAS/Research/Competition/Index.shtml, forthcoming in C-D Ehlermann and I Atana-
siu (eds), European Competition Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart
Publishing, 2007), at 13.
18
US Sentencing Commission, Guidelines Manual (Nov 2005), available at www.ussc.gov/
2005guid/TABCON05.htm. Since the judgment of the US Supreme Court in United States v Booker,
125 S Ct 738 (2005), the US Sentencing Guidelines are no longer mandatory but only advisory. As to
the impact of this change see SD Hammond, ‘Antitrust Sentencing in the Post-Booker Era: Risks
Remain High For Non-Cooperating Defendants’, address to the American Bar Association Section of
Antitrust Spring Meeting (Washington, DC, 30 Mar 2005), available at www.usdoj.gov/atr/public/
speeches/208354.pdf. See also US Antitrust Modernization Commission, Criminal Remedies Discus-
sion Memorandum (4 May 2006), available at www.amc.gov/pdf/meetings/
CrimRem_DiscMemo_060504-fin.pdf.
19
See in particular Sect 2R1.1 of the Sentencing Guidelines.
INTRODUCTION 117
into the cooperation agreement, or unless there is a breach of the cooperation

agreement by the defendant. This possibility (which amounts to the grant of
partial immunity) is used by the US Department of Justice in situations where a
company that cannot benefit from (full) immunity under the Corporate Leni-
ency Policy because it is not the first to come in, as part of its cooperation
pursuant to a plea agreement reveals that the suspected cartel was broader than
had been previously identified—either in terms of its duration or of the prod-
ucts, contracts or commerce affected.20
359. Sections 5K1.1 and 8C4.1 of the Sentencing Guidelines provide that the
court may depart from the guidelines to impose a sentence below the minimum
guideline range, upon a motion from the government that the defendant has
provided substantial assistance in the investigation or prosecution of another
defendant. The appropriate reduction is to be determined considering the
significance and usefulness of the assistance, taking into account the govern-
ment’s evaluation of the assistance rendered, its nature and extent, and its
timeliness. The US Department of Justice’s practice is to agree cooperation
discounts with an average in the order of 30 to 35 per cent for companies that are
the second to cooperate. In one extreme case, a 59 per cent discount was
granted.21 A third cooperating company receives a significantly lower discount,
the fourth even less, and so on.22 Moreover, the US Department of Justice’s
practice for the second cooperating company is to take as the starting point for
the discount the bottom of the guideline fine range, unless the company had a
significant leadership role in the cartel, whereas a higher starting point is taken
for subsequent cooperating companies.23
360. There are however no fixed discounts for cooperation outside the
Corporate Leniency Policy and the Leniency Policy for Individuals. The size of
the discount is determined in each specific case.24
361. Finally, the US Department of Justice has a practice (called ‘Amnesty
Plus’) of granting an additional discount to a cooperating company that reports
an unrelated antitrust violation (in addition to the immunity which the company
may obtain for this other violation under the Corporate Leniency Policy).25
20
SD Hammond, ‘Measuring the Value of Second-In Cooperation in Corporate Plea Negotia-
tions’, address to the 54th Annual American Bar Association Section of Antitrust Law Spring Meeting
(Washington, DC, 29 Mar 2006), available at www.usdoj.gov/atr/public/speeches/215514.pdf, at 3–4.
21
Ibid, at 5 and 2; the one extreme case was United States v Crompton Corporation, No CR
04–0079 MJJ (ND Cal 2004).
22
Ibid, at 11–12.
23
Ibid, at 5–7. Apart from the case of a significant leadership role, a higher starting point is also
taken if the company failed to discover and report the offence when it was earlier investigated for a
cartel in another market; the latter practice is called ‘Penalty Plus’.
24
Ibid, at 2. Cf the percentage brackets in the European Commission’s Leniency Notice, below,
para 369.
25
Ibid, at 9–10 and 14; see further below, paras 458–463.
5.1.2.2 The European Commission
362. The European Commission has a Leniency Notice, which applies to secret
cartels between two or more competitors involving practices such as fixing prices,
production or sales quotas, sharing markets including bid-rigging, restricting
imports or exports, and/or anti-competitive actions against other competitors.26
363. The Commission will grant immunity from any fine which would
otherwise have been imposed on an undertaking disclosing its participation in a
cartel if that undertaking is the first to submit information and evidence which in
the Commission’s view will enable it to carry out a targeted inspection in
connection with the cartel, and if, at the time of the application for immunity, the
Commission did not have sufficient evidence to adopt a decision to carry out an
inspection and had not yet carried out such an inspection.27
364. If no undertaking has been granted immunity on the above ground, the
Commission will alternatively grant immunity to the undertaking that is the first
to submit information and evidence which will enable the Commission to find
an infringement of Article 81 EC in connection with the cartel, provided that the
Commission did not, at the time of the submission, have enough evidence to
make such a finding.28
365. A number of additional conditions must be met in any case to qualify for
immunity: the undertaking must cooperate genuinely, fully, on a continuous
basis and expeditiously from the time it submits its application throughout the
Commission’s administrative procedure; and the undertaking must have ended
its involvement in the cartel immediately following its application, except for
what would, in the Commission’s view, be reasonably necessary to preserve the
integrity of the inspections.29 Moreover, an undertaking which took steps to
coerce other undertakings to join the cartel or to remain in it is not eligible for
immunity.30
366. Contrary to the US Department of Justice’s Corporate Leniency Policy,
immunity under the European Commission’s Leniency Notice does not mean
absence of prosecution. The Commission will normally still adopt a decision
26
Commission Notice on immunity from fines and reduction of fines in cartel cases [2006] OJ
C298/17; see Ch 4 above.
27
Ibid, pts 8(a) and 10. The inspections referred to are those provided for in Art 20(4) of Council
Reg 1/2003 of 16 Dec 2002 on the implementation of the rules on competition laid down in Arts 81
and 82 of the Treaty [2003] OJ L1/1. On the evidence required to conduct an inspection see the
Judgment of the ECJ of 22 Oct 2002 in Case C–94/00 Roquette Frères [2002] ECR I–9011; see further
CS Kerse and N Khan, EC Antitrust Procedure, 5th edn (London, Sweet & Maxwell, 2005); L Ortiz
Blanco (ed), EC Competition Procedure, 2nd edn (Oxford, Oxford University Press, 2006); and Ch 1
above.
28
Notice, above n 26, pts 8(b) and 11.
29
Ibid, pts 12 and 22. If the undertaking has not met these conditions, it cannot qualify either for
a fine reduction. See also Case T–12/06 Deltafina v European Commission, pending before the CFI.
30
Ibid, pt 13. It may however still qualify for a fine reduction if it fulfills the relevant requirements
and meets all the conditions therefor.
INTRODUCTION 119
finding an infringement of Article 81 EC. The fact that the undertaking cooper-
ated with the Commission during its administrative procedure will be indicated
in the decision, so as to explain the non-imposition of a fine.31
367. As the European Commission currently has no powers to impose penal-
ties on individuals other than undertakings,32 the grant of immunity to an
undertaking does not concern its directors or employees, and there is no other
immunity policy for individuals.
368. Undertakings disclosing their participation in a cartel that do not meet
the conditions for immunity may still under the Leniency Notice be eligible for a
reduction of any fine that would otherwise have been imposed if they provide the
European Commission with evidence which represents significant added value
compared to the evidence already in the European Commission’s possession, and
provided that they fulfill the same conditions as to genuine, full cooperation and
as to termination of the infringement as required for the immunity applicants.33
369. The first undertaking to provide such significant added value will receive
a reduction of 30 to 50 per cent of the fine which would otherwise have been
imposed; the second undertaking a reduction of 20 to 30 per cent, and subse-
quent undertakings a reduction of up to 20 per cent.34
370. The degree of added value depends on the extent to which the evidence
provided strengthens, by its very nature and/or its level of detail, the European
Commission’s ability to prove the infringement. The degree of corroboration
from other sources required for the evidence submitted to be relied upon against
other undertakings involved in the cartel will have an impact on the value of that
evidence, so that compelling evidence will be given a greater value than uncor-
roborated or simply corroborating statements.35
371. Guidance on the normal amount of the fine, and thus on the size of the
reduction in absolute terms, can be found in the European Commission’s Fining
Guidelines.36 The basic amount of the fine normally corresponds to a proportion
of the value of the undertaking’s sales of goods or services to which the
infringement directly or indirectly relates during the last full business year of its
participation in the infringement, multiplied by the number of years of the
infringement. The proportion depends on the gravity of the infringement. As a
general rule, it will be set at a level of up to 30 per cent. For secret cartels it will
generally be set at the higher end of the scale. Irrespective of the duration of the
31
Ibid, pt 38.
32
It would however be possible under the current EC Treaty to modify Reg 1/2003 so as to
provide for the imposition of penalties on individuals other than undertakings; see below, paras
618–624, and Judgment of the ECJ of 13 Sept 2005 in Case C–176/03 European Commission v EU
Council [2005] ECR I–7879.
33
Leniency Notice, above n 26, pts 23 and 24; and above, para 365.
34
Ibid, pt 26.
35
Ibid, pt 25.
36
Guidelines on the method of setting fines imposed pursuant to Art 23(2)(a) of Reg 1/2003
[2006] OJ C210/2; for a detailed analysis see Ch 4 above.
infringement, a sum of between 15 and 25 per cent of the value of sales will in
addition be included in the basic amount.37 The basic amount may then be
increased or reduced to take into account aggravating or mitigating circum-
stances.38 One of the aggravating circumstances listed in the Fining Guidelines is
the refusal to cooperate with or obstruction of the European Commission in
carrying out its investigations.39 Another aggravating circumstance mentioned is
the role of leader in, or instigator of, the infringement; the European Commis-
sion will also pay particular attention to any steps taken to coerce other
undertakings to participate in the infringement and/or any retaliatory measures
taken against other undertakings with a view to enforcing the practices constitut-
ing the infringement.40
372. Finally, the European Commission’s Leniency Notice also provides for
partial immunity like Section 1B1.8 of the US Sentencing Guidelines.41 If an
applicant for a fine reduction is the first to submit compelling evidence which the
European Commission uses to establish additional facts increasing the gravity or
the duration of the infringement, the European Commission will not take such
additional facts into account when setting any fine to be imposed on the
undertaking which provided this evidence.42
373. Contrary to the US Department of Justice’s Corporate Leniency Policy,43
the European Commission’s Leniency Notice does not contain any condition of
restitution being made, where possible, to injured parties. As regards the impact
of the granting of leniency on follow-on actions for damages, the Leniency
Notice states that the fact that immunity or reduction in respect of fines is
granted cannot protect an undertaking from the civil law consequences of its
participation in an infringement of Article 81 EC.44
374. Whereas the Leniency Notice only applies to secret cartels between
competitors, the European Commission also regularly uses its powers to impose
fines on undertakings for other types of infringements of Article 81 EC, such as
vertical agreements restricting imports or exports between EU Member States,45
or for abuses of a dominant position prohibited by Article 82 EC.46
375. For these other types of infringements, cooperation can be rewarded
with a fine reduction under the Fining Guidelines,47 which provide that the basic
37
Ibid, pts 13–26.
38
Ibid, pts 27–29.
39
Ibid, pt 28, second indent.
40
Ibid, pt 28, third indent.
41
42
Leniency Notice, above n 26, pt 26, last sentence.
43
44
Leniency Notice, above n 26, pt 38; see further below n 114.
45
See for instance Commission Dec of 5 Oct 2005 in Cases COMP/E2?36623, 36820 and 37275,
Automobiles Peugeot SA and Peugeot Nederland BV [2006] OJ L173/20.
46
See for instance Commission Dec of 24 Mar 2004 in Case COMP/C-3/37.792 Microsoft,
available at http://ec.europa.eu/comm/competition/antitrust/cases/decisions/37792/en.pdf.
47
INTRODUCTION 121
amount of the fine may be reduced where the European Commission finds that
mitigating circumstances exist, such as: where the undertaking concerned has
effectively cooperated with the European Commission outside the scope of the
Leniency Notice and beyond its obligation to do so; and where the undertaking
has terminated a non-secret infringement as soon as the European Commission
intervened.48
376. The Leniency Notice and the Fining Guidelines currently provide for the
grant of immunity from fines or the reduction of fines only in exchange for the
provision of intelligence and evidence of the antitrust violations, not in exchange
for the recognition of the infringement and acceptance of the reduced fine.49
377. Under an earlier version of its Leniency Notice,50 the European Commis-
sion also granted a reduction of the fine (typically of 10 per cent) where, after
receiving a statement of objections, the undertaking informed the Commission
that it did not substantially contest the facts on which the European Commission
based its allegations. The European Commission is currently considering intro-
ducing a new settlement policy further along this line, offering a fine reduction in
exchange for not merely a non-contestation of facts, but rather a recognition of
the infringement and a commitment to pay the fine.51
5.1.2.3 Competition Authorities of the EU Member States
378. The competition authorities of most EU Member States operate a leniency

programme for undertakings.52 These programmes are in many respects similar
to the European Commission’s leniency programme.
379. A number of differences however exist. For instance, in the United
Kingdom, the Office of Fair Trading’s leniency programme for undertakings53
applies not only to horizontal cartels but also to vertical retail price maintenance
agreements.54 Contrary to the European Commission’s Leniency Notice and to
the leniency programmes in all other EU Member States, it also contains a
48
Fining Guidelines, above n 36 pt 29, fourth and first indents.
49
Cf the plea agreements by the US Department of Justice above, paras 354–356.
50
Commission Notice on the non-imposition or reduction of fines in cartel cases [1996] OJ
C207/4. This notice was replaced in 2002: see above n 26.
51
See speech by Commissioner N Kroes, ‘Reinforcing the Fight against Cartels and Developing
Private Antitrust Damage Actions: Two Tools for a More Competitive Europe’ (Commission/IBA
Joint Conference on EC Competitiona Policy, 8 Mar 2007), available at http://ec.europa.eu/
commission_barroso/kroes/speeches_en.html; see also sect 2.5.3 above.
52
A list of all the competition authorities of the EU Member States that operate a leniency
programme is available at http://ec.europa.eu/comm/competition/antitrust/legislation/
authorities_with_leniency_programme.pdf.
53
Part 3, ‘Lenient treatment for undertakings coming forward with information in cartel activity
cases’, of OFT’s Guidance as to the Appropriate Amount of a Penalty (London, OFT, Dec 2004),
supplemented by Leniency and No-action—OFT’s Draft Final Guidance on the Handling of Applica-
tions (London, OFT, Nov 2006), available at www.oft.gov.uk.
54
Ibid, nn 23 and 8.
provision similar to the ‘Amnesty Plus’ practice of the US Department of

Justice.55 In Germany, the leniency programme of the Bundeskartellamt excludes
from immunity not only cartel participants that coerced others to participate in
the cartel, but also the cartel participant that was the only ringleader of the
cartel.56
380. In a number of EU Member States, the cartel prohibitions are enforced
not only by fines on undertakings, but also by criminal or other penalties on
individuals.57 Some of these Member States also have corresponding leniency
provisions for individuals. In the United Kingdom, for instance, directors of
companies that have committed competition law breaches can be subjected to
director disqualification orders, and individuals involved in cartel conspiracies
can be punished with imprisonment and/or fines under the criminal cartel
offence. The Office of Fair Trading has a policy of no-action letters for individu-
als, which grant immunity from prosecution under the cartel offence to cooper-
ating individuals, comparable with the US Department of Justice’s Leniency
Policy for Individuals.58 An undertaking can apply for no-action letters on behalf
of named employees, directors, ex-employees or ex-directors when seeking
leniency under the Office of Fair Trading’s leniency programme for undertakings,
or in conjunction with an application for leniency from the European Commis-
sion in accordance with the latter’s Leniency Notice.59 The Office of Fair Trading
will also not seek director disqualification orders against individuals who benefit
from no-action letters or who are directors of companies that benefit from
leniency under the Office of Fair Trading’s leniency programme or under the
European Commission’s Leniency Notice.60 On the other hand, in Germany, for
instance, a specific criminal offence, punishable with imprisonment, exists for
bid-rigging, but there is no leniency programme for this offence.
5.1.3 History
381. The contemporary practice of leniency in antitrust enforcement is generally

considered to have been started by the US Department of Justice in 1978, when it
adopted its first Corporate Leniency Policy. The basic idea of granting immunity
55
Ibid, para 3.16; see above, para 361, and below, paras 458–463.
56
Notice 9/2006 of the Bundeskartellamt on the immunity from and reduction of fines in cartel
cases—Leniency Programme—of 7 Mar 2006, available at www.bundeskartellamt.de/wEnglisch/
download/pdf/06_Bonusregelung_e.pdf.
57
See Ch 6 below, in particular paras 548–543.
58
Office of Fair Trading, The Cartel Offence—Guidance on the Issue of No-action Letters for
Individuals (London, OFT, Apr 2003), supplemented by Leniency and No-action—OFT’s Interim Note
on the Handling of Applications (London, OFT, July 2005), available at www.oft.gov.uk; see above, para
352.
59
Ibid, para 3.5; see above, paras 362–367.
60
Ibid, para 3.15 and OFT, Competition Disqualification Orders—Guidance (London, OFT, May
2003), available at www.oft.gov.uk, paras 4.12 and 4.27.
INTRODUCTION 123
or a reduced penalty in exchange for cooperation to help convict co-conspirators

is however probably as old as criminal prosecution.61
382. The US Department of Justice replaced its initial Corporate Leniency
Policy of 1978 with its current Corporate Leniency Policy in 1993. The three
main changes were that the grant of immunity before the start of an investigation
became automatic instead of being a matter of prosecutorial discretion, that
immunity also became available after an investigation had started, and that
immunity was automatically extended to cooperating directors and employees
coming forward together with the cooperating corporation. The US Department
of Justice also added its Leniency Policy for Individuals shortly afterwards in
1994.62 The most important change since then was introduced by the Antitrust
Criminal Penalty Enhancement and Reform Act of 2004, which grants partici-
pants in the US Department of Justice’s leniency programme the benefits of
single instead of treble damages and of the non-application of joint and several
liability with co-conspirators in follow-on private actions for damages.63
383. The European Commission’s first Leniency Notice was adopted in 1996,
and was clearly inspired by the US Department of Justice’s Corporate Leniency
Policy of 1993.64 The European Commission had however already in a number of
cases before 1996 reduced fines or even abstained from imposing fines in
recognition of cooperation received.65 The idea of granting immunity from fines
in exchange for cooperation with the European Commission also underlay
Article 15(5)(a) of Regulation 17, in force from 1962 to 2004, under the terms of
which no fine could be imposed in respect of acts taking place after notification
of an agreement to the European Commission and before its decision as to the
61
See NK Katyal, ‘Conspiracy Theory’ (2003) 112 Yale Law Journal 1307 at 1330–1, referring to
practices in English law going back to the early 12th century, as well as in early American law.
62
See above, sect 5.1.2.1 and AK Bingaman, ‘Antitrust Enforcement: Some Initial Thoughts and
Actions’, Address to the ABA Antitrust Section (New York, 10 Aug 1993), available at www.usdoj.gov/
atr/public/speeches/0867.pdf, at 7–9, and ‘Report from the Antitrust Division— Spring 1994’, Address
to the ABA Antitrust Spring Meeting (Washington, DC, 8 Apr 1994), available at www.usdoj.gov/atr/
public/speeches/0110.pdf, at 8–9, and SD Hammond, ‘Cornerstones of an Effective Leniency Pro-
gram’, Presentation at the ICN Workshop on Leniency Programmes (Sydney, 22–23 Nov 2004),
available at www.usdoj.gov/atr/public/speeches/206611.pdf, at 3–4.
63
See above, para 351
64
See above n 50, and WPJ Wils, ‘The Commission Notice on the Non-Imposition or Reduction
of Fines in Cartel Cases: A Legal and Economic Analysis’ (1997) 22 European Law Review 125. The
origin was even more visible in the earlier draft notice which was published by the European
Commission as part of the consultation procedure preceding the adoption of the 1996 Leniency
Notice [1995] OJ C341/13.
65
Examples of reduced fines include the Dec of 19 Dec 1984, in Wood Pulp [1985] OJ L85/1, para
148; Dec of 24 Apr 1986, in Polypropylene [1986] OJ L230/1, para 109; and Dec of 13 July 1994, in
Cartonboard [1994] OJ L243/52, paras 169–171. In its Dec of 1 Apr 1992, in Franco-West-African
Shipowners Committees [1992] OJ L134/1, para 74, sub (e), the European Commission decided that
‘there are grounds for exempting from fines the four shipping companies which although members of
the committees contributed in drawing the attention of the Commission to the practices dealt with in
this decision’; see also the Dec of 21 Dec 1994, in Tretorn [1994] OJ L378/45, para 77.
exemptability of the agreement under Article 81(3) EC, provided the acts fell
within the limits of the activity described in the notification.66
384. The European Commission replaced its initial Leniency Notice by a new
one in 2002. The main changes were that immunity became automatic, and that
fine reductions became more strictly aligned to the timing of the cooperation.67
In 2006 the European Commission again amended its Leniency Notice, mainly to
clarify the threshold for immunity in terms of information to be provided and to
clarify the duty of cooperation of all leniency applicants.68 As already mentioned
above, the European Commission is also considering introducing a new policy of
‘direct settlements’, under which fine reductions would be granted in exchange
for a recognition of the infringement and a commitment to pay the fine.69
385. The leniency programmes in the EU Member States have generally been
adopted following the example of the European Commission, and thus indirectly
of the US Department of Justice. The United Kingdom and Germany were the
first EU Member States to do so in 2000. Both have already supplemented or
amended their leniency programmes since.70
386. These relatively frequent modifications reflect a process of experimenta-
tion, and of emulation between jurisdictions.71 Competition authorities have
been learning how to design more effective leniency policies by trying them out.
They may also have to adjust leniency programmes in reaction to learning on the
part of the cartel participants and of the specialised lawyers defending them.
Indeed, leniency is a game played at two stages between the competition
authorities on the one hand and the cartel participants and their lawyers on the
other hand. At the stage of the application for leniency, the cartel participants’
lawyers will try to obtain for their clients immunity from penalties or as great a
reduction as possible, while giving as little evidence as possible of the antitrust
violations. Even if they obtain full immunity, it is generally against their interest
66
Council Reg 17 [1962] JO 13/204, [1959–62] OJ Spec Ed 87, replaced by Reg 1/2003, above n 27,
which abolished the entire notification system; see WPJ Wils, ‘Notification, Clearance and Exemption
in EC Competition Law: An Economic Analysis’ (1999) 24 European Law Review 139 and ch 1 of WPJ
Wils, Principles of European Antitrust Enforcement (Oxford, Hart Publishing, 2005).
67
See above, paras 362–377and F Arbault and F Peiró, ‘The Commission’s New Notice on
Immunity and Reduction of Fines in Cartel Cases: Building on Success’, EC Competition Policy
Newsletter, June 2002, 15–22, available at http://ec.europa.eu/comm/competition/publications/cpn/
cpn2002_2.pdf.
68
See above, paras 362–377, in particular n 26; O Guersent, ‘The EU Model of Administrative
Enforcement against Global Cartels—Evolving to Meet Challenges’, paper presented at the 11th EUI
Competition Law and Policy Workshop (Florence, 2–3 June 2006), available at www.iue.it/RSCAS/
Research/Competition/Index.shtml, forthcoming in C-D Ehlermann and I Atanasiu (eds), European
Competition Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart Publishing, 2007);
and press release IP/06/1705 and MEMO/06/469 of 7 Dec 2006, available at http://europa.eu/rapid/.
69
70
71
See WE Kovacic, ‘Background Note: Using Evaluation to Improve the Performance of Compe-
tition Policy Authorities’, in OECD, Evaluations of the Actions and Resources of Competition Authorities
(Paris, OWCD, 16 Dec 2005), DAF/COMP(2005)30, 19, at 22, available at www.oecd.org/dataoecd/7/
15/35910995.pdf.
CONTRIBUTION TO OPTIMAL ANTITRUST ENFORCEMENT 125
to have the violation fully established, as this can only increase their liability in
follow-on private actions for damages.72 At the stage of the formation and
maintenance of cartels, cartel participants will try to adapt their organisation to
the leniency policies, so as to mimimise the destabilising effect, or even (if the
leniency policies are badly designed) to exploit the leniency policies to their
advantage.73
387. As already mentioned, competition authorities in different jurisdictions
have been emulating each other in designing effective leniency programmes.
Systematic exchanges of experience have been taking place in the framework both
of the OECD and of the ICN (International Competition Network).74 In Europe,
the ECN (European Competition Network), which groups together the European
Commission and the competition authorities of the EU Member States, has
recently launched an ‘ECN Model Leniency Programme’.75
5.2 CONT R IBUTION TO OPTIMAL ANTIT R UST ENFORCEMENT
5.2.1 Optimal Antitrust Enforcement and the Imposition of Penalties
388. Before examining how the granting of immunity from penalties for anti-
trust violations or the reduction of such penalties can contribute to optimal
antitrust enforcement, it may be useful to recall the role of the imposition of
penalties in the optimal enforcement of the antitrust prohibitions.
389. As I have argued in more detail in Chapter 3 above,76 the imposition of
penalties on companies and individuals that are found to have breached the
antitrust prohibitions can in several ways contribute to the enforcement of these
prohibitions. First, the imposition of penalties may have a deterrent effect, in that
it creates a credible threat of prosecution and punishment which weighs suffi-
ciently in the balance of expected costs and benefits to deter calculating com-
panies and individuals from committing antitrust violations. Secondly, in the
case of collective violations such as cartels, the differentiation of penalties
72
See below, paras 451–454.
73
On cartel organisation and learning in general see MC Levenstein and VY Suslow, ‘What
Determines Cartel Success?’ (2006) 44 Journal of Economic Literature 43; on the effects of leniency on
cartel stability see below, paras 404–408 and 422–425.
74
See OECD, Fighting Hard-Core Cartels—Harm, Effective Sanctions and Leniency Programmes
(Paris, OECD, 2000), available at www.oecd.org/dataoecd/41/44/1841891.pdf; Hammond, above n
000 and ICN, Anti-Cartel Enforcement Manual, Chapter 2: Drafting and Implementing an Effective
Leniency Program, 2nd edn (Apr 2006), available at www.internationalcompetitionnetwork.org/
capetown2006/FINALFormattedChapter2-modres.pdf.
75
This Model Leniency Programme was posted on the ECN website at http://ec.europa.eu/comm/
competition/antitrust/ecn/ecn_home.html on 29 Sept 2006; see further C Gauer and M Jaspers,
‘Designing a European Solution for a ′one stop leniency shop′’ [2006] European Competition Law
Review 685, and K Dekeyser and M Jaspers, ‘A New Era of ECN Cooperation—Achievements and
Challenges with Special Focus on Work in the Leniency Field’ (2007) 30 World Competition 3.
76
See in particular above, paras 167–182
according to the role played by the different co-conspirators can have the effect of
raising the cost of setting up and running cartels.77 Thirdly, the public punish-
ment of those who violate the antitrust prohibitions may also have a moral effect,
in that it sends a message to the spontaneously law-abiding, reinforcing their
moral commitment to the antitrust prohibitions.78 Indeed, corporate managers
are not necessarily just maximisers of profits for themselves and their principals.
They may feel a moral responsibility to live within the law whether or not they
are likely to be caught, and this normative commitment could trump their
interest calculus.79 Psychological research suggests that normative commitment is
generally an important factor explaining compliance with the law.80 Moreover,
the detection and prosecution of antitrust violations with a view to the imposi-
tion of penalties may also have the additional direct benefit of bringing violations
to an earlier end. Finally, the imposition of penalties may contribute to corrective
justice, in the form of disgorgement of unjust enrichment,81 and possibly also in
the form of compensation, either because the imposition of public penalties
facilitates follow-on private actions for damages or because the making of
restitution to injured parties is somehow taken into account in the imposition of
the public penalties.82
390. The imposition of penalties for antitrust violations will however also
always have a cost. Two types of costs could be distinguished. First, it could have
undesirable side-effects. For instance, errors or the risk of errors in the imposi-
tion of penalties could lead to lawful and economically desirable conduct being
deterred. Secondly, the detection, prosecution and punishment of antitrust
violations will always have an administrative cost, which includes both the cost
borne by the public sector (cost of competition authorities, prosecutors and
courts) and the cost borne by the businesses or individuals concerned (cost of
lawyers and experts, management time).
391. The presence of cost has two implications. First, it raises the question
how much avoidance of antitrust violations (and how much corrective justice) is
worth the cost of attaining it. The answer to this question depends on the one
77
78
Criminal Law & Criminology 338 at 341–3; KG Dau-Schmidt, ‘An Economic Analysis of the Criminal
Law as a Preference-Shaping Policy’ [1990] Duke law Journal 1; CR Sunstein, ‘On the Expressive
Function of the Law’ (1996) 144 University of Pennsylvania Law Review 2021; DM Kahan, ‘Social
Influence, Social Meaning, and Deterrence’ (1997) 83 Virginia Law Review 349; NK Katyal, ‘Deter-
rence’s Difficulty’ (1997) 95 Michigan Law Review 2385; GE Lynch, ‘The Role of Criminal Law in
Policing Corporate Misconduct’ (1997) 60 Law and Contemporary Problems 23; DM Kahan, ‘Social
Meaning and the Economic Analysis of Crime’ (1998) 27 Journal of Legal Studies 609; and KG
Economics and the Law (Basingstoke, Macmillan, 1998) 84.
79
80
81
See above, para 182, n 38.
82
See below, para 413 and 414.
hand on how high the cost is and on the other hand on how much value society
attaches to the avoidance of antitrust violations (and to the pursuit of corrective
justice).83 In all likelihood something less than full enforcement is optimal.
Secondly, whatever level of avoidance of antitrust violations (and of corrective
justice) is aimed at, one should try to achieve that goal at the least possible cost.
5.2.2 Positive Effects of Leniency
392. Leniency can in several ways contribute to optimal antitrust enforcement,

