Séminaire Concurrence « Class action & damage claims ......12. En 2005, le directeur général de la concurrence, de la consommation et de la répression des fraudes et le directeur

Séminaire Concurrence « Class action & damage claims »

Mardi 14 mai 2013

Documentation

Intervenants

Liza Bellulo – Autorité de la Concurrence, Chef de Service du Président

Liza Bellulo occupe, depuis septembre 2011, les fonctions de chef du service du Président de l’Autorité de la concurrence, après avoir rejoint cette institution en 2009 en qualité de conseiller Europe. Elle est responsable, à ce titre, des affaires institutionnelles, européennes et internationales.

De 2007 à fin 2008, elle était chargée, en tant que conseillère technique, des institutions (traité de Lisbonne), de la justice, des affaires intérieures et du droit des sociétés au sein du cabinet de Jean-Pierre Jouyet, alors secrétaire d'Etat chargé des affaires européennes. De 2003 à 2007, au ministère de la Justice (adjointe du chef du bureau des questions institutionnelles et du contentieux du service des affaires européennes, puis chargée de mission auprès du directeur des affaires civiles et du sceau, Marc Guillaume), elle assurait la coordination des dossiers européens (affaires institutionnelles, droit économique, coopération judiciaire, droit des contrats, professions réglementées, contentieux et transpositions).

Ancienne élève de l'ENA (promotion René Cassin, 2003), Liza Bellulo est titulaire d'un Master of Business Administration de l'Ecole supérieure des sciences économiques et commerciales (ESSEC, 2000) et lauréate de l'Institut d'études politiques de Paris (1998). Elle enseigne le droit du marché intérieur dans le cadre du master II de droit européen des affaires de l’université Paris II-Panthéon Assas.

Eddy De Smijter - European Commission - DG Competition - Deputy Head of Unit A1, Private Enforcement

Eddy De Smijter joined the European Commission in 2001 as a member of the negotiation team on Regulation 1/2003 (procedural antitrust regulation). After the adoption of the Regulation, he co-authored an accompanying package of Commission notices, with particular responsibility for the cooperation with national courts. In 2005, he moved to the team that is responsible for the Commission policy on antitrust damages actions. In that capacity, he co-authored the Commission's Green and White Paper. He is now Deputy Head of the Unit dealing with enforcement of EU antitrust rules by national courts.

Before joining the European Commission, Eddy De Smijter was teaching and doing research in European law at the universities of Leuven ( Belgium ) and Tilburg (the Netherlands).

Sommaire

Annexe 1 – Ministère de l’économie et des finances – Projet de loi relatif à la consommation

Annexe 2 – Conseil de la Concurrence – Avis du 21 septembre 2006 relatif à l’introduction de l’action de groupe en matière de pratiques anticoncurrentielles

Annexe 3 – Commission européenne – Consultation publique relative à une approche européenne cohérente en matière de recours collectifs

Annexe 4 – Autorité de la Concurrence – Réponse à la consultation publique relative à une approche européenne cohérente en matière d'actions collectives

Annexe 5 – European Commission – White paper on damages actions for breach of the EC antitrust rules

Annexe 6 – European Commission – Joint Information Note of 5 October 2010 / Towards a Coherent European Approach to Collective Redress: Next Steps

Annexe 7 – European Parliament – Resolution of 2 February 2012 on the coherent European approach to collective redress

Annexe 8 – European Commission – Draft guidance paper Quantifying harm in actions for damages based on breaches of article 101 or 102 of the treaty on the functioning of the European Union

Annexe 9 – Resolution of the Meeting of Heads of the European Competition Authorities of 23 May 2012 / Protection of leniency material in the context of civil damages actions

Annexe 1

Annexe 2

RÉPUBLIQUE FRANÇAISE

Avis du 21 septembre 2006 relatif à l’introduction de l’action de groupe en matière

de pratiques anticoncurrentielles

Par lettre, en date du 24 janvier 2006, le directeur général de la concurrence, de la consommation et de la répression des fraudes (DGCCRF) et le directeur des affaires civiles et du Sceaux (DACS) ont demandé au Conseil de la concurrence de leur faire connaître ses observations sur le rapport relatif à l’action collective contre certaines pratiques abusives observées sur certains marchés.

Après en avoir délibéré, le Conseil souhaite faire, sur ce sujet, les observations suivantes :

INTRODUCTION : LA PLACE RESPECTIVE DE L’ACTION PUBLIQUE ET DE L’ACTION PRIVÉE

1. A la différence des Etats-Unis, la mise en œuvre de la répression des pratiques anticoncurrentielles est confiée à titre principal, en Europe, à des autorités spécialisées, dotées d’un pouvoir de sanction administrative (action publique). Ces autorités spécialisées sont la Commission européenne au niveau communautaire et les autorités nationales de concurrence (ANC) au niveau de chaque État. Celles-ci sont désormais réunies au sein du Réseau européen de concurrence (ci-après le Réseau) ( )1 depuis l’entrée en vigueur du règlement 1/2003. Cette primauté de l’action publique correspond à une pratique désormais bien établie et permet de mettre en œuvre une véritable politique de la concurrence à l’échelle de l’Union européenne, dont l’objectif est de créer un marché unique sur lequel s’exerce une concurrence non faussée. Dans un tel contexte, le recours direct aux juridictions (actions privées) ne peut avoir la même efficacité que l’action publique ainsi organisée. A ce stade de l’intégration du marché européen, il n’apparaît pas nécessaire de modifier ces équilibres. Le système doit reposer sur une mise en oeuvre du droit de la concurrence dans un cadre public fort, même si son efficacité peut être améliorée ou renforcée par une action privée maîtrisée.

2. La Commission européenne, dans son Livre vert sur les actions en dommages et intérêts, insiste également sur l’importance du rôle joué par les autorités publiques de la concurrence : « L’application du droit de la concurrence par les pouvoirs publics est

1 Le Réseau européen des autorités de concurrence existe depuis le 1er mai 2004. La langue de travail la plus souvent utilisée étant l’anglais, ce réseau est dénommé ECN (European Competition Network).

1

indispensable pour garantir une protection effective des droits conférés et une application effective des obligations imposées par le Traité » ( )2 . De surcroît, l’efficacité du Réseau est encore accrue par le développement dans de nombreux Etats membres de nouveaux instruments tels que les programmes de clémence permettant de mieux mettre à jour des pratiques secrètes ou les procédures d’engagements qui donnent la possibilité d’organiser une régulation plus rapide des marchés. Ces moyens sont spécifiques à l’action des autorités publiques et ne peuvent être mis en œuvre par la voie judiciaire.

3. Dans le cadre d’une réflexion sur les actions de groupe, il convient donc de garder à l’esprit la distinction entre l’action répressive et régulatrice menée à l’initiative des autorités publiques de concurrence et l’action privée à vocation réparatrice, engagée par les victimes des pratiques anticoncurrentielles devant les tribunaux. L’émergence d’une action privée plus efficace ne doit donc se concevoir que comme un complément du dispositif propre à l’action publique. Dans cette perspective, il convient tout particulièrement de veiller à la cohérence globale du système en s’assurant que la mise en œuvre des actions privées ne se fasse pas au détriment de l’efficacité de l’action publique, mais s’inscrive en complément positif des mécanismes de régulation publique des marchés déjà en place.

4. Le Conseil de la concurrence n’est pas compétent pour s’exprimer sur la totalité des questions soulevées par l’introduction éventuelle d’une procédure d’action de groupe dans le droit français et il se prononcera, dans le cadre du présent avis, uniquement sur l’impact que pourrait avoir la mise en place d’une action collective sur la mise en œuvre et l’efficacité du droit de la concurrence.

5. A cet effet, dans un premier temps, le Conseil procèdera à un rappel du contexte juridique et européen dans le cadre duquel se sont développées les réflexions sur l’opportunité du développement des actions collectives.

6. Dans un deuxième temps, le Conseil de la concurrence exposera les raisons pour lesquelles il considère qu’il est à la fois utile et souhaitable de disposer, au sein des actions civiles, d’actions de groupe afin d’organiser de manière plus juste et efficace le contentieux indemnitaire des pratiques anticoncurrentielles. Il attirera toutefois l’attention sur les points de rencontre, en nombre assez limité, entre les actions de groupe et les pratiques anticoncurrentielles et soulignera que le développement de l’action collective ne permettra pas de régler, par lui-même, tous les problèmes propres à l’action civile.

7. Dans un troisième temps, il s’attachera à démontrer que s'il est favorable à une réforme ayant pour objet d'offrir un espace plus large aux actions privées, et parmi elles, aux actions de groupe, une telle réforme ne peut être entreprise qu’à la condition que certaines précautions soient prises en particulier pour ce qui concerne la protection des programmes de clémence et l’organisation dans le temps des procédures publiques et privées.

2 Voir Livre vert de la Commission européenne « Actions en dommages et interêts pour infraction aux règles communautaires sur les ententes et les abus de position dominante », point 1-1.

2

I. Le cadre juridique

A. LE CONTEXTE JURIDIQUE FRANÇAIS

8. L’introduction d’une forme d’action en réparation de préjudices de masse, autrement nommée « action de groupe », « action collective » ou « class action » en référence au droit américain, qui consiste à intenter une action en justice au nom d’un groupe de victimes ayant chacune subi un dommage similaire, afin d’obtenir collectivement réparation du préjudice subi, est une question controversée.

9. La procédure d’action de groupe permet en pratique de renforcer les droits des consommateurs, qui n’ont souvent qu’un intérêt financier très limité à agir individuellement en justice concernant des litiges portant sur des dommages d’un faible montant. C’est là un moyen de rééquilibrer les rapports entre professionnels et consommateurs au profit de ces derniers.

10. Selon des modalités très variées, ce type de procédure est pratiqué aux Etats-Unis et au Canada mais aussi dans certains pays d’Europe. En effet, la Suède, la Norvège, le Portugal, l’Espagne, les Pays-Bas et le Royaume-Uni se sont dotés de ce type de procédure, avec des particularismes marqués selon les systèmes juridiques concernés. En Italie, un projet de loi a été présenté au Parlement, tandis que l’Allemagne envisage également d’introduire une telle action.

11. En France, c’est notamment dans le domaine de la consommation que les réflexions se sont succédé dans le sens d’une amélioration des recours ouverts aux victimes de préjudices collectifs. Ces réflexions ont abouti en 1992 à l’introduction, non pas d’une « action de groupe » au sens qui vient d’être défini, mais d’une action dite en représentation conjointe ouverte aux associations de consommateurs à des conditions très restrictives.

12. En 2005, le directeur général de la concurrence, de la consommation et de la répression des fraudes et le directeur des affaires civiles et du sceau ont été chargés de créer un groupe de travail en vue de faire des propositions allant de l'amélioration du dispositif de l’action en représentation conjointe jusqu’à la mise en place de mécanismes nouveaux qui ouvriraient aux associations de consommateurs le droit d’agir, dans certains types de litiges, au nom d’un groupe de consommateurs, pour obtenir le respect des règles de droit et la réparation des préjudices individuels.

13. Le rapport du groupe de travail a été remis le 16 décembre 2005 au ministre de l’économie, des finances et de l’industrie et au ministre de la justice. Ce groupe a précisé les axes d’une possible réforme, en laissant ouvertes les questions de la forme juridique d’une action de groupe et de son champ d’application.

14. Le président du Conseil de la concurrence, lors de son audition par le groupe de travail, a fait état d’une position ouverte de l’autorité de concurrence pour intégrer, dans le champ d’application de cette procédure, les préjudices subis par les consommateurs du fait de pratiques anticoncurrentielles, sous réserve de traiter des difficultés spécifiques juridiques et pratiques, qui feront l’objet des développements ci-après.

3

15. De nombreux membres de la communauté juridique et économique et du monde des affaires se sont exprimés parmi lesquels le premier président de la Cour de Cassation ( )3 . Ce dernier a notamment considéré que l’introduction d’une action collective était inéluctable et que ce type d’actions était la meilleure solution pour régler les problèmes d’asymétrie de puissance entre les grands groupes et les consommateurs multiples et isolés qui subissent individuellement des préjudices minimes qui globalement représentent des sommes considérables. Selon le Président Canivet, il ne faut pas exclure de cette action certains domaines tels que l’environnement, la concurrence et la santé publique. Enfin, sur la nature des dommages et intérêts, il a considéré qu’il ne fallait pas faire de la réparation du préjudice une sanction en adoptant des dommages à caractère punitif.

B. LE CONTEXTE JURIDIQUE EUROPÉEN

16. Dans le cadre de l’OCDE, des débats se sont tenus, à partir de l’expérience des Etats-Unis en la matière, sur les avantages et les inconvénients résultant d’une incitation donnée aux actions privées en réparation des dommages causés par les pratiques anticoncurrentielles. Ces réflexions sur l’action privée en général ont conduit à débattre de l’action de groupe qui constitue un développement possible de l’action en réparation.

17. A l’échelon européen, alors même qu’à première vue le droit de la responsabilité civile apparaît bien éloigné des préoccupations du Traité en matière de concurrence, le droit communautaire s’y intéresse depuis de nombreuses années. Ceci s’explique par le fait que les actions privées sont désormais perçues, à côté de l’action publique, comme un instrument supplémentaire de lutte contre les pratiques les plus nocives. Les actions en indemnisation n’ont plus seulement une fonction réparatrice, mais aussi un caractère dissuasif permettant de renforcer l’application du droit de la concurrence.

18. Plusieurs événements ont récemment contribué à conférer un regain d’actualité et d’intérêt à cette question de la place et du rôle des actions privées et plus spécifiquement, à la question de la réparation du préjudice causé par une pratique anticoncurrentielle.

19. Tout d’abord, la Cour de Justice, dans son arrêt « Courage » du 20 septembre 2001 ( )4 , a renforcé la place donnée au droit de la responsabilité dans le cadre de la sanction attachée aux pratiques anticoncurrentielles. En effet, elle énonce dans cet arrêt, qu’au nom de l’effet utile du Traité, toute personne doit pouvoir « effectivement » demander la réparation du préjudice que lui aurait causé un comportement restrictif de concurrence. Par suite, le règlement 1/2003 et la communication sur la coopération avec les juridictions évoquent l’octroi de dommages et intérêts aux victimes de pratiques anticoncurrentielles. Après l’affaire « Courage », plusieurs réformes substantielles visant à permettre une réelle indemnisation à la suite d’infractions aux règles prohibant les cartels et les abus de position dominante, ont été mises en place dans certains Etats-membres. Ainsi, au Royaume-Uni, l’Enterprise Act de 2002 accorde une place de choix aux actions privées et notamment aux actions collectives de consommateurs pour une même infraction ( )5 .

3 Voir La Tribune, 16 mai 2006, Guy Canivet, premier Président de la Cour de Cassation « Introduire l’action collective est une évolution inéluctable ». 4 Voir Courage /Crehan, affaire C-453/99, arrêt de la Cour de Justice du 20 septembre 2001, points 26 et 27. 5 Les sections 47A et B de l’Enterprise Act autorisent les actions collectives des consommateurs appelées « Group Litigation order ». L’action doit être exercée par un organisme agréé (specified body) qui agit pour le compte d’au moins deux consommateurs ayant donné leur consentement.

4

20. C’est dans ce contexte que la Commission européenne a annoncé son intention de favoriser les actions en réparation civile des dommages causés par des infractions au droit de la concurrence. A cet effet, elle s’est attelée à une vaste réflexion sur les moyens de faciliter les demandes d’indemnisation pour infraction aux règles de concurrence dans l’Union. L’action de groupe, en tant que développement possible de l’action privée, fait partie intégrante de cette analyse.

21. Afin de connaître les moyens les plus efficaces pour encourager les victimes à agir directement en dommages et intérêts devant les juridictions nationales, elle a fait procéder à un état des procédures existantes pour obtenir la réparation d’un dommage causé par une pratique anticoncurrentielle et à une quantification des actions effectivement mises en œuvre dans chaque Etat de l’Union européenne.

22. L’étude comparative réalisée par le Cabinet Ashurst auprès des 25 Etats-membres, publiée en septembre 2004, a révélé un très faible développement des actions privées en réparation d’une pratique anticoncurrentielle et « une étonnante diversité d’approche » des systèmes juridiques des Etats. Aussi, l’étude Ashurst a t-elle dressé le constat suivant : « alors que le droit communautaire exige un système efficace de traitement des demandes d’indemnisation faisant suite à des infractions aux règles sur les ententes, ce domaine du droit est caractérisé, dans les 25 Etats-membres par un total sous-développement ».

23. A partir du diagnostic établi par le rapport Ashurst et des premières discussions, la Commission a identifié les obstacles à la mise en place d’un système plus efficace d’introduction des actions en dommages et intérêts par les victimes de pratiques anticoncurrentielles dans le cadre d’un Livre vert publié le 21 décembre 2005 (ci-après « le Livre Vert », avec l’objectif suivant : « le fait de faciliter les demandes d’indemnisation doit permettre aux consommateurs et aux entreprises victimes d’une infraction aux règles sur les pratiques anticoncurrentielles, d’être plus facilement indemnisés par l’auteur de l’infraction, mais renforcera aussi l’application de ce droit. En ayant la possibilité d’introduire effectivement une demande d’indemnisation, chaque citoyen européen, que ce soit une entreprise ou un consommateur, deviendra plus proche des règles de concurrence et pourra participer plus activement à leur application » ( )6 .

II. L’intérêt des actions de groupe dans le domaine des pratiques anti-concurrentielles

24. S’il apparaît utile de disposer au sein des actions civiles, de procédures d’actions de groupe de consommateurs propres à rendre plus effectif le contentieux indemnitaire des pratiques anticoncurrentielles, l’expérience du Conseil de la concurrence montre que les cas dans lesquels des actions de groupe pourraient être utilement engagées contre des pratiques d’ententes ou d’abus de position dominante sont limités. En outre, le Conseil observe que l’action collective ne saurait régler certaines difficultés liées à la mise en œuvre des procédures civiles en elles-mêmes, difficultés qui, au contraire, pourront être accrues du fait de la spécificité du contentieux des pratiques anticoncurrentielles.

6 Voir Livre vert « Actions en dommages et intérêts pour infraction aux règles communautaires sur les ententes et les abus de position dominante », point 1-1.

5

A. L’UTILITÉ DES ACTIONS CIVILES EN GÉNÉRAL ET DES ACTIONS DE GROUPE EN PARTICULIER

25. Deux raisons principales militent en faveur du renforcement des actions civiles et du développement des actions de groupe en matière de concurrence : une meilleure réparation des préjudices subis par les consommateurs et leur association à la politique de concurrence dont le caractère dissuasif en serait globalement renforcé.

26. En premier lieu, ces actions peuvent contribuer à une meilleure reconnaissance du droit des consommateurs à obtenir une réparation du préjudice qu’ils ont subi du fait de pratiques anticoncurrentielles. Elles peuvent également réduire l’asymétrie entre la capacité d'agir des entreprises souvent constituées de grands groupes et celle des consommateurs, par nature isolés, qui subissent des préjudices au plan individuel minimes mais qui, globalement, représentent des sommes considérables.

27. Or, lorsque les victimes des pratiques anticoncurrentielles sont les consommateurs, les mécanismes d’indemnisation existants ne leur permettent pas d’intenter une action en réparation de manière effective. En effet, dans le cas d’atteinte au fonctionnement du marché affectant une catégorie de consommateurs, à concurrence de montants faibles pour chaque individu, le consommateur individuel est dissuadé d’agir compte tenu de la complexité et du coût des procédures au regard de la faiblesse du gain escompté. Une vraie procédure d’action de groupe aurait le mérite d’inciter de nombreux consommateurs victimes ayant subi individuellement un dommage mineur mais globalement élevé à entreprendre une action. Ce mécanisme d’agrégation des plaintes permettrait non seulement aux consommateurs de réaliser des économies d’échelle en regroupant leurs moyens, mais également aux autorités judiciaires d’optimiser leurs ressources et d’éviter le risque de décisions contradictoires.

28. S’agissant des actions des consommateurs, le Conseil de la concurrence constate que les saisines de la part des associations de consommateurs restent encore peu nombreuses. Comme il ressort du tableau ci- après, entre 1997 et 2005, 25 affaires contentieuses et 4 demandes d’avis ont été apportées par les organismes représentants les consommateurs.

6

7

Année Nombre des

saisines Associations des consommateurs

saisissantes Secteur d’activité Nature de la pratique Décisions et avis Type de la décision

1997 2

- Confédération syndicale du cadre de vie - Confédération syndicale du cadre de vie

- Avocats du Barreau des Hautes-Alpes - Avocats du Barreau de Clermont-Ferrand

Fixation des honoraires des avocats

- Décision n° 97-D-29 du 14.05.1997 - Décision n° 97-D-30 du 14.05.1997

Condamnation, injonction et sanction pécuniaire

1998 3

- Confédération syndicale du cadre de vie - Confédération syndicale du cadre de vie - Confédération syndicale du cadre de vie

- Avocats du Barreau des Alpes de Haute-Provence - Avocats du Barreau d’Aurillac - Avocats du Barreau de Marseille

- Fixation des honoraires des avocats - Fixation des honoraires des avocats - Etablissement de barèmes indicatifs

- Décision n° 98-D-01 du 7.01.1998 - Décision n° 98-D-02 du 7.01.1998 - Décision n° 98-D-07 du 14.01.1998

Condamnation, injonction et sanction pécuniaire

1999 5

- Fédération des familles de France - UFC Que Choisir - Comité des abonnés au câble de Metz - UFC du Val d’Oise - UFC Que Choisir

- Secteur de la parfumerie - Pratiques mises en œuvre par l’INC - Secteur du câble - Secteur d’hôtellerie - Le marché des produits du tabac

L. 420-1 et L. 420-2

- Décision n° 99-D-24 du 6.04.1999 - Décision n° 99-D-26 du 13.04.1999 - Décision n° 99-D-28 du 11.05.1999 - Décision n° 99-D-35 du 22.06.1999 - Décision n° 99-D-40 du 22.06.1999

- Retrait de saisine - Sursis à statuer - Irrecevabilité (association de consommateurs non agréée) - Sursis à statuer - Classement

2000 3

- UFC du Val-d’Oise - Confédération syndicale du cadre de vie 3) UFC Que Choisir

- Distribution de produits de consommation courante aux prisonniers - Avocats du Barreau de Bonneville - Marché de la presse consumériste

- L. 420-2 - L. 420-1 - L. 420-2

- Décision n° 00-D-27 du 13.06.2000 - Décision n° 00-D-23 du 31.05.2000 - Décision n° 00-D-54 du 28.11.2000

- Non-lieu - Injonction, publication - Sanction pécuniaire

2001 4

- UFC de Cergy-Pontoise - Association Force ouvrière consommateurs - UFC - Association des usagers de l’eau de Saint-Martin d’Uriage

- Convention d’affermage relative au service public du chauffage urbain - Matériaux de construction - Secteur bancaire - Contrat d’affermage pour la fourniture d’eau et l’assainissement

- L. 420-1 et L. 420-2 - L. 420-1 et L. 420-2 - Conditions d’une concertation entre les associations de consommateurs et la profession bancaire - L. 420-2

- Décision n° 01-D-48 du 26.07.2001 - Décision n° 01-D-74 du 5.11.2001 - Avis n° 01-A-13 du 19.06.2001 - Décision n° 00-D-89 du 20.02.2001

- Irrecevabilité - Rejet - Irrecevabilité (association de consommateurs non agréée)

2002 4

- Association St’Halt Tort Tue - Association de défense des justiciables - UFC Que Choisir de Nantes - Bouygues Télécom, UFC Que Choisir

- Secteur pharmaceutique - Secteur bancaire et prêts - Pratiques mises en œuvre par l’Institution de retraite nationale interprofessionnelle de salariés - Postes et télécommunications

- L. 420-1 et L. 420-2 - L. 420-1 et L. 420-2 - L. 420-1 et L. 420-2 - L. 420-2

- Décision n° 02-D-30 du 14.05.2002 - Décision n° 02-D-50 du 4.09.2002 - Décision n° 02-D-51 du 4.09.2002 - 02-D-69 du 26.11.2002

- Irrecevabilité (association de consommateurs non agréée) - Irrecevabilité (association de consommateurs non agréée) - Rejet - Rejet des saisines au fond et des mesures conservatoires

8

Année Nombre des

saisines Associations des consommateurs

saisissantes Secteur d’activité Nature de la pratique Décisions et avis Type de la décision

2003 1

- Confédération de la consommation, du logement et du cadre de vie

- Conditions de commercialisation du porte-monnaie électronique Monéo

Avis n° 03-A-17 du 18.09.2003

2004 1

UFC Que Choisir Conditions de la concurrence dans le secteur de la grande distribution non spécialisée

- Revente à perte et les accords de coopération commerciale après la loi du 1.07.1996 dite « loi Galland »

Avis n° 04-A-18 du 18.10.2004

2005 6

- Association La Quiétude - Confédération du logement et du cadre de vie (CLCD) - Confédération du logement et du cadre de vie (CLCD) - Que Choisir - UFC Que Choisir - Confédération du logement et du cadre de vie (CLCD)

- Prévoyance funéraire - Avocats du Barreau d’Evry - Avocats du Barreau de Montpellier - Distribution de l’eau potable en Ile-de-France - Téléphonie mobile - Services bancaires

- L. 420-1, L. 420-2, 81 et 82 CE - L. 462-7 - L. 462-7 - L. 420-2 - L. 420-1 - L. 462-1

- Décision n° 05-D-21 du 17.05.2005 - Décision n° 05-D-56 du 21.10.2005 - Décision n° 05-D-57 du 21.10.2005 - Décision n° 05-D-58 du 3.11.2005 - Décision n° 05-D-65 du 30.11.2005 - Avis n° 05-A-08 du 31.03.2005

- Rejet de la demande des mesures conservatoires - Non-lieu - Non-lieu - Sanction pécuniaire - Sanction pécuniaire

Total de saisines de 1997 à 2005 : 29 dont 4 pour avis

29. En second lieu, les actions privées en général et les mécanismes d’action de groupe en particulier peuvent contribuer à renforcer l'efficacité de la régulation concurrentielle en faisant de la victime et particulièrement du consommateur un véritable acteur et un allié des autorités publiques dans la lutte contre les pratiques anticoncurrentielles, notamment les cartels qui sont les plus dommageables. Cette volonté de s’appuyer davantage sur le juge judiciaire au moyen des actions privées pour assurer l’effectivité du droit de la concurrence a été affichée par la Commission européenne dans son Livre vert précité. Celle-ci considère, en effet, que le fait de faciliter les demandes d’indemnisation permettra non seulement aux consommateurs victimes d’une pratique anticoncurrentielle d’être plus facilement indemnisés par l’auteur de l’infraction mais renforcera aussi l’application du droit de la concurrence et garantira ainsi la pleine efficacité des dispositions du traité en contribuant à dissuader les comportements anticoncurrentiels.

30. En effet, il ressort de nombreuses études juridiques et économiques réalisées récemment, qu’en dépit d’une augmentation depuis une dizaine d’années, le niveau des amendes infligées par les autorités antitrust aux entreprises auteurs des pratiques anticoncurrentielles, particulièrement de cartels, apparaît encore insuffisamment dissuasif au regard du gain illicite que procurent ces pratiques et de la probabilité limitée de leur détection.

31. A côté des sanctions publiques qui, eu égard à leurs montants désormais élevés, contribuent à dissuader les entreprises de se livrer à des pratiques prohibées, les actions de groupe peuvent être un facteur de dissuasion supplémentaire, puisqu’elles augmentent le risque financier pour l’auteur de l’infraction.

32. En définitive, les actions de groupe, en tant que développement des actions en dommages et intérêts dans le contentieux indemnitaire des pratiques anticoncurrentielles pourraient tout à la fois participer de la réhabilitation de la place dévolue aux victimes, en l’espèce aux consommateurs ayant subi un dommage du fait d’une pratique anticoncurrentielle et concourir à la préservation de l’ordre public concurrentiel.

B. AUTRES ÉLÉMENTS DE DIAGNOSTIC

1. LE CHAMP DE L’ACTION DE GROUPE EN MATIÈRE DE PRATIQUES ANTICONCURRENTIELLES

33. Le champ de rencontre entre les pratiques anticoncurrentielles et les actions de groupe n’est pas aussi étendu qu’on pourrait l’imaginer dans la mesure où toutes les affaires ne se prêtent pas à une action collective.

34. Il faut d’abord rappeler qu’un grand nombre de cas soumis à l’appréciation des autorités de concurrence concerne des biens intermédiaires offerts sur un marché de gros qui se situe en amont du marché de détail sur lequel s’exprime la demande des consommateurs.

35. En effet, si les dommages à l’économie des pratiques anticoncurrentielles, notamment des cartels, et donc le nombre des victimes, apparaissent potentiellement élevés, les effets pour le consommateur final sont le plus souvent indirects, ce qui limite la possibilité effective de ce dernier de rechercher l’indemnisation du préjudice qu’il a subi.

9

36. A cet égard, la règle de transitivité du dommage, communément désignée par le terme anglais de « passing on » ( )7 consistant à permettre à l’auteur des pratiques d’invoquer le fait que la victime directe (entreprise) ne puisse pas prétendre à une indemnisation dans la mesure où elle aurait répercuté le préjudice subi en aval (sur le consommateur notamment), peut apparaître à première vue séduisante, en ce qu’il semble, en privant la victime directe d’indemnisation, ouvrir au consommateur final une possibilité de réparation des dommages indirects.

37. En réalité, l’acceptation de ce moyen conduit, en pratique, à compliquer considérablement le traitement des dossiers par le juge civil. On notera, à cet égard, que les tribunaux fédéraux américains refusent de recevoir l’argument du « passing on »8 : ils considèrent en effet, s’agissant des cartels par exemple, qu’accepter ce moyen conduirait à diluer la responsabilité des auteurs de l’entente sans pour autant permettre aux consommateurs victimes indirectes de rapporter la preuve de la réalité de leur préjudice. Effectivement, dans cette hypothèse, on perçoit aisément la difficulté pour le consommateur final de prouver que, si la victime directe a augmenté le prix de ses produits, c’est parce qu’elle a répercuté le surcoût illégal sur les tiers et d’établir le lien de causalité entre le comportement anticoncurrentiel et le dommage. Une telle démonstration requiert la possession de données relatives au fonctionnement du marché qu’il lui est quasiment impossible de rassembler (éléments relatifs à la demande, à la structure du marché et à la fonction de coût).

38. Dans ces conditions, on perçoit que les actions de groupe effectivement envisageables concernent des pratiques commises par des entreprises qui sont dans une relation directe avec le dommage subi par le consommateur final. De ce point de vue, les affaires traitées par le Conseil et susceptibles de donner lieu à une action de groupe, peuvent être ainsi regroupées : certaines ententes horizontales portant sur des biens de grande consommation, les ententes verticales qui visent à augmenter artificiellement le prix de détail et les abus de position dominante dans le cadre desquels un opérateur a pu retarder l’arrivée de concurrents ou exclure certains d’entre eux. Ces affaires ne sont pas si nombreuses, si on observe les statistiques de ces dix dernières années, mais elles sont souvent des affaires ayant conduit à un montant élevé de sanctions du fait de l’importance des chiffres d’affaires en jeu.

39. De ce fait, il apparaît au Conseil de la concurrence, que le choix de l'action de groupe s'avèrera utile essentiellement dans les cas où la victime est un consommateur direct, ce qui accroît la probabilité d’établir un lien clair entre le comportement anticoncurrentiel et le dommage. A cet égard, l’exemple le plus explicite est celui de l’entente des banques sur la renégociation des emprunts immobiliers (décision n° 00-D-28 du 19 septembre 2000) ou celle des opérateurs français de téléphonie mobile (décision n° 05-D-65 du 30 novembre 2005).

7 Sur cette question du « passing on » ou « pass throught », voir notamment l’article du professeur E. Combes « quelles sanctions pour les cartels ? une perspective économique », in RIDE 2006.

8 Sur l’interdiction du « passing on defence » de manière defensive (pzr l’auteur des pratiques anticoncurrentielles), voir l’arrêt de la Cour suprême des Etats-Unis « Hanover Schoe & co V.United Schoe Machinery Corporation ( 1968) ». Sur l’interdiction du « passing on defense » de manière offensive par la victime indirecte, voir l’arrêt de la Cour suprême « Illinois Brick co V. Illinois (1977) ». Sur l’atténuation de cette interdiction du « passing on » pour des raisons tenant à la double législation ( fédérale et étatique), voir l’arrêt de la Cour suprême, « California V.ARC America Corp. (1989) ».

10

2. LES DIFFICULTÉS PROPRES À L’ACTION CIVILE

40. Il faut également relever que l’indemnisation du préjudice résultant d’une pratique anticoncurrentielle dans le cadre classique de la procédure civile n’est pas toujours une chose aisée et ce n’est pas l’introduction de l’action collective qui va, par elle-même, résoudre ces problèmes spécifiques. Ces difficultés ne doivent pas être occultées si l’on souhaite introduire un mécanisme efficace d’action de groupe en matière de droit de la concurrence. Elles supposent des choix de politique juridique destinés à renforcer l’efficacité de l’action en dommage et intérêts.

41. De fait, conformément à l’article 1382 du code civil, l’action au civil nécessite que le plaignant, individuellement ou au sein d’un groupe de victimes, démontre l’existence de la faute, justifie la réalité du préjudice subi et fasse également la démonstration du lien de causalité entre les deux.

42. En matière de contentieux de la concurrence, pour assurer la démonstration de ces éléments, l’action privée doit surmonter quatre difficultés majeures : la maîtrise de l’expertise requise en droit de la concurrence, le standard de la preuve, l’établissement du lien de causalité et les problèmes liés à la quantification du dommage.

2-1 S’agissant de l’expertise économique et du standard de la preuve

43. Le droit des pratiques anticoncurrentielles est un droit spécifique, de caractère technique, qui requiert une capacité d’expertise des tribunaux. L’expertise en analyse économique du fonctionnement des marchés est en effet atypique par rapport aux grands blocs de légalité (droit commercial et droit civil) qui font le quotidien des juridictions.

44. Cette technicité particulière, indispensable au juge de la concurrence devra être renforcée, compte tenu de l’évolution en cours sur la manière d’appréhender les pratiques d’abus de position dominante, qui confirme le renforcement des raisonnements économiques pour la qualification des pratiques et l’analyse de leurs effets. Ce changement n’est pas sans incidence sur les actions privées en matière de concurrence dans la mesure où on s’éloigne d’une analyse fondée sur des catégories juridiques prédéfinies (« approche par la forme ») pour adopter une analyse impliquant la démonstration des effets anticoncurrentiels, actuels ou potentiels de la pratique dénoncée (« approche par les effets»).

45. Aussi, si le recentrage envisagé sur l’examen des effets des pratiques et la volonté de mieux mettre en phase la notion de position dominante et celle de réduction de l’intensité de la concurrence doivent être accueillis favorablement, force est de reconnaître que cette nouvelle approche ne facilitera pas la tâche du juge dans l’examen d’une plainte autonome d’un groupe de consommateurs, portant seul le contentieux sans le soutien d’une autorité de concurrence et sans pouvoir s’appuyer sur une décision déjà prononcée.

46. A ce titre, la spécialisation des tribunaux en matière de contentieux de la concurrence acquise très récemment ( )9 et les formations des juges en matière économique mises en place par la Cour de Cassation sont des facteurs très positifs qui contribueront à renforcer l’expertise économique du juge.

9 La spécialisation des juridictions a été introduite par la loi n° 2001-420 du 15 mai 2001 et le décret du 27décembre 2005.

11

47. Toutefois, s’il est essentiel que le juge dispose d’un volant d’affaires lui permettant d’acquérir une expérience en matière de traitement de contentieux de la concurrence, la question se pose néanmoins des moyens mis à sa disposition par le code de procédure civile pour ordonner des expertises et pour obtenir la production des éléments de preuve nécessaires à la démonstration d’une collusion ou d’une domination. A cet égard, on peut estimer que les règles civiles d’administration de la preuve sont suffisantes lorsque les pratiques examinées reposent sur des éléments contractuels ostensibles, la discussion portant alors sur l’analyse des clauses litigieuses de l’accord. En revanche, elles le sont beaucoup moins lorsque les pratiques sont occultes et supposent la recherche active d’éléments de preuve non identifiés par le demandeur ou lorsqu’il s’agit pour le juge de faire la démonstration de l’existence d’un abus de position dominante mis en œuvre au moyen, par exemple, d’une politique de prix.

48. En droit français, le juge civil ne dispose ni de l’équivalent de la procédure anglo-saxonne de « discovery » qui permet à l’expert ou au juge d’examiner l’ensemble des informations pertinentes pour la solution du litige, ni des pouvoirs d’enquête coercitifs du juge pénal ou du Conseil de la concurrence.

49. En l’absence de tels moyens d’investigation, on perçoit aisément la difficulté qu’aurait eu le juge civil, saisi à titre principal, c’est-à-dire sans qu’une décision préalable du Conseil ne soit intervenue, d’une action de groupe selon la procédure actuellement en vigueur de la représentation conjointe, de démontrer l’existence d’une entente.