depending on the type of cooperation concerned (provision of intelligence and
evidence of the antitrust violations and/or recognition of the violation and
acceptance of the penalty) and on whether the type of violations concerned are
committed by single offenders or are collective violations like cartels.
5.2.2.1 Improved Collection of Evidence and Intelligence
393. In order to be able to impose penalties or to seek the imposition of

penalties for antitrust violations, competition authorities need intelligence of the
existence of such violations, and evidence sufficient for the penalties to be upheld
in court.
394. Competition authorities can try to obtain the necessary intelligence and
evidence from three possible sources. First, they could themselves monitor
markets, observing publicly available information and data, and possibly use
economic analysis of these data to try to detect and prove violations. In particular
for cartels, this first method has not been very important in practice. Economic
evidence is generally insufficient in court to establish collusion, and screening for
cartels appears difficult, given current methods of economic analysis.84
395. Secondly, the authorities could obtain information from third parties.
Customers or competitors harmed by antitrust violations may bring complaints
to the authorities,85 and third parties may otherwise volunteer to provide
information. As Donald Baker has pointed out, ‘[t]he desire for revenge is more
picturesque, but it is still very much present. In my experience, disgruntled
83
See generally, with regard to deterrence, GJ Stigler, ‘The Optimum Enforcement of Laws’ (1970)
78 Journal of Political Economy 526 and KG Elzinga and W Breit, The Antitrust Penalties: A Study in
Law and Economics (Princeton, NJ, Yale University Press, 1976) at 9–15.
84
This may of course change with the development of better methods: see JE Harrington,
‘Behavioral Screening and the Detection of Cartels’, paper presented at the 11th EUI Competition Law
and Policy Workshop (Florence, 2–3 June 2006), available at www.iue.it/RSCAS/Research/
Competition/Index.shtml, forthcoming in Ehlermann and Atanasiu (eds), above n 17.
85
See European Commission’s Notice on the handling of complaints by the Commission under
Arts 81 and 82 of the EC Treaty [2004] OJ C101/5.
current employees, fired employees, former trade association officials, and even
ex-spouses and ex-lovers may be anxious to finger the individuals who they think
have done them in.’86
396. The third, and usually best, source of information is the companies and
individuals that have committed the antitrust violations themselves. For certain
types of violations, in particular secret price cartels, the undertakings that have
committed the violations and their staff may be the only ones holding the
information which the competition authorities need to detect and punish the
violations.
397. As I have analysed in more detail elsewhere,87 there are essentially three
ways in which competition authorities can obtain information from the com-
panies and individuals that have committed the antitrust violations: direct force,
compulsion and leniency. First, they could use direct, physical force or covert
intrusion to obtain evidence in the hands of the companies and their staff.
Indeed, depending on the applicable legal regime, competition authorities may
have powers to conduct inspections (so-called ‘dawn raids’) at business premises,
and inspections at private homes, or carry out directed surveillance or use covert
human intelligence sources.88 The use of inspections has at least two drawbacks.
First, it can only be used to obtain existing documents or other existing physical
or electronic evidence. Secondly, without help from the undertakings or staff
concerned, and unless the competition authorities already have precise intelli-
gence as to what documents or records they can find at what exact place, the
method is a very expensive one, in that the authorities will have to go through
many documents, files or records before finding relevant information.
398. In order to avoid the limitations and costs of the first method of
obtaining information through the direct use of force, competition authorities
could try to make companies and individuals cooperate in handing over the
information sought by using compulsion in the form of threatened sanctions for
refusal to cooperate. These could be of two kinds. Refusal to cooperate with
investigations could be punished as a separate offence, and such refusal could also
be punished through an increase in the amount of the penalty imposed for the
antitrust violations established following the investigation.
399. Compared with the first method of direct force, the second method of
making the company or its staff cooperate through the threat of punishment for
non-cooperation has two advantages. First, its use is not limited to obtaining
information contained in already existing documents or records. Secondly, the
method is less costly, in that the company and its staff will be able more easily to
86
DI Baker, ‘The Use of Criminal Law Remedies to Deter and Punish Cartels and Bid-Rigging’
(2001) 69 George Washington Law Review 693 at 708.
87
See WPJ Wils, ‘Self-incrimination in EC Antitrust Enforcement: A Legal and Economic
Analysis’ (2003) 26 World Competition 567, and ch 5 in Wils, Principles, above n 66.
88
For an overview of the investigative powers of the European Commission and the competition
authorities of the EU Member States see Ch 1 above.
locate or bring together the required elements of information than the competi-
tion authorities’ inspectors searching without help through a mass of business
records with which they are unfamiliar. The method of inducing cooperation
through the threat of sanctions also has its problems, however. First, as in the case
of the first method of direct force, the competition authorities already need to
have some level of intelligence concerning the antitrust violation. Secondly, it still
imposes a significant cost on the companies and individuals concerned, including
the economic cost in terms of resources spent on complying with the requests for
cooperation, as well as the psychic cost where individuals are subjected to
on-the-spot oral questions. Thirdly, the problem of reliability constitutes a
specific drawback of the second method of obtaining information under com-
pulsion. Information obtained under compulsion may be unreliable for two sets
of reasons. On the one hand, companies or individuals that have indeed commit-
ted antitrust violations or their staff may give misleading answers when ques-
tioned, in the hope of wrongfooting the investigating authorities and thus
escaping detection and punishment. The threat of sanctions for untruthful
answers may not be effective to the extent that the authorities are believed not to
be able to obtain the necessary information to establish their untruthfulness. On
the other hand, where the companies or persons being investigated are in reality
innocent, there is a risk that ingenious questioning will lead less sophisticated
respondents to make seemingly inculpatory statements, or, in situations of
on-the-spot oral questioning where psychological or physical pressure is created
or felt, the individuals questioned may end up making untruthful inculpatory
statements to escape from the pressure. Because of these costs and risks, the law
imposes some limits on the possible use of compulsion, in particular against
individuals, under the privilege against self-incrimination.89
400. Leniency is the third method of obtaining information from the com-
panies and individuals that committed the antitrust violations. Compared to the
first and second method, direct force and compulsion, this third method has clear
advantages. Indeed, contrary to the first method, it can be used to obtain all kinds
of information, not just existing documents or other existing physical evidence.
Like the second method, it saves on search costs, in that the collecting of relevant
information is done by the undertaking and its staff, who are most familiar with
it. But, contrary to the second method, it does not suffer from the same reliability
problems, as there is no clear incentive at least for immunity applicants to
provide unreliable information, given that they risk losing the benefit of immu-
nity if they provide misleading information and given that immunity from
punishment is the highest benefit they can obtain.90
89
For a detailed discussion see Wils, above n 87, as well as Ch 1 above; see also the Judgment of
the ECJ of 29 June 2006 and the Opinion of AG Geelhoed of 19 Jan 2006 in Case C–301/04 P
Commission v SGL Carbon [2006] ECR I–5915.
90
See also Opinion of AG Mischo of 18 May 2000 in Case C–298/98 P Finnboard v Commission
[2000] ECR I–10159, paras 24–27; see however below, paras 464–468. There can however be problems
401. It is however important to note that leniency is not a substitute for but a
complement to the other methods of collecting intelligence and evidence of
antitrust violations.91 Indeed, leniency can only work if the companies and
individuals concerned perceive a risk that the competition authorities will detect
and establish the antitrust violation without recourse to leniency. The risk could
be a specific one, where the competition authority is already collecting or
receiving information of the antitrust violation by other means (or is believed to
be doing so, or to be doing so in the near future), or a more general one, where a
competition authority, as a result of many other recent cases of successful
detection and prosecution, is believed to be good at it. In the case of collective
violations such as cartels, and if leniency policies are well designed, in that
immunity is granted only to the first co-conspirator to come forward, and
reductions in penalties are linked to the timing of the cooperation as compared
to the other co-conspirators, companies and individuals may decide to cooperate
out of fear that a co-conspirator may do so before them. Such a ‘race to
cooperate’ may amplify the positive effects of leniency, but again such a race can
only start if there is a risk that the competition authorities will detect and
establish the antitrust violation without recourse to leniency, or at least a belief by
at least one of the conspirators that at least one of the other co-conspirators may
believe that there is such a risk.92
402. Leniency not only allows for cheaper and more reliable collection of
intelligence and evidence of antitrust violations than would be possible if only
the other investigative methods were available. It also adds an incentive for cartel
participants to create and keep more evidence in the first place.93 This increases
the possibilities for competition authorities to find evidence through other
methods, for instance through ‘dawn raids’, and this increased possibility in turn
increases the incentives to apply for leniency, thus creating a virtuous circle.
403. More generally, the larger amount and better quality of evidence which
competition authorities may get hold of thanks to leniency (both because of the
reduced costs of collecting it and because more evidence is created and kept in
the first place) will allow for higher penalties to be upheld in court. This again
increases the incentives to apply for leniency, thus creating another virtuous
circle.94
with the reliability of information provided by applicants for reductions in penalties, as they may try
to obtain such reductions by wrongly incriminating other parties.
91
See Wils, above n 87, at 588, and Harrington, above n 84, at 14–16.
92
See also Wils, above n 64, at 133.
93
See C Aubert, P Rey and WE Kovacic, ‘The Impact of Leniency and Whistleblowing Programs
on Cartels’ (2006) 24 International Journal of Industrial Organization 1241 at 1260–4.
94
See also OECD, Prosecuting Cartels Without Direct Evidence of Agreement (Paris, OECD, 15 Feb
2006), DAF/COMP/GF(2006)3, at 4.
5.2.2.2 Increased Difficulty of Creating and Maintaining Cartels
404. Setting up and maintaining a successful cartel requires effort. The cartel
members have to select and coordinate their behaviour on mutually consistent,
collusive strategies, allowing the cartel participants as a group to increase profits,
and providing for a fair distribution of profits between them. They also need to
develop mechanisms to discourage cheating, involving monitoring, rewards and
punishments. In a dynamic economy, successful cartels may have to develop an
organisational structure that allows them to solve these problems continuously.95
405. A well-designed leniency programme can make these tasks more diffi-
cult. The possibility for a cheating cartel member to apply for leniency increases
the benefits of cheating, thus making collusion more difficult to sustain.96 It
increases uncertainty, making it more difficult for cartel participants to reach
agreement, diminishing trust among them, and increasing the need for costly
monitoring.97
406. This effect of raising the cost of setting up and maintaining cartels98 is
not unique to leniency. The same logic underlies the legal rule that cartels or
other restrictive agreements are legally unenforceable,99 and the practice of
imposing higher penalties for cartel members that play active roles in setting up
and running the cartel, and lower penalties for cartel members that are exclu-
sively passive.100
407. This effect of leniency may lead to cartels breaking down earlier than
they would otherwise or even not being formed in the first place.101
408. The size of this effect of leniency on cartel formation and stability (as
well as the size of the effect of leniency in terms of improved collection of
intelligence and evidence102) depends essentially on three factors:103 the expected
leniency discount, the probability of detection and punishment in the absence of
95
Levenstein and Suslow, above n 73, at 44–5; see also JE Harrington, ‘How Do Cartels Operate?’
(June 2006), available at www.econ.jhu.edu/People/Harrington/CartelOperations-6.28.06.pdf.
96
JE Harrington, ‘Optimal Corporate Leniency Pograms’ (Nov 2005), available at
www.econ.jhu.edu/People/Harrington/amnesty11–05.pdf, at 3, has called this the ‘Deviator Amnesty
Effect’ of leniency.
97
Katyal, above n 61, at 1342–50, and G Spagnolo, ‘Divide et Impera: Optimal Leniency
Programmes’, CEPR Discussion Paper No 4840 (Dec 2004), available at www.cepr.org/pubs/dps/
DP4840.asp, at 6.
98
Katyal, above n 61, at 1363, calls this ‘cost deterrence’.
99
See Art 81(2) EC.
100
This makes the setting up and functioning of cartels more difficult, because, faced with the
prospect of higher fines, there will be fewer volunteers to play active roles. Those who nevertheless
agree to do so are likely to want compensation in the form of a larger part of the gain, which is likely
to be difficult to agree on, and to weaken the sense of solidarity and mutual trust within the group:
see Katyal, above n 61, at 1341–6 and 1363–7; and above, paras 224–226, 309–313 and 324–327.
101
The two are of course linked, as the earlier breakdown means that the cartel is less profitable
and thus less likely to be entered into in the first place: see Wils, above n 64, at 130, 133 and 140, and
Harrington, above n 96, at 5–6.
102
103
See also Hammond, above n 62; Wils, above n 87, at 587; and Wils, above n 64, at 140.
cooperation, and the profitability of the cartel. The greater the discount from the
otherwise applicable penalty that is offered to the cartel member who defects and
cooperates with the competition authority, the stronger the incentive to cheat
and report will be. This discount is to be compared to the leniency available in
the event of cooperation at a later point in time, or of a lesser degree. For optimal
effect, leniency programmes should thus make the leniency discount dependent
on the timing of the cooperation (including in comparison with other cooperat-
ing cartel members) and its extent. Higher penalties for antitrust violations allow
for more effective leniency, as they allow for larger benefits to be offered to
leniency applicants.104 Clarity and certainty as to the leniency discounts on offer
are also important, as they reduce the risk for the potential leniency applicant of
opting for cooperation with the competition authority.105 As already mentioned
above, the effectiveness of leniency also crucially depends on the credibility of the
competition authority’s enforcement activity, more specifically on the risk, as
perceived by the potential leniency applicant, that the competition authority will
be able to detect, establish and punish the antitrust violation if that potential
leniency applicant decides not to apply for leniency. While this risk can be
enhanced by the good design of the leniency programme, creating a race to
cooperate between the different cartel members, it ultimately depends on (the
perception of) the competition authority’s ability to detect and punish the
antitrust violation without the help of leniency.106 Finally, the effect of leniency
will of course also depend on the profitability of the cartel. Cartels that are
already less profitable will of course be the first to break down, or not even come
into existence, as a result of leniency.
5.2.2.3 Lower Costs of Adjudication
409. Apart from its positive effects at the investigative stage (improved collection
of intelligence and evidence),107 leniency can also lead to cost savings at the
adjudicative stage,108 if the cooperation rewarded consists in the recognition of
the violation and acceptance of the penalty.
104
Conversely, higher penalties that cannot be waived or reduced as part of leniency make
leniency less effective: see below, below, paras 448, 452 and 456.
105
See OECD, above n 74, at 8, and Hammond, above n 62, at 19–20.
106
See above, para 401; OECD, above n 74, at 4; ICN, above n 74, at 2; and J Rosenstok, ‘Analysis
of Self-reporting in Law Enforcement against Cartels in the Netherlands’, Master Thesis, Maastricht
University (Mar 2005), available at www.encore.nl/publications/Rosenstok.pdf.
107
108
On the different systems of adjudication in antitrust enforcement in the EU and the US see
WPJ Wils, ‘The Combination of the Investigative and Prosecutorial Function and the Adjudicative
Function in EC Antitrust Enforcement: A Legal and Economic Analysis’ (2004) 27 World Competition
201, and sect 1.2.10 and ch 6 of Wils, Principles, above n 66.
410. In the case of the US Department of Justice, for instance, the practice of
plea agreements allows the saving of the costs of full litigation before the federal
district court, as well as the costs of further appeal.109
411. As for the European Commission, it currently adopts a reasoned decision
in every cartel case (typically of dozens or even a few hundred pages), finding the
infringement and imposing the fines. This decision can be appealed by each of
the companies concerned before the Court of First Instance, with the possibility
of a further appeal, limited to points of law, before the Court of Justice. On
average, three to four appeals are brought against every cartel decision.110
Significant cost savings could thus be made if a mechanism could be introduced
under which companies would recognise the violation and commit to paying an
accepted fine, as the European Commission’s decision would then no longer need
to contain the same detailed reasoning, and there would no longer be systematic
appeals to the EC Courts. As already mentioned above, the European Commis-
sion is currently considering the introduction of such a settlement mechanism.111
412. The savings in the costs of full litigation, or of fully reasoned decisions
and frequent appeals, allow or would allow the antitrust enforcement authorities
to investigate and punish more violations with their available resources, thus
increasing the probability of detection and punishment, and hence increasing
deterrence.
5.2.2.4 Restitution to Injured Parties
413. As already mentioned, one of the conditions for obtaining immunity under
the US Department of Justice’s Corporate Leniency Policy is that, where possible,
the corporation makes restitution to injured parties.112 Furthermore, under the
US Sentencing Guidelines, the voluntary payment of restitution can reduce the
defendant’s culpability and thus the amount of the penalties imposed.113 In the
United Kingdom, the Office of Fair Trading recently brought to an end its
investigation in the Independent Schools case. As part of the settlement, the
schools under investigation agreed to make ex gratia payments totalling £3
109
See above, paras 354 and 355. As has been pointed out by GM Grossman and ML Katz, ‘Plea
Bargaining and Social Welfare’ (1983) 73 American Economic Review 749, plea bargaining not only
saves resources, but can also function as an insurance and as a screening device. These positive effects
must however be weighed against possible negative effects, in particular the lowering of the penalty
level: see below, paras 415–418.
110
N Kroes, ‘The First Hundred Days’, speech at the 40th Anniversary of the Studienvereinigung
Kartellrecht (Brussels, 7 Apr 2005), available at http://ec.europa.eu/comm/competition/speeches/
index_speeches_by_the_commissioner.html, at 4; see also Case Associates, Casenote ‘Penalties for
Price-fixers—A Survey of Fines Imposed in 43 Cartels by the EC Commission’ (May 2006), available
at http://casecon.com/data/pdfs/casenote41.pdf, at 2, reporting that 87% of fines are appealed.
111
112
113
million to a charitable fund to benefit pupils who attended the schools, in

addition to the payment of a nominal penalty of £10,000 per school.114
414. Requiring restitution to injured parties as a condition for leniency may
contribute to the optimal pursuit of the enforcement objective of corrective
justice,115 in that it may otherwise not have been feasible for the injured parties to
obtain compensation through a follow-on private action for damages, or by
saving the costs of such follow-on litigation.
5.2.3 Risks of Negative Effects
5.2.3.1 Lowering of the Penalty Level
415. Waiving or reducing penalties to reward cooperation inevitably has a

negative effect on the penalty level, and thus on deterrence.116 It is thus crucial to
design and apply leniency policies in such a way that this negative effect is
outweighed by the positive effects discussed above,117 and that no more leniency
is granted than is strictly necessary to obtain these positive effects.118
416. The need to avoid unnecessary penalty reductions explains the limitation
of leniency policies that offer full immunity to secret, horizontal cartels, given
that these are the types of antitrust violations that are most difficult to detect.119
It also explains the limitation of the benefit of full immunity to the single first
114
Decision of the OFT No CA98/05/2006 Exchange of information on future fees by certain
independent fee-paying schools (20 Nov 2006; Case CE/2890–03); see also Decs of the European
Commission of 21 Oct 1998, in Pre-Insulated Pipe Cartel [1999] OJ L24/1, rec and of 30 Oct 2002, in
Nintendo [2003] OJ L255/33, recs 440 and 441, granting reductions of the fines imposed on
respectively ABB and Nintendo for violations of Art 81 EC in recognition of the fact that ABB had
paid ‘substantial compensation to Powerpipe and its previous owner’ and that Nintendo had ‘offered
substantial financial compensation to third parties identified in the Statement of Objections as having
suffered financial harm as a result of [Nintendo]’s violation’. The European Commission has refused
to grant such reductions in other cases, and the CFI has recently confirmed that addressees of fining
decisions have no right to obtain such reductions: see the Judgment of 27 Sept 2006 in Case T–59/02
Archer Daniels Midland v Commission, not yet reported, paras 349–355.
115
On the different objectives of antitrust enforcement and their relationship see above, paras
167–182.
116
Wils, above n 64, at 130; M Motta and M Polo, ‘Leniency Programs and Cartel Prosecution’
(2003) 21 International Journal of Industrial Organization 347; Harrington, above n 96, who has called
this the ‘Cartel Amnesty Effect’ of leniency; and Case Associates, above n 110, reporting that cartel
members have been receiving an average fine reduction of 42% through the European Commission’s
leniency programme (and a further 18% by disputing the fine in court).
117
118
Wils, above n 64, at 130 and 134; see also JM Connor, Global Price Fixing—Our Customers are
the Enemy (Boston, Mass, Kluwer Academic Publishers, 2001) at 387–8 and 561.
119
Ibid, at 131.
cartel member to cooperate.120 In order to establish the antitrust violation in

court, the evidence provided by the first leniency applicant may however not be
sufficient, and corroborating evidence may be needed from a second source.121
To the extent that this is the case, substantial leniency towards the second
cooperating cartel member may be necessary.122 Much more questionable is the
need to grant substantial leniency to the third, fourth, fifth, sixth or seventh
cooperating cartel member.123
417. One way to counteract the penalty-lowering effect of leniency may be
first to raise the level of penalties applicable in the absence of cooperation.124 If,
for instance, leniency leads to an average penalty reduction of 33 per cent, this
effect can be neutralised by first raising the penalty level applicable in the absence
of cooperation by 50 per cent. To some extent, this may indeed have happened
historically. In the US, leniency was introduced in 1978, and made more generous
in 1993 and 2004, whereas Congress raised the maximum jail term for antitrust
offences in 1974 and 2004, and the maximum fines for corporations in 1974,
1990 and 2004. The three changes in 2004 were part of the same Act.125 The
European Commission adopted its first Leniency Notice in 1996 and its first
Fining Guidelines, which reflected an increase in the level of fines, in early 1998.
The European Commission’s current reflections on introducing a new settle-
ments policy follow the publication of its new Fining Guidelines in 2006, which
120
Ibid, at 132–3; and OECD, above n 74, at 27. As already indicated above, paras 401–408 the
limitation of immunity to the single first leniency applicant also creates a race to be the first to
cooperate, and this ‘Race to the Courthouse Effect’ is a countervailing force to the ‘Cartel Amnesty
Effect’: see Harrington, above nn 96, and 116.
121
See generally B Vesterdorf, ‘Burden of Proof in Cartel Cases’, paper presented at the Nordic
Lawyers Academy—Challenges in Nordic Cartel Cases (Helsinki, 4–5 May 2006) at 10.
122
See for instance European Commission Dec of 30 Oct 2002 in Case COMP/E-2/37.784 Fine art
auction houses [2005] OJ L200/92, where, out of two cartel participants, the first received immunity
and the second a 40% reduction of the fine. One may observe that in this case the cartel had already
been detected by the US Department of Justice prior to the European Commission’s investigation: see
C Mason, The Art of the Steal—Inside the Sotheby’s-Christie’s Auction House Scandal (New York,
Putnam, 2004).
123
See for instance European Commission Dec of 17 Dec 2002 in Case COMP/37.667 Specialty
Graphite [2006] OJ L180/20, where, out of 8 cartel members, one received immunity, 6 a fine
reduction of 35% and the last a fine reduction of 10%. It should be noted that this decision applied
the 1996 version of the European Commission’s Leniency Notice, which was replaced in 2002, inter
alia to make the fine reductions more strictly dependent on their timing and usefulness: see above,
paras 368, 369 and 384.
124
Such general raising of the level of penalties in the absence of cooperation, which is
transparent and non-discriminatory, should be distinguished from an opportunistic raising, in a
specific case where cooperation has taken place, of the penalty from which the leniency discount is
deducted. The latter practice would undermine the clarity and certainty of the leniency policy, and
thus its effectiveness: see M Bloom, ‘Immunity/Leniency/Finalncial Incentives/Plea Bargaining’, paper
presented at the 11th EUI Competition Law and Policy Workshop (Florence, 2–3 June 2006), available
at www.iue.it/RSCAS/Research/Competition/Index.shtml, forthcoming in C-D Ehlermann and I
Atanasiu (eds), European Competition Law Annual 2006: Enforcement of Prohibition of Cartels
(Oxford, Hart Publishing, 2007), at 4; and above, para 408.
125
again reflect an increase in the level of fines.126 A similar pattern can be found in
the EU Member States. In France, for instance, a leniency programme was
introduced in 2001 by the same Act of Parliament which also increased the
maximum fine.127 One could argue that these general penalty increases could
have been introduced anyway, irrespective of the introduction of a leniency
policy, so that the fact remains that the leniency policy has the effect of lowering
the penalty level. On the other hand, the existence of a leniency policy, which
allows antitrust offenders to reduce their exposure to penalties by cooperating
with the authorities and which results in more and better evidence of the
antitrust violations being discovered, may precisely allow the higher penalty
levels, because the antitrust offenders’ failure to take advantage of the ability to
cooperate increases their culpability, and hence the acceptability of higher
penalties, and because the stronger evidence equally makes higher penalties
acceptable.
418. Another measure which could be taken to reduce the penalty-lowering
effect of leniency is to raise the requirements as to the quality of evidence that has
to be provided by the first cooperating cartel member applying for immunity. It
is often argued that the threshold as to the evidence to be provided by the first
immunity applicant should be kept low, so as maximise the number of leniency
applications.128 The ultimate objective of a leniency policy is however not a high
number of leniency applications, but stronger deterrence.129 Requiring very little
evidence from the first leniency applicant means that in more cases it will be
necessary to grant substantial leniency also to a second leniency applicant (and
possibly a third etc) in order to obtain enough evidence to be able to establish the
antitrust violation in court. This multiplication of leniency tends to reduce
deterrence. Moreover, requiring more evidence from the first leniency applicant
does not necessarily reduce the number of applications. It may instead increase
the incentive for cartel participants to create and keep more and stronger
evidence, which in turn increases the effectiveness of leniency (directly, as better
evidence is obtained through leniency, as well as indirectly, by increasing the
possibility for competition authorities to collect such evidence without resorting
to leniency, which again increases the incentive to seek leniency).130
5.2.3.2 Exclusive Reliance on Leniency
419. As explained above, a leniency policy can only work if, in the case of a
violation committed by a single offender, the offender fears that his violation will
126
127
See D Voillemot, Gérer la clémence (Brussels, Bruylant, 2005) at 59.
128
See for instance MJ Reynolds and DG$ Anderson, ‘Immunity and Leniency in EU Cartel Cases:
Current Issues’ [2006] European Competition Law Review 82 at 85.
129
See also sect 5.2.4 below.
130
See above, text to nn 93, 94 and 119–123. It should however be predictably clear what the
required threshold of evidence is: see above, text to nn 102–106.
be detected and punished by the antitrust enforcement authorities if he does not

apply for leniency. In the case of a cartel or other collective violation, leniency
will work only if one cartel member either believes that the cartel risks being
detected and punished without leniency or fears that at least one other cartel
member may hold such belief.131
420. A leniency policy will thus start working only if the antitrust enforce-
ment authority concerned has first built up a sufficient level of credibility as to its
capacity to detect and punish antitrust violations on its own. For the authorities
that have managed to do so, and subsequently operate a successful leniency
programme, there may be a risk in relying too much on this success. If for a
prolonged period they no longer detect and prosecute any cases other than
through leniency, they may lose their capacity to do so, or at least antitrust
offenders may start doubting their continued capacity to do so. If this were to
happen, the success of their leniency programmes would risk coming to an end.
421. It may thus be useful that antitrust enforcement authorities continue, at
least from time to time, detecting and punishing antitrust violations without the
help of leniency, and are seen to do so.132
5.2.3.3 Facilitation of the Creation and Maintenance of Cartels
422. As already mentioned above, successful cartels tend to be sophisticated