50. Et ce déficit de moyens ne saurait légitimer l’existence d’un standard de preuve moins exigeant pour le juge civil. En effet, tout abaissement du niveau de preuve requis pour les actions civiles poserait un véritable problème de compatibilité avec le standard de preuve auquel les autorités nationales de concurrence et la Commission européenne doivent se conformer. Il n’est pas envisageable de prévoir un standard de preuve différent selon que l’action est portée directement devant le juge civil ou bien, en premier lieu, devant une autorité de concurrence, puis devant un juge civil pour l’obtention de dommages et intérêts.

51. Force est donc de reconnaître que si la dualité des voies d’action doit être maintenue, pour certains types de pratiques nécessitant des investigations lourdes (cartels secrets) et une expertise économique approfondie (abus de domination au moyen, par exemple d’une politique de prix), le choix est en réalité plus limité qu’il n’apparaît en théorie. Le traitement au civil ne s’avère en réalité efficace que lorsque l’affaire a été préalablement traitée par le Conseil de la concurrence. Dans ce cas, en effet, le problème de l’expertise et de la preuve ne se pose pas dans les mêmes termes, eu égard aux pouvoirs d’enquête et aux moyens d’expertise économique dont dispose l’autorité spécialisée.

52. Ces interrogations sur l’efficacité de l’action civile sont très importantes dans la mesure où une action civile en complément de l’action publique ne contribuera à une indemnisation réelle de consommateurs que si le juge dispose des moyens suffisants et adéquats pour sanctionner et réparer les pratiques ayant un caractère réellement anticoncurrentiel. Dans cette hypothèse, les demandeurs ne seront pas incités à détourner l’action civile de son objet. Dans le cas inverse, si les tribunaux peinent à faire le partage entre les pratiques anticoncurrentielles et celles qui ne le sont pas, ajouter la possibilité d’une action civile ne va pas nécessairement dans le sens de l’intérêt général.

12

2-2 S’agissant du lien de causalité

53. A côté des difficultés tenant à l’expertise économique et à la preuve, se pose également la question de la détermination du lien de causalité direct et immédiat entre le dommage subi et la pratique dénoncée.

54. Comme expliqué ci-dessus, ces difficultés ne sont pas propres au contentieux en matière de droit de la concurrence ni à la procédure d’action de groupe mais tiennent pour une large part à l’absence de moyens procéduraux suffisants mis à la disposition des victimes et du juge civil. En effet, si la causalité peut être prouvée par tout moyen, dans la pratique, les indices et les présomptions graves, précises et concordantes selon les termes de l’article 1353 du code civil qui doivent permettrent au juge de forger sa conviction ne peuvent être rassemblés faute de réel pouvoir d’investigation. C’est une des raisons pour lesquelles les victimes préfèrent saisir la juridiction pénale devant laquelle une véritable instruction peut être menée.

55. En matière de concurrence, la question du lien de causalité peut être délicate dans la mesure où l’on se situe le plus souvent dans le domaine du probable. Les économistes expliquent par exemple que les prix peuvent être influencés par une entente mais aussi par d’autres facteurs tels qu’un choc sur les coûts, des fluctuations de la demande, des changements de la structure du marché. La difficulté va consister à démêler ces causalités multiples afin de ne pas, au moment de la quantification du dommage, surévaluer ou surestimer l’impact de la pratique. A l’évidence, une telle analyse implique le recueil des données relatives au fonctionnement du marché en cause.

56. A cet égard, dans le cas où une décision du Conseil de la concurrence aurait été prise préalablement à une action civile, les analyses menées par cette autorité spécialisée pourraient constituer des éléments utiles à l’appréciation du lien de causalité entre le dommage et la pratique anticoncurrentielle et être utilisées par le juge, le cas échéant, comme des présomptions de causalité.

2-3 S’agissant de la difficulté d’évaluation des dommages

57. A supposer que l’on ait résolu les questions de preuve de la pratique et de lien de causalité en optant, s’agissant des cas les plus complexes, pour les actions consécutives (« follow on »), une phase d’évaluation du dommage s’ouvre, et suscite de nouvelles difficultés en matière de concurrence.

58. En pratique, la mission d’évaluation des dommages incombera le plus souvent à un expert nommé par le juge. Ce type de mission n’est pas sans poser problème. Il est évident que le préjudice résultant de prix décidés par les opérateurs de manière concertée ne s’apprécie pas de la même manière que le préjudice né d’un fait dommageable plus courant. Comme l’explique le professeur Combes ( )10 , « l’évaluation du préjudice par la victime d’un cartel doit passer par la comparaison de la situation pendant la période de cartellisation du marché avec celle qui aurait prévalu en situation de concurrence. Mais l’estimation de l’équilibre concurrentiel et du prix qui lui est associé (dénommé « but for price ») n’est pas chose aisée ».

10 Voir l’article d’Emmanuel Combes précité

13

59. Dans ces conditions, une des options proposée par la Commission dans son Livre Vert sur les actions en en réparation de dommages causés par des pratiques anticoncurrentielles, consistant à élaborer un guide d’évaluation du montant des dommages et intérêts, devrait retenir l’attention. Ce guide pourrait par exemple présenter différents modèles économiques utiles à l’évaluation du dommage ainsi que des exemples de leur mise en œuvre. Il pourrait également éclairer le juge sur les spécificités des infractions aux règles de la concurrence ainsi que sur les paramètres à prendre en considération lors de l’évaluation du préjudice. Le Conseil de la concurrence est quant à lui favorable à des lignes directrices non contraignantes de la Commission sur ce sujet.

60. Comme l’observait le Président Canivet « l’enjeu est d’importance car lorsque les entreprises percevront nettement le risque certain d’avoir à désintéresser complètement les victimes de pratiques anticoncurrentielles qu’elles auront mises en œuvre et que les victimes percevront les chances sérieuses qu’elles ont d’être intégralement indemnisées, alors le recours aux juridictions civiles sera un mode utile et efficace de régulation des marchés » ( )11 .

61. S’agissant de la forme que doivent prendre les dommages et intérêts obtenus par une procédure collective, le Conseil de la concurrence considère que lorsque les montants à allouer sont individuellement faibles et dans la mesure où il s’agit de pratiques qui se situent dans un contexte de relations contractuelles entre le consommateur et l’entreprise auteur des pratiques, il conviendrait d’envisager une réparation soit sous forme de dommages-intérêts soit en nature.

III. Des précautions nécessaires en cas d’introduction des actions de groupe

62. Si l’évolution en matière de développement des actions privées et en particulier des actions de groupe appelle une réflexion ouverte, celle-ci doit néanmoins rester empreinte de pragmatisme.

63. En premier lieu, l’action privée doit être un complément positif à l’action publique. En effet, si le développement des actions privées peut contribuer à renforcer la dissuasion à l’encontre des pratiques anticoncurrentielles, en faisant du citoyen, entreprise ou consommateur, un allié dans la lutte contre ces pratiques, cette nouvelle étape ne peut être franchie qu’à la condition de veiller à ce que la mise en œuvre des actions privées ne se fasse pas au détriment de l’efficacité des actions publiques de régulation de la concurrence. L’attention sera ici particulièrement portée sur la protection des programmes de clémence mis en place par de nombreuses autorités de la concurrence.

64. En second lieu, la succession dans le temps des procédures publiques et privées, notamment les risques de sollicitation systématique des autorités de la concurrence par les juges de l’action en dommages et intérêts, doivent également retenir l’attention dans le cadre d’une réflexion sur l’articulation optimale entre l’action publique et l’action privée en matière de concurrence.

11 Voir Canivet, in l’expertise en droit de la concurrence.

14

A. LA PROTECTION DES PROGRAMMES DE CLÉMENCE

65. Ces programmes permettent, par les réductions de sanctions qu’ils prévoient, d’inciter les entreprises, parties à un cartel, à révéler elles-mêmes ces infractions néfastes pour l’économie, lesquelles n’auraient pas été forcément découvertes par les moyens d’enquête classiques à la disposition des autorités de concurrence. En outre, ces nouveaux instruments constituent un moyen de renforcer la démonstration et la régularité des preuves des pratiques en cause dans la mesure où une partie de ces preuves est apportée par le contrevenant demandeur à la procédure de clémence.

66. En 2001 ( )12 , le législateur français a considéré qu’il est de l’intérêt de l’économie française, et notamment des consommateurs, de faire bénéficier d’un traitement favorable les entreprises qui informent les autorités de concurrence de l’existence d’ententes illicites et qui coopèrent avec elles afin d’y mettre fin. Aussi, l’article L. 464-2-IV du code de commerce prévoit que l’entreprise qui, la première, demande la clémence et qui satisfait aux conditions prévues par la loi, obtient du Conseil de la concurrence une exonération totale des sanctions pécuniaires normalement encourues. Dans ces conditions, il apparaît tout à fait essentiel de veiller à ce que l’attractivité des programmes de clémence ne soit pas affaiblie par le développement des actions privées, en particulier par l’introduction d’actions de groupe de consommateurs.

67. S’agissant des infractions susceptibles de faire l’objet de demandes de clémence telles que les cartels secrets, la question se pose, en effet, de l’impact d’actions privées efficaces - notamment d’actions de groupe - sur les demandeurs de clémence potentiels. En effet, si les participants au cartel anticipent qu’une collaboration active avec les autorités de concurrence risque de conduire à une action civile individuelle ou collective, mais surtout peuvent permettre de faciliter l’aboutissement de ces actions à leur détriment, alors l’efficacité des programmes de clémence pourrait être compromise.

68. Cette question du risque que ferait peser sur les programmes de clémence la mise en œuvre d’actions privées en réparation a fait l’objet d’un débat intense aux Etats-Unis et a conduit le Congrès américain à adopter récemment une loi ( )13 qui limite le montant des dommages et intérêts dus aux victimes par le premier demandeur de clémence auprès de l’autorité chargée de l’action publique, le ministère de la justice.

12 Le programme de clémence français a été introduit par la Loi du 15 mai 2001 et figure à l’article L. 464-2 IV du code de commerce. A l’occasion du prononcé de la première décision de fond mettant en jeu la procédure de clémence, le Conseil de la concurrence a publié un communiqué de procédure dans lequel il précise les conditions de mise en œuvre de la procédure de clémence. 13 Voir L’Antitrust Criminal Penalty Enhancement and Reform Act, section 213 b, 15 USC, juin 2004 opère ce qu’on appelle un “de-trebling”, c’est-à-dire une impossibilité pour le juge civil de condamner le signataire d’un accord de clémence avec le DOJ au paiement des dommages et intérêts triples. Parallèlement, ce texte de loi relève les amendes maximum pour les entreprises de 1 à à 100 millions de dollars. Il augmente également le niveau maximal de l’amende pouvant être infligée à une personne physique de 300 000 à un million de dollars et porte la peine maximale de prison de 3 à 10 ans.

15

69. Il est vrai que la situation outre Atlantique est très différente de celle qui existe en Europe et particulièrement en France. Aux Etats-Unis, compte tenu du fort développement des actions civiles ( )14 et parmi elles de véritables « class action » ( )15 et eu égard à l’existence de la règle des triples dommages ( )16 , les actions privées représentent une menace réelle et forte pour les entreprises membres d’un cartel et susceptibles de venir demander la clémence.

70. Dans l’Union européenne, comme expliqué dans la première partie, les actions de groupe n’existent pas ou bien ont été introduites sous une forme différente de la « class action » américaine ; quant aux dommages et intérêts, ils ne sont pas punitifs mais conservent un caractère compensatoire malgré de récentes propositions françaises pour l’introduction de dommages punitifs ( )17 . Le Conseil de la concurrence reste attaché au maintien de la notion de dommages et intérêts compensatoires, fortement ancré dans le système juridique français et dans celui de nombreux Etats de l’Union européenne. Toutefois, même si les dommages et intérêts conservent leur caractère compensatoire, le renforcement d’actions privées jusqu’alors très peu développées et l’introduction d’une action de groupe en droit français, risquent de rendre moins attractives les procédures de clémence.

71. Certes, il ne s’agit pas de faire échapper le demandeur de clémence à ses obligations en matière de réparation du dommage qu’il a pu causer à ses concurrents et aux consommateurs. Toutefois, il convient de rappeler que pour envisager une réparation civile dans les affaires susceptibles de faire l’objet de demandes de clémence (cartels secrets), il est indispensable que le demandeur de clémence n’ait pas été dissuadé de venir dénoncer la pratique devant le Conseil ; en effet, si aucune entreprise ne vient demander la clémence (parce qu’elle en est dissuadée par les conséquences de sa demande sur une éventuelle action civile), la pratique ne sera pas révélée et les victimes, par définition, ne pourront pas demander réparation, mais plus encore, elles continueront à subir les dommages générés par le cartel, sans que ni l’autorité publique de concurrence, ni le juge, ne puissent faire cesser la pratique faute d’en avoir connaissance.

72. Aussi, parce qu’ils permettent la révélation des comportements anticoncurrentiels les plus dommageables aux consommateurs, lesquels seraient toujours actifs en l’absence de procédure d’auto révélation, les programmes de clémence, instrument très puissant dans les mains des autorités publiques de concurrence, doivent être sauvegardés afin de préserver leur caractère attractif.

14 Selon l’auteur américain d’une étude comparative qui fait référence, les actions privées représenteraient à la fin du XXè siècle 90 % de toutes les affaires de droit antitrust fédéral, voir C. Jones, Private Enforcement of Anti-trust law in the EU, UK and USA, Oxford University press, 1999, cité par le professeur L. Idot in « Regards sur les droits étrangers », Colloque, Paris, 29 avril 2004. 15 Les principales caractéristiques de la procédure de « class action » américaine qui la différencient des procédures de groupe existant en Europe sont principalement « l’opt in, » les « contengency fees » et l’existence d’un jury populaire. 16 La règle des dommages triples (« treble damages ») a été introduite aux Etats-Unis par les sections 7 et 4 du Sherman Act et du Clayton Ac : elle prévoit que toute personne qui subit une atteinte dans ses droits de propriété ou dans ses affaires du fait de la violation d’une règle de droit de la concurrence, dispose d’un droit d’action, afin d’obtenir une somme équivalente à trois fois le montant des dommages dont elle a souffert. 17 Voir le rapport de la Commission présidée par le Professeur P. Catala, sur l’avant projet de réforme du droit des obligations et de la prescription, remis au garde des sceaux le 22 septembre 2005. Cet avant projet, propose notamment l’introduction de la notion de dommages punitifs.(cf projet d’article 1371 du Code civil).

16

73. Pour limiter l’impact négatif que pourrait avoir l’action collective sur l’action publique, il sera donc nécessaire de garantir la confidentialité des déclarations faites par les entreprises demanderesses de la clémence devant les autorités de concurrence, afin qu’elles ne soient pas utilisées à d’autres fins. En d’autres termes, l’entreprise qui a obtenu le bénéfice de la clémence devant le Conseil ne doit pas, au minimum, être mise dans une situation moins favorable devant le juge civil, que les entreprises qui n’ont pas choisi de dénoncer l’existence d’un cartel et de coopérer. Par conséquent, l’autorité de la concurrence doit pouvoir refuser la communication de certaines pièces en relation avec les programmes de clémence, notamment les déclarations au soutien d’une demande de clémence.

74. C’est ce que la Commission européenne a prévu, dans le cadre du Règlement 1/2003, dans sa communication sur la coopération avec les juridictions nationales ( )18 . Si la Communication prévoit un devoir de transmission des informations aux juridictions nationales par la Commission, elle prévoit plusieurs exceptions à ce principe général. La première limitation est tirée des dispositions combinées des articles 10 et 287 du Traité CE et entraîne une interdiction relative pour la Commission de communiquer des informations couvertes par le secret professionnel aux juridictions nationales ( )19 . La seconde exception à la divulgation figure au paragraphe 26 de la Communication et concerne notamment les programmes de clémence : « La Commission peut refuser de communiquer des renseignements aux juridictions nationales pour des raisons essentiellement liées à la nécessité de préserver les intérêts de la Communauté ou pour éviter toute interférence dans son fonctionnement et avec son indépendance, notamment de nature à compromettre l’accomplissement de sa mission. La Commission s’abstiendra, par conséquent, de transmettre aux juridictions nationales des renseignements fournis volontairement par l’auteur d’une demande de clémence sans avoir obtenu l’accord de celui–ci ».

B. L’ARTICULATION DES PROCÉDURES DANS LE TEMPS

75. Comme expliqué supra, à la différence du modèle américain, les Etats de l’Union européenne connaissent une distinction bien marquée entre les autorités spécialisées de concurrence ( )20 et les juridictions de droit commun. Le règlement 1/2003 de modernisation des procédures d’application des articles 81 et 82 a renforcé cette distinction. En effet, avec l’abandon du régime centralisé mis en place par le Règlement 17/62, le Règlement 1/2003 renforce le rôle des autorités de concurrence, mais aussi celui des juridictions en leur reconnaissant une fonction essentielle dans l’application des règles communautaires de concurrence. Dans son considérant 6, notamment, ce texte énonce que les juridictions nationales « préservent les droits subjectifs prévus par le droit communautaire lorsqu’elles statuent sur des litiges entre particuliers, notamment en octroyant des dommages et intérêts aux victimes des infractions. Le rôle des juridictions nationales est, à cet égard, complémentaire de celui des autorités de concurrence des Etats-membres. Il convient dès lors de leur permettre d’appliquer pleinement les articles 81 et 82 du Traité ».

18 Voir la Communication de la Commission sur la coopération avec les juridictions nationales pour l’application des articles 81 et 82 du Traité CE. 19 Cf. § 24 de la Communication 2004/C 101/04. 20 Ces autorités spécialisées sont dénommées ANC ou NCA en anglais dans le nouveau vocabulaire issu du Règlement 1/2003. Elles sont membres de l’ENC, le Réseau des autorités de concurrence.

17

76. Eu égard à l’existence de ces instances parallèles, une réflexion sur l’opportunité de l’introduction d’actions privées telles que les actions de groupe, conduit nécessairement à se poser la question de l’articulation dans le temps des compétences entre les autorités spécialisées et le juge de droit commun.