organisations, capable of learning. It is thus safe to assume that cartel participants
will try to adapt their organisation to leniency policies, not only so as to minimise
the destabilising effect, but also, where possible, to exploit leniency policies to
facilitate the creation and maintenance of cartels.133 This raises the question
whether there could be features of leniency programmes that risk being exploited
to perverse effect.134
423. The limitation of immunity to the single first cooperating cartel partici-
pant is clearly important in avoiding perverse effects.135 If this condition did not
exist, one could imagine all the members of a cartel collectively denouncing it to
the antitrust enforcement authorities and claiming leniency, only after having
fully implemented the cartel agreement up to its time of natural death.136
424. In situations where the same companies participate in a number of
cartels in different markets, or repeatedly form cartels over time, one could
imagine a system in which cartel participants take turns to apply for leniency,
131
132
On the European Commission see Case Associates, above n 110, reporting that in 4 out of 39
cartel cases between 1999 and 2004 no leniency reductions were given.
133
134
Other than the general penalty-lowering effect discussed above, paras 415–418.
135
See Wils, above n 64, at 133.
136
Cartels tend to die natural deaths, be it only after a long life, because of the entry or expansion
of non-cartel members: see Levenstein and Suslow, above n 73.
every time one of the cartels is (about to be) detected by the competition
authority.137 In jurisdictions where cartels are not only punished by fines on
companies, but also by imprisonment of individuals, such a system is most
unlikely to be attractive,138 but in jurisdictions without such individual penalties
it may work.139
425. Another potential perverse scenario to be considered is the use of
leniency as a mechanism to punish deviations from the cartel agreement. Again
the limitation of immunity to the single first cooperating cartel participant very
much reduces the risk of such perverse effect, as it makes it impossible for all the
other cartel members together to denounce the cartel to the antitrust enforce-
ment authority, obtain immunity themselves, and have the deviator penalised by
the authority. One may however wonder whether it would be possible to set up a
system in which one cartel member played a central role in organising and
administering the cartel, and was the only one to have all the evidence of the
cartel, and could use this position to deter the other cartel members from
deviating from the cartel agreement under the threat of a leniency application. It
would however seem very unlikely that the other cartel members would be
willing to enter into such an agreement, as they would have to trust the ringleader
to act exclusively in the common interest, and would risk being the co-victims of
deviations by one of the other non-ringleaders. This would be different only if
the ringleader could coerce the other cartel members into the agreement. In all
jurisdictions, however, leniency programmes exclude coercers from immunity.140
137
The cartel participants could then also try to make use of the possibilities offered for partial
immunity to later leniency applicants that provide evidence of a longer duration of the infringement:
see above, paras 358 and 372 . A first cartel participant would apply for immunity, providing evidence
of the last third of the cartel period, while saying that the cartel may have started earlier, but that it
cannot retrieve evidence of that. A second cartel member would then come forward with evidence of
the other two thirds of the cartel period.
138
See also Aubert, Rey and Kovacic, above n 93, at 1250.
139
One (at least partial) solution could be to exclude repeat offenders from leniency. In one EU
Member State, Greece, recidivists are excluded from immunity. However, excluding repeat offenders
from leniency would mean that leniency would no longer work at all if the same group of companies
that was found to have formed a cartel originally subsequently enters a new cartel. If all these
companies cannot apply for leniency because they are repeat offenders, this second cartel will thus be
more stable than the first. Paradoxically, excluding recidivists from leniency may thus encourage
recidivism.
140
See above, paras 348, 349, 352, 365 and 379. In some jurisdictions, including the US and
Germany, leniency programmes go further in that they also exclude non-coercing sole ringleaders.
Such an exclusion generally makes sense in that it deters potential ringleaders from volunteering to
take on this role, in the same way as higher penalties for ringleaders: see above, para 406; K Mehta,
‘Comments on Switgard Feuerstein’s “Collusion in Industrial Economics—A Survey”’ (2005) 5
Journal of Industry, Competition and Trade 217 at 220; and Katyal, above n 61, at 1343 and 1365. On
the other hand, allowing (non-coercing) ringleaders to apply for immunity may have the effect of
making the other cartel members more cautious in following the ringleader’s initiative; see also
Aubert, Rey and Kovacic, above n 93, at 1250. Finally, the exclusion of the ringleader as a potential
immunity applicant reduces the number of such potential applicants, and thus reduces the potential
of a race to be the first to cooperate: see above, paras 401 and 408. The first version of the European
Commission’s Leniency Notice in 1996 excluded all ringleaders (even if there were several of them):
5.2.3.4 Negative Moral Effects
426. As already mentioned above,141 corporate managers are not necessarily just
maximisers of profits for themselves and their principals. They may feel a moral
responsibility to live within the law whether or not they are likely to be caught,
and this normative commitment could trump their interest calculus. The public
punishment of those who violate the antitrust prohibitions has thus not only a
deterrent effect, in that it helps create a credible threat of punishment for those
who would be willing to commit violations on the basis of a profit calculation,
but also has moral effects, in that it sends a message to the spontaneously
law-abiding, reinforcing their moral commitment to the rules.
427. For people to have such moral commitment to the law, however, it is
important that the law and its enforcement are perceived to be fair.142 Leniency,
in particular in its strongest form of immunity, may raise two (closely related)
concerns in this respect.143
428. First, there may be concerns about the retributive injustice of an antitrust
offender escaping punishment. The main response to this concern is to design
and apply leniency programmes in such a way as to ensure that no more leniency
is granted than is strictly necessary to obtain the positive enforcement effects,144
and to stress the condition for any beneficiary of leniency, and in particular of
immunity, to provide genuine and full cooperation to the enforcement authori-
ties. This appears to be reflected, for instance, in the case law of the Court of
Justice, which has emphasised that, for companies to benefit from leniency under
the European Commission’s Leniency Notice, their conduct must reveal ‘a
genuine spirit of cooperation’.145 Requiring restitution to injured parties as a
condition for leniency may also help to assuage justice concerns.146
429. A second (closely related) concern focuses on the unequal treatment
between the beneficiary of immunity or leniency and the other cartel partici-
pants, who receive full punishment for the same antitrust violation.147 Again the
main response is to ensure that leniency is granted only to the extent that the
see above n 50 and Judgment of the CFI of 15 Mar 2006 in Case T–15/02 BASF [2006] ECR II–497,
paras 316, 321 and 535–536. This was however changed in 2002, and the current Leniency Notice only
excludes coercers: see above, text to nn 29–30.
141
At para 389.
142
See Tyler, above n 80, as well as the literature listed above in n 78.
143
See also OECD, above n 74, at 26.
144
As argued above, in para 415, limiting leniency to what is strictly necessary to obtain its
enforcement benefits is also indicated to counter the penalty-reducing effect of leniency.
145
Judgments of the ECJ of 28 June 2005 in Joined Cases C–189/02 P etc Dansk Rorindustri and
Others v Commission (Pre-insulated Pipes) [2005] ECR I–5425, paras 388–403, and of 29 June 2006 in
Case C–301/04 P Commission v SGL Carbon [2006] ECR I–5977, para 68. See also, in the UK, the
Judgment of the Competition Appeals Tribunal of 19 May 2005 in Cases 1019/1/1/03 etc. Umbro and
Others v Office of Fair Trading [2005] CAT 22, paras 210–234.
146
See above, paras 413 and 414. One might also consider excluding recidivists from immunity, in
particular if the recidivism followed an earlier grant of immunity; see however above n 139.
147
See also Voillemot, above n 127, at 99–100.
company or individual has genuinely and effectively cooperated with the compe-
tition authority, thus objectively distinguishing its situation from the other cartel
participants that have not done so, or not to the same extent, or at the same early
point in time. It is equally important to ensure that leniency policies are applied
in a transparent and consistent way, thus providing an equal chance for all cartel
participants to benefit from them.148 In this respect, the Court of First Instance
has stressed that, under the general principle of equal treatment, the European
Commission must ensure that it does not distort the conditions of competition
between undertakings in a (potential) race to be the first to cooperate, by
contacting one of them, or giving them unequal access when they contact the
European Commission.149
430. Both concerns, as to the guilty remaining unpunished and as to
co-conspirators being treated unequally, appear to be particularly acute in the case
of immunity being granted to the ringleader of a cartel, in particular if that
ringleader has coerced other cartel members into committing the infringement. As
already discussed above, in all jurisdictions coercers are excluded from immunity,
and in several jurisdictions non-coercing sole ringleaders are also included.150
5.2.4 Measuring the Effects of Leniency
431. The most obvious measurement of the effects of leniency is the number of
leniency applications received by the competition authorities, and the amounts of
the penalties imposed with the help of these leniency applications. Indeed, both
the US Department of Justice and the European Commission have been regularly
reporting the success of their leniency policies, in that they receive several
applications for leniency per month, and that these applications have helped
them in imposing high amounts of penalties.151
148
On could try to argue that a company that is willing to cooperate but detains less evidence that
it can offer in exchange for leniency is disadvantaged as compared to another willing applicant with
more evidence at its hands. This argument should however be rejected (and has indeed been rejected
by the CFI in its Judgment of 18 July 2005 in Case T–241/01 SAS v Commission [2005] ECR II–2917,
paras 211–221), because its distinguishing according to the usefulness of the cooperation appears
objectively justified. Moreover, as pointed out above in para 402, granting leniency in proportion to
the evidence provided creates a desirable incentive for all cartel participants to create and keep
evidence.
149
Judgments of the CFI of 13 Dec 2001 in Joined Cases T–45/98 etc Krupp Thyssen and Others v
Commission [2001] ECR II–3757, paras 237–248, and of 15 Mar 2006 in Case T–15/02 BASF v
Commission [2006] ECR II–497, para 504. The European Commission would thus not be allowed to
adopt the US Department of Justice’s practice of ‘affirmative amnesty’, under which the latter
approaches a company—which may at the time not even know that it or its competitors are under
investigation—to cooperate and seek leniency: see ICN, above n 74, at 5, and Hammond, above n 20,
at 11.
150
See above, para 425 and n 140. See also above nn 139 and 146 as to the case of recidivists.
151
See for instance Hammond, above n 62, at 2–4; O Guersent, ‘The Fight against Secret
Horizontal Agreement in the EC Competition Policy’ in BE Hawk, 2003 Fordham Corporate Law
432. On its own, the number of leniency applications is not a good indicator
of the success of a leniency policy, as it may reflect excessive generosity and thus a
weakening rather than a strengthening of deterrence.152 The imposition of more
and higher penalties, without an increase in the competition authority’s
resources, is however clearly a sign of success. Indeed, more and higher penalties
means more deterrence and hence fewer antitrust violations.
433. The only doubt one could have about the overall positive effect of
leniency policies would be that this positive effect would be outweighed by
perverse effects of leniency, either in that cartel participants find ways to exploit
leniency programmes so as to facilitate the creation or maintenance of cartels, or
in that leniency programmes have negative moral effects leading to less respect
for the antitrust prohibitions.153 However, as explained above,154 leniency pro-
grammes in all jurisdictions appear to be designed in ways that appear to make it
very unlikely that they could be exploited in a way that facilitated the creation
and maintenance of cartels, certainly in those jurisdictions that impose not only
fines on companies but also imprisonment on individuals,155 and good design
and application of leniency programmes should also be able to limit negative
moral effects.
434. Of course, it would be nice if we were able to measure the overall effect of
leniency by directly observing its impact on the cartel population. But, given that
cartels hide themselves, we have no way directly to observe the population of
cartels. We can observe the population of discovered cartels, but this is a very
unreliable indicator of the cartel population, as an increase or decrease in the
number of detected cartels can be due as much to improved or worsened
detection as to an increase or decrease in the cartel population. In a recent paper,
Joseph Harrington has shown that the change in the duration of discovered
cartels can be informative about the change in the rate of cartels.156 In order to
use this proxy to measure the effect of a leniency policy, one would however also
have to be able to separate the effect of the leniency policy from the effect of
other changes in antitrust enforcement during the same period, such as changes
in the competition authority’s resources or in the applicable penalties.157
Institute International Antitrust Law & Policy (Huntington, NY, Juris Publishing, 2004) 43 at 45–9; and
European Commission, Report on Competition Policy 2005, SEC(2006)761, para 175.
152
Wils, above n 64, at 139.
153
154
Ibid.
155
The risk for jurisdictions without such penalties for individuals is that, in situations where the
same companies participate in a number of cartels in different markets, or repeatedly form cartels
over time, cartel participants may take turns to apply for leniency, every time one of the cartels is
(about to be) detected by the competition authority: see above, para 424.
156
JE Harrington, ‘Modelling the Birth and Death of Cartels with an Application to Evaluating
Antitrust Policy’ (June 2006), available at www.econ.jhu.edu/People/Harrington/CartelBirthDeath-
6.06.pdf.
157
See above, para 417, as to the phenomenon of penalties being raised in parallel with the
introduction of leniency programmes.
5.3 O BSTACLES TO THE INT RODUCTION OF LENIENCY POLICIES
435. As described above,158 many jurisdictions have introduced leniency policies

in the last decade, following the initial example of the US Department of Justice.
In most jurisdictions, however, the introduction of leniency programmes, in
particular programmes offering immunity from penalties, has not been uncon-
tested, and their use has not yet spread everywhere. In Germany, for instance, a
leniency programme for companies has existed since 2000, but there is no
leniency programme for individuals who are punishable with imprisonment for
the specific criminal offence of bid-rigging.
436. Whereas initially doubts about their effectiveness may have slowed down
the introduction and spread of leniency policies, such doubts appear no longer to
play a significant role, as leniency policies are widely credited with having
contributed to an important extent to the increase in the number and the size of
antitrust penalties obtained or imposed by the US Department of Justice and by
the European Commission.159 Objections of principle and institutional problems
may however still constitute obstacles to the introduction of leniency policies.
5.3.1 Objections of Principle
437. The two strongest objections of principle that could be raised against
leniency, in particular in its strongest form of immunity, namely the retributive
injustice of antitrust offenders escaping punishment and the unequal treatment
between the beneficiaries of leniency and the other offenders, have already been
discussed above.160 Provided that no more leniency is granted than is strictly
necessary to obtain the positive enforcement results, that beneficiaries of leniency
are required to provide genuine and full cooperation, that leniency policies are
applied in a transparent and consistent way, and that coercers are excluded from
immunity, the granting of leniency does not appear unfair.
438. Leniency has also raised objections akin to the use of informants or
spies.161 While one may readily applaud, in the light of historical experience with
totalitarian regimes, vigilance against the use by the state of neighbours, friends
and family to spy and inform on each other’s private livee, especially for political
offences, in the antitrust context this objection seems rather to reflect an ‘honour
among thieves’ culture.162 In the end, the issue appears to be whether or not, or to
what extent, one considers antitrust offences, in particular secret price-fixing
158
159
See above, paras 431–434and the references above in n 74.
160
161
In French, the emotionally charged term ‘délation’ is used in this context: see Voillemot, above
n 127, at 13 and 93–5.
162
Cf D Brault, ‘Jusqu’où augmenter la prime aux repentis?’ [2005] Revue Lamy de la Concurrence
113, at 115, referring to the ‘mutual confidence between contracting parties’ (‘la confiance mutuelle
entre co-contractants’).
OBSTACLES TO LENIENCY POLICIES 143
cartels, to be reprehensible. Strong antitrust enforcement requires a societal

consensus that antitrust offences are bad.163
439. Objections have also been raised to the compatibility of leniency with
respect for the rights of the defence, but none of these objections appear convinc-
ing. As the Court of Justice has held in relation to the European Commission’s
Leniency Notice: ‘admission of an alleged infringement is a matter entirely within
the will of the undertaking concerned. The latter is not in any way coerced to admit
the existence of the agreement. It must therefore be considered that the fact that the
Commission took account of the degree of cooperation with it shown by the
undertaking concerned, including admission of the infringement, for the purpose
of imposing a lower fine does not constitute any breach of its rights of defence’.164
5.3.2 Institutional Problems
440. In some jurisdictions the question has arisen whether the competition
authority has the power to introduce a leniency policy, or whether the legislator
has to introduce such a policy or has to create a new empowerment for the
competition authority to do so. In the case of the European Commission, which
itself adopted its Leniency Notice, the Court of Justice has confirmed that it had
the power to do so.165
441. In some jurisdictions there is a principle of mandatory prosecution,
under which the public prosecution office is obliged to take action in relation to
all criminal offences which may be prosecuted, provided there is sufficient
evidence.166 This could be considered to create a problem for the introduction of
leniency, at least in its stronger form of immunity, with regard to criminal
penalties. The problem does not appear to arise if the immunity granted is not
immunity from prosecution but only immunity from the imposition of penalties.
In any event, the principle of mandatory prosecution usually only applies ‘unless
otherwise provided by law’,167 which means that a leniency policy can be
introduced, provided that the introduction is done through a legislative act. It
may also be noted that the Council of Europe has since 1987 recommended to its
Member States that they introduce or extend the application of the principle of
163
See also below, paras 598–600.
164
Judgment of 14 July 2005 in Joined Cases C–65/02 P etc ThyssenKrupp Stainless and Others v
Commission [2005] ECR I–6773, paras 52–53; see also the Opinion of AG Léger of 28 Oct 2004 in the
same case, at paras 134–146; Wils, above n 64, at 137–8; Wils, above n 87, at 580; and Judgment of the
CFI of 15 Mar 2006 in Case T–15/02 BASF v Commission [2006] ECR II–497, para 58.
165
Judgment of 16 Nov 2000 in Case C–298/98 P Metsä-Serla (Finnboard) v Commission [2000]
ECR I–10171, paras 56 and 57; see also Wils, above n 64, at 135–6.
166
See Council of Europe, Proceedings of the 5th session of the Conference of Prosecutors
General of Europe, ‘Discretionary Powers of Public Prosecution: Opportunity or Legality Principle—
Advantages and Disadvantages’ (Celle, 23–25 May 2004), available at www.coe.int/t/e/legal_affairs/
legal_co-operation/conferences_and_high-level_meetings/european_public_prosecutors/
2004(Celle).asp#TopOfPage.
167
This is the case, for instance, in Germany: see s 152(1) of the Code of Criminal Procedure.
discretionary prosecution, wherever possible, or otherwise devise measures hav-

ing the same purpose.168 In any event, concerns to avoid arbitrariness and
unequal treatment, which appear to be underlying the idea of mandatory
prosecution, should not apply to leniency policies that are set out and applied in
a transparent and consistent way, and that provide no more leniency than is
strictly necessary to obtain the positive enforcement results.169
442. Finally, in some jurisdictions, the investigation of antitrust violations and
their prosecution are carried out by separate authorities. A problem may then
arise in that the prosecuting authority will resist giving up its prosecutorial
discretion and becoming bound by the investigating authority’s decisions on
leniency applications. This problem can be overcome by involving the prosecut-
ing authority in the introduction and application of the leniency programme,170
or by imposing it through a higher legislative act, making it binding on the
prosecuting authority.
5.4 SOME FUR THER ISSUES
5.4.1 Leniency and Criminal Penalties on Individuals
443. As set out above,171 whereas in the US the cartel prohibition is enforced not
only by fines on companies but also by imprisonment of individuals, the
European Commission and the competition authorities of most EU Member
States can currently impose fines only on undertakings. In the last decade, prison
sanctions for individuals have been introduced in Ireland and the United
Kingdom,172 and they may be introduced in the Netherlands in the near
future.173 Prison sanctions also exist in some other EU Member States, for
instance in Germany for bid-rigging. This trend towards criminalisation raises
the question how the addition of criminal penalties on individuals impacts upon
the effectiveness of leniency.
168
Recommendation R (87) 18 concerning the simplification of criminal justice, adopted by the
Committee of Ministers of the Council of Europe on 17 Sept 1987, again referred to in Recommen-
dation Rec (2000) 19 on the role of public prosecution in the criminal justice system, both available at
www.coe.int/t/e/legal_affairs/legal_co-operation/conferences_and_high-level_meetings/
european_public_prosecutors/00_0E_Reference%20Texts.asp#TopOfPage. As has also been pointed
out by the OECD, above n 74, at 26, specifically with regard to antitrust enforcement, few, if any,
enforcement agencies could function effectively without discretion. Some prioritising and balancing
of costs and benefits in the enforcement process is inevitable.
169
See also above, paras 426–430 and 437–439.
170
See ICN, above n 74, sect 2.6.7.
171
See above, paras 349, 352, 367 and 380; see also above n 32.
172
See below, paras 538–543.
173
The issue was debated in the Second Chamber of the Dutch Parliament on 15 June 2006, and
broad support for the introduction of prison sanctions for cartels emerged. See also below, paras
534–538.
SOME FURTHER ISSUES 145
444. If, as is the case in the US, Ireland and the UK, leniency (including
immunity) is also available for individuals, and the leniency policies for com-
panies and for individuals are well integrated, in that, when a company is the first
to apply for immunity, it can also obtain immunity for its cooperating staff, but
that immunity is normally no longer available for a company if an individual has
applied for it earlier (and is similarly normally unavailable for an individual if a
company has been first and the individual is not part of the company’s applica-
tion), leniency will be more effective, for three reasons.174
445. First, at least to the extent that the companies care for their staff, the
higher overall penalty discount available will make it more attractive for com-
panies to apply for leniency.
446. Secondly, additional races to be the first to cooperate are created between
the companies and their staff, as well as between individuals. As Scott Hammond
has explained with regard to US Department of Justice’s practice of combining
corporate and individual leniency programmes:
the individual amnesty program helps prevent companies from covering up their
misconduct. The real value and measure of the Individual Leniency Program is not in
the number of individual applications we receive, but in the number of corporate
applications it generates. It works because it acts as a watchdog to ensure that
companies report the conduct themselves. … So long as one of its employees has
individual exposure, the company remains at great risk. If the company self-reports the
conduct under the Corporate Leniency Policy, then the company and all of its
cooperating executives will avoid criminal prosecution. However, if the company delays
or decides not to report, then the company puts itself in a race for leniency with its own
employees. In this example, if the company does not report the conduct first, then the
executive may come forward on his own and report the conduct for his own protection,
thereby potentially leaving the company out in the cold. … If the secretary gets
nervous, say, after talking to a relative who convinces her that she has real criminal
exposure for her own conduct, she may decide to report the conduct.175
As Donald Baker has further explained, also describing US practice:
The same process works between individuals at different companies involved in the
suspected conspiracy. There certainly have been cases where an individual was indicted
for a one-on-one conversation with someone who worked for a competitor and who
then turned the defendant in. Of course, as we saw with the Lysine cartel, an individual
who turns herself in to the government may become a government informant and
continue to participate in conspiracy meetings, armed with recording devices in order
to entrap the other conspiracy members.176
447. Thirdly, the presence of criminal penalties for individuals strongly reduces
the risk of perverse effects of leniency policies.177 It would in particular appear to
174
See also Bloom, above n 124.
175
Hammond, above n 62, at 12–14.
176
Baker, above n 86, at 708.
177
eliminate the risk that, in situations where the same companies participate in a
number of cartels in different markets, or repeatedly form cartels over time,
cartel participants may take turns to apply for leniency every time one of the
cartels is (about to be) detected by the competition authority.178
448. To the contrary, if leniency (including immunity) is not available for
individuals, the addition of criminal penalties for individuals will decrease the
effectiveness of leniency. At least to the extent that they care about their staff,
companies will be less inclined to apply for leniency, as this will expose their staff
to punishment. Even if they do not care about their staff, companies may not be
able to make useful leniency applications, as their staff will not be willing to
cooperate for fear of incriminating themselves.
449. It thus appears essential, when adding criminal penalties for individuals
to corporate fines, to provide also for leniency (including immunity) for indi-
viduals.179 This conclusion is confirmed, on the one hand, by the experience of
the successful combined application by the US Department of Justice of its
Corporate Leniency Policy and its Leniency Policy for Individuals,180 and, on the
other hand, by the example of Germany, where criminal penalties for individuals
exist for bid-rigging, without a corresponding possibility of leniency, with the
result that the German competition authority, which has an otherwise function-
ing leniency programme for companies, has never received a leniency application
in a bid-rigging case.
450. The introduction of leniency, in particular immunity, for criminal pen-
alties might be considered more problematic than in the case of fines on
companies.181 However, as argued above,182 it would appear that all the concerns
relating to retributive justice, equal treatment and legality which could be raised
can be adequately addressed, provided that no more leniency is granted than is
strictly necessary to obtain the positive enforcement results, that beneficiaries of
leniency are required to provide genuine and full cooperation, that coercers are
178
179
See further below, paras 592–595.
180
As has been pointed out by Margaret Bloom, above n 124, at 9, when comparing the statistics
of the success of the leniency policies of the US Department of Justice and for instance the European
Commission, one should observe that roughly half of the immunity applications received by the
European Commission follow leniency applications to the US Department of Justice, and might thus
never have taken place if it were not for the strong attraction of the US Department of Justice’s
leniency policy, linked to the threat of imprisonment; see also A Stephan, ‘An Empirical Assessment of
the 1996 Leniency Notice’, CCP Working Paper 05–01 (Sept 2005), available at www.ccp.uea.ac.uk/
public_files/workingpapers/CCP05–
10.pdf#search=%22Empirical%20Assessment%20Leniency%20Andreas%20Stephan%22.
181
In Sweden, where currently only corporate fines exist, with a leniency programme, a Govern-
mental Committee of Inquiry in 2004 produced a proposal to introduce criminal penalties for
individuals, but without the possibility of leniency, referring to the general rule of mandatory
prosecution in the Swedish Code of Judicial Procedure: see Konkurrensbrott—En lagstiftningsmodell,
SOU 2004:131, available at www.regeringen.se/sb/d/264/a/36225. See above, para 441 for arguments
as to why the principle of mandatory prosecution should not constitute an insurmountable obstacle
to introducing leniency.
182
At paras 426–430 and 435–442.
excluded from immunity, that leniency policies are applied in a transparent and
consistent way, and, in those jurisdictions where this would be necessary, are laid
down in a legislative act.
5.4.2 Leniency and Follow-on Private Actions for Damages
451. The European Commission published in 2005 a Green Paper on damages

actions for breach of the EC antitrust rules, opening a wide debate in the EU
Member States on options to facilitate private damages actions, including
follow-on actions in cases where penalties have already been imposed by the
European Commission or the competition authority of an EU Member State.183
Such actions are currently less common in the EU Member States than in the US.
This debate raises the question how follow-on private actions for damages impact
upon the effectiveness of leniency.
452. To some extent the impact of follow-on private damages actions appears
similar to that of the addition of criminal penalties on individuals.184 If a
leniency applicant, and in particular an immunity applicant, could also be offered
leniency or immunity from damages, the effect of follow-on private damages
actions would be to strengthen the effectiveness of leniency, because potential
leniency applicants would be attracted by the larger benefit available to them. To
the contrary, if immunity applicants or other leniency applicants cannot be
offered corresponding immunity or leniency as regards damages, follow-on
private damages actions weaken the effectiveness of leniency, as the prospect of
having to pay damages, or at least of having to defend oneself in follow-on
private actions, adds a negative element to be weighed in the potential leniency
applicant’s decision whether or not to apply for leniency.185
183
The Green Paper COM(2005)672 and the Commission Staff Working Paper SEC(2005)1732
annexed to it, both of 19 Dec 2005, as well as many subsequent contributions to the debate, are
available at http://ec.europa.eu/comm/competition/antitrust/others/actions_for_damages/
index_en.html; see also D Waelbroeck and D Slater, ‘The Commission’s Green Paper on Private
Enforcement: “Americanisation” of EC Competition Law Enforcement?’, and J Lawrence, ‘Seeking the
Perfect Balance—Some Reflections on the Commission Green Paper: Damages Actions for Breach of
the EC Antitrust Rules (19 Dec 2005)’, papers presented at the 11th EUI Competition Law and Policy
Workshop (Florence, 2–3 June 2006), available at www.iue.it/RSCAS/Research/Competition/
Index.shtml, forthcoming in C-D Ehlermann and I Atanasiu (eds), European Competition Law Annual
2006: Enforcement of Prohibition of Cartels (Oxford, Hart Publishing, 2007), and Wils, above n 76.
184
185
One might argue that leniency applicants will still come forward for fear that another
co-conspirator would otherwise apply for leniency first. This argument is only partially convincing,
however, because the negative effect of the prospect of follow-on damages actions also applies to all
the other co-conspirators, and any potential leniency applicant will understand this. It should also be
noted that the negative effect exists even if the fact of cooperating with the authority does not place
the leniency applicant in a worse position in respect of follow-on damages actions than other cartel
members that do not cooperate. Competition authorities generally try to avoid putting the leniency
applicant in a worse position than the other cartel members, in particular by providing for an oral
procedure for leniency applications, so to avoid corporate statements being discoverable: see Bloom,
453. As already mentioned above, in the US, following a statutory modifica-

tion in 2004, companies that have been granted immunity by the US Department
of Justice under its leniency programme and that are found by the court in
follow-on actions for damages to have provided satisfactory cooperation with the
private claimants are only liable for single damages based on their own share of
the commerce affected by the violation, whereas the other cartel members remain
jointly and severally liable for treble damages based on the harm caused by the
entire conspiracy.186
454. In the EU, successful leniency applicants do not currently benefit from
any immunity or reduction of liability with regard to follow-on private damages
actions.187 Following the US example, the European Commission’s Green Paper
on damages actions for breach of the EC antitrust rules puts forward for
discussion the options of giving successful leniency applicants a rebate on
damages and of removing joint and several liability.188 As the Green Paper also
floats the idea of double damages for cartels,189 the proposed rebate could take
the form of single instead of double damages.190
5.4.3 Leniency in Multiple Jurisdictions
455. International cartels may be subject to the imposition of penalties in

multiple jurisdictions.191 This raises the question how the fact that penalties may
also be imposed in other jurisdictions impacts upon the effectiveness of leniency.
above n 124, at 17–21; Reynolds and Anderson, above n 128, at 82–4; and K Nordlander, ‘Discovering
Discovery—US Discovery of EC Leniency Statements’ [2004] European Competition Law Review 646.
186
187
188
Commission Staff Working Paper, above n 183, paras 235 and 236 (Options 29 and 30).
189
Ibid, para 150 (Option 16).
190
Ibid, para 235. The initial reactions to the Green Paper, however, do not appear to indicate
much support for the idea of double damages: see the contributions referred to above in n 183, and U
Böge and K Ost, ‘Up and Running, or Is It? Private Enforcement—the Situation in Germany and
Policy Perspectives’ [2006] European Competition Law Review 197 at 201–2, referring to constitutional
concerns in Germany. On punitive damages see also the Judgment of the ECJ of 13 July 2006 in Joined
Cases C–295/04 etc Manfredi and Others [2006] ECR I–6619, paras 92–96, and the Judgment of the
CFI of 27 Sept 2006 in Case T–59/02 Archer Daniels Midland v Commission, not yet reported, paras
349–355.
191
The general rule in international law is that each jurisdiction can impose the penalties
provided for in its laws for the violation of its antitrust laws on its territory, without any obligation to
take into account the penalties imposed (or not imposed) in the other jurisdictions: see Judgment of
the ECJ of 29 June 2006 in Case C–308/04 P SGL Carbon v Commission [2006] ECR I–5977, paras
29–39, as well as the Opinion of 19 Jan 2006 of AG Geelhoed in the same case. As the ECJ pointed out
in para 30 of the same judgment, the legal situation is ‘completely different’ inside the EU. On the
relationship between the European Commission and the competition authorities of the EU Member
States see Wils, ‘The Principle of Ne Bis in Idem in EC Antitrust Enforcement: A Legal and Economic
Analysis’ (2003) 26 World Competition 131, and Wils, above n 66, chs 1, 2 and 3. See also above, para
387 and n 75 as to the ECN Model Leniency Programme.
456. This impact appears similar to that of follow-on private damages actions,
and thus to some extent also to that of the addition of criminal penalties on
individuals.192 If the leniency applicant, and in particular the immunity appli-
cant, can also obtain leniency or immunity in the other jurisdictions, the
effectiveness of leniency will be increased, because potential leniency applicants
will be attracted by the larger benefit available to them. To the contrary, if
immunity applicants or other leniency applicants cannot obtain corresponding
immunity or leniency in the other jurisdictions, the effectiveness of leniency will
be weakened, as the prospect of penalties in the other jurisdictions adds a
negative element to be weighed in the potential leniency applicant’s decision
whether or not to apply for leniency.
457. This impact helps to explain the efforts of those antitrust enforcement
authorities that were the first to introduce leniency policies, in particular the US
Department of Justice, in convincing the authorities in other jurisdictions to
adopt similar policies.193
5.4.4 ‘Amnesty Plus’
458. As already mentioned above, the US Department of Justice and the UK