77. Pour la solution d’un litige relatif à des pratiques anticoncurrentielles, qu’il s’agisse d’une action individuelle ou d’une action de groupe, deux cas de figure principaux peuvent se présenter : en premier lieu, l’action en réparation et l’action publique peuvent être mises en œuvre pour juger la même pratique anticoncurrentielle. Dans cette hypothèse, on distingue deux situations : tout d’abord le cas où l’action privée est exercée en complément de l’action publique d’une autorité nationale ou communautaire de la concurrence : l’action publique précède alors la demande en réparation. On parle d’actions complémentaires ou « follow on ». Ensuite, le cas dans lequel l’action publique et l’action privée sont exercées en même temps. On parlera alors d’actions parallèles. En second lieu, l’action en dommages et intérêts peut être exercée de manière autonome par un consommateur ou un groupe de consommateurs à titre principal devant le juge civil sans qu’une autorité de concurrence ne se soit prononcée préalablement sur les pratiques ou, au contraire, avant une telle décision. On parle dans ce dernier cas d'action autonome (ou « stand alone »).

1. La procédure d’action complémentaire ( dite de « follow on »)

78. La procédure dite d’action complémentaire est celle qui, en terme d’articulation dans le temps, (au moins pour les pratiques les plus difficiles à démontrer) apparaît la plus satisfaisante : dans un premier temps, l’autorité de la concurrence adopte une décision de condamnation et inflige des sanctions pécuniaires aux entreprises contrevenantes ; dans un second temps, la victime, en action individuelle ou en action de groupe, forte de cette décision constatant une pratique anticoncurrentielle, saisit le juge civil pour obtenir réparation de son préjudice. Même si, en droit français, la décision de condamnation prise par le Conseil de la concurrence, même devenue définitive, ne s’impose pas au juge de la réparation, celui-ci pourra utiliser cette dernière comme un élément utile qui contribuera à emporter sa conviction sur l’existence et la qualification de l’infraction.

79. Dans le cas d’actions consécutives de l’autorité de concurrence et du juge civil, l’Allemagne a consacré un véritable allégement des conditions de la responsabilité civile en reconnaissant une autorité de chose décidée aux décisions du Bundeskartellmamt et depuis l’entrée en vigueur de la loi GWB sur les restrictions de concurrence ( )21 , une autorité de chose décidée à toutes les décisions des autorités de concurrence, membre du réseau ECN, pourvu qu’elles soient devenues définitives.

80. Une telle solution, si elle était adoptée en France au moins pour les décisions du Conseil de la concurrence devenues définitives ( )22 , contribuerait à faciliter la tâche du demandeur et du juge dans la preuve du comportement anticoncurrentiel.

81. Au niveau communautaire, dans le but essentiellement de prévenir les conflits de décisions, le Règlement 1/2003 dans son article 16-1 dispose que lorsque les juridictions nationales statuent en application des articles 81 et 82 du traité sur des pratiques qui font déjà l’objet

21 voir l’article 33,alinéa 4 de la GWB (« Gesetz gegeb Wettbewerbsbeschränkungen »). 22 Si cette voie était suivie en France, il conviendrait de déterminer avec précision quelles sont les décisions de l’autorité de concurrence qui seraient couvertes par cette autorité de chose décidée.

18

d’une décision de la Commission, elles ne peuvent prendre de décisions qui iraient à l’encontre de la décision adoptée par la Commission ( )23 .

82. Toutefois, si les actions consécutives présentent de nombreux avantages en terme d’articulation des procédures, elles peuvent présenter l’inconvénient d’un allongement substantiel de la durée du contentieux. Dans tout contentieux et sûrement plus encore en matière de contentieux économique, le facteur temps est un élément fondamental tant pour la réunion des preuves que pour l’efficacité de la procédure. Dans le passé, les affaires « Peugeot c/Eco system » et « Mors Labinal », par leur durée de traitement de plus de dix ans, ont illustré ce problème ( )24 . Aussi, afin de ne pas perdre les bénéfices qui peuvent être tirés de procédures consécutives (« follow on »), il conviendrait de réfléchir, dans le cadre de la récente spécialisation des tribunaux, à la mise en place d’un traitement prioritaire des affaires dont les pratiques ont déjà été condamnées par l’autorité spécialisée.

83. Cette réflexion a été menée par les Britanniques à l’occasion de l’adoption, en 2002, de « l’Enterprise Act » en matière de concurrence. Dans le but de mettre en place une procédure plus rapide et moins coûteuse en matière de réparation des dommages causés par de pratiques anticoncurrentielles, ce texte modifie le rôle du « Competition Appeal Tribunal » (CAT) ( )25 , qui exerce à titre principal un contrôle juridictionnel sur les décisions prises en matière de pratiques anticoncurrentielles par l’autorité de la concurrence britannique, l’OFT (Office of Fair Trading). Sur le fondement de la section 47 A de l’ «Enterprise Act », il est désormais possible de demander directement devant le CAT des dommages et intérêts en réparation d’un préjudice causé par des pratiques anticoncurrentielles, tant nationales que communautaires. Une telle action en dommages et intérêts ne peut intervenir qu’à la suite d’une décision de condamnation devenue définitive émanant de l’OFT, du CAT ou de la Commission européenne ( )26 . Il convient de relever ici que l’existence de cette procédure particulière en réparation devant le CAT n’empêche pas la victime d’agir devant les juridictions de droit commun.

84. Ces dispositions peuvent contribuer à ouvrir les actions en réparation aux acheteurs indirects. On voit effectivement l’intérêt que peut recouvrir, pour les consommateurs victimes d’une pratique anticoncurrentielle, cette voie d’action supplémentaire.

85. S’agissant particulièrement des actions collectives de consommateurs telles qu’introduites en droit britannique en 2002 ( )27 , dans le cas où l’OFT a préalablement rendu une décision de condamnation, le fait, pour un groupe de consommateurs, de pouvoir porter une action directement devant le CAT, constitue certainement un avantage.

2. Les actions parallèles

86. Dans le cas d’actions parallèles entre l’autorité de la concurrence et le juge civil, deux difficultés principales peuvent se présenter.

23 L’article 16-1 du Règlement opère une codification des principes dégagés par la Cour de Justice dans son arrêt Masterfood du 14 décembre 2000. 24 Ces affaires, citées dans le rapport Ashurst, ont donné lieu à une décision de la Commission puis ont fait l’objet d’actions en DI qui ont duré plus de 10 ans. 25 Le Competition Appeal Tribunal a été créé par le Competition Act de 1998. 26 Voir la section 47 A (6) à (8) de l’Enterprise Act de 2002. 27 Voir la section 47 B de l’Enterprise Act, l’action collective des consommateurs doit être exercée par un organisme agréé (specified body) qui agit pour le compte d’au moins deux consommateurs, ayant donné leur consentement.

19

87. Tout d’abord, la concomitance des procédures peut conduire à des risques de décisions divergentes entre l’autorité publique et le juge de la réparation. En effet, en droit français, aucune règle ne recommande au juge civil d’éviter de prendre une décision qui irait à l’encontre d’une décision envisagée dans une procédure pendante devant le Conseil de la concurrence. Le droit communautaire de la concurrence, quant à lui, a prévu, à l’article 16-1 du Règlement 1/2003 comme expliqué ci-dessus, des dispositions particulières ayant pour objet de prévenir les conflits entre une décision prise par la Commission et un jugement que le juge s’apprête à rendre. L’article 16-1 va plus loin en considérant que les juridictions nationales doivent également éviter de prendre des décisions qui iraient à l’encontre de la décision envisagée dans une procédure intentée par la Commission. A cet effet, ce texte préconise que la juridiction nationale qui a connaissance d’une instruction de la Commission concomitante à sa propre instruction, évalue s’il est nécessaire de suspendre sa procédure. Dans ce cas, afin d’éviter un allongement de la procédure, la Commission s’engage à traiter l’affaire en cause de manière prioritaire ( )28 .

88. Dans les affaires où il existe un risque réel de divergence entre la décision à venir du Conseil de la concurrence et celle du juge civil, il pourrait être recommandé au juge de suspendre la procédure, à charge pour le Conseil de faire toute diligence pour accélérer la procédure pendante devant lui. Une telle solution paraît raisonnable et moins radicale que celle adoptée par le droit espagnol consistant à remettre en cause le système de compétences parallèles en ne permettant au juge d’intervenir qu’à la condition que l’autorité de la concurrence ait préalablement rendu une décision ( )29 . Par ailleurs, le droit belge encadre également la compétence du juge civil mais de manière un peu moins stricte : toute juridiction devant laquelle la solution du litige dépend du caractère licite d’une restriction de concurrence, doit surseoir à statuer et poser une question préjudicielle à la Cour d’Appel de Bruxelles, qui est aussi exclusivement compétente pour examiner les recours contre les décisions de l’autorité de la concurrence belge ( )30 .

89. La règle adoptée au niveau communautaire semble donc la plus souhaitable en ce qu’elle préserve la compétence naturelle du juge de droit commun en matière de concurrence tout en permettant d’assurer une cohérence dans l’application de la règle de droit.

90. Il convient de noter qu’avec la récente spécialisation des tribunaux en matière de concurrence, les risques de conflit seront réduits et en tout état de cause, limités dans le temps, dans la mesure où ne subsiste en appel qu’une seule Cour, la cour d’appel de Paris qui juge également les recours contre les décisions du Conseil de la concurrence.

91. A côté du risque de conflit de décisions, une autre difficulté, liée à l’articulation, peut naître de la concomitance des procédures publiques et privées. Par exemple, si le plaignant dévoile l’existence de la procédure devant le Conseil en faisant une demande en dommages et intérêts devant le juge civil, alors même que le Conseil de la concurrence, saisi des mêmes pratiques, est en train d’organiser des visites et saisies, il existe alors un risque d’alerte des opérateurs et de déperdition des preuves, et, in fine, d’inefficacité de l’action de l’autorité publique non compensée par l'action efficace du juge civil. Lorsque l’autorité spécialisée est

28 Voir point 12 de la Communication de la Commission sur la coopération avec les juridictions pour l’application des articles 81 et 82 du traité CE. 29 En droit espagnol, ces limitations de la compétence du juge ne s’appliquent pas en cas de violation du droit communautaire. En effet, une telle limitation irait à l’encontre de l’effet direct reconnu par la Cour de Justice aux articles 81 et 82 du traité CE. Par ailleurs, le droit espagnol de la concurrence est en voie de révision. 30 Voir l’article 42 bis de la loi belge sur la protection de la concurrence économique du 1er juillet 1999.

20

saisie, dans certains cas, une phase de discrétion s’impose donc sauf à anéantir les chances de succès des procédures, aussi bien publiques que privées.

3. Les actions autonomes ( dites en « stand alone »)

92. Enfin, dans l’hypothèse d’actions civiles en dommages et intérêts exercées à titre principal et de manière autonome par rapport à l’action publique, les difficultés d’articulation entre les procédures publiques et privées tiennent principalement aux insuffisances du système de la responsabilité civile lui-même.

93. En effet, le contentieux des pratiques anticoncurrentielles est un contentieux technique qui nécessite une expertise spécifique et des moyens d’investigation particuliers. Et lorsqu’un opérateur sur le marché ou un consommateur fait le choix d’engager directement une procédure judiciaire, sans saisir au préalable le Conseil de la concurrence ou la Commission européenne, il va se trouver placé face à une double contrainte : celle de faire la preuve des pratiques qu’il allègue, preuve qui en matière de concurrence peut être une charge très lourde, notamment si les pratiques en cause sont occultes. La seconde difficulté consiste à démontrer l’atteinte causée à la concurrence par les faits établis. Cette démonstration suppose que le plaignant fournisse au juge ou que celui-ci puisse obtenir et utiliser les éléments d’informations permettant une analyse du marché et la mesure de l’incidence économique des comportements dénoncés sur ce marché.

94. Certes, eu égard aux spécificités du contentieux de la concurrence et à son caractère d’ordre public, le droit communautaire comme le droit français de la concurrence ont prévu des dispositions permettant au juge de demander l’assistance des autorités de la concurrence. En droit français, le ministre chargé de l’économie a le pouvoir d’intervenir devant les juridictions civiles ou pénales en déposant des conclusions et en produisant des procès-verbaux ou des rapports utiles à la solution du litige. Le Conseil de la concurrence peut être consulté par les juridictions sur les pratiques anticoncurrentielles nationales ou communautaires relevées dans les affaires dont elles sont saisies (art. L. 462-3 du code de commerce).

95. En droit communautaire, avant même la réforme opérée par le Règlement 1/2003 dans le cadre du processus de modernisation du droit de la concurrence, le Traité de Rome a donné au juge la possibilité de faire appel à la Commission pour l’éclairer dans la mise en œuvre du droit communautaire en lui fournissant notamment, lorsque ces renseignements sont disponibles, des données relatives à la situation des marchés ou des secteurs économiques concernés ( )31 . Pour l’application du droit communautaire, le juge peut également recourir au mécanisme de la question préjudicielle de l’article 234 du traité CE.

96. Toutefois, afin que le système combinant action publique et action privée soit équilibré et, par-là même, efficace, de telles sollicitations par les juges agissant dans le cadre d’actions privées individuelles ou collectives, de l’avis et de l’expertise des autorités publiques de la concurrence pour qualifier les pratiques qui leur sont soumises, ne doivent pas prendre un caractère systématique.

97. En effet, d’une part, les autorités de la concurrence ont pour mission, avec les ressources qui restent limitées, de mener une véritable action publique en matière de concurrence et ne doivent pas être limitées dans la mise en œuvre de cette mission première par un nombre trop important de demandes d’avis visant à trancher un litige privé.

31 Voir l’article 5 du Traité de Rome et la Communication du 13 février 1993 relative à la coopération entre la Commission et les juridictions nationales pour l’application des articles 85 et 86 du Traité CE.

21

98. D’autre part, les mécanismes de coopération existants, tant au niveau national que communautaire, n’ont pas pour objet de reporter sur les autorités de la concurrence le travail d’instruction et d’analyse qui incombe au juge. En réalité, et les dispositions introduites par le Règlement 1/2003 en matière de coopération des autorités de la concurrence avec les juges en témoignent, les procédures d’assistance au juge s’inscrivent essentiellement dans une volonté de garantir une application cohérente du droit de la concurrence. A cet égard, le Règlement 1/2003, à côté de la possibilité pour le juge de solliciter l’assistance des autorités de la concurrence, introduit dans son article 15-3 le droit, pour les autorités nationales de la concurrence, agissant d’office, en d’autres termes sans avoir été préalablement sollicitées par le juge, d’intervenir dans un litige civil ou un procès pénal « au sujet de l’application des articles 81 et 82 du Traité » ( )32 . Le Règlement qui confère ce même pouvoir d’intervention d’office à la Commission est explicite sur le but de cette coopération. Il précise en effet que la Commission peut agir d’office « lorsque l’application cohérente des articles 81 et 82 du traité l’exige ».

99. Le nouveau système de mise en œuvre des articles 81 et 82 du traité CE instauré par le Règlement 1/2003, met l’accent sur la fonction essentielle que remplissent les juridictions nationales dans la préservation des droits subjectifs prévus par le droit communautaire, en complément du rôle joué par les autorités nationales. A ce titre, le Règlement 1/2003 fait désormais obligation aux juridictions d’appliquer le droit de la concurrence communautaire lorsqu’elles appliquent les règles nationales de concurrence aux pratiques susceptibles d’affecter le commerce entre Etats-membres.

100. Dans ce contexte, on comprend encore mieux que les autorités de la concurrence n’ont nullement vocation à se substituer au juge en donnant des avis ou des expertises.

101. L’ensemble de ces éléments montre que si l’on souhaite que le juge joue pleinement le rôle qui lui est assigné en matière d’action privée et que les autorités spécialisées continuent efficacement leur action publique en matière de concurrence, ces dernières ne doivent pas être amenées à consacrer une part trop importante de leur activité à répondre au juge dans le cadre de demandes d’avis trop systématiques. Celles-ci doivent principalement agir auprès des juridictions pour contribuer à assurer la cohérence dans l’application des règles de concurrence dans une perspective « publique et privée ».

102. A cet effet, le juge doit pouvoir disposer des moyens suffisants pour mettre en œuvre une action privée efficace en matière de concurrence. Au sein des actions privées, l’action collective de consommateurs victimes de pratiques anticoncurrentielles, requiert des moyens spécifiques, d’autant plus importants que cette action est menée à titre principal devant le juge civil (stand alone). A cet effet, comme expliqué aux § 81 et suivants, les difficultés propres à l’action civile devraient faire l’objet de choix de politique juridique permettant le renforcement de l’efficacité de l’action en dommages et intérêts au sein de laquelle une place pourrait être faite à l’action collective des consommateurs.

32 Voir l’article 15-3 du Règlement 1/2003.

22

Conclusion 103. Dans le cadre du mouvement tendant au renforcement ou à la création de mécanismes de

recours fondés sur l’action privée dans les espaces juridiques nationaux et en Europe, le Conseil de la concurrence est favorable à ce que soient permises en France les actions de groupe des consommateurs en matière de concurrence, afin de faire de ces derniers de véritables acteurs de la politique de concurrence elle-même recentrée sur la protection de leur bien-être. Nul doute que si l'on veut renforcer la confiance des consommateurs dans l'économie de marché, encore fragile et parfois vacillante en France comme le montrent certaines études récentes, il faut donner à ceux qui les représentent les moyens de pouvoir lutter eux-mêmes, par les voies juridiques les plus appropriées, contre les dérives ou les abus constatés sur les marchés et de permettre au consommateur individuel de toucher concrètement les bénéfices d'une telle politique.

104. Toutefois, cette démarche nécessite des orientations et un encadrement des modalités de mise en œuvre. A cet égard, si pour les raisons évoquées ci-dessus, on en vient à favoriser les actions complémentaires, sans toutefois exclure les procédures entièrement autonomes, il est nécessaire de prendre toutes les mesures utiles à la préservation de la complémentarité de l’action publique et des actions privées, chacune ayant une mission indispensable à la sauvegarde du bien-être du consommateur.

105. Une attention particulière doit, dans le même esprit, être portée à la protection des programmes de clémence, qui constituent un outil indispensable de lutte contre les ententes et les cartels. Sur tous ces points, il existe de multiples options juridiques et procédurales qui autorisent la création d’une forme d’action de groupe capable de poursuivre avec efficacité un double but de dissuasion et de réparation, tout en évitant les effets pervers et la judiciarisation excessive de la vie économique qui a pu être observée dans d’autres pays.

Délibéré sur le rapport oral de Mme Daudret-John, par M. Lasserre, président, Mmes Aubert, Perrot, et M. Nasse, vice-présidents.

Le rapporteur général,

Thierry Dahan

Le président,

Bruno Lasserre

© Conseil de la concurrence

23

Annexe 3

FR FR

COMMISSION EUROPÉENNE

Bruxelles, le 4 février 2011 SEC(2011) 173 final

DOCUMENT DE TRAVAIL DES SERVICES DE LA COMMISSION

CONSULTATION PUBLIQUE:

Renforcer la cohérence de l'approche européenne en matière de recours collectifs

FR 2 FR

1. INTRODUCTION

1.1. Le recours collectif comme instrument pouvant permettre de renforcer le contrôle de l'application du droit de l'Union européenne

1. L'application effective du droit de l'Union est d'une importance capitale pour les citoyens comme pour les entreprises. Ainsi que la stratégie Europe 20201 et le programme de Stockholm2 le soulignent, l'Union européenne doit faire en sorte que ses citoyens et ses entreprises, notamment les petites et moyennes entreprises (PME), puissent exploiter dans la pratique les possibilités que leur offrent le marché unique et l'espace européen de justice. Des droits qui ne pourraient être exercés dans la pratique n'auraient pas de raison d'être. En cas de violation de droits substantiels de l'Union, les citoyens et les entreprises doivent être en mesure de faire respecter les droits que leur confère la législation de l'Union.

2. Aux fins de l'application effective du droit de l'Union en pareils cas, le contrôle public qu'exerce la Commission européenne (par exemple, dans les recours en manquement ou les procédures en matière de concurrence) et qui a souvent comme points de départ des plaintes de citoyens ou d'entreprises, constitue un instrument essentiel. En tant que gardienne des traités, la Commission doit veiller à ce que non seulement l'intérêt des personnes privées, mais aussi l'intérêt public et, au sens large, celui de l'Union soient pris en compte. Les autorités nationales jouent également un rôle important dans le contrôle public de l'application du droit de l'Union, notamment dans les domaines du droit de la concurrence, du droit des consommateurs et du droit de l'environnement, et la législation de l'Union existante renforce les possibilités de coopération transfrontière entre ces mêmes autorités pour lutter contre les infractions3.

3. Sous l'effet de l'élargissement de l'Union européenne, le nombre de cas de non-respect du droit de l'Union a sensiblement augmenté du fait de l'extension du champ d'application territorial du droit de l'Union, rendant ainsi d'autant plus nécessaire un contrôle plus décentralisé de l'application du droit de l'Union. Ce phénomène a aussi mis à l'ordre du jour la question de savoir si d'autres mécanismes de contrôle par la sphère privée devraient compléter le système actuel de voies de recours de l'Union afin de renforcer l'application du droit de l'Union.

4. Le recours individuel est le premier moyen à la disposition de la sphère privée pour faire respecter le droit de l'Union: des personnes physiques ou morales peuvent engager, à titre individuel, une action en justice pour faire valoir les droits que leur confère le droit de l'Union. De récents règlements européens ont institué des procédures accélérées qui permettent aux parties d'obtenir rapidement un titre

1 COM(2010) 2020 du 3.3.2010. 2 Document n° 17024/09 du Conseil, adopté par le Conseil européen des 10 et 11 décembre 2009. 3 Aussi le règlement (CE) n° 2006/2004 relatif à la coopération en matière de protection des

consommateurs établit-il un cadre général aux fins de la coopération entre les autorités publiques chargées de l'application de la législation. Dans le domaine du droit de la concurrence, un réseau européen de la concurrence a été inauguré lors de l'entrée en vigueur du règlement (CE) n° 1/2003 relatif à la mise en œuvre des règles de concurrence prévues aux articles 81 et 82 du traité [désormais articles 101 et 102 du TFUE].

FR 3 FR

exécutoire dans les petits litiges transfrontaliers et pour les créances incontestées4. En outre, des garanties procédurales sont accordées aux parties qui tentent de résoudre à l'amiable leur litige transfrontalier en recourant à la médiation5. Des normes minimales communes relatives à l'aide judiciaire garantissent un accès effectif à la justice dans les litiges transfrontaliers, indépendamment des ressources financières du demandeur6. Cependant, lorsqu'un grand groupe de citoyens et d'entreprises sont victimes de la même infraction au droit de l'Union, les recours individuels ne constituent pas, dans bien des cas, une voie de recours efficace pour mettre un terme aux pratiques illégales en cause ou obtenir réparation du préjudice causé par ces pratiques: les citoyens et les entreprises hésitent souvent à engager des poursuites contre des pratiques illégales, notamment lorsque le préjudice individuel n'est guère important par rapport au coût d'une action en justice. C'est ainsi que des pratiques illégales perdurent et font subir un grave préjudice cumulé aux citoyens et entreprises de l'Union. De surcroît, ainsi que le reconnaît la stratégie numérique pour l'Europe7, l'application du droit de l'Union dans l'environnement numérique semble parfois difficile en raison du manque de clarté quant aux règles applicables, en particulier pour les consommateurs. L'insécurité juridique et la difficulté supposée des recours sapent la confiance de ces derniers et constituent ainsi un obstacle au développement du commerce électronique transfrontière.

5. En outre, lorsque des infractions au droit de l'Union provoquent des recours individuels en chaîne, le droit procédural de nombreux États membres, bien souvent, ne permet pas aux juridictions de traiter tous les recours efficacement et dans un délai raisonnable. Ce constat vaut pour les actions collectives en cessation, mais aussi et surtout pour les actions en réparation.

6. Pour ces motifs, des mécanismes de recours collectif pourraient être envisagés afin de remédier aux carences observées actuellement dans le contrôle de l'application du droit de l'Union.

1.2. Qu'entend-on par «recours collectif»?

7. Les citoyens et les entreprises de l'Union devraient pouvoir intenter une action en justice lorsqu'ils subissent un préjudice du fait d'une infraction à une législation de l'UE instituant des droits substantiels. Lorsque des citoyens et des entreprises sont victimes de la même infraction, commise par la même entreprise, la jonction de leurs plaintes dans une procédure de recours collectif unique, ou leur représentation en justice, aux fins de ce recours, par une entité représentative ou un organisme agissant dans l'intérêt public, pourraient simplifier la procédure et réduire les frais. Le «recours collectif» est une notion large englobant tout mécanisme tendant à faire cesser ou à prévenir des pratiques commerciales illégales affectant un grand nombre de plaignants ou encore à obtenir la réparation du préjudice causé par de telles

4 Le règlement (CE) n° 805/2004 portant création d'un titre exécutoire européen pour les créances

incontestées, le règlement (CE) n° 1896/2006 instituant une procédure européenne d’injonction de payer et le règlement (CE) n° 861/2007 instituant une procédure européenne de règlement des petits litiges.

5 Cf. directive 2008/52/CE sur certains aspects de la médiation en matière civile et commerciale. 6 Directive 2003/8/CE visant à améliorer l'accès à la justice dans les affaires transfrontalières par

l'établissement de règles minimales communes relatives à l'aide judiciaire accordée dans le cadre de telles affaires.

7 COM(2010) 245 du 19.5.2010.

FR 4 FR

pratiques. Il existe essentiellement deux types de recours collectif: les actions en cessation, par lesquelles les plaignants cherchent à obtenir la cessation d'un comportement illégal, et les actions en réparation, par lesquelles ils cherchent à obtenir des dommages-intérêts à titre d'indemnisation du préjudice subi. Les procédures de recours collectif peuvent revêtir diverses formes, telles que des mécanismes extrajudiciaires de règlement des litiges ou le recours à un organisme public ou une autre entité représentative pour intenter l'action collective.

1.3. Formes de recours collectif existant dans l'Union européenne

8. La notion de recours collectif n’est pas neuve dans l’Union. La législation de l'UE et les accords internationaux existants imposent aux États membres de prévoir la possibilité d'intenter des actions collectives en cessation dans certains domaines. En conséquence, tous les États membres ont mis en place des procédures permettant de solliciter le prononcé d'une injonction de cesser des pratiques illégales. Dans le domaine du droit des consommateurs, la directive sur les actions en cessation8 a conféré aux autorités de protection des consommateurs et aux associations de consommateurs la qualité pour intenter des actions en cessation contre les pratiques qui enfreignent les règles nationales et européennes en la matière dans tous les États membres. Dans le domaine de l'environnement, la convention Aarhus fait obligation aux États membres de garantir l'accès à la justice en cas d'infraction aux normes environnementales. Tous les États membres s'y sont conformés en instaurant l'une ou l'autre forme de recours collectif en cessation, conférant aux organisations non gouvernementales la qualité pour contester les décisions administratives dans le domaine de l'environnement.

9. La majorité des États membres ont mis en place, dans certains domaines, des procédures de recours collectif en réparation. Les mécanismes existant pour l'indemnisation d'un groupe de victimes lésées par des pratiques commerciales illégales diffèrent largement au sein de l'Union9. Fondamentalement, chaque régime national d'indemnisation est unique, et il n'est pas deux dispositifs identiques dans ce domaine. Certaines procédures ne s'appliquent qu'à des secteurs bien définis (par exemple, la récupération des pertes sur investissements en Allemagne ou les préjudices causés par les pratiques anticoncurrentielles au Royaume-Uni), tandis que d'autres ont une portée plus vaste (les procédures espagnoles d'action collective, par exemple). Une deuxième différence porte sur la qualité pour agir dans les procédures de recours en réparation, un certain nombre d'États membres ayant dévolu le pouvoir d'engager des procédures dans certains domaines à des autorités publiques (par exemple, le médiateur en Finlande), alors que d'autres confèrent la qualité pour agir à des organisations privées telles que les associations de consommateurs (cas de la Bulgarie, par exemple) ou à des particuliers agissant au nom d'un groupe (au Portugal). De nombreux États membres appliquent une combinaison de plusieurs règles en la matière. Une différence supplémentaire réside dans la catégorie de victimes autorisées à intenter ces recours collectifs en réparation. La plupart des régimes nationaux évoqués plus haut ouvrent ces recours aux consommateurs, mais

8 JO L 166 du 11.6.1998, p. 51. 9 Cf., par exemple, l'étude de 2008 intitulée «Evaluation of the effectiveness and efficiency of collective

redress mechanisms in the European Union», commandée par la Commission européenne et disponible à l'adresse suivante: http://ec.europa.eu/consumers/redress_cons/collective_redress_en.htm#Studies

http://ec.europa.eu/consumers/redress_cons/collective_redress_en.htm#Studies

FR 5 FR

seuls quelques-uns étendent cette possibilité à d'autres victimes, comme les petites entreprises. Les différences se rapportent aussi à l'effet du jugement sur les membres du groupe concerné: dans la majorité des États membres, la décision n'est contraignante que pour ceux qui ont expressément choisi d'être partie à la procédure (cas de la Suède et de l'Italie, par exemple), tandis que, dans un petit nombre de cas, elle lie la totalité des membres du groupe, à moins qu'ils n'aient exprimé leur volonté de ne pas s'associer à la procédure (Portugal, Danemark, Pays-Bas). En outre, il existe des différences en ce qui concerne le moment auquel les personnes ayant le droit d'engager une action sont identifiées individuellement; dans certains États membres, cette identification doit avoir lieu lorsque l'action représentative est entamée (au Royaume-Uni par exemple), alors qu'elle peut intervenir ultérieurement dans d'autres cas (Pologne, Espagne). On note enfin des différences sensibles dans le financement des actions collectives, la répartition des indemnités obtenues et le recours aux modes alternatifs de règlement des litiges. L'incidence de toute mesure européenne (éventuelle) sur les systèmes juridiques nationaux dépendrait de l'existence ou non d'un régime de recours collectifs dans l'État membre concerné et des caractéristiques propres de ce régime.

1.4. Renforcer la cohérence de l'approche européenne en matière de recours collectifs

10. Eu égard à la diversité des régimes nationaux existants et à leurs différents degrés d'efficacité, un manque de cohérence dans l'approche européenne en matière de recours collectifs pourrait empêcher les citoyens et les entreprises de l'Union de jouir pleinement de leurs droits et entraîner des disparités dans l'exercice de ces droits. L'adoption d'un cadre européen cohérent, s'appuyant sur les différentes traditions nationales, pourrait faciliter le renforcement du recours collectif (actions en cessation et/ou en réparation) dans des domaines ciblés. En tout état de cause, ce cadre devrait définir des principes communs que toute initiative (éventuelle) de l'Union en matière de recours collectifs devrait respecter, quel que soit le domaine concerné. L'objectif est de garantir, dès le début, que toute proposition (éventuelle) en la matière, tout en visant à assurer une application plus effective du droit de l'Union, s'inscrive bien dans la tradition juridique de l'Union et l'ensemble des voies de recours déjà disponibles à cet effet.

11. Les services de la Commission européenne mènent depuis plusieurs années des travaux afin de définir des normes européennes en matière d'actions collectives en réparation dans les domaines du droit des consommateurs et du droit de la concurrence. La Commission a adopté un livre vert sur les actions en dommages et intérêts pour infraction aux règles communautaires sur les ententes en 200510 et un livre blanc en 200811. En 2008, elle a également publié un livre vert sur les recours collectifs pour les consommateurs12. La position des parties prenantes sur de nombreux aspects du problème est connue: la plupart des associations de consommateurs sont favorables aux mécanismes de recours collectif en réparation au niveau de l'UE, tandis que de nombreux représentants de l'industrie craignent une multiplication des recours abusifs. Cependant, les parties prenantes ont aussi mis la Commission en garde contre une possible incohérence entre ses différentes initiatives

10 COM(2005) 672 du 19.12.2005. 11 COM(2008) 165 du 2.4.2008. 12 COM(2008) 794 du 27.11.2008.

FR 6 FR

en matière de recours collectifs, ce qui appelle donc un renforcement de la cohérence à cet égard.

12. C'est pourquoi la Commission lance une consultation publique horizontale intitulée «Renforcer la cohérence de l’approche européenne en matière de recours collectifs». Cette consultation a notamment pour objet de dégager des principes juridiques communs en matière de recours collectifs. Elle devrait également permettre de réfléchir à la manière d'intégrer ces principes communs dans le système juridique de l'Union et dans les ordres juridiques des vingt-sept États membres, et de déterminer dans quels domaines différentes formes de recours collectif (actions collectives en cessation et/ou actions collectives en réparation) pourraient apporter une valeur ajoutée pour l'amélioration du contrôle de l’application de la législation de l'Union ou pour un renforcement de la protection des droits des victimes. La série de principes qui en découlera devrait guider toute initiative législative éventuelle de l'Union en matière de recours collectifs.

2. VALEUR AJOUTEE POTENTIELLE DU RECOURS COLLECTIF POUR AMELIORER LE CONTROLE DE L'APPLICATION DU DROIT DE L'UNION

13. Il y a lieu d'examiner avec attention la question de savoir si une initiative européenne apporterait une valeur ajoutée aux fins du renforcement du contrôle de l'application du droit de l'Union et - dans l'affirmative - dans quels domaines, et si d'autres voies pourraient être envisagées pour combler les éventuelles lacunes du système actuel. Dans ce contexte, les récentes avancées législatives européennes qui sont exposées ci-dessus devraient être prises en compte. Il conviendrait également d'examiner si les lacunes actuelles pourraient être comblées en étendant à d'autres domaines la possibilité d'action collective en cessation qui existe en matière de protection des consommateurs. En outre, devrait être posée la question de savoir si le rôle des organismes publics nationaux (tels que le médiateur) et/ou celui des organisations représentatives privées dans le contrôle de l'application du droit de l'Union pourraient êtres renforcés, dans le respect des modèles nationaux existants.

14. Toute initiative nouvelle devrait également respecter les principes de subsidiarité et de proportionnalité énoncés à l'article 5 du traité sur le fonctionnement de l'Union européenne. Certains mécanismes d'action collective en cessation sont déjà en place dans tous les États membres dans le domaine de la protection des consommateurs, tandis que plusieurs États membres connaissent, à des degrés divers, d'autres formes de recours collectif (notamment l'action en réparation). Il conviendrait d'examiner la nécessité, dans ces conditions, d'une action au niveau de l'Union pour assurer une application effective du droit de l'Union. Qui plus est, toute action au niveau de l'Union devrait tenir compte de la dimension transfrontière spécifique du recours collectif (action en cessation et/ou action en réparation).

Questions:

Q 1 Quelle valeur ajoutée apporterait l'instauration de nouveaux mécanismes de recours collectif (action en cessation et/ou action en réparation) aux fins du contrôle de l'application du droit de l'Union?

FR 7 FR

Q 2 Le recours collectif par la sphère privée (ci-après «recours collectif privé») devrait-il être indépendant du contrôle public de l'application du droit, ou bien constituer un mécanisme complémentaire ou subsidiaire? Une coordination entre recours collectif privé et contrôle public de l'application du droit est-elle nécessaire? Dans l'affirmative, comment cette coordination peut-elle être assurée? À votre avis, existe-t-il des exemples dans les États membres ou dans des pays tiers que vous jugez particulièrement instructifs dans la perspective d'une éventuelle initiative européenne?

Q 3 L'Union européenne devrait-elle renforcer le rôle des organismes publics nationaux et/ou des organisations représentatives privées dans le contrôle de l'application du droit de l'Union? Dans l'affirmative, de quelle manière et dans quels domaines?

Q 4 À votre avis, que faudrait-il pour que l'action au niveau européen en matière de recours collectifs (actions en cessation et/ou actions en réparation) soit conforme aux principes du droit de l'Union, tels que la subsidiarité, la proportionnalité et l'efficacité? Votre réponse serait-elle différente selon le domaine considéré?

Q 5 Suffirait-il d'étendre à d'autres domaines le champ d'application des règles de l'Union existant en matière d'actions collectives en cessation? Conviendrait-il d'instaurer des mécanismes d'action collective en réparation au niveau de l'Union?

Q 6 Toute éventuelle action européenne devrait-elle être fondée sur une approche contraignante ou, au contraire, non contraignante (telle qu'un guide de bonnes pratiques)? Quels sont, selon vous, les avantages et les risques respectifs de chaque approche? Votre réponse différerait-elle en fonction du domaine considéré?

3. PRINCIPES GENERAUX POUVANT GUIDER D'EVENTUELLES INITIATIVES DE L'UNION EN MATIERE DE RECOURS COLLECTIFS

15. Sur la base des résultats des consultations précédentes, il est possible de dégager une première série de principes fondamentaux communs, susceptibles d'orienter d'éventuelles initiatives européennes en matière de recours collectifs (actions collectives en cessation et/ou actions collectives en réparation), parmi lesquels: (1) la nécessité de garantir un recours effectif et efficace; (2) l'importance de l'information et du rôle des entités représentatives; (3) la nécessité de tenir compte de la résolution consensuelle collective comme mode alternatif de règlement des litiges; (4) le besoin de garanties solides pour éviter les recours abusifs; (5) la disponibilité de mécanismes de financement appropriés, notamment pour les citoyens et les PME; (6) l'importance d'une application effective du droit de l'Union dans toute l'UE. Ces principes pourraient s'appliquer à toutes les formes de recours collectif (actions collectives en cessation et/ou actions collectives en réparation), bien que certains puissent être plus pertinents pour les actions collectives en réparation.