Office of Fair Trading have a policy, called ‘Amnesty Plus’, under which a
cooperating company that does not qualify for immunity as to a first cartel being
investigated but that uses the occasion of that first investigation to report a
second, distinct cartel will receive, in addition to the immunity it can obtain for
the second cartel, a further reduction of the fine for the first cartel.194
459. There would indeed appear to be an increased probability that com-
panies that participate in one cartel also participate in others. First, these
companies, or the responsible individuals within them, must have overcome the
normative or moral barrier which would prevent the law-abiding from engaging
in illegal activity.195 Secondly, by participating in one cartel, they will have gained
valuable experience, making it more likely that their participation in other cartels
will be successful.196 Thirdly, to the extent that ability to pay plays a role in
determining the amount of penalties for cartels, companies that participate in
several cartels may be punished proportionally less.197
192
See above, paras 451–454 and 443–450.
193
See above, paras 381–387 and n 75 concerning the ECN Model Leniency Programme.
194
See above, paras 361 and 379; see also ICN, above n 74, at 4, and D McAlwee, ‘Should the
European Commission Adopt “Amnesty Plus” in its Fight Against Hard-core Cartels?’ [2004]
European Competition Law Review 558.
195
On the importance of normative commitment to the law as a barrier to antitrust violations see
above, paras 389 and 426–430.
196
On the complexity of successful cartels and the importance of learning see above, paras 386,
404–408 and 422–425.
197
See for instance the case of SGL Carbon, which received from the European Commission a
33% discount from its fine for participation in the Electrical and mechanical carbon and graphite
460. It would also seem that, once the participation of a company in one
cartel has been detected by an antitrust enforcement authority, there is an
increased probability that the antitrust enforcement authority will also detect the
participation in the other cartels. Indeed, a rational competition authority will be
aware of the increased probability that companies that participate in one cartel
also participate in other cartels, and will thus spontaneously focus its investiga-
tions in that direction. Moreover, investigations into one cartel may incidentally
produce evidence of other cartels, for instance when the authority’s investigators
conducting a ‘dawn raid’ at the premises of a company to collect evidence of one
suspected cartel come across evidence of another cartel.198
461. Given this increased probability of detection of the second cartel, once
the participation of a company in a first cartel has been detected, the justification
for an ‘Amnesty Plus’ policy does not appear obvious. ‘Amnesty Plus’ comes down
to granting more than 100 per cent leniency for the second cartel. As further
discussed below, one could argue whether it is a good idea in general to grant
more than 100 per cent leniency, ie not only to waive the entire penalty but also
to give a positive financial reward to a cooperating company.199 But whatever one
may think about this idea in general, it is difficult to see the justification for
applying it only in the specific case of a company the participation of which in a
first cartel has already been detected. Given that in that specific situation, the
probability of detection is higher than normal, there should be less need to go
beyond 100 per cent leniency and grant a positive financial reward.
462. It has been pointed out that over half of the investigations of suspected
international cartels by the US Department of Justice were initiated as a result of
leads generated during the investigation of a separate market.200 It does not
follow however that this would not also have happened without ‘Amnesty Plus’, or
that similar phenomena could not be observed at competition authorities that do
not have an ‘Amnesty Plus’ policy.201
products cartel (Dec of 3 Dec 2003 in Case COMP/38.359, available at http://ec.europa.eu/comm/

competition/antitrust/cases/decisions/38359/en.pdf, paras 358 to 360) because of its bad financial
situation and the fact that it had already been fined for its participation in two other contemporane-
ous cartels; see further A Stephan, ‘The Bankruptcy Wildcard in Cartel Cases’, CCP Working Paper
06–5 (Mar 2006), available at www.ccp.uea.ac.uk/public_files/workingpapers/CCP06–5.pdf; C Hard-
ing, ‘Effectiveness of Enforcement and Legal Protection’, paper presented at the 11th EUI Competition
Law and Policy Workshop (Florence, 2–3 June 2006), available at www.iue.it/RSCAS/Research/
Competition/Index.shtml, forthcoming in C-D Ehlermann and I Atanasiu (eds), European Competi-
tion Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart Publishing, 2007); and
above, paras 206 and 341–343.
198
See the Judgment of the ECJ of 17 Oct 1989 in Case 85/87 Dow Benelux v Commission [1989]
ECR 3137, paras 17–19, holding that in a situation where the European Commission’s inspectors
discover evidence of a violation outside the subject-matter indicated in the decision ordering the
inspection, the European Commission can initiate a new investigation and use its investigative powers
to obtain that evidence in the framework of this new investigation.
199
See below, paras 467 and 468.
200
Hammond, above n 62, at 14.
201
On the European Commission see Opening exposé by Commissioner N Kroes at the 11th EUI
Competition Law and Policy Workshop (Florence, 2–3 June 2006), available at www.iue.it/RSCAS/
463. The obvious disadvantage of ‘Amnesty Plus’ is that it involves an addi-

tional lowering of the penalty.202 It could arguably make it more attractive for a
company already participating in one cartel also to join other cartels.203
5.4.5 Positive Financial Rewards or Bounties
464. Leniency as discussed above in this chapter consists in the reduction of

penalties or, in its strongest form, the grant of immunity from penalties for
antitrust violations in exchange for cooperation with the antitrust enforcement
authorities.204 William Kovacic and others have proposed the grant of positive
financial rewards or bounties to incentivise cooperation in the detection and
prosecution of cartels.205
465. In April 2005 the Korean Fair Trade Commission introduced an Inform-
ant Reward System to facilitate the detection of secret violations of competition
law.206 In June 2005 a first reward of 66.87 million won (about US$63,700) was
paid to an anonymous informant who had provided decisive evidence in a
welding rod cartel case, including the names of executives of the six cartel
member companies, meeting places and details of agreement.207
466. It may be useful to distinguish between three different types of positive
financial rewards or bounties, depending on whether the beneficiary is (1) a
company that participated in the cartel, (2) an individual that could otherwise be
penalised for its role in the cartel, or (3) another informant.
5.4.5.1 Rewards to Companies that Participated in Cartels
467. If the beneficiary of the positive financial reward is a company that

participated in the cartel, giving a positive financial reward comes down to
granting more than 100 per cent corporate leniency; instead of merely reducing
Research/Competition/Index.shtml, forthcoming in C-D Ehlermann and I Atanasiu (eds), European

Competition Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart Publishing, 2007)
at 7, pointing out that the Commission’s ex officio investigations almost always trigger immunity and
leniency applications in related geographical or product markets.
202
203
See Rosenstok, above n 106, at 64.
204
205
See WE Kovacic, ‘Private Monotoring and Antitrust Enforcement: Paying Informants to Reveal
Cartels’ (2001) 56 George Washington Law Review 766; Aubert, Rey and Kovacic, above n 93; Spagnolo,
above n 97; A Riley, ‘Beyond Leniency: Enhancing Enforcement in EC Antitrust Law’ (2005) 28 World
Competition 377; and WE Kovacic, ‘Bounties as Inducements to Identify Cartels’, paper presented at
the 11th EUI Competition Law and Policy Workshop (Florence, 2–3 June 2006), available at
www.iue.it/RSCAS/Research/Competition/Index.shtml, forthcoming in C-D Ehlermann and I Atana-
siu (eds), European Competition Law Annual 2006: Enforcement of Prohibition of Cartels (Oxford, Hart
Publishing, 2007).
206
See www.ftc.go.kr/data/hwp/rewardsystem.doc.
207
See www.ftc.go.kr/data/hwp/informant_reward.doc.
the penalty which would otherwise have been imposed on the company, or
waiving it entirely, the cooperating company is given a positive financial reward.
Such a system simply continues the logic of corporate leniency to a further point.
Compared with the mere granting of immunity, or only a reduction of the
penalty, such a system is likely to create stronger incentives to cooperate, as the
benefit offered is bigger.208 It however also increases the risks of negative effects,
in that it further lowers the penalty level,209 is likely to raise additional concerns
in terms of retributive justice and equal treatment and may thus have negative
moral effects,210 and increases the risk of cartel members finding ways to exploit
the leniency system in ways that facilitate the creation and maintenance of
cartels.211
468. As mentioned above,212 in situations where the same companies partici-
pate in a number of cartels in different markets, or repeatedly form cartels over
time, one could imagine a system in which cartel participants take turns to apply
for leniency, every time one of the cartels is (about to be) detected by the
antitrust enforcement authorities. In jurisdictions where cartels are not only
punished by fines on companies, but also by imprisonment of individuals, such a
system is most unlikely to be attractive, but in jurisdictions without such
individual penalties it might work.213 Positive financial rewards obviously
increase this risk. It would therefore not appear advisable for jurisdictions that do
not have criminal penalties for individuals to grant positive financial rewards for
companies that participated in cartels.214
5.4.5.2 Rewards to Individual Cartel Offenders
469. If the beneficiary of the positive financial reward is an individual cartel

offender, who could otherwise have been penalised for his role in the cartel
(assuming thus a jurisdiction in which not just companies but also individuals
are prosecuted for cartel offences), giving a positive financial reward comes down
to granting more than 100 per cent individual leniency; instead of merely
reducing the penalty which would otherwise have been imposed on the indi-
vidual, or waiving it entirely, the cooperating individual is given a positive
financial reward. Compared with the mere granting of immunity, or only a
reduction of the penalty, such a system is again likely to create stronger incentives
208
209
210
211
See above, para 422–425.
212
At para 424.
213
Ibid; Aubert, Rey and Kovacic, above n 93, at 1250.
214
Cf P Buccirossi and G Spagnolo, ‘Optimal Fines in the Era of Whistleblowers’, CEPR
Discussion Paper No 5465 (Jan 2006), available at www.cepr.org/pubs/dps/DP5465.asp, who precisely
propose such rewards as, according to those authors, a superior alternative to the introduction of
criminal penalties on individuals.
to cooperate, as the benefit offered is bigger.215 It also increases the risk of

negative effects, but less so than in the case of rewards to companies. Indeed,
while it has an additional effect of lowering the penalty level216 and is likely to
raise additional concerns in terms of retributive justice and equal treatment and
may thus have negative moral effects,217 there would not seem to be an increased
risk of cartel members finding ways to exploit the leniency system in ways that
facilitate the creation and maintenance of cartels.218
5.2.4.3 Rewards to Non-offender Informants
470. If the beneficiary of the financial reward is an informant who has not
herself committed a punishable offence, the system is unrelated to leniency. It
rather builds on one of the methods other than leniency which antitrust
enforcement authorities traditionally use to collect intelligence and evidence of
antitrust violations, namely to obtain information from volunteering third
parties.219 The reward adds an incentive to the other incentives, such as revenge,
or a sense of civic duty, which informants may have to volunteer information to
the antitrust enforcement authority. I understand this to be the system which was
introduced in Korea in 2005.220
215
216
See above,para 415.
217
See above, aras 426–430.
218
See above, paras 422–425 and 468.
219
220
See above, para 465; see further Kovacic, above n 205, as to various possible objections to such
a system and answers to these objections.
6
Is Criminalisation the Answer?
471. This chapter addresses five questions concerning criminalisation of EU
antitrust enforcement: First, what do we mean by ‘criminalisation’, or ‘criminal’
enforcement (as opposed to public enforcement of a ‘civil’ or ‘administrative’
nature)? Secondly, is there a tendency in the EU Member States to criminalise
antitrust enforcement (in comparison with US antitrust enforcement and with
antitrust enforcement at the level of the EU institutions)? Thirdly, is criminal
antitrust enforcement, more specifically imprisonment, desirable (in general,
irrespective of whether it takes place at the level of the Member States or of the
EU institutions, or whether it is harmonised at EU level)? Fourthly, is it
problematic that antitrust enforcement is criminalised at the level of individual
EU Member States without parallel criminalisation at the level of the EU
institutions or without EU harmonisation? Fifthly, would it be legally possible to
criminalise antitrust enforcement at the level of the EU institutions, or to have
EU harmonisation of criminal antitrust enforcement in the Member States?
6.1 WHAT DO WE MEAN BY ‘CR IMINALISATION’?
6.1.1 No Definition of ‘Criminal’ in EC or EU Law
472. Article 23(5) of Regulation 1/2003,1 the main regulation governing the
enforcement of Articles 81 and 82 EC, provides that decisions by which the
European Commission imposes fines on undertakings for violations of Articles
81 and 82 EC, or for obstruction of investigations into possible violations, ‘shall
not be of a criminal law nature’.2 The last sentence of recital 8 of the same
Regulation says that ‘this Regulation does not apply to national laws which
1
Council Reg 1/2003 of 16 Dec 2002 on the implementation of the rules of competition laid
OJ L269/1.
2
On the interpretation of this provision see below, paras 506 and 507, and further WPJ Wils,
Principles of European Antitrust Enforcement (Oxford, Hart Publishing, 2005), sect 1.1.4.3.3 (n 88).
156 IS CRIMINALISATION THE ANSWER?
impose criminal sanctions on natural persons except to the extent that such
sanctions are the means whereby competition rules applicable to undertakings
are enforced’.3
473. The EU Treaty has a Title VI concerning ‘Provisions on police and
judicial cooperation in criminal matters’. Article 30 (a) EU provides that common
action in the field of police cooperation shall include operational cooperation in
relation to the prevention, detection and investigation of ‘criminal offences’.
Article 31 EU provides for common action on judicial cooperation ‘in criminal
matters’.
474. EC or EU law does not, however, contain any definition of what is
criminal.
6.1.2 Distinguishing Characteristics of Criminal Law
475. Looking across legal systems (of the EU Member States, and of third
countries, in particular English-speaking third countries), criminal law enforce-
ment, as opposed to public law enforcement of a civil or administrative nature,
appears to have the following six distinguishing characteristics:4
6.1.2.1 Criminal Penalties
476. First, it appears that criminal law has a monopoly on the use of imprison-
ment. Whereas fines can be either criminal or civil or administrative, imprison-
ment appears to be essentially a criminal sanction. The possibility of a prison
sanction does not seem to be a necessary condition for a prohibited act or an
enforcement procedure to be criminal, but it is certainly a sufficient condition.
477. Compared with civil or administrative enforcement, criminal enforce-
ment tends to make relatively more use of penalties imposed on individuals
(natural persons), and relatively less of penalties imposed on corporate entities or
legal persons. Individual penalties are however neither a necessary nor a suffi-
cient condition for enforcement to be criminal. In Denmark, for instance,
violations of Articles 81 and 82 EC and of the parallel prohibitions under
national competition law are punished with fines, which are of a criminal law
nature but which are in practice only imposed on undertakings, not on natural
persons. In Germany, on the other hand, leaving aside the specific criminal law
provisions on bid-rigging, violations of Articles 81 and 82 EC and of the parallel
3
On the meaning of this sentence see below, para 532; see also below, para 619, on the use of the
word ‘criminal’ in the text of the EC Treaty.
4
See WPJ Wils, The Optimal Enforcement of EC Antitrust Law (The Hague, Kluwer Law
International, 2002), sect 8.7.2.1.
WHAT DO WE MEAN BY ‘CRIMINALISATION’? 157
national prohibitions are punished through fines, which are imposed on natural
persons and only derivatively also on undertakings, but which are of an admin-
istrative law nature.5
6.1.2.2 Criminal Intent
478. Secondly, the commission of a criminal offence usually requires that the
prohibited act be committed with a guilty state of mind (criminal intent), not by
mere negligence.
6.1.2.3 Moral Condemnation
479. Thirdly, the imposition of criminal sanctions carries, and is designed to

carry, a stigma effect. Criminal enforcement has a stronger message-sending role
or expressive function than civil or administrative enforcement.6
6.1.2.4 Less Strict Relationship between Penalty and Harm
480. Fourthly, in criminal law there appears to be a less strict relationship, if any,
between the size of the penalty and the size of the harm caused than in the setting
of civil sanctions. This appears to reflect the idea that criminal law does not seek
to price certain behaviour (by making the actor bear the external costs of his
behaviour) but rather to prohibit it (unconditionally, ie irrespective of the actual
size of the external costs).7
5
For an overview of antitrust enforcement in the 25 Member States see D Cahill and J Cooke
(eds), The Modernisation of EU Competition Law Enforcement—FIDE 2004 National Reports (Cam-
bridge, Cambridge University Press, 2004), and M Holmes and L Davey (eds), A Practical Guide to
National Competition Rules Across Europe (The Hague, Kluwer Law International, 2004).
6
Issues 17, and ‘General prevention revisited: research and policy implications’ (1975) 66 Journal of
Criminal Law & Criminology 338 at 341–3, as well as KG Dau-Schmidt, ‘An Economic Analysis of the
Criminal Law as a Preference-Shaping Policy’ [1990] Duke law Journal 1, CR Sunstein, ‘On the
Expressive Function of the Law’ (1996) 144 University of Pennsylvania Law Review 2021, DM Kahan,
‘Social Influence, Social Meaning, and Deterrence’ (1997) 83 Virginia Law Review 349, NK Katyal,
‘Deterrence’s Difficulty’ (1997) 95 Michigan Law Review 2385, GE Lynch, ‘The Role of Criminal Law
in Policing Corporate Misconduct’ (1997) 60 Law and Contemporary Problems 23, DM Kahan, ‘Social
Meaning and the Economic Analysis of Crime’ (1998) 27 Journal of Legal Studies 609, and KG
Economics and the Law (Basingstoke, Macmillan, 1998) 84.
7
See R Cooter, ‘Prices and Sanctions’ (1984) 84 Columbia Law Review 1523 and JC Coffee,
‘Paradigms Lost: the Blurring of the Criminal and Civil Law Models–and What Can be Done about It’
(1992) 101 The Yale Law Journal 1875.
6.1.2.5 Criminal Powers of Investigation
481. Fifthly, under criminal law, enforcement authorities tend to have stronger
investigative powers.8
6.1.2.6 Criminal Rights of Defence
482. Finally, criminal procedures tend to have stronger procedural protections,

designed to avoid false convictions. In particular, in criminal enforcement
systems the adjudicative or decision-making function is always separated from
the investigative and prosecutorial function or functions.9 Whereas it appears to
be a necessary condition of criminal enforcement that the body which imposes
the criminal penalties is separate and independent from the investigator and
prosecutor, it is however not a sufficient condition. In Finland, for instance,
violations of Articles 81 and 82 EC and of the parallel national prohibitions are
punished with fines on undertakings, which are of an administrative law nature,
but which are imposed by the Market Court, a specialised court which is entirely
separate and independent from the Competition Authority, which acts as inves-
tigator and prosecutor. In criminal enforcement systems, the standard of proof
for establishing violations also tends to be higher. In the English-speaking
countries, criminal convictions normally require proof ‘beyond reasonable
doubt’, whereas civil or administrative findings can normally be made ‘by a
preponderance of the evidence’ or a ‘balance of probabilities’.
6.1.2.7 Links between the Six Characteristics
483. These six characteristics listed above are clearly interrelated. The use of
imprisonment and the stigma effect create the stronger need for procedural
protections to avoid false convictions. These stronger procedural protections
create in their turn the need for stronger investigative powers so as to avoid
enforcement becoming ineffective. On the other hand, the message-sending role
of criminal enforcement explains the requirement of a guilty state of mind (as
this shows the need for wrong valuations to be repudiated) as well as the absence
of a strict relationship between punishment and actual harm, reflecting the desire
to deter the criminal behaviour unconditionally rather than merely to price it.
8
For an overview of the law and economics of the collection of intelligence and evidence by the
competition authorities from antitrust offenders see chr 5 of Wils, above n 2; see also Ch 1 above.
9
See Wils, above n 4, sect 8.7.4.5; and ch 6 of Wils, above n 2.
WHAT DO WE MEAN BY ‘CRIMINALISATION’? 159
6.1.3 The Wider Notion of ‘Criminal’ in the European Convention on Human

Rights
484. The notion of ‘criminal’ enforcement as described above and as used

throughout most of this chapter should be distinguished from the wider notion
of ‘criminal’ as conceived under Article 6 of the European Convention on Human
Rights (ECHR).10
485. The European Court of Human Rights has, within the framework of
Article 6 ECHR, developed an autonomous notion of ‘criminal’, which covers
proceedings of an administrative law nature which fulfill the following
conditions:11—the offences are defined by a general rule, applicable to all citizens
and not only to some of them;—the rule is linked to penalties in the event of
non-compliance;—the sanctions are intended not as pecuniary compensation for
damage but essentially as a punishment to deter reoffending; and—the sanctions
are severe.12
486. The European Court of Human Rights has never had the occasion to rule
on an antitrust enforcement procedure at the level of the EU institutions or in
one of the EU Member States.13 In 1989, the European Commission of Human
Rights (a body which has since been abolished) decided a case involving
competition proceedings before the French competition authority.14 It decided
that the sanctions imposed by that authority were of a criminal nature within the
meaning of Article 6 ECHR. The case was not brought before the European
Court of Human Rights.
487. In 2004, the European Court of Human Rights decided a case concerning
proceedings under the Russian Law on Competition and the Restriction of
Monopolies in the Commodity Markets, which had led to the confiscation of
profits obtained by a number of companies in violation of that law. The
10
See K Dekeyser and C Gauer, ‘The New Enforcement System for Articles 81 and 82 EC and the
Rights of Defence’, paper presented at the 31st Annual Fordham Corporate Law Institute Conference
on International Antitrust Law & Policy (New York, 8 Oct 2004), forthcoming in BE Hawk (ed), 2004
Annual Proceedings of the Fordham Corporate Law Institute—International Antitrust Law and Policy
(Huntington, NY, Juris Publishing. 2005) 537 at 540–2; Wils, above n 4, sect 8.7.2.3; and Wils, above n
2, sects 1.1.4.3.3 (n 88) and 6.1.4.1.
11
Restated in the words of Dekeyser and Gauer, above n 10, at 541.
12
Engel and Others v the Netherlands, Series A No 32, Judgment of 8 June 1976, para 82; Özturk v
Germany, Series A No 73, Judgment of 21 Feb 1984, para 50; Bendenoun v France, Series A No 284,
Judgment of 24 Feb 1994, para 47.
13
In App No 56672/00 Senator Lines GmbH v the 15 Member States of the European Union, the
question was raised. The European Commission had imposed a fine of €273 million on Senator Lines.
The EC Courts rejected the company’s request to suspend the payment of the fine. The company
alleged a violation of Art 6 ECHR (right of access to court), in that it was required to pay the fine
before a decision was taken in the substantive proceedings before the EC Courts. It claimed that this
would have resulted in the insolvency and liquidation of the company before the substantive issues
were determined. The application was declared inadmissible by the ECtHR after the CFI decided to
set aside the fine imposed by the European Commission.
14
European Commission of Human Rights, Stenuit v France, Series A No 232-A, Decision of 11
July 1989. Stenuit withdrew its application in 1991.
European Court of Human Rights declared the application brought by the

companies concerned for alleged violation of Article 6 ECHR inadmissible on the
ground that the proceedings at issue were not criminal within the meaning of
Article 6 ECHR.15
488. The European Court of Justice appears to have acknowledge since its
Hüls judgment of 1999 that the fines imposed by the European Commission then
under Regulation 17 and now Regulation 1/2003 may be qualified as ‘criminal’
within the meaning of Article 6 ECHR.16
489. Antitrust proceedings at the level of the EU institutions or in the EU
Member States which are, according to EU law or to the applicable national law,
of an administrative or civil nature may thus fall under the wider notion of
‘criminal’ within the meaning of Article 6 ECHR.
490. It should however be pointed out that, in its case law on the compatibil-
ity of specific proceedings with the substantive requirements of Article 6 ECHR,
the European Court of Human Rights appears to distinguish between enforce-
ment proceedings which are criminal within a more traditional, narrower sense
or which are considered as criminal under the applicable domestic law, and
proceedings which are only criminal within the wider meaning of Article 6
ECHR.
491. This distinction appears most clearly in the case law of the European
Court of Human Rights on the separation of the decision-making function from
the investigative and prosecutorial function or functions. Indeed, the European
Court of Human Rights has ruled that, for reasons of efficiency, the determina-
tion of civil rights and obligations or the prosecution and punishment of offences
which are ‘criminal’ within the wider meaning of Article 6 ECHR can be
entrusted to administrative authorities, provided that the persons concerned are
able to challenge any decision thus made before a judicial body that has full
jurisdiction and that provides the full guarantees of Article 6(1) ECHR.17 This
means that antitrust proceedings such as those at the level of the EU institutions
are not incompatible with Article 6 ECHR, even if the European Commission
combines the investigative and prosecutorial functions with the decision-making
function, given that the addressees of European Commission decisions imposing
fines can bring an action for annulment before the Court of First Instance, which
manifestly provides the full guarantees of Article 6(1) ECHR and which under-
takes a comprehensive review of the Commission’s decisions.18 This alternative
means of satisfying the requirements of Article 6(1) ECHR, however, does not
15
Decision of 3 June 2004 on App nos 69042/01 etc OOO Neste St Petersburg and Others v Russia.
16
Case C–199/92P Hüls v European Commission [1999] ECR 4287, para 150.
17
Judgments of 23 June 1981, Le Compte, Van Leuven and De Meyere v Belgium, Series A No 43,
para 51, of 1 Feb 1983, Albert and Le Compte v Belgium, Series A No 58, para 29, of 21 Feb 1984,
Öztürk v Germany, Series A No 73, para 56, and of 24 Feb 1994, Bendenoun v France, Series A No 284,
para 46.
18
See Judgment of the CFI of 15 Mar 2000 in Joined Cases T–25/95 etc Cimenteries CBR and
Others v Commission [2000] ECR II–700, para 719: ‘[w]hen the Court of First Instance reviews the
CRIMINAL ENFORCEMENT IN THE EU MEMBER STATES 161
appear to be available in more traditional areas of criminal law or in areas

considered criminal under domestic law. In those areas the case law of the
European Court of Human Rights appears to require separation of the adjudica-
tive function from the investigative and prosecutorial functions not only on
appeal but also at first instance.19
6.2 IS THERE A TENDENCY TO CR IMINALISE ANTIT R UST ENFORCEMENT IN THE EU