Questions:

Q 7 Partagez-vous l'opinion que toute initiative européenne en matière de recours collectifs (actions collectives en cessation et/ou actions collectives en réparation) devrait satisfaire à une série de principes communs définis au niveau de l'Union?

FR 8 FR

Quels devraient être ces principes? À quel principe attacheriez-vous une importance particulière?

Q 8 Comme indiqué ci-dessus, plusieurs États membres ont pris des initiatives en matière de recours collectifs. L'expérience acquise jusqu'ici par ces États membres pourrait-elle être mise à profit pour définir une série de principes au niveau de l'Union?

Q 9 Estimez-vous que toute initiative européenne devrait présenter certaines caractéristiques pour pouvoir garantir un accès effectif à la justice tout en tenant dûment compte des traditions juridiques de l'Union et des ordres juridiques des vingt-sept États membres?

Q 10 Avez-vous connaissance de l'existence de bonnes pratiques en matière de recours collectifs dans un ou plusieurs États membres dont l'Union européenne ou d'autres États membres pourraient tirer des enseignements? Veuillez expliquer pourquoi vous jugez ces pratiques particulièrement utiles. Existe-t-il, en revanche, des pratiques nationales qui ont posé ou posent problème et comment ces problèmes ont-ils pu ou pourraient-ils être surmontés?

3.1 La nécessité de garantir un recours effectif et efficace

16. Toute initiative européenne en matière de recours collectifs devrait, en premier lieu, garantir un fonctionnement effectif et efficace du système mis en place, quel qu'il soit. Le droit a un recours effectif constitue un droit fondamental: la Charte des droits fondamentaux de l'Union européenne dispose, en effet, que toute personne dont les droits et libertés garantis par le droit de l'Union ont été violés a droit à un recours effectif devant un tribunal13. En cas de plaintes multiples, la jonction de plaintes individuelles dans une procédure unique de recours collectif ou l'introduction d'une plainte par une entité représentative (par exemple, médiateur, association de consommateurs ou association professionnelle) devraient permettre de réduire les frais supportés par les parties et d'améliorer l'efficacité des recours tant judiciaires qu'extrajudiciaires. Il conviendrait d'éviter la répétition de recours identiques ou similaires et le risque de décisions contradictoires. Un système de recours collectifs qui entraînerait des procédures longues et coûteuses ne serait ni dans l'intérêt des consommateurs ni dans celui des entreprises et devrait donc être exclu. Pour être effectif et efficace, un système de recours collectif devrait par conséquent pouvoir garantir la sécurité juridique et l'adoption de décisions équitables dans un délai raisonnable, tout en respectant les droits de toutes les parties concernées.

Questions:

Q 11 Quelles devraient être, selon vous, les caractéristiques d'un système de recours collectif effectif et efficace? Un mécanisme de recours collectif ouvert aux PME devrait-il présenter des caractéristiques particulières?

13 Article 47, premier alinéa (JO C 364 du 18.12.2000, p. 1). Ce droit est encore renforcé par l'article 19,

paragraphe 1, du traité sur l'Union européenne, qui intègre le principe d'une protection juridictionnelle effective (JO C 306 du 17.12.2007).

FR 9 FR

Q 12 Comment peut-on garantir un recours effectif, tout en évitant des procédures longues et coûteuses?

FR 10 FR

3.2 L'importance de l'information et du rôle des entités représentatives

17. Il conviendrait de déterminer les caractéristiques que devrait présenter un mécanisme de recours collectif pour fonctionner d'une manière effective et efficace. L'information des victimes joue, de toute évidence, un rôle important. Pour être en mesure de regrouper leurs plaintes, les citoyens et les entreprises doivent savoir, d'une part, qu'ils ont été victimes de la même pratique illégale et, d'autre part, qu'ils ont la possibilité d'engager une action collective ou de devenir partie à une procédure en cours. Or, cela peut s'avérer particulièrement difficile lorsque les pratiques illégales affectent des victimes dans plusieurs États membres.

18. En outre, il conviendrait de déterminer le rôle que devraient jouer les associations représentant les intérêts des victimes en cas de plaintes multiples, notamment dans un contexte transfrontière. On pourrait ainsi considérer que, pour être efficace, une action collective (en cessation et/ou en réparation) doit être intentée par une entité représentative qui soit à même de représenter, dans son État membre, des victimes d'autres États membres. De même, une entité représentative pourrait être habilitée à représenter des victimes dans des procédures judiciaires ou extrajudiciaires dans un autre État membre.

Questions:

Q 13 Comment, quand et par qui les victimes d'infractions au droit de l'Union devraient-elles être informées des possibilités d'intenter une action collective (en cessation et/ou en réparation) ou de devenir partie à une procédure en cours? Quels seraient les moyens les plus efficaces de s'assurer qu'un maximum de victimes sont informées, notamment lorsque celles-ci sont domiciliées dans plusieurs États membres?

Q 14 Quelle est la meilleure façon de garantir une représentation efficace des victimes, notamment dans les situations transfrontières? Comment pourrait-on faciliter la coopération entre les différentes entités représentatives, notamment dans les affaires transfrontières?

3.3 La nécessité de tenir compte de la résolution consensuelle collective comme mode alternatif de règlement des litiges

19. Les mécanismes de résolution consensuelle collective des litiges complètent utilement les voies de recours judiciaires et peuvent, bien souvent, permettre aux victimes de régler plus rapidement et à un moindre coût leur litige. Les parties doivent dès lors avoir la possibilité de résoudre leur litige collectif par un mécanisme extrajudiciaire, soit en faisant intervenir un tiers (par exemple, au moyen d'un mécanisme de règlement alternatif des litiges, tel que l'arbitrage ou la médiation), soit sans une telle intervention (par exemple, règlement entre les parties concernées). Il y aurait lieu d'explorer les moyens de faciliter le recours aux modes alternatifs de règlement des litiges en cas de plaintes multiples. Il conviendrait également de se demander dans quelle mesure et dans quels domaines un recours judiciaire collectif pourrait être subordonné à une tentative préalable de résolution consensuelle collective du litige.

20. La résolution consensuelle collective d'un litige peut, dans bien des cas, déboucher sur une solution équitable pour toutes les parties en cause, aucune d'entre elles ne

FR 11 FR

devant se sentir contrainte d'accepter une solution qu'elle jugerait inéquitable. Le recours effectif à la résolution consensuelle et l'équité de la solution qui en découle dépendent toutefois, en grande partie, des facteurs pouvant inciter les parties à entamer une telle procédure. L'existence d'un système de recours judiciaire effectif devrait ainsi encourager les parties à trouver un arrangement extrajudiciaire, mécanisme susceptible de permettre le règlement d'un grand nombre de litiges en évitant le recours aux tribunaux. Une initiative sur les modes alternatifs de règlement des litiges, portant sur les modes alternatifs individuels et collectifs de règlement des litiges de consommation, est en cours d'élaboration.

Questions:

Q 15 En dehors de l'existence de voies de recours judiciaires, quels autres facteurs seraient nécessaires pour encourager le recours aux modes alternatifs de règlement des litiges en cas de plaintes multiples?

Q 16 Devrait-on subordonner toute action judiciaire collective en réparation à une tentative préalable de résolution consensuelle collective du litige?

Q 17 Quelle est la meilleure manière de garantir l'équité de la solution obtenue par résolution consensuelle collective? Ce contrôle de l'équité devrait-il incomber aux tribunaux?

Q 18 Conviendrait-il de conférer à la solution obtenue par résolution consensuelle collective un caractère contraignant pour les parties, dans les cas qui ne sont pas actuellement couverts par la directive 2008/52/CE sur certains aspects de la médiation en matière civile et commerciale?

Q 19 D'autres garanties sont-elles nécessaires dans le cadre de la résolution consensuelle collective des litiges pour assurer un accès effectif à la justice?

3.4 De solides garanties contre les recours abusifs

21. Toute initiative européenne sur l'action collective (en cessation et/ou en réparation) devrait, dès le début, prévenir tout risque de recours abusifs. De nombreuses parties prenantes ont exprimé le souhait d'éviter certains abus qui ont été constatés aux États-Unis avec le système des «class actions» (actions de groupe). À cause de ses fortes incitations économiques, ce système encourage en effet les parties à saisir la justice même si le recours n'est pas nécessairement bien fondé. Ces incitations sont le fruit d'une combinaison de facteurs, notamment l'existence de dommages-intérêts punitifs, l'absence de restrictions quant à la qualité pour agir (presque tout le monde pouvant intenter une action en justice pour le compte d'un groupe ouvert de parties lésées), la possibilité de subordonner les honoraires des avocats aux résultats («contingency fees») et la procédure étendue de découverte de la preuve («discovery procedure»). La Commission considère que ces éléments cumulés accentuent le risque de recours abusifs dans une mesure qui n'est pas compatible avec la tradition juridique européenne.

22. Toute initiative européenne en matière d'actions collectives (en cessation et/ou en réparation) ne devrait en aucune manière encourager, pour des raisons économiques,

FR 12 FR

le dépôt de plaintes abusives. Il conviendrait, en outre, de prévoir des garanties effectives pour prévenir les actions collectives abusives. Ces garanties devraient s'inspirer des systèmes nationaux de recours judiciaire dans les États membres de l'Union. Les mécanismes nationaux existants montrent en effet que différentes garanties ou des combinaisons de celles-ci sont possibles.

23. Une garantie courante est le principe selon lequel la partie qui succombe supporte les frais de procédure et les honoraires d'avocats des deux parties.

24. Le plein respect de l'intérêt légitime de toutes les parties constitue une autre garantie importante. Ces droits relèvent du droit à accéder à un tribunal impartial14 et doivent être également protégés dans le cas des recours collectifs.

25. Un recours collectif pouvant revêtir diverses formes, il y aurait lieu de prévoir des garanties permettant de prévenir les risques propres à chacune d'elles. Par exemple, dans les cas où des entités représentatives sont habilitées à intenter une action en justice, il conviendrait d'examiner les conditions d'octroi de la qualité pour agir à ces entités dans les procédures de recours collectif. C'est ainsi que, dans plusieurs États membres, conformément aux dispositions nationales mettant en œuvre la convention Aarhus, les ONG doivent remplir un certain nombre de conditions (certaine durée d'existence, étendue géographique des activités, nombre de membres, objectif de promotion de l'intérêt public, etc.) afin d'obtenir la qualité pour agir dans les procédures collectives.

26. S'agissant des garanties, il conviendra de trouver un juste équilibre entre la prévention des recours abusifs et la protection du droit à un accès effectif à la justice pour les citoyens et les entreprises de l'Union, en particulier les PME. Si cet équilibre s'avère difficile à atteindre d'une manière générale, il pourrait être nécessaire, en dernier ressort, de s'adresser à un juge pour établir cet équilibre dans une affaire pendante donnée.

Questions:

Q 20 Comment pourrait-on protéger d'une manière adéquate l'intérêt légitime de toutes les parties dans les actions collectives (en cessation et/ou en réparation)? Parmi les garanties existant dans les États membres ou les pays tiers, quelles sont celles que vous jugez particulièrement efficaces pour limiter les recours abusifs?

Q 21 Le principe selon lequel la partie qui succombe doit supporter les frais de procédure et les honoraires d'avocats des deux parties devrait-il s'appliquer aux actions collectives (en cessation et/ou en réparation) dans l'Union européenne? Existe-t-il, selon vous, des circonstances justifiant une dérogation à ce principe15? Dans l'affirmative, ces dérogations devraient-elles être rigoureusement circonscrites par la

14 Article 47, deuxième alinéa, de la Charte des droits fondamentaux de l’Union européenne. 15 Voir, par exemple, dans le domaine du droit de l'environnement, l'article 10 bis de la

directive 85/337/CEE et l'article 15 bis de la directive 96/61/CE (telles qu'elles ont, toutes deux, été modifiées par la directive 2003/35/CE) qui interdisent que le coût de ces procédures soit prohibitif pour les ONG. Voir également l'arrêt de la Cour dans l'affaire C-427/07, Commission/Irlande.

FR 13 FR

législation ou devraient-elles être laissées à l'appréciation au cas par cas des tribunaux, éventuellement dans le cadre d'une disposition juridique générale?16

Q 22 Qui devrait être autorisé à former un recours collectif? Le droit de former un recours collectif devrait-il être réservé à certaines entités? Dans l'affirmative, quels sont les critères que ces entités devraient remplir? Veuillez préciser si votre réponse varie en fonction du type de mécanisme de recours collectif et du type de victimes (par exemple, consommateurs ou PME).

16 Article 16 du règlement (CE) n° 861/2007 instituant une procédure européenne de règlement des petits

litiges: «La partie qui succombe supporte les frais de la procédure. Toutefois, la juridiction n'accorde pas à la partie qui a eu gain de cause le remboursement des dépens qui n'étaient pas indispensables ou qui étaient disproportionnés au regard du litige».

FR 14 FR

Q 23 Quel devrait être le rôle du juge dans les procédures de recours collectif? Lorsque des entités représentatives ont qualité pour agir, devraient-elles être reconnues comme «entités représentatives» par un organisme public compétent, ou cette question devrait-elle être laissée à l'appréciation au cas par cas des tribunaux?

Q 24 Quelles autres garanties devraient être prévues par une éventuelle initiative européenne en matière de recours collectifs?

3.5 Des mécanismes appropriés de financement des recours collectifs, notamment pour les citoyens et les PME

27. Les citoyens et les entreprises, en particulier les PME, ne devraient pas être privés de l'accès à la justice, faute de ressources financières suffisantes. Il conviendrait, dès lors, d'examiner dans quelle mesure il existe des mécanismes de financement appropriés aux fins des recours collectifs. Les mécanismes de financement des recours collectifs devraient permettre de financer les recours fondés, sans encourager l'introduction de recours non fondés.

Questions:

Q 25 Comment pourrait-on organiser un financement approprié des actions collectives (en cessation et/ou en réparation), compte tenu notamment de la nécessité d'éviter les recours abusifs?

Q 26 Pourrait-on concevoir des modes de financement privés (tels qu'un financement par des tiers ou une assurance «protection juridique») pour assurer un juste équilibre entre la nécessité de garantir l'accès à la justice et la nécessité d'éviter tout détournement de procédure?

Q 27 Les entités représentatives qui forment des recours collectifs devraient-elles pouvoir recouvrer les frais de procédure, y compris les frais administratifs, auprès de la partie succombante? Existe-t-il par ailleurs d'autres moyens de couvrir les coûts des entités représentatives?

Q 28 Conviendrait-il d'examiner d'autres aspects du financement des recours collectifs pour garantir un accès effectif à la justice?

3.6 Application effective du droit de l'Union dans toute l'UE

28. Dans un marché intérieur des entreprises et des consommateurs, les règles de droit procédural européen en matière civile et les règles de détermination du droit applicable devraient fonctionner efficacement, dans la pratique, pour les actions collectives (qu'il s'agisse d'actions en cessation ou d'actions en réparation), et les décisions devraient être exécutoires dans toute l'Union. Se pose dès lors la question de savoir si les règles européennes actuelles en matière de compétence judiciaire, de reconnaissance et d'exécution des décisions17, ainsi que sur la détermination du droit applicable sont suffisantes à cette fin, ou si une approche européenne cohérente en matière d'actions collectives (en cessation et/ou en réparation) exigerait l'adoption de

17 Cf. le règlement (CE) n° 44/2001 du Conseil sur la compétence judiciaire, la reconnaissance et

l'exécution des décisions en matière civile et commerciale.

FR 15 FR

règles spécifiques supplémentaires en matière de droit applicable et/ou de compétence. Jusqu'ici, les parties prenantes n'ont signalé à la Commission aucun problème pratique dans ce domaine. La présente consultation publique a donc pour objectif de recueillir des avis et des informations sur les difficultés particulières qui pourraient se poser en matière de compétence et de détermination du droit applicable en cas de recours collectif, compte tenu des disparités actuelles entre les systèmes juridiques nationaux, notamment en matière d'actions collectives en réparation, de la nécessité d'un contrôle transfrontière effectif de l'application du droit et d'une prévention des recours abusifs, y compris la course au mieux-disant judiciaire («forum shopping»).

Questions:

Q 29 Avez-vous connaissance de problèmes transfrontières spécifiques qui se seraient posés dans l'application pratique des règles en matière de compétence, de reconnaissance ou d'exécution des décisions? Quelles ont été les conséquences de ces problèmes et quelles solutions ont été finalement trouvées?

Q 30 Des règles particulières en matière de compétence, de reconnaissance et d'exécution des décisions et/ou de détermination du droit applicable sont-elles requises pour les recours collectifs, afin de garantir une application effective du droit de l'Union dans toute l'Union européenne?

Q 31 D'autres règles particulières vous paraissent-elles nécessaires en matière de recours collectifs dans les situations transfrontières, par exemple pour la résolution consensuelle collective des litiges ou pour les infractions au droit de l'Union commises par des fournisseurs de biens et services en ligne?

3.7 Autres principes possibles

29. La liste des principes communs qui ont été recensés jusqu'ici pour guider une approche européenne en matière de recours collectifs n'est pas exhaustive. L'Union européenne pourrait en établir d'autres.

Question:

Q 32 D'autres principes communs devraient-ils être ajoutés par l'Union européenne?

4. PORTEE D'UNE APPROCHE EUROPEENNE COHERENTE EN MATIERE DE RECOURS COLLECTIFS

30. En ce qui concerne l'indemnisation du préjudice subi en raison d'une infraction au droit de l'Union, il conviendrait d'examiner l'opportunité d'étendre à d'autres domaines (tels que le droit de l'environnement et le droit des services financiers) les travaux actuels de la Commission en matière d'actions collectives (en cessation et/ou en réparation) dans les domaines de la concurrence et de la protection des consommateurs et des passagers. Indépendamment de sa portée, l'approche européenne devra être cohérente.

Questions:

FR 16 FR

Q 33 Les travaux de la Commission en matière d'actions collectives en réparation devraient-ils être étendus à d'autres domaines du droit de l'Union en plus de ceux de la concurrence et de la protection des consommateurs? Dans l'affirmative, quels seraient ces domaines ? Ces domaines présentent-ils des particularités à prendre en compte?

Q 34 Une éventuelle initiative européenne en matière de recours collectifs devrait-elle être de portée générale, ou serait-il préférable d'envisager des initiatives limitées à certains domaines?