MEMBER STATES?
6.2.1 US Antitrust Enforcement
492. To the extent that there is currently a tendency in the EU Member States, or
even at the level of the EU institutions,20 to criminalise antitrust enforcement,
this is undoubtedly inspired by US antitrust enforcement, even if criminal
antitrust legislation in the US may historically in turn have been inspired by older
European precedents.21
493. Since the original 1890 version of the Sherman Act, violations of sections
1 and 2 of the Act have been punishable under US federal law by criminal fines
for both enterprises and individuals and by imprisonment for individuals.22
Initially the maximum term of imprisonment was one year.23 In 1974 Congress
raised the maximum term to three years24 and in 2004 to 10 years.25 The
maximum criminal fines for enterprises were similarly increased to $1 million in
1974,26 to $10 million in 1990,27 and to $100 million in 2004.28
legality of a decision finding an infringement of [Article 81(1) and/or Article 82 EC], the applicants
may call upon it to undertake an exhaustive review of both the Commission’s substantive findings of
fact and its legal appraisal of those facts’.
19
See Judgments of 26 Oct 1984, De Cubber v Belgium, Series A No 86, paras 31–32, and of 25 Feb
1997, Findlay v United Kingdom, Reports 1997–I, para 79.
20
See below, paras 538–543 on EU Member States and below, para 510, on the EU institutions.
21
In particular English 16th century legislation, which had however fallen in disuse by the 19th
century: see N Green, ‘The Road to Conviction—The Criminalisation of Cartel Law’ in BE Hawk
(ed), 2003 Annual Proceedings of the Fordham Corporate Law Institute International Antitrust Law &
Policy (Huntington, NY, Juris Publishing, 2004) 13 at 13–22. In France, following the Loi Le Chapelier
of 2–17 Mar 1791, a criminal cartel prohibition was also contained in Art 419 of the Napoleonic Penal
Code.
22
For an overview of the history of US criminal antitrust enforcement from 1890 to 2001 see DI
Baker, ‘The Use of Criminal Law Remedies to Deter and Punish Cartels and Bid-Rigging’ (2001) 69
George Washington Law Review 693 at 694–6.
23
Sherman Act, ch. 647, 26 Stat 209 (1890).
24
Antitrust Procedures and Penalties Act, Pub L No. 93–528, § 3, 88 Stat 1706, 1708 (1974).
25
Antitrust Criminal Penalty Enhancement and Reform Act of 2004, HR 1086.
26
Above n 24.
27
Antitrust Amendments Act of 1990, Pub L No 101–588, § 4, 104 Stat 2879, 2880 (1990).
28
Above n 25.
494. More than half of the US states also have criminal antitrust statutes that
provide for jail sentences, and many of these statutes predate the federal Sherman
Act.29
495. It was not until 1959, however, that prison sentences were imposed
against businessmen for price fixing without acts or threats of violence.30 Since
then, enforcement has increased very strongly. During the 1990s, the US Depart-
ment of Justice successfully prosecuted an average of more than 35 individuals
each year. During 1990–9, 132 individuals were sentenced to prison terms
averaging almost nine months, and 154 were sentenced to periods of confine-
ment in a halfway house or community treatment centre averaging four
months.31
496. In his 2004 overview of US antitrust enforcement, Joel Davidow
reported:
In the last fiscal year, defendants in [Department of Justice Antitrust] Division cases
were sentenced to more than 10,000 jail days—a record—with an average sentence of
more than 18 months. In the last four years, a total of over 75 years of imprisonment
have been imposed on antitrust defendants, with more than 30 defendants receiving
prison sentences of one year or longer. It is not just US executives who are facing prison
sentences, but foreign executives as well. Business people from Canada, France, Ger-
many, Sweden, and Switzerland have now served term in US prisons for violating US
antitrust laws. Extradition of cartel leaders is now frequently sought. The Justice
Department is conducting over 40 criminal investigations of international cartel
activity.32
497. Although the text of the criminal provisions in the Sherman Act covers all
types of restrictive agreements as well as unilateral monopolistic behaviour, by
long tradition the US Department of Justice limits criminal prosecution to
hard-core violations of section 1 of the Sherman Act such as price fixing, bid
rigging, market division or customer allocation schemes among horizontal
competitors.33
29
KG Dau-Schmidt, ‘Sentencing Antitrust Offenders: Reconciling Economic Theory with Legal
Theory’ (1983) 9 William Mitchell Law Review 75 at 75–6.
30
JC Gallo, KG Dau-Schmidt, JL Craycraft and CJ Parker, ‘Criminal Penalties Under the Sherman
Act: A Study of Law and Economics’ (1994) 16 Research in Law and Economics 25 at 40.
31
‘Antitrust Deterrence in the United States and Japan’, remarks by SM Chemtob at a Conference
on Competition Policy in the Global Trading System (Washington, DC, 23 June 2000), available at
www.usdoj.gov/atr/public/speeches/5076.pdf, at 7–8; for more statistics see JC Gallo, K Dau-Schmidt,
JL Craycraft and CJ Parker, ‘Department of Justice Antitrust Enforcement, 1955–1997: An Empirical
Study’ (2000) 17 Review of Industrial Organization 75.
32
J Davidow, ‘Recent US Antitrust Developments of International Relevance’ (2004) 27 World
Competition 407 at 409; see also SD Hammond, ‘An Overview Of Recent Developments In The
Antitrust Division’s Criminal Enforcement Program’, address to the American Bar Association
Midwinter Leadership Meeting (Kona, Hawaii, 10 Jan 2005), available at www.justice.gov/atr/public/
speeches/207226.htm.
33
SW Waller, ‘The Incoherence of Punishment in Antitrust’ (2003) 78 Chicago-Kent Law Review
207 at 216, referring in n 48 to US v Dunham Concrete Prods, Crim No 1842 (ED La 1969) as the last
498. As the US Department of Justice has no powers to impose or to seek the

imposition of civil or administrative fines, it only seeks prospective, injunctive
relief against other violations of section 1 and all violations of section 2 of the
Sherman Act. The Federal Trade Commission has relatively recently applied in
competition cases its power, which it had traditionally used in consumer protec-
tion cases, to seek disgorgement of unlawful profits.34 Otherwise deterrence and
punishment are left to private treble-damages actions.35
499. The imposition of criminal penalties for violation of section 1 of the
Sherman Act requires a finding of criminal intent. It also carries a rather strong
moral condemnation. Indeed, in the US, antitrust values enjoy a very wide
consensus36 and price-fixing is considered to be immoral, like theft.37
500. To investigate suspected criminal antitrust violations, the US Depart-
ment of Justice can make use of a full arsenal of criminal investigatory powers.38
The US Department of Justice also operates a very successful Corporate Leniency
Policy and an Individual Leniency Policy, which allow corporations and individu-
als to avoid criminal prosecution if they are the first to cooperate with the
Department.39
501. Defendants in turn benefit from a full range of criminal rights of
defence—less extensive however for corporations than for individuals—40 and
violations have to be proven ‘beyond reasonable doubt’.
criminal monopolisation indictment discovered by that author; and US Antitrust Modernization

Commission, Report and Recommendations (Washington, DC, Antitrust Modernization Commission,
Apr 2007) at 296–7 and 303.
34
S 13 (G) of the FTC Act; see FTC v Mylan Labs Inc, 62 F Supp 2d 25, 36 (DDC 1999) and Waller,
above n 33, at 218.
35
See ibid, at 210 and 230–2; S Calkins, ‘Corporate Compliance and the Antitrust Angencies’
Bi-Modal Penalties’ (1997) 60 Law and Contemporary Problems 127 at 136–9 and 150; and letter dated
5 Jan 2005 from RH Pate, Assistant Attorney General, Department of Justice Antitrust Division, to DA
Garza, Chair, Antitrust Modernization Commission, ‘Suggested Topics for Antitrust Modernization
Commission Study’, available at www.usdoj.gov/atr/public/comments/207122.pdf, at 2. The Antitrust
Modernization Commission did not propose any change in this respect: see US Antitrust Moderniza-
tion Commission, above n 33, at 285–91.
36
In the words of Robert Reich, ‘antitrust is popular with the public’ and ‘has a compelling
ideologic attraction’, ‘the ideal of competition continuing to receive broad support from a wide
spectrum of American public opinion’: RB Reich, ‘The Antitrust Industry’ (1980) 68 Georgetown Law
Journal 1035 at 1054 and n 1.
37
See Wils, above n 4, sect 8.6.7, and Baker, above n 22, at 714.
38
See below, paras 564 and 565 and 592–594.
39
US Department of Justice Corporate Leniency Policy (10 Aug 1993) and US Department of
Justice Leniency Policy for Individuals (10 Aug 1994): see above, paras 347–361 and below, para 594.
40
Corporations can in particular not invoke the privilege against self-incrimination enshrined in
the 5th Amendment to the US Constitution: see White v US, 322 US 694 (1911).
6.2.2 Antitrust Enforcement at the Level of the EU Institutions
502. As regards the enforcement of Articles 81 and 82 EC at the level of the EU

institutions, Article 23(2) of Regulation 1/2003 empowers the European Com-
mission to impose fines for violations of Articles 81 and 82 EC only on
‘undertakings’. Neither the Treaty nor Regulation 1/2003 contains any definition
of the term ‘undertaking’. The Court of Justice has held that ‘in competition law,
the term “undertaking” must be understood as designating an economic unit …
even if in law that economic unit consists of several persons, natural or legal’.41
503. In practice the Commission does not simply impose fines on undertak-
ings thus defined as economic units. For reasons of enforceability, it rather
addresses its fining decisions to entities possessing legal personality, typically
companies, to which the violation committed by the undertaking is imputed.42
As regards the choice of the person or persons to whom a violation is to be
imputed, the general rule formulated by the Community Courts is that ‘when …
a violation is found to have been committed, it is necessary to identify the natural
or legal person who was responsible for the operation of the undertaking at the
time when the violation was committed, so that it can answer for it’.43
504. If the undertaking found to have committed a violation were found to
consist of an unincorporated business (a single trader, with or without employ-
ees, who has not incorporated his or her business, or a professional exercising his
or her profession alone and unincorporated, or several natural persons operating
a single business without any employment relationship between them and
without any form of legal person), the Commission would necessarily have to
address its fining decision to the natural person or persons operating the
business. However, in the more than 100 decisions in which the Commission has
imposed fines under Regulation 17 and Regulation 1/2003 hitherto, this situation
has never occurred: all fines have been imposed on companies or other legal
persons. Indeed, the undertakings found to have committed violations of Article
81 or 82 EC invariably either consisted of a single company or other legal person,
which was the obvious addressee of the fining decision, or consisted of a group of
companies, in which case fines were imposed on one or more of the companies in
the group.44
41
Judgment of 12 July 1985 in Case 170/83 Hydrotherm v Compact [1985] ECR 3016, para 11. For
a detailed analysis of the notion of undertaking and the relationship between undertaking and legal
or natural person see Wils, above n 4, ch 7.
42
See Opinion of Judge Vesterdorf, acting as AG, in Case T–1/89 Rhône-Poulenc v Commission
[1991] ECR II–869 at 916, Judgment of the CFI of 20 Apr 1999 in Joined Cases T–305/94 etc
Limburgse Vinyl Maatschappij and Others v Commission [1999] ECR II–945, para 978.
43
Judgment of the CFI of 17 Dec 1991 in Case T–6/89 Enichem Anic v Commission [1991] ECR
II–1695, para 236, and Judgment of the ECJ of 16 Nov 2000 in Case C–279/98 P Cascades v
Commission [2000] ECR I–9709, para 78.
44
In Pre-Insulated Pipes, one of the undertakings was controlled and managed by a natural
person, Dr W Henss, but the Commission appears to have made an effort (for reasons unexplained)
to avoid imposing the fine on him, identifying instead a collection of companies which it held jointly
505. According to Article 23(2) of Regulation 1/2003, fines can be imposed on

undertakings where they infringe Article 81 or 82 EC ‘either intentionally or
negligently’. Criminal intent is thus clearly not required.
506. Article 23(5) of Regulation 1/2003 expressly provides that the decisions
by which the European Commission imposes fines ‘shall not be of a criminal law
nature’. As explained above,45 this provision appears to have some relevance not
for the question whether Article 6 ECHR is applicable to the antitrust enforce-
ment proceedings at the level of the EU institutions, but rather for the question
of the compatibility of those proceedings with the substantive requirements of
Article 6 ECHR. The EU Council, when it adopted Regulation 1/2003, may also
have considered that this provision had some relevance to the necessary legal
basis under the EC Treaty.46 It may also have some importance as regards the
status of such fines under national law, for instance whether they are tax
deductable.47
507. The affirmation that fining decisions ‘shall not be of criminal nature’
could also be read as reflecting the absence of moral condemnation of violations
of Articles 81 and 82 EC. On the other hand, there appears in recent years to have
been an increase on the part of the European Commission in the rhetoric of
condemnation of cartels in particular. For instance, in a speech in 2000, Mario
Monti, the then Competition Commissioner, declared that ‘cartels are cancers on
the open market economy, which forms the very basis of our [European]
Community’.48 It may also be noted that the European Commission’s Directorate
General for Competition has in recent years started using the term ‘antitrust’
alongside the official term ‘competition law’.49 As Christopher Harding has
pointed out, the North American term ‘antitrust’ reflects the proscriptive and
negative character of US antitrust law, as opposed to the historically more
tolerant European approach.50
508. As regards the relationship between the size of the penalty for antitrust
violations and the size of the harm caused by them, Christopher Harding has
made the following observations, starting from the ranking of ‘seriousness’ in the
European Commission 1998 Guidelines on the method of setting fines:
and severally liable for the fine: see recs 157–160 and Art 3(d) of the Decision of 21 Oct 1998 in
Pre-Insulated Pipe Cartel [1999] OJ L24/1, and para 105 of the judgment of the CFI of 20 Mar 2002 in
Case T–9/99 HFB and Others v Commission [2002] ECR II–1530, in which the Court found that the
European Commission had ‘intentionally’ not established Dr Henss’s liability.
45
Paras 488–491.
46
See further paras 618–624 and n 192 below.
47
See Wils, above n 2, sect 1.1.4.3.3, n 88, and CS Kerse and N Khan, EC Antitrust Procedure, 5th
edn (London, Sweet & Maxwell, 2005) at para 7–090.
48
‘Fighting Cartels Why and How? Why Should We be Concerned with Cartels and Collusive
Behaviour?’, speech at the 3rd Nordic Competition Policy Conference (Stockholm, 11–12 Sept 2000),
available at http://europa.eu.int/comm/competition/speeches.
49
See the website of DG Competition: http://europa.eu.int/comm/competition/index_en.html.
50
C Harding, ‘Business Cartels as a Criminal Activity: Reconciling North American and European
Models of Regulation’ (2002) 4 Maastricht Journal of European and Comparative Law 393 at 404.
What is especially notable in this ranking is the focus on elements of market impact,
and especially—at the higher level of seriousness—impact on the functioning of the
single market. There is no explicit mention of the ‘bad attitude’ elements which are so
prominent in the American rules, so that the European ‘offence’ is very much an offence
of outcome. A further examination of the Commission’s practice in determining the
severity of fines (its ‘sentencing practice’) reveals that market impact factors supply the
cardinal points on its sentencing tariff, while attitudinal factors (such as organized
collusive activity, furtiveness and secrecy, antitrust awareness, recidivist behaviour)
appear as aggravating factors, or ordinal points, moving up and down the tariff within a
range founded upon a base gravity derived from market impact.51
The Commission’s more recent 2006 Guidelines on the method of setting fines
appear different in this respect, in particular in that the impact on the market is
no longer mentioned as a criterion, and that repeated infringement becomes an
important criterion.52
509. As regards the European Commission’s powers of investigation, Regula-
tion 17 already gave the European Commission significant powers, in particular
the power to conduct inspections at business premises, including the ability to
overcome opposition with the assistance of the police or other authorities of the
Member State where the inspection takes place. But these powers are clearly
much more limited than those of the US Department of Justice or those which
are common for criminal investigations in the EU Member States.
510. Since 1996, the European Commission has been operating a leniency
programme which, in its current version dating from 2006, offers full immunity
to the first cartel member to cooperate with the Commission’s investigation.53
Such a leniency policy, which was clearly adopted following the US example,
would traditionally in the EU Member States, if known at all, be reserved for the
most serious forms of organised crime.54 Of course, since the European Commis-
sion can only impose fines on undertakings, the leniency programme only
concerns them, not individuals (unless the individual is an undertaking55).
511. Regulation 1/2003 has strengthened the European Commission’s powers
of investigation in two respects.56 First, Article 21(1) empowers the Commission
to conduct investigations in premises other than business premises, including the
homes of directors and employees of the undertakings concerned, ‘if a reasonable
suspicion exists that books or other records related to the business and to the
subject-matter of the inspection, which may be relevant to prove a serious
violation of [Article 81 or 82 EC], are being kept in [those] premises’. Article
51
Harding, above n 50, at 412; see Ch 4, n 2 above.
52
See Ch 4 above, in particular n 84 and paras 293–308.
53
European Commission Notice on immunity from fines and reduction of fines in cartel cases
[2006] OJ C298/17; see Ch 5 above, in particular paras 362–370.
54
For instance the mafia in Italy.
55
56
For a detailed analysis of these changes see Dekeyser and Gauer, above n 10.
21(3) adds that a decision to conduct such inspection can only be executed with
prior authorisation from the national judicial authority of the Member State
where the inspection takes place.
512. Secondly, Article 20(2)(e) of Regulation 1/2003 provides that, during an
inspection at the business premises of an undertaking, the inspectors are empow-
ered ‘to ask any representative or member of staff of the undertaking … for
explanations on facts or documents relating to the subject-matter and purpose of
the inspection and to record the answers’.57 However, Regulation 1/2003 does not
threaten the individuals thus questioned with any sanctions for refusal to answer,
nor for giving incorrect, incomplete or misleading answers. Article 23(1)(d) only
provides for fines on undertakings that ‘fail or refuse to provide a complete
answer on facts’, ‘give an incorrect or misleading answer’ or ‘fail to rectify within a
time-limit set by the Commission an incorrect, incomplete or misleading answer
given by a staff member’.
513. The defendant companies’ rights of defence in antitrust proceedings
conducted by the European Commission have significantly increased over time.58
The European Commission combines the investigative and prosecutorial func-
tion with the adjudicative function in the first instance, but its decisions are
subject to exhaustive review by the Court of First Instance.59 The standard of
proof which the European Commission has to satisfy to find a violation of Article
81 or 82 EC and impose a fine according to Chris Kerse and Nicholas Khan ‘is a
high one, although it would be going too far to say that the infringement must be
proven to the criminal standard of “beyond reasonable doubt”’.60
6.2.3 The EU Framework for Antitrust Enforcement in the EU Member States
514. Regulation 1/2003 frames antitrust enforcement in the EU Member States

in two ways. First, it allows Member States to provide for criminal enforcement of
Articles 81 and 82 EC. Secondly, it restricts the possibilities for Member States to
enforce their national competition laws in isolation from the enforcement of
Articles 81 and 82 EC.
57
Under Reg 17, the Commission officials could only seek explanations relating to the books and
records under examination: see Judgment of the ECJ of 26 June 1980 in Case 136/79 National
Panasonic v Commission [1980] ECR 2056, para 15, and Order of the CFI of 9 June 1997 in Case
T–9/97 Elf Atochem v Commission [1997] ECR II–919, para 23.
58
See K Lenaerts and I Maselis, ‘Procedural Rights and Issues in the Enforcement of Articles 81
and 82 of the EC Treaty’ (2001) 24 Fordham International Law Journal 1615, and Kerse and Khan,
above n 47.
59
See above nn 18 and 9.
60
Kerse and Khan, above n 47, para 8–038. The standard formula used by the EC Courts is that
‘the Commission must produce sufficiently precise and consistent evidence to support the firm
conviction that the alleged infringement took place’; see for instance the Judgment of the CFI of 8 July
2004 in Joined Cases T–67/00 etc JFE Engineering Corp v European Commission [2004] ECR II–2501,
para 179. See further B Vesterdorf, ‘Burden of Proof in Cartel Cases’, paper presented at the Nordic
Lawyers Academy—Challenges in Nordic Cartel Cases (Helsinki, 4–5 May 2006).
6.2.3.1 Member States can Criminalise their Enforcement of Articles 81 and 82 EC
515. Article 5 of Regulation 1/2003 provides that ‘[t]he competition authorities

of the Member States shall have the power to apply Articles 81 and 82 of the
Treaty in individual cases. For this purpose, acting on their own initiative or on a
complaint, they may take … decisions … imposing fines, periodic penalty
payments or any other penalty provided for in their national law’.
516. Member States can thus in their national law provide for administrative
or civil fines on companies for violations of Article 81 or 82 EC, or criminal fines,
or fines on individuals, of an administrative, civil or criminal nature, or impris-
onment, or indeed ‘any other penalty’.
517. That ‘any other penalty’ also covers penalties on individuals, including
imprisonment, is further confirmed by Article 12(3) of Regulation 1/2003.
Whereas Article 12(1) provides that the European Commission and the compe-
tition authorities of the Member States can exchange all kinds of information,
and use the information thus exchanged in evidence, Article 12(3) adds the
following limitation:
to an infringement of [Article 81 or 82 EC] or, in the absence thereof,
cannot be used by the receiving authority to impose custodial sanctions
518. This provision confirms what is already apparent from the terms ‘any other
penalty’ in Article 5 of Regulation 1/2003, namely that Member States can indeed
provide for sanctions on natural persons, including imprisonment, for violations
of Articles 81 and 82 EC.
519. By thus allowing Member States to provide for criminal enforcement of
Articles 81 and 82 EC, the EU Council, which adopted Regulation 1/2003
unanimously, would appear to have rejected the view, which has sometimes been
suggested, according to which Articles 81 and 82 EC would somehow be
inherently non-criminal, and that the EC Treaty would thus exclude criminal
enforcement of Articles 81 and 82 EC not only at the level of the EU institutions
but also at the level of the Member States.61
61
See remarks by Gerard Hogan in Roundtable Debate Session IV—Rights, Privileges and Ethics
in Competition Cases, the 31st Annual Fordham Corporate Law Institute Conference on Interna-
tional Antitrust Law & Policy (New York, 8 Oct 2004), in BE Hawk (ed) 2004 Annual Proceedings of the
Fordham Corporate Law Institute—International Antitrust Law and Policy (Huntington, NY, Juris
Publishing, 2005).
520. This view appears mistaken. It is of course correct that the text of Articles
81 and 82 EC only mentions ‘undertakings’. But, as is apparent from Article 83
EC, the prohibitions on restrictive agreements and abuse of dominant positions
by undertakings, as set out in Articles 81 and 82 EC, are only ‘principles’, and the
appropriate regulations or directives to give effect to these principles are to be
laid down by the Council. Nothing in the EC Treaty suggests that the Council
could thus not allow Member States to provide for penalties for individuals,
including imprisonment, to enforce Articles 81 and 82 EC, as it did in Regulation
1/2003.62
521. The possibility offered by Regulation 1/2003 for Member States to
provide for criminal enforcement of Articles 81 and 82 EC has been used by a
number of Member States.63 Ireland and Estonia have provided only for criminal
enforcement, through fines on companies and individuals, as well as imprison-
ment.64 In the United Kingdom, non-criminal fines are provided for companies,
but also director disqualification for individuals, as well as imprisonment and/or
criminal fines for individuals specifically for price-fixing, limitation of supply or
production, market-sharing and bid-rigging.65 In France, Cyprus and the Slovak
Republic, the main sanctions provided for are non-criminal fines on companies,
but imprisonment for individuals is also provided for, at least in theory.66 In
Denmark and Malta, essentially only fines on companies are provided for, but
these have a criminal character. In Malta, directors may be held jointly and
severally liable with the company for these fines. In Greece, non-criminal fines on
companies (with joint and several liability for managers) are complemented by
criminal fines for individuals. In Germany, non-criminal fines are provided for
individuals, and derivatively also for companies, as well as imprisonment specifi-
cally for bid-rigging. In Austria, non-criminal fines are provided only for com-
panies, but also imprisonment specifically for bid-rigging.67
6.2.3.2 Member States Cannot Criminalise the Enforcement of Only their National
Competition Laws
522. The possibilities for Member States to enforce their national competition
laws in isolation from the enforcement of Articles 81 and 82 EC are limited in
three ways. First, Article 3(2) of Regulation 1/2003 forbids Member States to
prohibit under national competition law agreements or concerted practices
which may affect trade between Member States but which do not restrict
62
See also below, para 619 .
63
See above n 5.
64
See also below, para 539 on Ireland.
65
See also ibid and below, paras 539 and 582 on the United Kingdom.
66
See also below, para 534 on France.
67
See also below, para 534 on Austria.
competition within the meaning of Article 81(1) EC, or which fulfill the
conditions of Article 81(3) EC or which are covered by a block exemption
regulation applying Article 81(3) EC.68
523. Secondly, it follows from the principle of equivalence, a general principle
of Community law developed in the case law of the Court of Justice,69 that the
penalties which Member States provide for violations of Articles 81 and 82 EC
pursuant to Article 5 of Regulation 1/2003 must at least be equivalent in
effectiveness and dissuasiveness to the sanctions they provide for comparable
violations of their national competition laws. If a Member State thus provides, for
instance, imprisonment sanctions to enforce a cartel prohibition under national
law, it must also provide for such sanctions to enforce the cartel prohibition in
Article 81 EC.
524. Thirdly, Article 3(1) of Regulation 1/2003 provides:
Where the competition authorities of the Member States or national courts apply
national competition law to agreements, … or concerted practices within the meaning
of [Article 81(1) EC] which may affect trade between Member States within the
meaning of that provision, they shall also apply [Article 81 EC] to such agreements, …
or concerted practices. Where the competition authorities of the Member States or
national courts apply national competition law to any abuse prohibited by [Article 82
EC], they shall also apply [Article 82 EC].
525. Unless the practices concerned do not affect trade between Member States,
national competition authorities and national courts thus cannot apply national
competition law without simultaneously applying Articles 81 and 82 EC. This
means that they cannot avoid the application of the mechanisms of cooperation
with the European Commission and/or within the EU network of competition
authorities as provided for in Articles 11 and 15 of Regulation 1/2003.70
526. Article 3(3) of Regulation 1/2003 clarifies that, ‘[w]ithout prejudice to
general principles and other provisions of [European] Community law’, Article
3(1) and (2) ‘do [not] preclude the application of provisions of national law that
predominantly pursue an objective different from that pursued by [Articles 81
and 82 EC]’. Recital 9 of Regulation 1/2003 explains that Articles 81 and 82 EC
‘have as their objective the protection of competition on the market’, and gives
‘national legislation that prohibits or imposes sanctions on acts of unfair trading
practice’ as an example of legislation protecting ‘other legitimate interests’.
68
On Art 3 of Reg 1/2003 and on block exemption regulations see Wils, above n 2, sects 1.2.8.2
and 1.2.1 respectively.
69
See, inter alia, Judgment of the ECJ of 15 Sept 1998 in Case C–231/96 Edis [1998] ECR I–4990,
paras 34 and 36–37; for a more detailed analysis see P Oliver, ‘Le règlement 1/2003 et les principes
d’efficacité et d’équivalence’, in [2005] Cahiers de droit européen 343.
70
On these cooperation mechanisms see D Schnichels, ‘The Network of Competition Authorities:
How Will it Work in Practice?’ in D Geradin (ed), Modernisation and Enlargement: Two Major
Challenges for EC Competition Law (Antwerp, Intersentia, 2004) 99–121 and Wils, above n 2, chs 1 and
2.
527. The three limitations mentioned above can be illustrated by the cartel
offence in the United Kingdom. Section 188 of the Enterprise Act 2002 provides
that an individual is guilty of this criminal offence, punishable by imprisonment
and/or a fine, if he or she dishonestly agrees with one or more other persons that
undertakings will engage in one or more of the prohibited cartel activities,
namely price-fixing, limitation of supply or production, market-sharing and
bid-rigging.71
528. First, given that price-fixing, limitation of supply or production, market-
sharing and bid-rigging are, when they actually or potentially affect trade
between Member States, also prohibited under Article 81 EC, this prohibition
under UK law is compatible with Article 3(2) of Regulation 1/2003.
529. Secondly, section 188 of the Enterprise Act 2002 does not distinguish in
any way between, on the one hand, price-fixing, limitation of supply or produc-
tion, market-sharing or bid-rigging that may affect trade between Member States,
and thus falls under Article 81 EC, and, on the other hand, prohibited cartel
activities that do not affect trade between Member States, and thus only fall
under the Chapter I prohibition of the Competition Act 1998, the provision of
UK national competition law comparable to Article 81 EC. Unless the UK
authorities were to develop an administrative practice of investigating and
prosecuting only cases which concern prohibited cartel activities that do not
affect trade between Member States, the principle of equivalence is thus fully
respected.
530. Thirdly, when a case which concerns prohibited cartel activities that may
affect trade between Member States is investigated, prosecuted and punished by
the competent UK authorities and courts, it would in my view follow from
Article 3(1) of Regulation 1/2003 that they will have to establish that the
prohibited cartel activities indeed may affect trade between Member States and is
thus contrary to Article 81 EC, and they will have to comply with the mecha-
nisms of cooperation with the European Commission and within the EU
network of competition authorities as provided for in Articles 11 and 15 of
Regulation 1/2003.
531. No prosecution has yet been brought under section 188 of the Enterprise
Act 2002, but the UK authorities have indicated that they consider that Regula-
tion 1/2003 would not apply to the cartel offence.72 In the view of the UK
authorities, ‘[t]he cartel offence in Part 6 of the Enterprise Act and the prohibi-
tions of Article 81 or Chapter I are aimed at different legal persons and the
71
Enterprise Act 2002, available at www.legislation.hmso.gov.uk/acts/acts2002/20020040.htm.
72
This view has also been expressed by an official of the European Commission’s Directorate
General for Competition, speaking in a personal capacity: see remarks by Kris Dekeyser in Roundta-
ble Debate Session IV—Rights, Privileges and Ethics in Competition Cases, the 31st Annual Fordham
Corporate Law Institute Conference on International Antitrust Law & Policy (New York, 8 Oct 2004),
in BE Hawk (ed), 2004 Annual Proceedings of the Fordham Corporate Law Institute—International
Antitrust Law and Policy (Huntington, NY, Juris Publishing, 2005). The European Commission has to
my knowledge never taken a position on this matter.
former is concerned with what is essentially dishonest activity. The cartel offence
aims to catch dishonest activity by individuals whereas the civil prohibitions are
directed at undertakings. [Regulation 1/2003] does not apply to national laws,
which impose criminal sanctions on natural persons, except to the extent that
such sanctions are the means whereby competition rules applying to undertak-
ings are enforced’.73
532. As already argued above,74 I believe that the idea that Article 81 EC is
somehow inherently non-criminal, and could not be enforced by penalties on
individuals, including penalties of a criminal law nature, is mistaken, and
contradicted by Regulation 1/2003, as well as by the practice of other Member
States, in particular Ireland and Estonia, but also France, Cyprus, the Slovak
Republic, Denmark, Malta and Greece. The UK authorities’ view that Regulation
1/2003 would not apply to the criminal cartel offence apparently also relies upon
the very last sentence of recital 8 of Regulation 1/2003, which says that ‘this
Regulation does not apply to national laws which impose criminal sanctions on
natural persons except to the extent that such sanctions are the means whereby
competition rules applying to undertakings are enforced’.75 However, this sen-
tence in my view merely specifies, with regard to national laws which impose
criminal sanctions on natural persons, what already follows from Article 3(3) of
Regulation 1/2003, namely that Article 3(1) and (2) of this Regulation does not
preclude the application of provisions of national law that predominantly pursue
an objective different from that pursued by Articles 81 and 82 EC. Article 3(1)
and (2) of the Regulation would however appear to apply to all national laws that
do not predominantly pursue an objective different from that pursued by Articles
81 and 82 EC, irrespective of whether those laws impose criminal sanctions on
natural persons.76 I have difficulty seeing how one could consider that the UK
cartel offence predominantly pursues an objective different from that pursued by
Article 81 EC. Indeed, it is perfectly clear from the history of the Enterprise Act
2002 that the cartel offence was introduced in UK law because it was considered
that the existing fines on companies were insufficient to deter hard-core cartels
prohibited by Chapter I of the Competition Act 1998 and Article 81 EC, and that
imprisonment was the most effective additional deterrent.77 The cartel offence in
the UK should thus in my view be considered a means whereby competition rules
applying to undertakings are enforced.
73
Para 10.16 of the UK Department of Trade and Industry’s Consultation on Modernisation
(London, DTI, Apr 2003), available at www.dti.gov.uk/ccp/consultpdf/compmodcon.pdf.
74
Above, paras 519–521; see also below, para 619.
75
Whereas the Italian, Spanish and Dutch versions of this sentence use the equivalent of the
words ‘the means’ in the English version (gli strumenti, el medio, het middel), the French and German
versions rather use the equivalent of ‘a means’ (un moyen, Mittel).
76
See Wils, above n 2, sect 1.2.8.2.
77
See, inter alia, Department of Trade and Industry, ‘A World Class Competition Regime’,
presented to Parliament by the Secretary of State for Trade and Industry, July 2001, available at
www.archive.official-documents.co.uk/document/cm52/5233/523301.htm.
6.2.4 Decriminalisation and Criminalisation of Antitrust Enforcement in the

EU Member States
533. Looking at antitrust enforcement in the EU Member States over the last 20
years, one can detect a double tendency of decriminalisation and criminalisa-
tion.78
6.2.4.1 Decriminalisation on the One Hand
534. The tendency to decriminalise antitrust enforcement is illustrated by the

examples of France, the Netherlands, Austria and Luxembourg. In France,
antitrust enforcement under the Competition Acts of 1953 and 1958 was exclu-
sively criminal. After a first reform in 1977, the Competition Act of 1986
introduced the current French system, which relies essentially on administrative
fines on companies, imposed by the Competition Council, a specialised admin-
istrative authority. Criminal sanctions, including imprisonment, are however still
provided for in French law. In the Netherlands, the current Competition Act,
which came into effect in 1997, replaced an earlier system of criminal enforce-
ment by a purely administrative system with only fines on undertakings, compa-
rable to the enforcement system at the level of the EU institutions. In Austria, the
2002 competition law reform also replaced a system of criminal enforcement by a
system of administrative fines on companies. A specific criminal offence of
bid-rigging was however maintained. In Luxembourg, the 2004 Competition Act
abolished an earlier competition regime with criminal enforcement by a new
regime with only administrative fines on companies.
535. These movements to decriminalise antitrust enforcement reflect, I
believe, two phenomena. The first explanation is that some of the older compe-
tition laws with criminal enforcement which were abolished in these Member
States were in fact not comparable to Articles 81 and 82 EC, because they were
not based on the prohibition principle. In Luxembourg, for instance, before the
2004 reform, there were in fact no sanctions at all for the mere fact of having
concluded and operated a price cartel. Only if a cartel had been investigated and
the undertakings had been ordered to put an end to it were the owners, managers
and directors of those undertakings threatened with criminal sanctions if they
continued the cartel notwithstanding the order to put an end to it. Luxembourg
can thus hardly be considered to be leading a new trend towards decriminalisa-
tion. Rather, it was the last to follow an earlier trend started by the European
Communities and Germany in the 1950s.
536. Indeed, looking at the whole of Europe over close to a century, specifi-
cally with regard to cartels, it would seem that most jurisdictions are historically
78
See also above, para 521 on the current situation in various Member States.
moving through similar phases in the same direction. In a first phase, the one
which Luxembourg abandoned only in 2004, there was no real cartel prohibition,
only a system of control of abuse under which, once an abuse had been
established, the undertakings concerned would have been ordered to put an end
to this abuse for the future. Failure to respect this specific order for the future
may have been punishable with criminal sanctions, including imprisonment. In a
second phase, first reached by the European Coal and Steel Community in the
early 1950s, and subsequently by Germany and by the European Economic
Community in the late 1950s, cartels were prohibited, but the direct effect of this
prohibition was softened through a notification system which allowed temporary
immunity from sanctions,79 and violations of the prohibition were only punish-
able with fines on undertakings. In a third phase, reached at the EU level through
Regulation 1/2003, the deterrent effect of the cartel prohibition is enhanced
through the abolition of the notification system. In a fourth phase, reached long
ago by the United States80 and recently entered into by the United Kingdom,81 the
threat of individual sanctions, in particular imprisonment, is added, so as to
increase deterrence for those types of cartel infringements which are most
profitable and most difficult to detect, namely secret price-fixing, bid-rigging and
market-sharing arrangements.
537. The second explanation for the observed decriminalisation in some EU
Member States over the past 20 years is that, even if some of these Member States
had already moved to a system based on the prohibition principle, their criminal
enforcement did not work in practice because it was badly conceived. Indeed, as
will be argued further in this chapter,82 criminal antitrust enforcement, in its
strongest form, relying on imprisonment, can only work or only makes sense if it
is limited to hard-core cartels, complementing other penalties for companies and
for lesser antitrust violations, if a dedicated investigator and prosecutor and
sufficient powers of investigation are provided for, and if judges or juries are
willing to convict, all of which requires a sufficiently broad political or societal
consensus that cartels are really bad and thus deserve severe punishment. These
conditions would not appear to have been fulfilled in those Member States at the
relevant time.83
79
See Wils, above n 2, sect 1.1.4.4, and G Marenco, ‘Does a Legal Exception System Require an
Amendment of the Treaty?’ in C-D Ehlermann and I Atanasiu (eds), European Competition Law
Annual 2000: The Modernisation of EC Antitrust Policy (Oxford, Hart Publishing, 2001) 145 at 174–5.
80
81
See above, para 527, and below, para 538.
82
At paras 586–602.
83
See further below, para 591. In the Netherlands, more recently, the situation has developed, and
criminal enforcement may again be introduced, but now on a sounder basis. The issue was debated in
the Second Chamber of the Dutch Parliament on 15 June 2006, and broad support for the
introduction of prison sanctions for cartels emerged. In a letter to Parliament of the following day
(Tweede Kamer 2005–6, 30 071, no 27), the Minister of Economic Affairs announced that the
Government would submit a legislative proposal.
6.2.4.2 Criminalisation on the Other Hand
538. The tendency to criminalise antitrust enforcement is primarily illustrated

by the examples of Ireland and the United Kingdom, but also by other phenom-
ena, such as the introduction of leniency programmes in an increasing number of
EU Member States.84
539. In Ireland, since entry into force of the Competition Act 1996, later
replaced by the Competition Act 2002, violations of the national law prohibitions
equivalent to Articles 81 and 82 EC, as well as violations of Articles 81 and 82 EC,
are criminal offences, punishable with fines on companies, and imprisonment
and/or fines for individuals. A comparable system has been adopted in Estonia.
In the United Kingdom, as already mentioned above,85 the Enterprise Act 2002
has added a criminal cartel offence, limited to hard-core cartels, punishable with
imprisonment and/or fines for individuals only, as well as a sanction of director
disqualification for individuals,86 to a system which otherwise relies on adminis-
trative fines on companies for breach of Articles 81 and 82 EC and the equivalent
prohibitions under national law.
540. The introduction of criminal penalties in Ireland and the United King-
dom has been accompanied by the grant of corresponding criminal powers of
investigation to the investigating authorities. In the United Kingdom, for
instance, the Office of Fair Trading has been given new powers of covert
surveillance for the purpose of investigating the criminal cartel offence.87 The
Irish and UK competition authorities also operate leniency policies similar to
those of the US Department of Justice.88
541. As part of a wider trend to criminalise antitrust enforcement, it should
also be mentioned that an increasing number of Member States have followed the
European Commission in operating a leniency programme for undertakings that
help the competition authority to detect cartels.89 As mentioned above,90 such
leniency programmes would traditionally in the EU Member States, if known at
all, typically be reserved for the most serious forms of organised crime.
542. Similarly to what was pointed out above with regard to antitrust enforce-
ment at the level of the EU institutions, it would also appear that in those EU
Member States where antitrust enforcement remains based on administrative
84
See also above, para 510 and n 83 as to recent developments in the Netherlands.
85
Para 527..
86
See further below, para 582 on director disqualification.
87
Ss 199–200 of the Enterprise Act 2002: see also Office of Fair Trading, ‘Powers for Investigating
Criminal Cartels—Guidance’ (London, OFT, Jan 2004), available at www.oft.gov.uk/.
88
See above, para 500 and, as to the UK, Office of Fair Trading, ‘The Cartel Offence—Guidance
on the Issue of No-action Letters for Individuals’ (London, OFT, Apr 2003), supplemented by
Leniency and No-action—OFT’s Interim Note on the Handling of Applications (London, OFT, July
2005), available at www.oft.gov.uk.
89
See the list at http://europa.eu.int/comm/competition/antitrust/legislation/
authorities_with_leniency_programme.pdf; see also above, paras 378 and 385.
90
Para 510.
fines on companies, there is a more general, broad tendency to strengthen both

the powers of investigation of the competition authorities and the defendant
companies’ rights of defence in antitrust proceedings, even if either level gener-
ally remains below the level observed in fully criminalised systems. On the
specific issue of the separation of investigative, prosecutorial and adjudicative
functions, several EU Member States, including France, Belgium, Sweden and
Finland, have to varying extents separated these functions, even if they only
impose fines of an administrative law nature.91
543. The introduction of criminal penalties in Ireland and the United King-
dom, as well as the introduction of leniency programmes at the level of the EU
institutions and in many EU Member States, has clearly been inspired by the US
example.92 That Europe follows the US example in matters of antitrust is of
course not new. Indeed, the history of antitrust in modern Europe,93 and in
particular the history of Article 81 EC and similar provisions in national law, is
very much one of American influence. The prohibition on abuse of a dominant
position, first in the ECSC Treaty of 1951 and subsequently in Article 86 of the
EEC Treaty of 1957 (now Article 82 EC), was not without precedent: in Germany,
for instance, an ‘Ordinance against the Abuse of Economic Power’ had been
enacted in 1923.94 However, the prohibition on restrictive agreements first laid
down in Article 65 of the ECSC Treaty and then also in Article 85 of the EEC
Treaty (now Article 81 EC) was ‘a fundamental innovation in Europe’.95 Before
the Second World War, cartels were a widespread and highly esteemed institution
throughout Europe.96 The insertion of the prohibition on restrictive agreements
in European law, as well as around the same time in German national law, was
due to American influence, if not pressure.97 As Joel Davidow has recently
observed, ‘In the early years of the 21st Century, it has become evident that US
antitrust, once nearly the sole source of influential antitrust enforcement, has
91
See Wils, above n 2, sect 1.2.10.
92
See PF Kunzlik, ‘Globalization and Hybridization in Antitrust Enforcement: European “borrow-
ings from the U.S. approach’ [Summer 2003] The Antitrust Bulletin 319; and above, para 492.
93
See above n 21 on older European precedents.
94
K Nörr, ‘Law and Market Organization: The Historical Experience in Germany From 1900 to
the Law Against Restraints of Competition (1957)’ (1995) 151 Journal of Institutional and Theoretical
Economics/Zeitschrift für die gesamte Staatswissenschaft 5 at 10.
95
J Monnet, Mémoires (Paris, Fayard, 1976) at 413; see however above n 21 on the older European
precedents.
96
HG Schröter, ‘Cartelization and Decartelization in Europe, 1870–1995: Rise and Decline of an
Economic Institution’ (1996) 25 Journal of European Economic History 129 at 137, who also writes at
140 that Yugoslavia was the only European country where cartels were prohibited at that time.
97
See Monnet, above n 95, at 356–7 and 411–13, D Spierenburg and R Poidevin, The History of the
High Authority of the European Coal and Steel Community (London, Weidenfeld and Nicholson, 1994)
at 26–8, and, with a detailed description of the strong resistance to be overcome in Germany, V
Berghahn, The Americanization of West German Industry 1945–1973 (New York, Berg, 1986); see also
G Marenco, ‘The Birth of Modern Competition Law in Europe’ in A von Bogdandy, PC Mavroidis and
Y Mény (eds), European Integration and International Co-ordination—Studies in Transnational
Economic Law in Honour of Claus-Dieter Ehlermann (New York, Kluwer Law International, 2002) 279,
and Wils, above n 4, sects 5.3.2 and 9.5.
IS CRIMINALISATION OF ANTITRUST ENFORCEMENT DESIRABLE? 177
become part of a “dynamic duo”, along with the European Union. The EU and its
Member States have eclipsed the United States in terms of number of profession-
als employed, sometimes in size of fines and certainly in terms of countries
imitating the basic enforcement scheme’.98 However, ‘[t]he United States, even if
no longer the world’s largest jurisdiction, remains the most influential in terms of
policy and theory’.99 The trend to criminalise antitrust enforcement reflects this
influence.
6.3 IS CR IMINALISATION OF ANTIT R UST ENFORCEMENT DESIR ABLE?
544. In this third section I want to address the question whether, why or under
what conditions criminalisation of antitrust enforcement is desirable. The ques-
tion is addressed in general, irrespective of whether such criminalisation takes
place at the level of the EU Member States or of the EU institutions, or whether it
is harmonised at EU level. The question whether it is problematic that antitrust
enforcement is criminalised at the level of individual EU Member States without
parallel criminalisation at the level of the EU institutions or without EU
harmonisation will be dealt with afterwards in the fourth section of this chapter.
545. As described above,100 criminalisation is a phenomenon with different
faces, including for instance the spread of leniency programmes in jurisdictions
which continue to rely exclusively on administrative fines on companies. I will
however focus on its strongest variant, namely the introduction of individual
penalties, and more specifically imprisonment, as exemplified by Ireland and the
United Kingdom.
546. I will first consider five arguments in favour of imprisonment as a
penalty for antitrust offences, then consider for what types of antitrust offences
these arguments are valid, examine whether there are no equally effective
alternatives to imprisonment, and finally discuss five conditions which need to be
fulfilled for criminal antitrust enforcement to be effective, not only in theory but
also in practice.
6.3.1 Five Arguments in Favour of Imprisonment
6.3.1.1 Effective Deterrence with only Fines on Companies would Require

Impossibly High Fines
547. In a paper presented at the sixth Competition Law and Policy Workshop at
the European University Institute (EUI) in 2001, I argued that the minimum level
98
Davidow, above n 32, at 407.
99
Ibid.
100
Paras 509–513 and 541 and 542.
of fines required generally to deter price cartels and other antitrust offences of
comparable profitability and ease of concealment would be in the order of 150
per cent of the annual turnover in the products concerned by the violation.101
548. This argument was based on the following: I took the figure of 10 per
cent of the selling price as an estimate of the average price increase from
price-fixing. This estimate had been relied upon by the US Sentencing Commis-
sion when drafting its Sentencing Guidelines and appeared generally accepted in
the American literature, no European studies being available. As the price
increase caused by the cartel will normally depress demand for the cartel
members’ products, I assumed conservatively that a price increase of 10 per cent
would lead to an increase in profits of 5 per cent of turnover. On the basis of the
findings of duration in a number of cartel decisions of the European Commis-
sion, and of estimates in the American literature, I assumed, again conservatively,
a cartel duration of five years. Finally, I assumed a probability of detection and
punishment of 16 per cent, which I considered again a conservative estimate,
given that the one existing study in the American literature, by Peter Bryant and
Woodrow Eckard, based on a sample of 184 price-fixing cases for the period from
1961 to 1988, had produced an estimate of between 13 and 17 per cent,102 and
that European competition authorities have weaker investigative powers than
their American counterparts. Assuming a 10 per cent price increase, and a
resulting increase in profits of 5 per cent of turnover, a five-year duration and a
16 per cent probability of detection and punishment, the floor below which fines
will generally not deter price-fixing would be in the order of 150 per cent of the
annual turnover in the products concerned by the violation.103
549. A slight variation on this argument was subsequently used by the UK
government in its White Paper setting out its case for the introduction of a
criminal cartel offence.104 The argument was also referred to by the OECD in its
2002 report on hard-core cartels.105
550. Five years later, would these estimates still appear appropriate? As regards
the average price increase from price-fixing, the OECD reported in 2002 the
results of a survey it had carried out. In the limited data (14 cases) of that survey,
101
WPJ Wils, ‘Does the Effective Enforcement of Articles 81 and 82 EC Require Not Only Fines on
Undertakings But Also Individual Penalties, In Particular Imprisonment?’, paper presented at the 6th
Competition Law and Policy Workshop at the European University Institute (Florence, 1–2 June
2001), later published in C-D Ehlermann and I Atanasiu, European Competition Law Annual 2001:
Effective Private Enforcement of EC Antitrust Law (Oxford, Hart Publishing, 2003) 411 and in Wils,
above n 4, 188–237.
102
PG Bryant and EW Eckhard, ‘Price Fixing: The Probability of Getting Caught’ [1991] Review of
Economics and Statistics 531.
103
I added that the figure would be even higher if one were to take into account the fact that fines
are only paid several years after the gain from the violation is obtained. Indeed, my calculations
assumed a zero interest rate. With a more realistic higher interest rate, the minimum fine necessary to
achieve deterrence would be even higher.
104
DTI, above n 77, at 40 (box 7.3).
105
OECD, ‘Report on the Nature and Impact of Hard Core Cartels and Sanctions against Cartels
under National Competition Laws’, DAFFE/COMP(2002)7 (9 Apr 2002), at 19, n 20.
the median overcharge was between 15 and 20 per cent.106 Most recently, John
Connor has made a survey of hundreds of published social-science studies of
hard-core cartels that contained 674 observations of long-run overcharges. His
primary finding was that the median cartel overcharge for all types of cartels over
all time periods throughout the world is 25 per cent. He also conducted a survey
of 24 final verdicts in decided US horizontal collusion cases, revealing an average
median overcharge of 21 per cent and an average mean overcharge of 30 per
cent.107 On the basis of these new surveys, I should have taken 20 per cent rather
than 10 per cent as an estimate of the average price increase from price-fixing in
my argument. Keeping the same conservative assumption about the price elastic-
ity of demand for the cartel members’ products, I would then take 10 per cent of
turnover as an estimate of the increase in profits resulting from the price-fixing.
551. As regards the average duration of price-fixing cartels, the literature
published in the last five years would not alter my conservative estimate of five
years.108
552. As regards the probability of detection and punishment, it has correctly
been pointed out that the detection rate in the US has presumably improved
since the period 1961 to 1988 on which Peter Bryant and Woodrow Eckard based
their estimate of between 13 and 17 per cent,109 because of the use of more
effective investigative methods, in particular the use of optimised leniency
programmes.110 Some have opined that nevertheless ‘there is little reason to
believe that the true probability of detection is outside this range’.111 It would in
any event still remain true that the probability is probably lower in Europe than
in the US.112 I believe I would remain very safely on the conservative side by
assuming the probability of detection and punishment in Europe not to be above
33 per cent.
553. Assuming then a 20 per cent price increase and a resulting increase in
profits of 10 per cent of turnover, a five-year duration and a 33 per c ent
probability of detection and punishment, I would thus still today reach the
106
OECD, above n 105, at 9.
107
Connor, above n 33.
108
See SJ Evenett, MC Levenstein and VY Suslow, ‘International Cartel Enforcement: Lessons
from the 1990s’ [2001] World Economy 1221 at 1226; Connor, above n 33, at 12; and MC Levenstein
and VY Suslow, ‘What Determines Cartel Success?’ (2006) 44 Journal of Economic Literature 43.
109
Above n 102.
110
See OECD, above n 105, at 13; and D Bush, JM Connor, JJ Flynn, S Ghosh, W Grimes, JE
Harrington, N Hawker, R Lande WG Shepherd and S Semerano, ‘How to Block Cartel Formation and
Price-Fixing’, Amicus Curiae brief in the US Supreme Court in Hoffman-La Roche v Empegran, 15 Mar
2004, available at http://aei-brookings.org/admin/authorpdfs/page.php?id=931, at 28; the point was
first made to me by Bill Bishop shortly after I presented my 2001 EUI paper: see also Wils, above n 2,
ch 5.
111
Bush et al, above n 110, at 28.
112
See also OECD, above n 105, at 13.
conclusion that the floor below which fines on companies will generally not deter
price-fixing would be in the order of 150 per cent of the annual turnover in the
products concerned by the violation.113
554. As I argued in more detail in my 2001 EUI paper,114 raising the general
level of fines to such a high level would be impossible or unacceptable for various
reasons. First (and least importantly, as statutory ceilings could be raised), such
higher fines are likely regularly to breach the statutory ceilings on the amount of
fines which can legally be imposed.115
555. Secondly, and most importantly, such high fines are likely often to exceed
the undertakings’ ability to pay, with as a result not only incomplete deterrence
but also costly side-effects. The fact that the companies that have engaged in a
price-fixing cartel have profited financially from the violation offers no guarantee
that they will also be able to pay the minimum fine required for effective
deterrence. Indeed, in the estimated average case of a price cartel leading to extra
profits of 10 per cent of turnover for five years, with a probability of detection
and punishment of 33 per cent, the cartel members would each have gained over
the whole duration of the cartel 50 per cent of their annual turnover in the
products concerned by the violation. Even if they had retained these profits until
the fine was imposed, these would only pay for one third of the fine of 150 per
cent of annual turnover in the products concerned. It is more likely that these
profits would not have been retained but rather would have been paid out in
taxes, dividends, salaries and wages.116
556. Even liquidating the assets of the firms concerned is often unlikely to
generate enough revenue to pay the fine required for effective deterrence. Indeed,
many companies’ annual turnover exceeds their assets. A fine of 150 per cent of
annual turnover in the products concerned by the violation could thus only be
113
I would again add that the figure would be even higher if one were to take into account the fact
that fines are only paid several years after the gain from the violation is obtained: see above n 103. I
would also add that the figure is probably also too low because it ignores the likely effects of
overconfidence bias. Indeed, it appears that people consistently tend to overestimate the probability of
good things happening to them, and underestimate the probability of bad things happening to them,
in particular when the event is question is perceived to be controllable: see ND Weinstein, ‘Optimistic
Biases About Personal Risks’ (1989) 246 Science 1232 and RB Korobkin and TS Ulen, ‘Law and
Behavioral Science: Removing the Rationality Assumption from Law and Economics’ (2000) 88
California Law Review 1051 at 1091–5. Applied to deterrence through antitrust fines, this suggests that
companies and individual decision-makers within them will tend to overestimate the gain from the
antitrust violation and to underestimate the probability of being caught: see also ML Denger, ‘Too
Much or Too Little’, American Bar Association, Section of Antitrust Law, Antitrust Remedies Forum,
available at www.abanet.org/antitrust/remedies, at 5: ‘most cartel participants don’t think they will get
caught’; see also above, paras 195, 196 and 243.
114
Above n 101.
115
See, eg, on fines imposed by the European Commission, the ceiling of 10% of total turnover of
the undertaking concerned in Art 23(2) of Reg 1/2003, above n 1.
116
GJ Werden and MJ Simon, ‘Why Price Fixers should Go to Prison’ [Winter 1987] The Antitrust
Bulletin 917 at 928 n 35, referring to empirical estimates that unions are able to capture most of the
monopoly profits earned by US manufacturing firms.
paid out of the assets of large, diversified companies, which have taken part in
cartels only for a small part of their activities,117 or companies with very high
asset-to-sales ratios.118
557. If such high fines were really imposed, many of the companies concerned
would be forced into bankruptcy.119 This would, however, entail undesirable
social costs because, in the absence of perfect markets, it would hurt not only
managers and shareholders, on whom the bankruptcy may be considered to have
a desirable deterrent effect, but also all other stakeholders in the firm: employees,
suppliers, customers, creditors and tax authorities.120
558. Even if they stay below the level of inability to pay, the imposition of
such high fines is likely to be costly. Very high fines may have undesirable
side-effects. In the absence of perfect markets, very high fines imposed on
companies will have an incidence on all the stakeholders in the firm. Bondholders
and other creditors will suffer a diminution in the value of their securities.
Employees may suffer from cost-cutting campaigns induced by the need to pay
the fine. Tax receipts will be reduced. Finally, consumers may end up suffering.
Indeed, if the firm competes in a product market characterised by imperfect
competition (as will often be the case), the fine may be partly recovered from
consumers in the form of higher prices.121
6.3.1.2 Fines on Companies do not Always Guarantee Adequate Incentives for

Responsible Individuals within the Firm
559. As I also argued in more detail in my 2001 EUI paper,122 the logic of fining
undertakings or companies for antitrust violations is that the threat of the
corporate sanction, if sufficiently high, will induce the undertaking or company
to try to make its agents respect the law. This shifting of the enforcement
function from the authorities to the firm makes sense because in general the firm
has a relatively good ability to influence the behaviour of its agents. However, in a
number of circumstances the firm may not, in reality, be capable of adequately
controlling the behaviour of its agents, with the result that exclusive reliance on
117
When looking at the names of companies found to have been involved in cartels, one often sees
the same names coming back, suggesting that companies found to have been involved in a cartel in
one market may have been active in cartels in their other markets as well.
118
Werden and Simon, above n 116, at 928–9.
119
C Craycraft, JL Craycraft and JC Gallo, ‘Antitrust Sanctions and a Firm’s Ability to Pay’ (1997)
12 Review of Industrial Organization 171, in an empirical study based on a sample of 386 firms
convicted of price fixing in the US between 1955 and 1993, have estimated that 58% of the firms
would not have been able to survive the imposition of an optimal fine without becoming technically
bankrupt. Because of data restrictions, the firms in their sample were the larger firms, which means
that their result may very well underestimate the extent of the problem of inability to pay.
120
RH Kraakman, ‘Corporate Liability Strategies and the Costs of Legal Controls’ (1984) 93 Yale
Law Journal 857 at 882.
121
JC Coffee, ‘”No Soul to Damn: No Body to Kick”: an Unscandalized Inquiry into the Problem
of Corporate Punishment’ (1981) 79 Michigan Law Review 387 at 401–2.
122
Above n 101.
corporate sanctions will not lead to effective deterrence. This situation can arise
because firms may not be able to impose sufficiently serious sanctions on their
agents. It can also arise when firms are management controlled. Finally, the
managers responsible for the violation may have left the firm by the time the
violation is detected, and there may be little a company can do to a former
employee.
6.3.1.3 Increased Effectiveness of Leniency and Whistle Blowing Programmes
560. Sanctions against individuals have another beneficial effect, in that they
provide an incentive for individuals to offer cooperation in cartel investigations,
against the interests of their employers.123
561. As Scott Hammond has explained with regard to US Department of
Justice’s practice of combining corporate and individual leniency programmes:
the individual amnesty program helps prevent companies from covering up their
misconduct. The real value and measure of the Individual Leniency Program is not in
the number of individual applications we receive, but in the number of corporate
applications it generates. It works because it acts as a watchdog to ensure that
companies report the conduct themselves. … So long as one of its employees has
individual exposure, the company remains at great risk. If the company self-reports the
conduct under the Corporate Leniency Policy, then the company and all of its
cooperating executives will avoid criminal prosecution. However, if the company delays
or decides not to report, then the company puts itself in a race for leniency with its own
employees. In this example, if the company does not report the conduct first, then the
executive may come forward on his own and report the conduct for his own protection,
thereby potentially leaving the company out in the cold. … If the secretary gets
nervous, say, after talking to a relative who convinces her that she has real criminal
exposure for her own conduct, she may decide to report the conduct.124
562. As Donald Baker has further explained, also describing US practice:
The same process works between individuals at different companies involved in the
suspected conspiracy. There certainly have been cases where an individual was indicted
for a one-on-one conversation with someone who worked for a competitor and who
then turned the defendant in. Of course, as we saw with the Lysine cartel, an individual
who turns herself in to the government may become a government informant and
continue to participate in conspiracy meetings, armed with recording devices in order
to entrap the other conspiracy members.125
563. Donald Baker has also pointed out that the system of imposing criminal
liability on individuals generates not only incentives for whistle blowing based on
123
124
SD Hammond, ‘Cornerstones of an Effective Leniency Program’, paper presented at the ICN
Workshop on Leniency Programs (Sydney, 22–23 Nov 2004), available at www.usdoj.gov/atr/public/
speeches/206611.htm, at 12–14; see also Ch 5 above.
125
fear, but also incentives for whistle blowing based on the desire for revenge: ‘[t]he
desire for revenge is more picturesque, but it is still very much present. In my
experience, disgruntled current employees, fired employees, former trade associa-
tion officials, and even ex-spouses and ex-lovers may be anxious to finger the
individuals who they think have done them in’.126
564. As the OECD has pointed out, ‘There is an offsetting consideration
relating to individual sanctions, however. In many countries, sanctioning indi-
viduals requires criminalisation of the conduct. In that context individuals have
rights against self-incrimination, which makes it more difficult to obtain evi-
dence from them unless they willingly co-operate. In this sense, criminalisation
of cartel conduct for individuals makes voluntary co-operation more possible,
because of the threat of personal sanctions, but also makes it more necessary,
because a right against self-incrimination attaches’.127
565. The US experience,128 though, suggests that the net result is an increase
in the effectiveness of investigations. As will be discussed further in this chap-
ter,129 replication of the successful US experience in Europe would however
require that the competition authorities concerned can make similar use of
leniency instruments to that made by the US Department of Justice.
6.3.1.4 Imprisonment is a Very Effective Deterrent
566. There is ample evidence, based in particular on the US experience, that