5. CONSULTATION PUBLIQUE

Toutes les parties concernées sont invitées à présenter leur contribution avant le 30 avril 2011. Ces contributions doivent être envoyées, si possible sous format électronique, à [email protected], ou par écrit à l'adresse suivante:

Commission européenne «Consultation sur le recours collectif/Consultation on collective redress» Avenue du Bourget, 1-3 B-1140 Bruxelles (Evere) Belgique

Chaque contribution doit clairement porter la mention «Consultation sur le recours collectif/Consultation on collective redress». Dans un souci de transparence, les organismes (notamment les ONG, les associations professionnelles et les entreprises commerciales) sont priés de rendre publique toute information utile les concernant en s'inscrivant au registre des représentants d'intérêts et en souscrivant à son code de conduite.

Afin de susciter un véritable débat sur cette question, la Commission a publié le présent document de consultation sur son site internet Europa, à l'adresse suivante:

http://ec.europa.eu/justice/news/consulting_public/news_consulting_public_en.htm

Toutes les contributions reçues seront publiées sur ce site. Les auteurs peuvent toutefois demander que leur point de vue demeure confidentiel, en indiquant expressément sur la première page de leur réponse qu’ils s’opposent à sa publication.

Déclaration de confidentialité

Objet et étendue du traitement des données à caractère personnel

La Commission enregistrera et traitera vos données personnelles dans la mesure où elles sont nécessaires au suivi de votre réponse à la consultation publique sur le recours collectif. Ces données seront traitées conformément au règlement (CE) n° 45/2001 relatif à la protection des personnes physiques à l’égard du traitement des données à caractère personnel par les institutions et organes communautaires et à la libre circulation de ces données. Elles seront enregistrées et stockées aussi longtemps que des actions complémentaires seront nécessaires concernant votre réponse. Dans un souci de transparence, les contributions – y

mailto:[email protected]


FR 17 FR

compris votre nom et votre poste au sein de votre organisation – seront communiquées au public, notamment par l’intermédiaire du site internet de la Commission, à l'adresse suivante:


Droit de rectification et responsable du traitement des données à caractère personnel

Pour de plus amples informations concernant le traitement des données à caractère personnel ou l’exercice de vos droits (par exemple, l’accès à ces données et la rectification des données inexactes ou incomplètes), veuillez écrire à l’adresse suivante: [email protected]

Vous disposez à tout moment du droit de saisir le contrôleur européen de la protection des données à l’adresse suivante: [email protected].




Annexe 4

1

RÉPUBLIQUE FRANÇAISE

Réponse à la consultation publique de la Commission européenne relative

à une approche européenne cohérente en matière d’actions collectives

SYNTHESE :

1. La réparation des préjudices causés par les infractions aux règles de concurrence répond à une exigence constitutionnelle et européenne.

2. L’absence de standards européens est à l’origine d’une inégalité au détriment des consommateurs et PME résidant dans les Etats qui ne disposent pas de procédures d’action collective. Cette situation ne fait par ailleurs nullement échec à l’assignation des entreprises commettant des pratiques anticoncurrentielles de portée européenne établies dans les juridictions des autres Etats membres. Ces entreprises sont en effet de plus en plus susceptibles d’être attraites dans ces derniers Etats, même si les pratiques n’ont pas produit leurs effets à titre principal sur leur territoire.

3. Dans ce contexte, un modèle européen d’action collective peut être défini pour assurer de façon effective les droits des victimes de pratiques anticoncurrentielles tout en prévoyant un ensemble de garde-fous contre les recours abusifs. Des propositions récentes de parlementaires français ont contribué, à l’issue d’une mûre réflexion, à proposer des avancées en ce sens.

4. Un modèle européen efficace devrait répondre à trois conditions. Premièrement, il pourrait privilégier la formation d’actions en réparation consécutives soit à une décision de la Commission européenne, soit à une décision d’une autorité nationale de la concurrence, si elle établit des pratiques anticoncurrentielles, à condition que cette dernière soit devenue définitive. Cette orientation constitue également une garantie contre les procédures et recours abusifs. Deuxièmement, ce dernier type de décision devrait voir son statut renforcé afin de simplifier l’engagement et le traitement de l’action en réparation, comme c’est déjà le cas dans un certain nombre d’Etats membres. Troisièmement, il serait nécessaire de protéger la confidentialité des déclarations et demandes de clémence. SYNTHESIS :

1. The award of damages caused by an infringement to competition rules is a legal requirement that springs from French constitutional law and EU law.

2. The absence of common EU principles creates unequal conditions between SMEs and consumers in the Member States where no collective action mechanism is available, who cannot be compensated for their harm, while those in other Member States are entitled to damages. This situation does not thwart altogether private litigations against undertakings which engage into anticompetitive practices at EU level and which are established in the former Member States. Courts are prone indeed to assert their jurisdiction in the latter Member States, even if they are not those in whose territory the effects have mainly taken place.

3. In this context, European standard principles for collective actions may be defined to the benefit of victims while providing for adequate safeguards that prevent abusive litigation. French MPs have already put forward possible solutions in this field after a careful examination of the issues at stake.

4. An efficient and effective European model could respond to three major conditions. First, it should foster primarily actions following a European Commission decision or a NCA decision, where they find anticompetitive practices, provided that the latter has become final. Second, in order to streamline private enforcement mechanisms, the latter type of decision could ground, in and in itself, the finding that a fault was committed, as this is already the case in a number of Member States. Last, it would be necessary to ensure the confidentiality of leniency applications and declarations.

2

Introduction : La Commission européenne a invité, le 4 février 2011 l’ensemble des intéressés à lui faire part de leurs observations sur un document permettant, d’une part, d’évaluer la valeur ajoutée d’actions collectives au niveau européen dans les domaines du droit de la consommation et du droit de la concurrence, et, d’autre part, de recueillir des propositions concernant les principes cadre qui pourraient guider la définition de standards européens en la matière. L’Autorité de la concurrence, qui est chargée d’assurer, en France, l’application des articles L. 420-1 et -2 du Code de commerce et 101 et 102 du traité sur le fonctionnement de l’Union européenne (ci-après « TFUE »), se prononcera dans le cadre de la présente réponse à la consultation sur les seuls aspects qui intéressent l’application du droit de la concurrence et le rôle des autorités nationales de concurrence (ci-après « ANC »). L’Autorité se concentrera plus particulièrement sur les développements intervenus postérieurement à son avis du 21 septembre 2006 relatif à l’introduction de l’action de groupe en matière de pratiques anticoncurrentielles1

.

Un droit effectif à réparation des personnes publiques ou privées victimes de pratiques anticoncurrentielles, qu’il s’exerce de manière individuelle ou collective, constitue, du point de vue de l’Autorité, le complément indispensable de l’action de protection de l’ordre public économique dont elle est chargée. Le droit à réparation de personnes ayant subi un dommage du fait d’une faute répond à une exigence constitutionnelle découlant de la Déclaration des droits de l’homme et du citoyen, ainsi que l’a rappelé le Conseil constitutionnel2

.

Dans le domaine du droit de la concurrence, ainsi que l’a jugé la Cour de justice de l’Union européenne, « la pleine efficacité du droit européen serait mise en cause si toute personne ne pouvait demander réparation du dommage que lui aurait causé un contrat ou un comportement susceptible de restreindre ou fausser le jeu de la concurrence »3. Le législateur national est compétent pour en aménager les modalités, tout en étant tenu d’assurer l’équivalence et l’effectivité des droits tirés de l’ordre juridique de l’Union européenne au regard de ceux tirés de l’ordre juridique national, et de s’assurer que la législation applicable ne rend pas excessivement difficile l’exercice du droit à réparation4

1 Le texte complet de cet avis peut être téléchargé sur le site internet de l’Autorité :

.

http://www.autoritedelaconcurrence.fr/doc/classactions.pdf 2 Conseil constitutionnel, Décision n° 2010-2 QPC du 11 juin 2010, Mme Viviane L., considérant 11 (« Considérant qu'aux termes de l'article 4 de la Déclaration de 1789 : « La liberté consiste à pouvoir faire tout ce qui ne nuit pas à autrui » ; qu'il résulte de ces dispositions qu'en principe, tout fait quelconque de l'homme qui cause à autrui un dommage oblige celui par la faute duquel il est arrivé à le réparer ; que la faculté d'agir en responsabilité met en œuvre cette exigence constitutionnelle ; que, toutefois, cette dernière ne fait pas obstacle à ce que le législateur aménage, pour un motif d'intérêt général, les conditions dans lesquelles la responsabilité peut être engagée ; qu'il peut ainsi, pour un tel motif, apporter à ce principe des exclusions ou des limitations à condition qu'il n'en résulte pas une atteinte disproportionnée aux droits des victimes d'actes fautifs ainsi qu'au droit à un recours juridictionnel effectif qui découle de l'article 16 de la Déclaration de 1789 »), décision n° 2010-8 QPC du 18 juin 2010, Époux L., considérant 10 ; décision n°2011-116 QPC du 8 avril 2011, M. Michel Z. et autre, considérants 3 et 4. Voir également les décisions n° 99-419 DC du 9 novembre 1999, loi relative au pacte civile de solidarité, et n° 82-144 DC du 22 octobre 1982, Loi relative au développement des institutions représentatives du personnel. 3 CJCE, 20 septembre 2001, Courage, C-453/99, Rec. p. I-6297 ; 13 juillet 2006, Manfredi, C-295 à -298/04, Rec. p. I-6619. 4 CJCE, Courage, précité, point 29 ; 10 juillet 1997, Palmisani, C-261/95, Rec. p. I-4025

http://www.autoritedelaconcurrence.fr/doc/classactions.pdf�

3

Ainsi que l’ont relevé plusieurs rapports publics en France5, l’exercice du droit à réparation est fortement contraint en pratique et a donné lieu à ce jour à un très faible nombre d’actions ayant débouché sur l’octroi de dommages-intérêts aux victimes6 à la suite de pratiques anticoncurrentielles. Le nouveau régime de la prescription pourrait néanmoins leur être plus favorable7

.

Dans ce contexte, des parlementaires ont déposé des propositions de loi instituant des actions collectives formées à l’initiative d’associations de consommateurs agréées dans le domaine du droit de la consommation8, ou incluant également dans leur champ les violations du droit de la concurrence9

5 Rapport d’information sur l’action de groupe, n° 499 (2009-2010), MM. BÉTEILLE et YUNG, 26 mai 2010, Sénat ; Rapport d’information, « la responsabilité civile : des évolutions nécessaires », MM. ANZIANI et BÉTEILLE, fait au nom de la commission des lois, déposé au Sénat le 15 juillet 2009; Rapport de la commission pour la libération de la croissance française, sous la présidence de M. ATTALI, janvier 2008 ; Rapport sur la dépénalisation de la vie des affaires, sous la présidence de M. COULON, janvier 2008; Rapport d’information déposé par la délégation de l’Assemblée Nationale pour l’Union européenne sur le Livre vert sur les actions en dommages et intérêts pour infraction aux règles communautaires sur les ententes et les abus de position dominante du 28 juin 2006 et présenté par M. LAFFINEUR ; Rapport sur l’action de groupe, sous la présidence de MM. CERUTTI et GUILLAUME, décembre 2005.

. Ils ont ainsi pour objectif de faire des consommateurs des acteurs à part

6 L’Autorité de la concurrence ne dispose pas de statistiques officielles en la matière en l’absence de nomenclature du ministère de la Justice et des Libertés permettant de retracer exclusivement les actions en réparation liées à une violation de l’article L. 420-1 ou -2 du Code de commerce et, le cas échéant, de l’article 101 et/ou 102 TFUE. Le rapport réalisé par Ashurst à la demande de la Commission européenne (« Study on the conditions of claims for damages in case of infringement of EC competition rules ») et publié sur le site de la DG Concurrence mentionnait en 2004, en ce qui concerne la France (http://ec.europa.eu/competition/antitrust/actionsdamages/study.html), douze décisions, dont la moitié étaient fondées sur le droit de l’Union européenne. Quatre réparations ont été accordées à la suite d’une décision du Conseil de la concurrence (auquel s’est substituée l’Autorité de la concurrence en 2009) et une autre à la suite d’une décision de la Commission européenne. Une décision judiciaire a fait suite à un avis du Conseil de la concurrence. Depuis la publication de ce rapport, l’Autorité n’a connaissance que d’une décision de la juridiction administrative accordant une réparation, à la suite de la constatation d’une entente par le Conseil de la concurrence en date du 29 novembre 1995. Le juge administratif connaît des demandes d’indemnisation des personnes publiques à la suite de manœuvres dolosives telles qu’une entente à l’occasion de la passation d’un marché public ou d’une délégation de service public. En l’espèce, la réparation a été fixée à un peu plus de 6 millions d’euros (TA Paris, 6e sect., 1ere ch., 27 mars 2009, SNCF ; CE, 27 juin 2008, Société Guintoli, N° 306627; CE, 19 décembre 2007, société Campenon-Bernard, n°324642). Deux décisions judiciaires, rendues à la suite d’une décision de la Commission européenne, n’ont pas accordé de réparation aux demandeurs (Tribunal de commerce de Nanterre, 11 mai 2006, SA Laboratoires Arkopharm c/ sociétés Roche et Hoffmann La Roche, 2004F02643 ; Tribunal de commerce de Paris, 26 janvier 2007, Laboratoires Juva c/ Hoffmann La Roche, RG03048044). 7 La loi n° 2008-561 du 17 juin 2008 portant réforme de la prescription en matière civile a modifié l’article 2224 du Code civil. Elle prévoit que les actions personnelles ou mobilières se prescrivent par cinq ans à compter du jour où le titulaire d’un droit a connu ou aurait dû connaître les faits lui permettant de l’exercer. Le report du point de départ, la suspension ou l’interruption de la prescription ne peuvent avoir pour effet de porter le délai de prescription extinctive au-delà de vingt ans à compter du jour de la naissance du droit. 8 Assemblée Nationale, Proposition de loi de M. CHATEL visant à instaurer les recours collectifs de consommateurs, n° 3055, enregistrée à la Présidence de l’Assemblée nationale le 26 avril 2006 9 Proposition de loi tendant à renforcer la protection des consommateurs par la création d’une action de groupe fondée sur l’adhésion volontaire, présentée par M. Laurent BETEILLE, enregistrée à la présidence du Sénat le 22 décembre 2010 ; Sénat, proposition de loi tendant à renforcer la protection des consommateurs par la création d’une action de groupe fondée sur l’adhésion volontaire, présentée par M. Richard YUNG, enregistrée à la présidence du Sénat le 22 décembre 2010 ; Sénat, proposition de loi sur le recours collectif, présentée par Mme Nicole BRICQ, M. Richard YUNG et plusieurs de leurs collègues, enregistrée à la Présidence du Sénat le 9 février 2010 ; Proposition de loi n°1897 (2008-2009) relative à la suppression du crédit revolving, à l'encadrement des crédits à la consommation et à la protection des consommateurs par l'action de groupe,

http://ec.europa.eu/competition/antitrust/actionsdamages/study.html�

4

entière du respect de ces règles et de renforcer leur confiance dans l’économie de marché, tout en veillant à concilier l’introduction d’actions collectives avec les principes directeurs de la procédure civile française et à éviter les dérives ayant pu être constatées dans des Etats tiers. La Commission européenne pourrait utilement tirer parti de ces réflexions pour définir, le cas échéant, des standards européens communs afin d’assurer un accès effectif à la justice aux citoyens et aux entreprises européennes, de participer à la construction d’un espace unifié de justice et de liberté, et de prévenir une évolution trop hétérogène des législations nationales qui nuirait à ces deux objectifs. Un socle de règles cohérentes et équilibrées, tirant bénéfice de l’expérience des Etats membres et d’une mûre réflexion des parlementaires européens et nationaux, serait aussi de nature à mieux prévenir les risques de dérives que ne pourraient le faire des législations nationales dispersées. Il limiterait également les risques du développement d’une concurrence entre les places de droit. Le droit de la concurrence apparaît comme un domaine privilégié de ces réflexions. Depuis l’entrée en vigueur du règlement (CE) n°1/2003, le 1er mai 2004, les ANC fonctionnent en effet dans le cadre d’un réseau intégré. Elles appliquent des normes substantielles communes dérivant directement du traité (les articles 101 et 102 TFUE) et faisant l’objet d’une interprétation uniforme par la Cour de justice de l’Union européenne du fait du contrôle juridictionnel que cette dernière exerce sur les décisions de la Commission européenne ou de son rôle dans le cadre de questions préjudicielles des juridictions nationales. La réalisation d’un degré élevé de cohérence et de convergence découle ainsi de l’existence même du réseau européen de la concurrence. L’Autorité estime dans ce contexte que la définition de standards européens en matière d’actions en réparation aurait une valeur ajoutée particulière en droit de la concurrence, afin d’assurer une égalité entre les consommateurs européens, et de respecter les principes d’effectivité et d’équivalence des règles de concurrence issues des traités européens10 (1). Afin d’assurer la complémentarité des actions publique et privée, l’office des autorités nationales de concurrence et de celui des juges nationaux doivent par ailleurs être soigneusement articulés, en privilégiant les actions en réparation formées à la suite de la décision d’une ANC et en protégeant les programmes de clémence11

(2).

1. LA DEFINITION DE STANDARDS EUROPEENS EN MATIERE D’ACTIONS

EN REPARATION SERAIT DE NATURE A FAIRE PROGRESSER LES DROITS DES VICTIMES TOUT EN PREVENANT LES RECOURS ABUSIFS

Les obstacles à la formation d’actions par les entreprises ou les consommateurs pour obtenir réparation des dommages causés par des pratiques anticoncurrentielles ont déjà fait l’objet d’analyses approfondies en France par les ministères compétents et les parlementaires12

enregistrée à la Présidence de l'Assemblée nationale le 2 septembre 2009 et rejetée le 15 septembre 2009; Proposition de loi n° 322 (2005-2006) sur le recours collectif, présentée par Mme Nicole BRICQ, M. Richard YUNG et plusieurs de leurs collègues, déposée au Sénat le 25 avril 2006

10 Réponse aux questions n°s 1, 4, 5, 6, 7, 8, 10, 11 et 12 11 Réponse aux questions n°s 2, 3, 10, 12, et 32 12 Voir note de bas de page 5, 8 et 9

5

français depuis 2005. Sans revenir sur ce diagnostic, qui conserve son actualité en France, il convient de relever que l’évolution du paysage européen depuis cette date a modifié la situation des victimes françaises, et, plus généralement de celles qui résident dans les Etats qui ne disposent pas de mécanismes d’actions collectives au regard de celles qui sont établies dans d’autres Etats membres. Cela rend nécessaire la définition de standards européens lorsque les dommages résultent de violations de l’article 101 et/ou 102 TFUE (1.1.), ainsi que la mise en place de garanties permettant d’éviter les recours abusifs (1.2.). Dans le domaine du droit de la concurrence, ces standards devront répondre à certaines conditions spécifiques pour assurer l’effectivité des recours (1.3.). 1.1.L’hétérogénéité des règles nationales est source d’inégalités pour les victimes de

pratiques anticoncurrentielles

Treize Etats membres de l’Union européenne sont aujourd’hui dotés de mécanismes d’actions collectives, dont les modes de fonctionnement et le champ d’application restent variables13

.