imprisonment is a very effective deterrent for potential antitrust offenders.
567. In Arthur Liman’s words, ‘For the purse snatcher, a term in the peniten-
tiary may be little more unsettling than basic training in the army. To the
businessman, however, prison is the inferno, and conventional risk-reward analy-
sis breaks down when the risk is jail. The threat of imprisonment, therefore,
remains the most meaningful deterrent to antitrust violations.’130 Or as Joseph
Bauer has more recently put it, ‘[t]he sight of A. Alfred Taubman, the extremely
wealthy chairman of the board of Sotheby’s, the world-famous auction house,
convicted and sentenced, at the age of 78, to a one-year term of imprisonment
126
Ibid.
127
128
See Baker, above n 22, and Hammond, above n 124, as well as paragraphs 592 to 595 below.
129
Below, paras 592–595.
130
AL Liman, ‘The Paper Label Sentences: Critique’ (1977) 86 Yale Law Journal 619 at 630–1; in
the same sense see DI Baker and BA Reeves, ‘The Paper Label Sentences: Critique’ (1977) 86 Yale Law
Journal 619 at 621, and MB Mukasey, ‘Comment on Lynch’ (1997) 60 Law and Contemporary Problems
71; see also Antoine Winckler’s remarks in the panel discussion on criminal sanctions at the 6th
Competition Law and Policy Workshop at the European University Institute (Florence, 1–2 June
2001), published in C-D Ehlermann and I Atanasiu, European Competition Law Annual 2001: Effective
Private Enforcement of EC Antitrust Law (Oxford, Hart Publishing, 2003), 387 at 404: ‘[m]y experience
as a legal practitioner is that criminal sanctions are an extremely effective deterrent’.
and a substantial fine for participating in a price-fixing conspiracy, doubtless sent

a message to other business executives about the risks and penalties for this kind
of behavior’.131
568. The success of the current criminal antitrust enforcement in the US is
shown by several indicators. First, according to Terry Calvani, in the US ‘there is
much less domestic cartel activity today than before’.132 The majority of cartels
detected and prosecuted by the US Department of Justice in recent years have
been international, often even global in membership and geographic spread.133
569. Secondly, as Scott Hammond has pointed out:
We [the US Department of Justice] are observing firsthand in some of our investiga-
tions how the threat of criminal prosecution in the United States has deterred a
significant number of global cartels from extending their conspiracy into the United
States. We have uncovered cartels that operated profitably and illegally in Europe, Asia,
and elsewhere around the world, but did not expand their cartel activity to the United
States solely because it was not worth the risk of U.S. sanctions. I am referring to cartels
that had every opportunity to target U.S. consumers, because they sold in the U.S.
market. Indeed, in some cases, the U.S. market was the largest and potentially most
profitable but the collusive conduct still ceased at the border. Why? The answer, from
the mouth of the cartel members and verified by our investigators, is that the executives
did not want to get caught and going to jail in the United States.134
570. Thirdly, as William Kolasky has indicated with regard to those international
cartels that nevertheless continue to target also US consumers, ‘[t]he power of
criminal enforcement is illustrated by another characteristic common to those
cartels in which the conspirators know that their conduct is unlawful—namely,
their efforts to stay away from the United States. International cartels often try to
minimize their contacts in the United States by conducting their meetings
abroad. This has been particularly true since 1995, when the lysine investigation
became public. In fact, cooperating defendants in several recent cases have
revealed that the cartels changed their practices and began avoiding contacts in
the United States at all costs once the [US Department of Justice Antitrust]
Division began cracking and prosecuting international cartels. Some cartel
members go so far as to try to keep their cartel activity secret from all U.S.-based
employees, even those responsible for carrying out their instructions as to the
firm’s output and prices’.135
131
JP Bauer, ‘Reflections on the Manifold Means of Enforcing the Antitrust Laws: Too Much, Too
Little, or Just Right?’ (2004) 16 Loyola Consumer Law Review 303 at 307.
132
T Calvani, ‘Enforcement of Cartel Law in Ireland’ in BE Hawk (ed), 2003 Annual Proceedings of
the Fordham Corporate Law Institute International Antitrust Law & Policy (Huntington, NY, Juris
Publishing, 2004), 1 at 6.
133
See Connor, above n 33, at 10; and Hammond, above n 32, at 2.
134
Hammond, above n 124, at 8–9.
135
W Kolasky, ‘Criminalizing Cartel Activity: Lessons from the U.S. Experience’, fifth lunch talk of
the Global Competition Law Centre (Brussels, 29 Sept 2004), available at www.gclc.coleurop.be/
lunchtalk20040929.htm, at 11–12.
571. Fourthly, as Margaret Bloom has pointed out, ‘Another factor that
illustrates the effect of criminal powers is the extent to which international cartels
are first uncovered through amnesty applications to the [US Department of
Justice]. In these cases, it is the threat of criminal sanctions that drives executives
to go first to the [US Department of Justice] – and only later to the European
Commission and other relevant authorities. Without the U.S. criminal sanctions
and active enforcement record, how many of these cases would be revealed by
leniency applications to the European Commission? It is, of course, impossible to
answer this question, but I suspect the answer might well be few’.136
6.3.1.5 Imprisonment Carries a Uniquely Strong Moral Message
572. Corporate managers are not necessarily just maximisers of profits for
themselves and their principals. They may feel a moral responsibility to live
within the law whether or not they are likely to be caught, and this normative
commitment could trump their interest calculus.137 Indeed, psychological
research suggests that normative commitment is generally an important factor
explaining compliance with the law.138
573. The public punishment of those who violate the antitrust prohibitions
has thus not only a deterrent effect, in that it helps to create a credible threat of
punishment for those who would be willing to commit violations on the basis of
a profit calculation, but also has moral effects, in that it sends a message to the
spontaneously law-abiding, reinforcing their moral commitment to the rules.139
574. Imprisonment is by far the most expressive sanction. Prison sanctions
not only send a special message not conveyed by fines, they also send it more
effectively, as prison sentences for businesspeople are much more newsworthy
than fines and will thus get more publicity and be more noted by other
businesspeople.140
6.3.2 For Which Types of Antitrust Violations is Imprisonment Desirable?
575. The above argument that effective deterrence with only fines on companies
would require impossibly high fines, and that such fines should be supplemented
by imprisonment to achieve effective deterrence, is based on the paradigmatic
case of a price cartel.141 The conclusions reached only apply to those types of
136
M Bloom, ‘The Great Reformer: Mario Monti’s Legacy in Article 81 and Cartel Policy’ (2005) 1
Competition Policy International, Volume 1, No 1 (Spring 2005), 55.
137
138
139
See above n 6; see also paras 171 and 181.
140
Werden and Simon, above n 116, at 934; Lynch, above n 6, at 47 and 49; and DM Kahan, ‘Social
Influence’, above n 6, at 383–4; see also above, para 567.
141
violations of Articles 81 and 82 EC which are comparable to price cartels as

regards their profitability and ease of concealment. For those types of infringe-
ments which are difficult to hide, or which are anyway not very profitable for the
undertakings committing them, fines on companies, or fines on companies
combined with director disqualification,142 could be sufficient to achieve effective
deterrence.
576. Secondly, because imprisonment is a much more onerous punishment
than fines, the costs of imposing it erroneously are all the greater. These costs
include not only the suffering of those unjustly punished, but also the risk-
bearing cost to those risking unjustified punishment, the risk of lawful behaviour
being undesirably deterred, as well as possible negative effects on other people’s
normative commitment to the law.143 Prison sanctions should therefore only be
imposed in case of clear-cut violations. Alternatively one could conclude that
imprisonment should only lie for wilful violations of the law where there is
evidence that the individuals knew that they were violating the law or acted with
flagrant disregard for the law, with wilfulness being presumed in cases of
clear-cut violations.144
577. Applying these two sets of criteria, it would seem that imprisonment is
desirable for horizontal, naked price fixing, bid rigging and market allocation
schemes. It seems that all other types of antitrust violations, including vertical
restraints, other horizontal agreements and violations of Article 82 EC, are
usually easier to detect, as the victims of the violation will very often be aware of
it, and that the risk of undesirably chilling lawful behaviour is much more
substantial, as the borderline between anti-competitive and pro-competitive
behaviour is often less obvious. For those other types of violations, fines on
undertakings, or fines on undertakings combined with director disqualification,
may suffice.145
6.3.3 Are there no Equally Effective Alternatives to Imprisonment?
578. The above argument in favour of imprisonment consists in saying, on the

one hand, that hard-core cartels cannot be deterred only by fines on companies,
and, on the other hand, that imprisonment constitutes an effective deterrent. But
could effective deterrence not also be reached by adding other types of sanctions
or remedies to fines on companies?
142
See below, paras 582 and 583, on director disqualification.
143
See also Kraakman, above n 120, at 865 and 887, and RD Blair, ‘A Suggestion for Improved
Antitrust Enforcement’ [1985] The Antitrust Bulletin 433 at 444, n 32.
144
See Baker and Reeves, above n 130, at 623–4.
145
The UK would thus in my view appear to have struck the balance right, in providing for a
criminal cartel offence added to fines on undertakings and director disqualification for all types of
antitrust violations: see above, paras 527 and 539 and below, para 602.
6.3.3.1 Fines on Individuals
579. Fines on individuals would not appear to be an equally effective alternative

to imprisonment. The main reason146 is that companies can relatively easily
indemnify their agents for any threat of fines or any fines effectively imposed,
thus taking away the deterrent effect of the penalty on the individuals concerned.
Companies can relatively easily compensate their agents in advance for taking the
risk of being fined and/or indemnify them ex post when they have to pay the fine.
The crucial advantage of imprisonment is that it is impossible to shift the penalty
ex post, and also more difficult to arrange for a premium to compensate the risk
in advance.147
580. As Donald Baker has reported a senior corporate executive telling him,
‘as long as you are only talking about money, the company can at the end of the
day take care of me—but once you begin talking about taking away my liberty,
there is nothing that the company can do for me’.148
581. Fines on individuals may nevertheless have at least some stigmatising
effect. They could thus be considered as a third-best option, to be used when the
stronger options of imprisonment and director disqualification are unavail-
able.149
6.3.3.2 Director Disqualification
582. As already mentioned above, in the UK the Enterprise Act 2002 introduced
not only the criminal cartel offence, but also Competition Disqualification
Orders (CDOs).150 On application from the Office of Fair Trading, the competent
court must make a CDO against a director of a company (including a de facto
director) if that company commits a breach of competition law (Article 81 or 82
EC or the corresponding prohibitions in UK competition law) and the court
considers that person’s conduct as a director makes him or her unfit to be
concerned in the management of a company. The maximum period of disquali-
fication under a CDO is 15 years. During the period in which a person is subject
146
As explained in my 2001 EUI paper, above n 101, another reason is that, just as for fines on
companies, the level of fines required for deterrence risks being impossibly high.
147
See Werden and Simon, above n 116, at 931; Blair, above n 143, at 440–1; Kraakman, above n
120, at 876–9; and CD Stone, ‘The Place of Enterprise Liability in the Control of Corporate Conduct’
(1980) 90 Yale Law Journal 1 at 45–65.
148
149
Specifically with regard to the enforcement of Arts 81 and 82 EC at the level of the EU
institutions, fines on individuals would have an advantage in that they could certainly be introduced
under Art 83 EC or Art III-163 of the Constitutional Treaty, provided that they are ‘not of a criminal
law nature’ in the same way as the current fines on undertakings: see above, para 506. It would suffice
to amend Reg 1/2003: see paras 618–634 below, and also Wils, above n 4, sect 8.7.4.5.
150
Ss 9A to 9E of the Company Directors Disqualification Act 1986, as amended by the Enterprise
Act 2002; see also Office of Fair Trading, ‘Competition Disqualification Orders—Guidance’ (London,
OFT, May 2003), available at www.oft.gov.uk/.
to a CDO, it is a criminal offence for him or her to be a director of a company or,

in any way, whether directly or indirectly, to be concerned or take part in the
management of a company.
583. Director disqualification would not appear to be as strong a deterrent as
imprisonment. Depending on the age of the person concerned, director disquali-
fication may simply be the occasion to move into comfortable retirement, and the
company could fully compensate its former director for any financial loss. It can
also only be used for directors, not for middle management. On the other hand,
when it can be used, director disqualification is certainly more effective than fines
on individuals. If the person concerned is not ready for retirement, being barred
from taking part in the management of a company may hurt him or her in a way
which the company cannot fully compensate financially. Director disqualification
may also have a stigmatic effect and send a moral message, not as much as
imprisonment would,151 but certainly more than fines. While not being an
equally effective alternative to imprisonment, director disqualification would
thus appear to be a useful complement, as well as a defensible second-best.152
6.3.3.3 Private Actions for Damages
584. Could the problem of hard-core cartels not being deterred by mere fines on
companies be solved by adding private follow-on actions for damages? Here the
answer is in my view clearly no.153
585. As explained above,154 the main problem with fines on companies is that
the floor below which fines will generally not deter hard-core cartels would be in
the order of 150 per cent of the annual turnover in the products concerned by the
violation, and that fines of such a high level would often exceed the companies’
ability to pay. This problem exists just as much for damages as for fines, as well as
for any combination of the two. Adding private actions for damages against
companies to fines on companies also does not bring with it the above explained
advantage of adding individual penalties, namely the increased effectiveness of
leniency and whistle blowing programmes.
6.3.4 Five Conditions for Criminal Antitrust Enforcement to be Effective not

only in Theory but also in Practice
586. For criminal antitrust enforcement, and in particular imprisonment, to

work as an effective deterrent, it does not suffice to enact legislation creating a
151
152
See also above, para 577.
153
See also below, para 602.
154
At paras 547–558.
criminal offence. The effective operation of such legislation much also be

ensured. Indeed, deterrence will only work if there is a credible threat of detection
and punishment.
587. As explained below, for imprisonment to be an effective deterrent for
hard-core cartel behaviour, four conditions would need to be fulfilled: (1) a
dedicated investigator and prosecutor and (2) sufficient powers of investigation
must be provided for, and (3) judges or juries must be willing to convict, all of
which requires (4) a sufficiently broad political and public consensus that cartels
are really bad and thus deserve severe punishment.
588. Moreover, given that imprisonment can only work or only makes sense if
it is limited to hard-core cartels and complements other penalties for companies,
it remains important to provide for such other penalties as well as for penalties to
deter antitrust offences other than hard-core cartels.
6.3.4.1 The Need for a Dedicated Investigator and Prosecutor
589. For the threat of imprisonment to be credible, there must be a sufficiently

well resourced and sufficiently dedicated body, or bodies, in charge of investigat-
ing and prosecuting criminal antitrust offences.
590. In the US, the Department of Justice Antitrust Division has a National
Criminal Enforcement Section in Washington, DC, as well as seven field offices,
in Atlanta, Chicago, Cleveland, Dallas, New York, Philadelphia and San Francisco,
dedicated to investigating and prosecuting criminal cases.
591. The absence of a sufficiently well resourced and sufficiently dedicated
investigator and prosecutor may to a significant extent explain the ineffectiveness
of criminal enforcement in those EU Member States which have decriminalised
their enforcement in the last two decades.155 Indeed, if antitrust offences have to
compete with manslaughter and child abuse cases for the attention of a (typically
already over-stretched) generalist public prosecutor’s office, they are most
unlikely to get much priority.
6.3.4.2 The Need for Adequate Powers of Investigation
592. The success of US criminal antitrust enforcement in recent years is also the
result of the strong powers of investigation which the US Department of Justice
Antitrust Division has at its disposal.156
593. First, as Terry Calvani has stated, ‘the Division began using special agents
of the Federal Bureau of Investigations (“F.B.I.”) to assist in its cartel investiga-
tions. F.B.I. agents now routinely staff these investigations. Indeed, agents are
155
156
Apart from the 3 elements mentioned in paras 593 to 595, Donald Baker also mentions the
grand jury system, a secretive process favouring the investigators: see Baker, above n 22, at 693, 697
and 708.
actually seconded to the Division field offices where they work together with
Division attorneys in jointly investigating cartels. [T]hese F.B.I. agents have
brought with them the investigative techniques used to combat serious crime.
Antitrust investigators now routinely employ computer forensics, “wire” wit-
nesses, conduct “ambush” interviews, etc., in the course of investigations’.157
594. Secondly, as already explained above, the US Department of Justice has
both a Corporate Leniency Policy and an Individual Leniency Policy, under
which it can guarantee immunity from prosecution to cooperating companies or
individuals.158 It can also make precisely calculated promises of reduced penalties
to those cooperating at a later stage that cannot benefit from immunity.159 The
Corporate Leniency Policy also includes a so-called ‘Amnesty Plus’ option, under
which a company can get a substantial additional discount in the fine it will have
to pay for a first cartel being investigated (where the company was not the first
leniency applicant and hence does not get immunity from prosecution) if it
reveals a second cartel in another market for which it can get immunity.160 The
Department of Justice also increases the pressure by asking the so-called ‘omni-
bus question’ at the conclusion of a witness interview or grand jury interrogation.
The question goes something like this: ‘do you have any information whatsoever,
direct or indirect, relating to price fixing, bid rigging, market allocation, etc. with
respect to other products in this industry or in any other industry?’. The witness
must answer the question and must answer it fully and truthfully, or he or she
not only loses the leniency protection but is also subject to the penalties of
perjury or of making false statements or declarations.
595. Thirdly, as already mentioned above, in the US companies cannot invoke
the privilege against self-incrimination enshrined in the Fifth Amendment to the
US Constitution.161
157
Calvani, above n 132, at 7–8. It should be noted that the existence of a leniency programme
does not absolve one from the need to have other effective powers of investigation. Indeed, leniency
can only work if potential leniency applicants know that the competition authorities are also able to
detect cartels without cooperation from leniency applicants: see above, paras 401 and 408 and Wils,
above n 2, sect 5.2.3.
158
See above, paras 500, 561and 562, and 347–361.
159
See SD Hammond, ‘Measuring the Value of Second-In Cooperation in Corporate Plea
Negotiations’, address to the 54th Annual American Bar Association Section of Antitrust Law Spring
Meeting (Washington, DC, 29 Mar 2006), available at www.usdoj.gov/atr/public/speeches/
215514.pdf.
160
See GR Spratling, ‘Making Companies an Offer They Shouldn’t Refuse’, paper presented at the
Bar Association of the District of Columbia’s 35th Annual Symposium on Associations and Antitrust
(Washington, DC, 16 Feb 1999), available at www.usdoj.gov/atr/public/speeches/2247.pdf; and Ham-
mond, above n 32, at 10.
161
See above n 40.
6.3.4.3 Judges and Juries Must be Willing to Convict
596. For the threat of imprisonment to be credible, it is also essential that judges
and juries are willing to convict antitrust offenders and sentence them to
prison.162
597. In the US, this is the case today, thanks at least in part to the Sentencing
Guidelines.163
6.3.4.4 The Need for Political and Public Support
598. Underlying the three above conditions is a more fundamental fourth one,
namely that there needs to be a sufficiently broad political and public consensus
that cartels are really bad and thus deserve severe punishment.
599. In Donald Baker’s words, ‘When all is said and done, criminal liability
ought to reflect the culture and values of the people of a country. The United
States has a long tradition of fairly broad and sometimes noisy support of
antitrust law and enforcement. After all, this is what created the Sherman Act in
1890 and the felony statute in 1974. Robber barons and cartel conspirators are
bad people in the American popular lexicon.’164
600. As is well known, Europe does not have the same history.165 We are
however not condemned to remaining stuck in our past. As John Coffee has
pointed out, ‘the limited empirical evidence on public attitudes toward white-
collar crimes suggests that the public learns what is criminal from what is
punished, not vice versa’.166 Leadership, such as that being shown by the Irish and
UK governments and competition authorities in criminalising cartel enforce-
ment, may thus work.167
162
See Kunzlik, above n 92, at 345–6 and Green, above n 21, at 26–8.
163
Baker, above n 22, at 694. It was not always the case in earlier times: see Baker and Reeves,
above n 130. See also above n 159.
164
165
See above, para 543; Wils, above n 4, sects 8.6.7 and 9.5; and C Harding and J Joshua,
Regulating Cartels in Europe (Oxford, Oxford University Press, 2003).
166
Coffee, above n 7, at 1889.
167
See, inter alia, the speech by John Vickers, Chairman of the Board of the UK Office of Fair
Trading at the British Chambers of Commerce on 31 Mar 2003, ‘Policy for Markets and Enterprise’,
available at www.oft.gov.uk/NR/rdonlyres/62E7FFA8–3803–440C-8119–33FF119C0967/0/
spe0203.pdf, at 4: ‘[s]ince hard core cartels are like theft, criminalisation makes the punishment fit
what is indeed a crime’.
6.3.4.5 Fines on Companies and Public Penalties for Antitrust Violations Other
than Hard-core Cartels Remain Necessary
601. Imprisonment as a penalty for hard-core cartels should only be used in

combination with fines on the companies concerned.168 Indeed, if firms were not
responsible for the antitrust violations committed by their agents, serious incen-
tive problems would result. Given that the firm normally benefits from the
antitrust violations committed by its agents, it would have an incentive to
encourage violations. Even if they risk individual penalties, the firm’s agents may
be willing to take the risk if they expect to be sufficiently rewarded in terms of
salary and promotion (or avoidance of demotion or dismissal). Assigning liability
to the firm tends to remove the incentive for both the firm and its agents to
commit antitrust violations on behalf of the firm.169 Firms are instead encour-
aged to strive for clarity in the messages conveyed to their agents as to the
behaviour they expect from them.170
602. As argued above,171 imprisonment appears desirable only for horizontal,
naked price fixing, bid rigging and market allocation schemes. However, all other
types of antitrust violations, including vertical restraints, other horizontal agree-
ments and violations of Article 82 EC, also need to be deterred. As already
mentioned above,172 in the US, apart from the possibility for the Federal Trade
Commission to seek disgorgement, deterrence of antitrust violations other than
hard-core cartels is left to private treble-damages actions. As appears to be
recognised also in the US,173 this could be considered problematic. Indeed, as I
have argued in detail elsewhere, private antitrust enforcement is inherently
inferior to public enforcement.174 The United Kingdom, with its combination of
168
See also above, paras 560–565 as to the increased effectiveness of leniency and whistle-blowing
programmes resulting from the combination of corporate and individual liability, and my 2001 EUI
paper, above n 101, for other reasons why imprisonment should only be used in combination with
corporate fines.
169
TS Ulen, ‘The Economic Case for Corporate Criminal Sanctioning’ in WS Lofquist, MA Cohen
and GA Rabe (eds), Debating Corporate Crime (Cincinnati, Anderson, 1997), 117 at 133, and Coffee,
above n 121, at 410.
170
DA DeMott, ‘Organizational Incentives to Care About the Law’ (1997) 60 Law and Contempo-
rary Problems 39 at 47.
171
Paras 575–577.
172
Paras 497 and 498.
173
See letter from the Assistant Attorney General, Department of Justice Antitrust Division, above
n 35, at 2 (‘Our system’s extreme weighting of the civil enforcement balance toward private lawsuits as
opposed to federal government enforcement is not necessarily ideal’).
174
See WPJ Wils, ‘Should Private Antitrust Enforcement Be Encouraged in Europe?’ (2003) 26
World Competition 473, and Wils, above n 2, ch 4; see also the leading article ‘MS.dosh settlement—
Microsoft’s pay-offs show why regulators are necessary’, Financial Times, 26 Nov 2004, 12; J Kay,
‘Antitrust concerns are not so easily dismissed’, Financial Times, 30 Nov 2004, 13; and B Groom,
European Comment, ‘Microsoft’s Brussels trouble not over yet’, Financial Times, 10 Nov 2004, 16: ‘[i]t
is a comfort that competition policy in Europe is pursued largely by public authorities rather than
through private lawsuits. Rival companies can be bought off, but the [European] Commission’s role is
to safeguard consumers.’
CRIMINALISATION AT MEMBER STATE LEVEL ONLY 193
a criminal cartel offence with a system of administrative fines and director

disqualification for all types of antitrust violations, would in my view appear to
have struck the balance exactly right.
6.4 IS IT PROBLEMATIC THAT ANTIT R UST ENFORCEMENT IS CR IMINALISED AT THE

LE VEL OF INDIV IDUAL EU MEMBER STATES W ITHOUT PAR ALLEL CR IMINALISATION
AT THE LE VEL OF THE EU INSTITUTIONS, OR W ITHOUT EU HAR MONISATION?
6.4.1 Regulation 1/2003 has been Designed to Accommodate Criminalisation

in Individual Member States
603. The fact that individual Member States, such as Ireland, the United King-
dom and Estonia, have criminalised their enforcement of Articles 81 and 82 EC,
without parallel criminalisation at the level of the EU institutions and without
EU harmonisation, is not problematic. Indeed, Regulation 1/2003 has precisely
been designed to accommodate criminalisation in individual Member States.
604. As already explained above,175 Article 5 of Regulation 1/2003 leaves each
Member State the choice whether or not to provide for criminal penalties to
enforce Articles 81 and 82 EC.
605. Article 12(3) of Regulation 1/2003, also already mentioned above,176
ensures that criminal enforcement in some Member States can coexist with
non-criminal enforcement in other Member States. Indeed, whereas Article 12(1)
of Regulation 1/2003 provides that the European Commission and the competi-
tion authorities of the Member States can exchange all kinds of information, and
use the information thus exchanged in evidence, Article 12(3) adds the following
limitation:
to an infringement of [Article 81 or 82 EC] or, in the absence thereof,
cannot be used by the receiving authority to impose custodial sanctions.
606. Without this limitation, the free circulation of information between the
competition authorities of all the Member States as well as the European
Commission would have caused problems both in the Member States which
provide for individual penalties, in particular imprisonment, and in the other
Member States as well as at the level of the EU institutions. Indeed, if in Ireland
175
At paras 515–518.
176
Ibid.
or the United Kingdom individuals could have been sent to prison on the basis of
evidence collected by the European Commission or for instance the Italian
competition authority in its administrative proceedings leading only to fines on
companies, the rights of defence of the individuals concerned would have been
weakened. Indeed, the Irish and UK courts would probably have refused the use
of such evidence. On the other hand, the risk of evidence collected by the
European Commission or the Italian competition authority in its administrative
proceedings leading only to fines on companies being subsequently used in
Ireland or the United Kingdom to send individuals to prison could have led the
EC and Italian courts or legislators to raise the level of the rights of the defence in
proceedings by the European Commission or the Italian competition authority to
the same level at that applying to criminal investigations in Ireland or the United
Kingdom. This could have seriously harmed the effectiveness of the investiga-
tions by the European Commission or the Italian competition authority.177
607. All these problems have been avoided as a result of Article 12(3), whereas
this provision does not create any obstacle to the competition authorities of
Ireland, the United Kingdom, Estonia and any other Member State which may
follow their example, exchanging all information between them and using it in
evidence to impose prison sanctions on individuals.
608. Article 12(3) of Regulation 1/2003 prevents information collected in a
jurisdiction that does not provide for imprisonment to be used in evidence in
another jurisdiction to seek imprisonment. It does not, however, prevent the
intelligence being transmitted. The UK competition authorities could thus for
instance, after having received information from the European Commission or
the Italian competition authority, start an investigation of their own themselves
to collect the same or related information and subsequently use it in evidence to
impose criminal sanctions on an individual. This possibility could have as a result
that individuals, as well as the undertakings wanting to protect them, may be less
willing to cooperate with the investigations of the European Commission or the
Italian competition authority. To the extent that cooperation under leniency
programmes is concerned, this problem has been anticipated and addressed in
the European Commission’s Notice on cooperation within the network of
competition authorities, which provides that information provided under a
leniency programme to the European Commission or a national competition
authority will not be transmitted to another competition authority unless the
latter commits itself not to use this information, nor any information obtained
following that information, to impose sanctions on the leniency applicant or on
any of its employees or former employees.178
177
See further Wils, above n 2, sect 1.2.12.
178
Commission Notice on cooperation within the Network of Competition Authorities [2004] OJ
C101/43, paras 37 to 42; see also Schnichels, above n 70, at 114–15, and Wils, above n 2, sect 1.2.14.
CRIMINALISATION AT MEMBER STATE LEVEL ONLY 195
6.4.2 What Would be the Advantages of Criminalising the Enforcement of EU

Antitrust Law in all Member States, through EU Harmonisation, as well as at
the Level of the EU Institutions?
609. The fact that Regulation 1/2003 has been designed to accommodate crimi-
nalisation in individual Member States, without EU harmonisation, and without
criminalisation of antitrust enforcement at the level of the EU institutions, with
the result that criminalisation in individual Member States is not problematic,
does not rule out that there may be advantages in criminalising the enforcement
of EU antitrust law in all Member States, through EU harmonisation, as well as at
the level of the EU institutions.
610. I would personally expect that criminal cartel enforcement for instance
in the United Kingdom will prove to be effective, even if enforcement in most
other Member States and at the level of the EU institutions continues to rely only
on fines on companies, in the same way as US criminal antitrust enforcement has
in recent years been highly successful,179 even if most of the rest of the world does
not have similar enforcement. I would however see three or four advantages in
criminalising the enforcement of EU antitrust law in all Member States, through
EU harmonisation, as well as at the level of the EU institutions.
611. First, as explained above,180 Article 12(3) of Regulation 1/2003 provides,
for very good reasons, that the competition authorities in Member States such as
the United Kingdom can only use information exchanged in the network of EU
competition authorities in evidence in order to impose prison sanctions on an
individual if the information has been collected by the competition authority of a
jurisdiction which equally provides for imprisonment. If all EU Member States
were to follow the UK example in criminalising their enforcement of hard-core
cartels, either of their own motion or as a result of EU harmonisation, and if
enforcement at the level of the EU institutions could be similarly criminalised,
there would be an obvious advantage, in that more competition authorities could
work more effectively together in investigating criminal cartel offences.
612. Secondly, more widespread or harmonised criminalisation throughout
the EU would enhance the effectiveness of sanctions, allowing for instance
individuals suspected or convicted of a criminal cartel offence to be extradited
between Member States, or the effect of director disqualifications to be EU-wide.
613. Thirdly, as explained above,181 Article 12(3) of Regulation 1/2003 pre-
vents information collected in a jurisdiction that does not provide for imprison-
ment being used in evidence in another jurisdiction to seek imprisonment, and
the European Commission’s Notice on cooperation within the network of
competition authorities provides that information provided under a leniency
programme to the European Commission or a national competition authority
179
180
Paras 605–608.
181
Ibid.
will not be transmitted to another competition authority unless the latter

commits itself not to use this information, nor any information obtained
following that information, to impose sanctions on the leniency applicant or on
any of its employees or former employees.
614. The risk of imprisonment in some Member States and the exchange of
intelligence within the network of competition authorities may however still have
an impact on the effectiveness of investigations by the competition authorities in
the other Member States or by the European Commission in the absence of
leniency applications.182 Indeed, the risk of imprisonment may lead cartel
participants to make even greater efforts to hide the evidence of their behaviour,
and the possibility of intelligence being transmitted to the competition authori-
ties in, for example, the United Kingdom may reduce the willingness of individu-
als, and of the undertakings wanting to protect them, to cooperate outside the
framework of leniency with investigations by the European Commission or, for
example, the Italian competition authority. Individuals may for instance be less
forthcoming in answering on-the-spot questions during inspections,183 or com-
panies may be more evasive in their answers to written requests for information.
615. Fourthly, criminalisation in some EU Member States, and not in others,
nor at the level of the EU institutions, could result in those Member States that
have criminal enforcement playing a disproportionate role, or carrying a dispro-
portionate share of the burden, in the overall task of enforcing EU competition
law, in the same way as the US in recent years has played a disproportionate role
in deterring and punishing world-wide cartels.184
616. Specifically with regard to the role of the European Commission in the
overall task of enforcing EU competition law, Peter Kunzlik has made the
following conjecture of what might happen if criminal enforcement in Ireland
and the United Kingdom were successful, and other EU Member States followed
their example:
the situation might well arise in which cartels will be uncovered and prosecuted more
efficiently at the national level than at the European level. In such a scenario the
European Commission’s view that its proper vocation is primarily in the enforcement
of Community competition law against hard-core cartels—a central premise of its own
182
The ability of competition authorities to detect cartels in the absence of leniency applications
is also important for the effectiveness of leniency programmes. Indeed, if companies were to believe
that the competition authorities would not be able to detect a cartel on their own, they would have no
reason to apply for leniency either: see above n 157, and above, paras 401, 408 and 419–421.
183
See above, para 512 on such questioning under Art 20(2)(e) of Reg 1/2003; see also JS Venit
and T Louko, ‘The Commission’s New Power to Question and its Implications on Human Rights’,
paper presented at the 31st Annual Fordham Corporate Law Institute Conference on International
Antitrust Law & Policy (New York, 8 Oct 2004), as well as remarks by several participants in
Roundtable Debate Session IV—Rights, Privileges and Ethics in Competition Cases at the same
conference, all forthcoming in BE Hawk (ed), 2004 Annual Proceedings of the Fordham Corporate Law
Institute—International Antitrust Law and Policy (Huntington, NY, Juris Publishing, 2005).
184
See Wils, above n 4, sect 8.7.3.3.
LEGAL POSSIBILITY OF EU CRIMINALISATION 197
modernization proprosals—may need to be revisited. By the time that the Commis-

sion’s investigations will have cranked into gear, the national systems may already have
taken the initiative. The Commission may in practice lose its central role in European
cartel enforcement to quicker, more effective, and better-resourced national competi-
tion authorities operating under [systems such as the ones in Ireland and the United
Kingdom].185
617. In Margaret Bloom’s words:
If criminal sanctions do provide far more effective deterrence, why has Commissioner
Monti apparently not considered introducing such sanctions as part of the fight against
cartels? Should such powers be considered now? One obvious answer is the fact that
only six out of the 25 Member States have such powers in their national law and these
have not yet been used on their own for any custodial sentences. Also, the European
Commission does not currently have criminal powers in any area. At this time, it seems
unlikely that there would be the necessary support within the European Union for such
a change. But if the United Kingdom or another Member State demonstrates effective
use of its powers and increases deterrence significantly as a result, the Commission
should consider criminal powers seriously.186
6.5 WOULD IT BE LEGALLY POSSIBLE TO CR IMINALISE ANTIT R UST ENFORCEMENT AT

THE LE VEL OF THE EU INSTITUTIONS, OR TO CR IMINALISE ANTIT R UST
ENFORCEMENT IN ALL EU MEMBER STATES THROUG H EU HAR MONISATION?
6.5.1 Under the Current EC Treaty

618. As I have argued in more detail in my 2001 EUI paper,187 under the current
EC Treaty it would be legally possible to criminalise antitrust enforcement at the
level of the EU institutions, and to provide for a system in which the Court of
First Instance, upon investigation and prosecution by the European Commission,
would impose prison sanctions on individuals for hard-core violations of Article
81 EC.188
619. Indeed, as is apparent from Article 47 EU (‘nothing in this Treaty shall
affect the Treaties establishing the European Communities’) and from the very
first words of Article 29 EU (‘without prejudice to the powers of the European
185
Kunzlik, above n 92, at 352–3; see also ibid, at 347–48: ‘[i]f the approach adopted in [Ireland
and the UK] were to be emulated by other member states then transatlantic cartel enforcement
cooperation might become substantially refocused upon the national rather than the European level.
This would, in turn, affect the prestige of the European Commission within the European system. It
might potentially take back from the Commission, and return to the authorities of the member states
concerned, the de facto role as primary collaborators with the U.S. in terms of an important aspect of
transatlantic cartel enforcement cooperation’.
186
Bloom, above n 136.
187
Above n 101.
188
The same conclusion was reached by Manfred Zuleeg in his paper for the same EUI Workshop,
‘Criminal Sanctions to Be Imposed on Individuals As Enforcement Instruments in European
Competition Law’, published in C-D Ehlermann and I Atanasiu, European Competition Law Annual
2001: Effective Private Enforcement of EC Antitrust Law (Oxford, Hart Publishing, 2003) 453.
Community’), the existence of the EU Treaty, with its Title VI on police and
judicial cooperation in criminal matters, does not affect the ability or inability to
introduce under the EC Treaty criminal sanctions to enforce EC competition law.
Apart from two specific provisions in Articles 135 and 280 EC, concerning
customs cooperation and fraud affecting the financial interests of the European
Community, the EC Treaty does not contain any provision excluding criminal
law from its scope. As already mentioned above,189 the use of the term ‘undertak-
ings’ in Article 81 EC leaves open the question what sanctions, on undertakings
and/or individuals, are to be provided for. This choice is to be made by the
Council, when using its broad legislative powers under Article 83(1) EC to lay
down ‘the appropriate regulations or directives to give effect to the principles set
out in Articles 81 and 82’.
620. More problematic is the explicit reference to fines in Article 83(2)(a) EC,
which provides that the regulations or directives referred to in Article 83(1) ‘shall
be designed in particular: (a) to ensure compliance with the prohibition laid
down in Article 81(1) and in Article 82 by making provision for fines and
periodic penalty payments’.
621. It is clear that the authors of the Treaty thought that compliance with the
antitrust prohibitions would be induced by the threat of financial penalties, not
imprisonment. However, one could argue on the basis of a teleological interpre-
tation that, if the Council were to come to the view that prison sanctions were
necessary for effective deterrence, the reference to fines in Article 83(2)(a) EC
would not prevent it from adding such sanctions, given the wide mandate in
Article 83(1) EC for laying down any ‘appropriate regulations or directives to give
effect to the principles set out in Articles 81 and 82’.
622. In any event, even if one were to conclude that Article 83 does not
provide a sufficient legal basis for introducing prison sanctions, the solution
would be to use Article 308 EC, which empowers the Council, acting unani-
mously on a proposal from the Commission and after consulting the European
Parliament, to take the appropriate measures whenever action by the Community
proves necessary to attain, in the course of the operation of the common market,
one of the objectives of the Community and the Treaty has not provided the
necessary powers.
623. Of course, the introduction of prison sanctions would require the
separation of the powers of investigation and decision currently held in common
by the Commission, and the transfer of the latter power to an independent
judicial body, such as the Court of First Instance. The Commission would then
act as a prosecutor before the Court of First Instance.190 Such transfer of the
decisional power from the Commission to the Court of First Instance would be
possible under the existing EC Treaty. Indeed, as already mentioned above,
189
Above, para 520.
190
This requirement of separation flows from both Arts 5 and 6 ECHR: see above, paras 488–491.
See also above n 9.
Article 83(1) EC gives an wide mandate to the Council, acting by a qualified

majority on a proposal from the Commission and after consulting the European
Parliament, to lay down ‘the appropriate regulations or directives to give effect to
the principles set out in Articles 81 and 82’. Article 83(2)(d) adds that the
Council’s regulations and directives ‘shall be designed in particular: (d) to define
the respective functions of the Commission and of the Court of Justice applying
the provisions laid down [under Article 83(1)]’. Moreover, Article 229 EC
provides that regulations adopted by the Council may give the Court of Justice
‘unlimited jurisdiction with regard to the penalties provided for in such regula-
tions’. That such ‘unlimited jurisdiction’ can take the form not only of an
unlimited review of a Commission decision, but also of imposition of penalties
directly by the EC Courts, is apparent from the German language version of
Article 229 EC which talks about giving the Court a competence ‘which includes
the power of unlimited control of judgment and the power to modify or impose
such penalties’.191
624. The above reasoning to the effect that Article 83 EC (if necessary
combined with Article 308 EC) would provide a sufficient legal basis for
criminalising antitrust enforcement at the level of the EU institutions can also be
transposed to criminalisation of antitrust enforcement in all Member States
through EU harmonisation, which thus appears equally possible. This conclusion
also finds support in the recent judgment of the Court of Justice confirming EC
competence as regards protection of the environment through criminal law.192
6.5.2 Under the Signed but Unratified Constitutional Treaty
625. It would appear doubtful whether the same ability to criminalise antitrust
enforcement at the level of the EU institutions would still exist under the signed
but unratified Constitutional Treaty.193
191
Emphasis added; my translation of ‘eine Zuständigkeit … welcher die Befugnis zu unbeschränk-
ter Ermessensnachprüfung und zur Änderung oder Verhängung solcher Maβnahmen umfaβt’. That the
Commission would prosecute undertakings before the Court appears also to have been the under-
standing of the Spaak Report preceding the adoption of the EEC Treaty: see Comité intergou-
vernemental créé par la Conférence de Messine, Rapport des Chefs de Délégation aux Ministres des
Affaires Etrangères, Mae 120f/56 corrigé (Brussels, 21 Apr 1956), at 56: ‘[p]our l’application concrète, il
conviendra d’établir une procédure qui évite autant que possible une multiplication de procès devant la
Cour. A cette fin, la Commission européenne constituerait un comité consultatif des ententes et
discriminations qui l’aiderait dans une tâche de conciliation et d’arbitrage. A défaut d’une solution
acceptée par les parties à l’expiration d’un délai qui pourrait être de deux mois, les Etats ou la
Commission européenne elle-même pourraient introduire une plainte devant la Cour. Dans tous les cas,
l’affaire sera instruite par la Commission européenne.’
192
Judgment of the ECJ of 13 Sept 2005 in Case C–176/03 European Commission v EU Council
[2005] ECR I–7879; see also Communication from the Commission to the European Parliament and
the Council on the implications of the Court’s judgment of 13 Sept 2005 (Case C–176/03 Commis-
sion v Council), COM(2005)583final/2 of 24 Sept 2005.
193
Treaty establishing a Constitution for Europe [2004] OJ C310/1.
626. The crucial difference from the current situation is that the Constitu-
tional Treaty integrates the areas currently covered by the EC Treaty and by the
other Titles of the EU Treaty into a single Treaty, without any hierarchy of the
former over the latter. It can thus no longer be said that the presence of
provisions on judicial cooperation in criminal matters (in Articles III-270 to
III-274 of the Constitutional Treaty) does not affect the scope of the provisions
on competition, and in particular of Article III-163 of the Constitutional Treaty,
which is the (for all practical purposes identical) successor to Article 83 EC.
627. Article III-271(2) of the Constitutional Treaty reads as follows:
If the approximation of criminal laws and regulations of the Member States proves
essential to ensure the effective implementation of a Union policy in a area which has
been subject to harmonisation measures, European framework laws may establish
minimum rules with regard to the definition of criminal offences and sanctions in the
area concerned.
628. Article III-271(1) further provides for the possibility for European frame-
work laws to ‘establish minimum rules concerning the definition of criminal
offences and sanctions in the areas of particularly serious crime with a cross-
border dimension resulting from the nature or impact of such offences or from a
special need to combat them on a common basis’.194
629. European framework laws, which correspond to directives under the
current EC Treaty, are, according to Article I-33(1) of the Constitutional Treaty,
legislative acts ‘binding, as to the result to be achieved, upon each Member State
to which it is addressed, but shall leave to the national authorities the choice of
form and methods’.
630. Whereas Article III-271 of the Constitutional Treaty could undoubtedly
be used to criminalise the enforcement of EU antitrust law in all the EU Member
States, the fact that this Article only allows the use of European framework laws
(and not European laws, the equivalent of regulations under the current EC
Treaty) appears to close the door to criminalisation of EU antitrust enforcement
at the level of the EU institutions.
631. Given that Article III-271 is the provision of the Constitutional Treaty
specifically providing a legal basis for criminalisation, it appears doubtful
whether this limitation to the instrument of European framework laws could be
overcome by using Article III-163 instead.
632. A possible way forward could be provided by Article III-274 of the
Constitutional Treaty, which provides for the possibility of establishing a Euro-
pean Public Prosecutor’s Office, and giving this new body the task of investigat-
ing and prosecuting, in the competent courts of the Member States, the
194
Cartels are not among the areas of crime listed in the second subpara of Art III-271(1), but
they could be added through use of the third subpara.
perpetrators of ‘serious crime having a cross-border dimension’. This provision

would however not allow the European Commission to act as prosecutor, or the
EU Courts to act as competent courts.
633. As already mentioned just above, criminalisation of antitrust enforce-
ment in all Member States through EU harmonisation, though limited to the
establishment of ‘minimum rules’, would clearly appear to be possible under
Article III-271 of the Constitutional Treaty.
634. In the end, of course, the main issue is whether the European Commis-
sion and the EU Member States would want to criminalise the enforcement of
EU competition law at the level of the EU institutions and, through harmonisa-
tion, in all Member States. The day there is a consensus that this outcome would
be desirable, it will also be possible to make whatever Treaty changes may be
needed.
Index
A possibility under the EC Treaty:
Access to the file: paragraphs 105, 108 to paragraphs 514 to 532, 618 to 634
116 powers of investigation: paragraphs 32
Amnesty: see Leniency to 34, 592 to 595
Availability bias: paragraphs 175, 195, 196 procedural rights and guarantees: 62,
501, 564, 565, 623
B types of infringement for which
Bias: desirable: paragraphs 575 to 577
availability bias: paragraphs 175, 195,
US example: paragraphs 492 to 501
196
wider notion of ‘criminal’ in ECHR:
overconfidence bias: paragraphs 197,
paragraphs 484 to 491
198, 243
Bounties: paragraphs 464 to 470 D
Damages: see Compensation, Corrective
C justice
Charter of Fundamental Rights of the Deterrence: paragraphs 172, 181, 184 to
European Union: paragraphs 56, 112, 207, 223 to 228, 243, 283 to 290, 299,
208 300, 324 to 327, 334 to 337, 393 to
Commitments: see Settlements 408, 415 to 425, 443 to 450, 464 to
Community Courts: 470
judicial review of settlements: Director disqualification: paragraphs 539,
paragraphs 151 to 153 582, 583
role as to procedural rights and
guarantees: paragraphs 51 to 60, 64 E
to 66, 75 Econometrics (use in setting fines):
Compensation: paragraphs 178, 179, 182, paragraphs 237 to 243, 285 to 287
348, 349, 373, 413 to 454, 584, 585 Effectiveness (principle of): paragraphs 22,
Competition authorities of the Member 38 to 40
States: see National competition Equality (principle of): paragraphs 233 to
authorities 236, 429, 430
Compliance programmes: paragraphs 218 Equivalence (principle of): paragraphs 22,
to 221, 301 (footnote 119) 37, 523, 529
European Commission:
Corrective justice: paragraphs 178, 179,
commitments: paragraphs 67 to 172
182, 261, 348, 349, 373, 413, 414
Court of Justice: see Community Courts cooperation with national competition
Criminalisation: paragraphs 471 to 634 authorities: 19 to 27, 43, 119
benefits: paragraphs 547 to 574, 609 to criminal powers: paragraphs 484 to 491,
617 502 to 513, 609 to 634
conditions for effectiveness: 586 to 602 guidelines on fines: paragraphs 244 to
definition: paragraphs 472 to 491 343
history: paragraphs 493 to 496, 534 to discretion as to fines: paragraphs 266 to
543, 598 to 600 274
leniency: paragraphs 443 to 450, 560 to enforcement priorities: paragraphs 92 to
565 103
national competition authorities: leniency policy: paragraphs 328 to 330,
paragraphs 514 to 543, 603 to 608 340, 362 to 377
Columns Design Ltd / Job: Wils_Antitrust / Division: Wils_Index /Pg. Position: 1 / Date: 14/1
204 INDEX
powers of investigation: paragraphs 5 to moral effect: paragraphs 171, 181, 189,

27, 44 to 60 192, 208 to 212, 233 to 236, 260,
settlements: paragraphs 67 to 172 426 to 430, 572 to 574
European Competition Network (ECN): optimal fines: paragraphs 163 to 243
assistance in investigations: paragraphs procedural fines: paragraphs 9, 10, 13,
12, 15, 19 to 26, 41 to 43, 67 to 75 14, 18
exchange of information: paragraphs proportionality: paragraphs 208 to 212,
27, 41, 43, 67 to 75, 603 to 614 233 to 236
generally: paragraphs 1, 297 purpose: paragraphs 167 to 182, 255 to
261
leniency: paragraph 387
reduction because of compliance
see also National competition authorities
programme: paragraphs 218 to
European Convention on Human Rights: 221, 302 (footnote 119)
paragraphs 53 to 56, 63, 75, 262
reduction because of cooperation:
(footnote 29), 484 to 491, 623
(footnote 190) paragraphs 214 to 217, 227 to 228,
European Court of Human Rights: 315 to 318, 328 to 330, 340; see
paragraphs 53 to 56, 63, 75 also Leniency
European Court of Justice: see Community settlements involving fines: paragraphs
Courts 159 to 162, 377, 354 to 360, 409 to
Exchange of information: paragraphs 27, 412
41, 43, 67 to 75, 603 to 614 use of econometrics to determine the
amount: paragraphs 237 to 243,
F 285 to 287
Financial rewards: paragraphs 464 to 470 Fundamental rights:
Fines: generally: paragraphs 44 to 75
ability to pay: paragraphs 207 to 212, legal professional privilege: paragraphs
229 to 236, 336, 338, 339, 341 to 45, 46, 52, 60, 65, 66, 73 to 75
343 rights of the defence: paragraphs 44 to
collection of fines: paragraph 23 75
deterrent effect: paragraphs 172, 181, self-incrimination (privilege against):
184 to 207, 223 to 228, 283 to 290, paragraphs 44 to 46, 52, 55, 57 to
324 to 327, 334 to 337, 393 to 408, 59
415 to 425, 547 to 571 see also Charter of Fundamental Rights
of the European Union
discretion: paragraphs 266 to 274
European Convention on Human Rights
equal treatment: paragraphs 233 to 236
General principles of Community law
European Commission guidelines:
paragraphs 244 to 343 G
fines for failure to comply with General principles of Community law:
commitments: paragraphs 87, 132 effectiveness: paragraphs 22, 38 to 40
fines on individuals: paragraphs 579 to equality: paragraphs 233 to 236, 429,
581 430
foreseeability: paragraphs 272 to 274 equivalence: paragraphs 22, 37, 523, 529
increase for repeated infringement: proportionality: paragraphs 208 to 212,
paragraphs 293 to 303 233 to 236
limits to high fines: paragraphs 203 to rights of the defence: paragraphs 44 to
212, 229 to 236, 338, 339, 341 to 75
343, 547 to 558 see also Fundamental rights
INDEX 205
Goals: see also General principles of Community

of antitrust enforcement: paragraphs law
93, 167 to 182, 179, 255 to 261 Fundamental rights
of antitrust law: paragraphs 179, 187, Moral effects of penalties
188
Guidelines: L
European Commission guidelines on Legal professional privilege: paragraphs
fines: paragraphs 244 to 343 45, 46, 52, 60, 65, 66, 73 to 75
purpose of guidelines: paragraphs 262 Leniency: paragraphs 214 to 217, 227, 228,
to 274 328 to 330, 340, 344 to 470
amnesty plus: paragraphs 361, 379, 458
to 463
H
History: beneficial effects: paragraphs 393 to
criminalisation: paragraphs 493 to 496, 414, 431 to 434, 443 to 450
534 to 543, 598 to 600 bounties: paragraphs 464 to 470
leniency: paragraphs 381 to 387 compared to other methods of
settlements: paragraphs 82 to 90 investigation: paragraphs 1 to 4,
Human rights: see Fundamental rights 393 to 403
compensation to injured parties:
I paragraphs 348, 349, 373, 413 to
Immunity: see Leniency 414, 451 to 454criminalisation:
Imprisonment: see Criminalisation paragraphs 443 to 450, 560 to 565
Informants: paragraphs 2, 395, 464, 465, equal treatment: paragraph 429
470, 560 to 565 European Commission: paragraphs 328
Inspections: paragraphs 11 to 25 to 330, 340, 362 to 377, 383, 384
see also Investigations exclusion of ringleaders: paragraphs
Investigations: 348, 350, 359, 365, 379, 425, 430
assistance by other authorities: history: paragraphs 381 to 387
paragraphs 12, 15, 19 to 26, 41 to measuring its effects: paragraphs 431 to
43 434
different methods: paragraphs 1 to 4, moral effects: paragraphs 426 to 430
393 to 403, 464 to 470 multiple jurisdictions: paragraphs 455
fines: 9, 10, 13, 14, 18, 23, 304 to 308 to 457
inspections: paragraphs 11 to 25 national competition authorities:
periodic penalty payments: paragraphs paragraphs 378 to 380, 385
9, 10, 14, 18, 23 negative effects: paragraphs 415 to 434
powers of investigation: paragraphs 5 to penalty plus: paragraph 359 (footnote
75, 592 to 595 23)
requests for information: paragraphs 6 positive financial rewards: paragraphs
to 10 464 to 470
see also Leniency rights of the defence: paragraph 439
US Department of Justice: paragraphs
J 347 to 361, 382
Justice:
corrective justice: paragraphs 178, 179, M
182, 261, 348, 349, 413, 414 Moral effect of penalties: paragraphs 171,
proportional justice: paragraphs 208 to 181, 189, 192, 208 to 212, 233 to 236,
212, 233 to 236 260, 426 to 430, 572 to 574
206 INDEX
N see also Investigations

National competition authorities: Rights of the defence:
criminal sanctions: paragraphs 32 to 34, access to the file: paragraphs 105, 108 to
62, 514 to 543, 603 to 617 116
impact of European Commission generally: paragraphs 44 to 75
settlements: paragraphs 141 to 150 legal professional privilege: paragraphs
leniency policies: paragraphs 378 to 45, 46, 52, 60, 65, 66, 73 to 75
380, 385 leniency: paragraph 439
powers of investigation: paragraphs 28 self-incrimination (privilege against):
to 47, 61 to 75 paragraphs 44 to 46, 52, 55, 57 to
settlements: paragraphs 157 to 158 59
see also European Competition Network see also Fundamental rights
(ECN)
National courts: S
enforcement of European Commission Self-incrimination (privilege against):
settlements: paragraphs 133 to 140 paragraphs 44 to 46, 52, 55, 57 to 59
impact of European Commission Settlements: paragraphs 67 to 162
settlements: paragraphs 141 to 150 access to the file: paragraphs 105, 108 to
Negligence: paragraphs 319 to 323 116
history: paragraphs 82 to 90
O impact on national competition
Overconfidence bias: paragraphs 197, 198, authorities: paragraphs 141 to 150
243 impact on national courts: paragraphs
133 to 150
P informal settlements: paragraphs 82 to
Periodic penalty payments: paragraphs 9,
87, 89, 154 to 156
10, 14, 18, 23
Plea-bargaining: see Settlements judicial review: paragraphs 151 to 153
Powers of investigation: paragraphs 5 to legal effect: paragraphs 125 to 150
43, 592 to 595 optimal use: paragraphs 91 to 105
Prison : see Criminalisation rights of third parties: paragraphs 117,
Private enforcement: see Compensation, 118, 133 to 140, 145 to 150, 153
National courts settlements involving fines: paragraphs
Proportionality: paragraphs 208 to 212, 159 to 162, 377, 409 to 412, 354 to
233 to 236
360
R types of commitments: paragraphs 122
Recidivism : see Repeated infringement to 124
Redress: see Compensation
Repeated infringement: paragraphs 293 to U
303 Undertaking (notion of): paragraphs 27
Requests for information: paragraphs 6 to (footnote 31),
10 see also Commitments, Settlements

Wills, Wouter PJ - Efficiency and Justice in European Antitrust Enforcement PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Wills, Wouter PJ - Efficiency and Justice in European Antitrust Enforcement PDF

Transféré par

Droits d'auteur :

Formats disponibles

JOBNAME: Wouter Wils PAGE: 1 SESS: 2 OUTPUT: Mon Jan 21 15:06:07 2008

EFFICIENCY AND JUSTICE IN EUROPEAN

Efficiency and Justice in

Published in North America (US and Canada) by

© Wouter PJ Wils 2008

Hart Publishing, 16C Worcester Place, OX1 2JW

British Library Cataloguing in Publication Data

Typeset by Columns Design, Reading

vi INTRODUCTION AND ACKNOWLEDGEMENTS

INTRODUCTION AND ACKNOWLEDGEMENTS vii

private damages actions, and of penalties in other jurisdictions, ‘Amnesty Plus’,

viii INTRODUCTION AND ACKNOWLEDGEMENTS

Introduction and Acknowledgements v

1 Powers of Investigation and Procedural Rights and Guarantees: The

2 Settlements of Antitrust Investigations: Commitment Decisions

2.2.3 Consultation, Adoption and Publication 36

3 Optimal Antitrust Fines: Theory and Practice 49

3.2.2 The Trade-off between the Level of Fines and the

4 The European Commission’s 2006 Guidelines on Fines 77

4.5.2.1 Termination as Soon as the Commission

5 Leniency: Theory and Practice 113

5.4.2 Leniency and Follow-on Private Actions for Damages 147

6 Is Criminalisation the Answer? 155

6.3.1.2 Fines on Companies do not Always Guarantee

xvi TABLE OF CASES

TABLE OF CASES xvii

xviii TABLE OF CASES

Case C-36/92 P SEP v Commission, 15 December 1993 [1994] ECR I-1914

Court of First Instance

TABLE OF CASES xix

TABLE OF CASES xxi

Case T-201/04 R Microsoft v Commission, 22 December 2004 [2004] ECR II-4463:

European Court of Human Rights

xxiv TABLE OF LEGISLATION

1.1 POWERS OF INVESTIGATION

1.1.1 Different Ways of Gathering Intelligence and Evidence

1. Under Regulation 1/2003,1 the European Commission (hereafter also ‘the

2 INVESTIGATION AND PROCEDURAL RIGHTS

violations, in particular exclusionary practices, but also non-covert exploitative

1.1.2 Cases Dealt with by the European Commission

1.1.2.1 Information-gathering by the Commission Itself

6. The Commission’s powers of investigation are set out in Chapter V of

4 INVESTIGATION AND PROCEDURAL RIGHTS

misleading information or do not supply information within the required

6 INVESTIGATION AND PROCEDURAL RIGHTS

power to examine all business-related records and to take or obtain copies or

1.1.2.2 Assistance by the (Competition) Authorities of the Member States

19. As already indicated above,24 when the Commission conducts an inspection

the Member States provide to overcome opposition to Commission inspections

8 INVESTIGATION AND PROCEDURAL RIGHTS

1.1.3 Cases Dealt with by a Competition Authority of a Member State

1.1.3.1 Information-gathering by the National Competition Authority Itself

29. The powers of investigation of a competiion authority of a Member State are

10 INVESTIGATION AND PROCEDURAL RIGHTS

or 82 EC but for a related offence,37 it can be expected that correspondingly

1.1.3.2 Assistance by the Competition Authorities of Other Member States

41. Article 22(1) of Regulation 1/2003 provides as follows:

12 INVESTIGATION AND PROCEDURAL RIGHTS

1.2 PROCEDUR AL RIG HTS AND GUAR ANTEES

1.2.1 Procedural Rights and Guarantees Circumscribing or Limiting Powers

44. A number of procedural rights and guarantees circumscribe or limit the

PROCEDURAL RIGHTS AND GUARANTEES 13

14 INVESTIGATION AND PROCEDURAL RIGHTS