En application du règlement (CE) n°44/2001 du 22 décembre 2000 concernant la compétence judiciaire, la reconnaissance et l'exécution des décisions en matière civile et commerciale (dit « Bruxelles I »), les actions en responsabilité extracontractuelle visant à obtenir la réparation de dommages liés à la violation de règles de concurrence peuvent être formées soit dans l’Etat membre dans lequel le défendeur a sa résidence habituelle (article 2), soit devant le tribunal de l’Etat dans lequel le fait dommageable s'est produit (article 5, paragraphe 3), soit, s'il y a plusieurs défendeurs, devant le tribunal du domicile de l'un d'eux, « à condition que les demandes soient liées entre elles par un rapport si étroit qu'il y a intérêt à les instruire et à les juger en même temps afin d'éviter des solutions qui pourraient être inconciliables si les causes étaient jugées séparément » (article 6, paragraphe 1). L’article 28 donne par ailleurs au juge compétent la faculté de regrouper des demandes connexes. Les victimes de pratiques anticoncurrentielles ont pu, en application de ces différentes options, introduire leurs actions dans les Etats membres dont les règles de procédure civile et

13 Allemagne (investissements financiers, loi dite KapMuG du 16 août 2005 ; par ailleurs des recours individuels en réparation peuvent, depuis 2008, être exercés en leur nom propre par des sociétés spécialisées ayant acquis des droits litigieux moyennant une rétrocession partielle des dommages-intérêts en cas de succès de l’action, comme cela a été le cas dans le cadre de la réparation accordée par la Cour d’appel de Düsseldorf dans le prolongement du cartel du ciment), Bulgarie (droit de la consommation ; article 188 de la loi sur la protection du consommateur et chapitre 33 du Code de procédure civile), Danemark (responsabilité contractuelle et extracontractuelle ; par. 254 a à k de la loi sur l’administration de la justice), Espagne (droit de la consommation ; sections 206.1.6, 221, 223, par. 3 et 519 du Code de procédure civile et article 11 de la loi n°15/2007 sur la concurrence), Finlande (droit de la consommation ; loi Ryhmäkannelaki n°444/2007 du 13 avril 2007), Hongrie (action d’intérêt public en droit de la concurrence pouvant ensuite fonder des actions en réparation individuelles ; article 92 de la loi sur la concurrence, telle qu’applicable à compter du 1er novembre 2005), Italie (droit de la concurrence et droit de la consommation ; article 140 bis du Code de la consommation), Pays-Bas (médiations collectives en matière contractuelle et extracontractuelle, soumises à validation par la Cour d’appel d’Amsterdam ; articles 7.907-910 du Code civil et 1013 à 1018 du Code de procédure civile), Pologne (responsabilité contractuelle et extracontractuelle, à l’exception des dommages corporels ; loi sur les actions de groupe du 17 décembre 2009), Portugal (droit de la consommation, santé publique, environnement ; loi n°83/1995 sur l’action populaire), Royaume-Uni (loi de portée générale applicable tant en matière contractuelle qu’extra-contractuelle ; article 19, III des Civil procedural rules), Roumanie (droit de la consommation ; ordonnance n° 21/1992), Suède (responsabilité contractuelle et extracontractuelle ; loi n°2002/599 sur les actions de groupe).

6

de responsabilité14

leur sont les plus favorables. Une certaine concurrence entre les places de droit tend ainsi à naître. Il est en effet aisé, pour une personne installée dans un Etat membre, de réunir les éléments de fait établissant qu’une pratique anticoncurrentielle commise dans cet Etat est susceptible d’avoir eu des effets dans un autre Etat où, compte tenu d’un cadre législatif plus propice à l’action en réparation, elle souhaite introduire une telle action.

Ainsi, dans l’affaire dite du cartel des vitamines15, la High Court britannique a enjoint16 à des groupes internationaux dont les sièges se situaient en France, en Suisse et en Allemagne mais dont des filiales étaient établies au Royaume-Uni et n’avaient pas participé directement aux pratiques de verser des dommages et intérêts à des filiales allemandes et britanniques d’un groupe néerlandais. La Cour s’est fondée sur l’article 6, paragraphe 1 du règlement « Bruxelles I » et sur la notion d’entreprise en droit de la concurrence pour établir sa compétence. Elle a par ailleurs admis le bien-fondé des demandes d’entreprises qui n’avaient effectué aucun achat auprès des filiales britanniques du groupe qui avait pris part à l’infraction dans la mesure où, en l’absence de cartel, ces filiales britanniques auraient probablement offert aux entreprises allemandes des produits à un meilleur prix. Dans un autre arrêt de juillet 2010 faisant suite à une action en réparation fondée sur la décision de la Commission européenne relative au cartel du caoutchouc17, la Cour d’appel de Londres a retenu sa compétence par un raisonnement analogue, après avoir vérifié que la filiale britannique ne pouvait ignorer que la société mère avait commis des pratiques anticoncurrentielles de la société mère, ce qui était le cas en l’espèce compte tenu de la portée européenne de celles-ci. Plus récemment, des filiales de Honda ont assigné une douzaine d’entreprises françaises, allemandes, italiennes et britanniques sur le fondement d’une décision de la Commission européenne de novembre 2007 dans le secteur du verre automobile18

. Une action vient également d’être formée au Royaume-Uni par un demandeur établi en Irlande contre des filiales britanniques de groupes internationaux dont les pratiques ont été sanctionnées par une ANC, en l’occurrence l’autorité de concurrence italienne.

Des médiations collectives en matière financière à l’initiative de représentants néerlandais d’investisseurs néerlandais et résidant dans d’autres Etats que les Pays-Bas se forment par ailleurs dans ce dernier pays à l’encontre de groupes établis dans un autre Etat19

. Les règles de compétence pourraient être interprétées de manière équivalente en droit de la concurrence. Compte tenu du développement encore récent des mécanismes d’actions collectives dans d’autres pays, il est probable que les actions internationales se développent bien davantage encore dans un avenir proche devant des fors étrangers.

Ainsi, aucune entreprise ayant participé à un cartel ou commis un abus de position dominante en violation des règles européennes de concurrence n’est exonérée de son obligation de 14 S’agissant de la violation des règles de concurrence européenne, toutes les juridictions de l’Union se réfèrent aux articles 101 et/ou 102 TFUE afin d’établir l’existence d’une faute ouvrant droit à réparation. Il est ici fait référence aux règles générales relatives à l’intérêt à agir et à la responsabilité délictuelle, dont l’application est déterminée, sauf convention fixant la loi applicable, par l’article 6, paragraphe 3 du règlement (CE) n°864/2007 sur la loi applicable aux obligations non contractuelles dit « Rome II ». Le domaine de la loi déterminée par cette règle de conflit englobe également la cession de créance de droits litigieux (article 15, point e). 15 COMP/E-1/37.512, 10 janvier 2003, JOUE 2003, L 6/1 16 High Court of Justice, 6 mai 2003, Provimi Ltd and ors v Aventis Animal Nutrition SA and ors, [2003] EWHC 961 (Comm). 17 Court of Appeal, London, 23 juillet 2010, Cooper Tire & Rubber Co v Shell Chemicals UK [2010] UKCLR 1097, EWCA Civ 864 18 COMP/39.125, 12 novembre 2008, JOUE 2009, C173/13 19 Cour d’appel d’Amsterdam, 12 novembre 2010, LJN: BO3908, 200.070.039/01 (sur la seule question de la compétence à ce stade)

7

réparer les dommages causés, quel que soit son lieu d’établissement. Bien plus, les entreprises installées dans un Etat dont la législation est moins propice à l’introduction de telles actions courent davantage de risques que celles-ci ne soient, compte tenu de ce vide juridique, portées devant des juridictions plus coutumières d’un tel contentieux. Dans le même temps, l’absence de mécanisme d’action collective dans certains Etats membres prive les victimes résidant dans ces Etats qui ne disposent pas de moyens analogues à ceux d’un grand groupe, c’est-à-dire au premier chef les PME et les consommateurs individuels, de la perspective concrète d’obtenir réparation de leur préjudice. Il en résulte une forte asymétrie entre les acteurs économiques et une forte inégalité entre les consommateurs européens, selon l’Etat dans lequel ils résident. La définition de standards européens communs serait de nature à remédier à cette difficulté. 1.2.La définition de garanties communes constitue la meilleure protection contre les

litiges abusifs

La définition de standards européens en matière d’actions collectives est ainsi nécessaire pour assurer un égal accès à la justice. Mais il convient de l’entourer de toutes les garanties nécessaires pour prévenir les litiges abusifs. L’Autorité de la concurrence est favorable aux propositions communes des parlementaires français20

en ce sens.

Quatre grands principes peuvent ainsi être dégagés. En premier lieu, les standards européens pourraient reposer sur le principe de dommages et intérêts exclusivement compensatoires, correspondant à une appréciation objective du préjudice subi, car la fonction punitive est déjà assurée par les sanctions administratives prononcées par les ANC. La Cour de cassation a en outre récemment jugé que la reconnaissance d’un jugement allouant des dommages-intérêts disproportionnés au regard du préjudice subi et des manquements aux obligations contractuelles du débiteur était contraire à l’ordre public international21

.

En deuxième lieu, les standards européens pourraient privilégier la formation d’actions consécutives aux décisions d’ANC devenues définitives, qui offrent des garanties procédurales renforcées et assurent une plus grande efficacité dans le traitement des demandes. En troisième lieu, afin d’acclimater progressivement les actions collectives dans les Etats qui ne connaissent pas à ce jour cette procédure, il paraît préférable, dans un premier temps, de réserver la formation du groupe à une entité agréée chargée de recueillir le consentement exprès de l’ensemble des parties concernées en usant de moyens de publicité appropriés sous le contrôle du juge. La rémunération du conseil auquel elles pourront avoir recours ne devrait pas dépendre du montant des dommages-intérêts obtenus. En quatrième lieu, si les modalités d’accès aux preuves doivent faciliter la démonstration de la responsabilité de l’auteur du dommage, il convient toutefois d’éviter que la procédure ne porte des atteintes disproportionnées à la vie des entreprises. Les mesures d’instruction pourraient ainsi rester à la discrétion du juge, d’autant que les actions en réparation formées à 20 Voir notes de bas de page 8 et 9 21 Civ. 1ere, 1er décembre 2010, n°09-13.303, P+B+R+I.

8

la suite de décisions d’une autorité de concurrence, qui contiennent déjà les constatations nécessaires à l’établissement de la faute, devraient être privilégiées (voir 2.). 1.3.Des standards européens communs ne seront efficaces en droit de la concurrence

que sous certaines conditions

Afin d’assurer l’efficacité des standards européens, certaines options proposées par la Commission européenne, qui conviennent dans d’autres domaines du droit, ne paraissent pas adaptées en droit de la concurrence. Ainsi, l’extension du pouvoir d’injonction au droit de la concurrence, tel que prévu par la directive n°98/27/CE relative aux actions en cessation en matière de protection des intérêts des consommateurs n’aurait qu’un très faible intérêt car les ANC disposent déjà d’un pouvoir équivalent en application de l’article 5 du règlement n°1/2003 précité. S’agissant de la médiation, elle ne doit pas devenir une condition préalable à l’action contentieuse mais devrait rester une option ouverte aux parties dans le cadre de l’action en réparation pour faciliter l’indemnisation des victimes des pratiques, conformément à la directive n°2008/52/CE du 21 mai 2008 sur certains aspects de la médiation en matière civile et commerciale. La médiation n’est en effet efficace, en particulier en droit de la concurrence, que si les victimes disposent d’un recours effectif, ou si, comme aux Pays-Bas, la transaction est obligatoirement homologuée par une juridiction spécialisée. Cette homologation est en outre de nature à faciliter la reconnaissance transfrontalière des droits obtenus. Il convient en outre de veiller à la bonne complémentarité de l’action publique et de l’action privée, comme l’illustrent les développements qui suivent.

2. LES ACTIONS EN REPARATION FORMEES A LA SUITE DE LA DECISION

D’UNE ANC GARANTISSENT LA COHERENCE ENTRE L’ACTION PUBLIQUE ET PRIVEE ET ASSURENT UNE PROTECTION CONTRE LES PROCEDURES ABUSIVES

Les ANC assument une responsabilité particulière dans le domaine du droit de la concurrence. Chargées de l’« action publique », elles assument un rôle de détection, de collecte de preuves et de décision en matière de pratiques anticoncurrentielles, que ce soit de leur propre initiative ou à la suite de plaintes d’organisations de consommateurs ou d’entreprises lésées, ou encore de demandes de clémence. Pour faciliter la mise en œuvre d’actions privées sans nuire à l’action publique, il convient donc de les articuler avec soin. De ce point de vue, les futurs standards européens pourraient utilement privilégier la formation des actions collectives consécutives au constat de violation des règles de concurrence par les ANC devenues définitives, dites actions « follow on », qui présentent des avantages pour l’ensemble des parties prenantes et constituent un rempart contre les abus (2.1.). A cet effet, le renforcement du statut des décisions d’ANC, s’inspirant de l’expérience d’une majorité d’Etats membres, serait indispensable (2.2.). Enfin, l’efficacité de l’articulation entre les actions publiques et privées dépend de la protection effective des déclarations

9

effectuées dans le cadre de programmes de clémence, qui jouent un rôle clé dans la détection des pratiques anticoncurrentielles (2.3.). 2.1.Les procédures consécutives à la décision d’une ANC, dites « follow on », présentent

des avantages pour l’ensemble des parties prenantes

La formation d’actions en réparation, individuelles ou collectives, à la suite de décisions d’une ANC établissant le constat d’une violation du droit de la concurrence est considérablement facilitée et présente des avantages tant pour les consommateurs et PME victimes de pratiques anticoncurrentielles (2.1.1.) que pour les défendeurs (2.1.2.), et pour le juge chargé de statuer sur la réparation (2.1.3.). 2.1.1. Les avantages pour les victimes de pratiques anticoncurrentielles Ainsi que l’a indiqué la France dans une note à l’OCDE22

22Note de la délégation de la France, Table ronde sur la quantification par les tribunaux nationaux et les autorités de concurrence des dommages causés à l’économie, 25 janvier 2011, DAF/COMP/WD(2011)16

: « l’ensemble des décisions de l’Autorité, y compris ses décisions constatant l’existence d’une infraction et imposant une sanction, sont publiques (hormis les mentions couvertes par le secret des affaires) et disponibles gratuitement, dès leur adoption et leur notification aux entreprises condamnées, sur son site Internet. Ensuite, ces décisions sont motivées de façon généralement très détaillée. Elles permettent donc à toutes les victimes d’avoir une idée précise des conséquences générales que le cartel ou l’abus qu’elles concernent a pu avoir sur les secteurs ou marchés concernés et sur les consommateurs (qu’il s’agisse de consommateurs finals, d’utilisateurs de biens intermédiaires, de distributeurs, de clients ou encore de concurrents). Elles peuvent aussi les renseigner sur l’incidence de cette pratique anticoncurrentielle sur le fonctionnement des secteurs ou marchés en aval et sur le fonctionnement plus général de l’économie. Enfin, les analyses et les estimations de surprix que l’Autorité peut être amenée à formuler dans certains cas où les données nécessaires pour les réaliser sont disponibles et fiables, sont accessibles aux victimes potentielles.[…] Lorsque le demandeur en réparation s’en prévaut, ces analyses, toujours très motivées et circonstanciées, constituent en effet des éléments de contexte utiles, qui donnent des indications générales aidant le juge à resituer les arguments invoqués par le demandeur pour prouver son préjudice dans le contexte des conséquences économiques plus larges de l’infraction. Enfin, les entreprises victimes, ainsi que les consommateurs par l’intermédiaire d’organisations agréées, ont la possibilité de jouer, plus en amont, le rôle de plaignant dans les procédures d’infraction menées par l’Autorité. L’introduction d’une saisine devant l’Autorité leur donne, à l’image des entreprises qui sont ultérieurement mises en cause, la possibilité de bénéficier du statut de partie devant l’Autorité, qui leur assure des droits procéduraux très étendus : accès au dossier, droit de présenter des observations suite à la notification des griefs et au rapport des services d’instruction, droit d’être entendu oralement par le collège durant la séance, en particulier. Ils ont donc l’opportunité de faire valoir leur point de vue notamment sur l’importance du dommage à l’économie, sur la base d’argumentaires économiques de nature qualitative ou quantitative. Les possibilités que la loi donne aux victimes de pratiques anticoncurrentielles de contribuer à la mission de préservation de l’ordre public économique de l’Autorité jouent un rôle très important en pratique, en aidant l’Autorité à détecter des ententes ou des abus de position dominante

10

potentiels, en produisant des informations utiles et en participant au débat contradictoire, aussi bien sur les faits que sur les éléments à prendre en considération pour déterminer le montant des sanctions. En particulier, les entreprises ou les consommateurs victimes, qui sont présents au quotidien sur le terrain, peuvent jouer un rôle déterminant en apportant à l’Autorité des informations et des données économiques utiles à la bonne compréhension des faits et de leurs conséquences, par exemple dans des secteurs tels que les services, la banque ou encore la grande distribution. »23

.

Les victimes, dont les associations de consommateurs, peuvent ainsi participer à l’établissement de la violation des règles de concurrence et bénéficient de ce fait d’un allègement de la charge de la preuve, grâce à l’instruction effectuée par l’Autorité de la concurrence. Cette dernière apparaît particulièrement précieuse pour les pratiques anticoncurrentielles qui présentent un caractère secret, tels que les cartels. 2.1.2. Les avantages pour les entreprises en cause

Les actions dites « follow on » présentent également des avantages pour les entreprises en cause.

En premier lieu, toutes les ANC sont dotées de pouvoirs qui leur permettent d’instruire des dossiers sans qu’il soit nécessaire d’instituer de nouvelles procédures de type « discovery », qui sont généralement coûteuses24

, et peuvent exposer les entreprises concernées, si elles ne sont pas suffisamment encadrées, à la divulgation de secrets d’affaires à des tiers.

Les ANC sont des autorités accoutumées à la gestion des secrets d’affaires25

23 Ainsi, dans une décision 10-D-28 du 20 septembre 2010, l'Autorité de la concurrence a sanctionné 11 banques pour une entente consistant en la mise en place de commissions interbancaires non justifiées lors du passage à la dématérialisation du traitement des chèques. Pour apprécier le dommage à l’économie, l’Autorité a notamment recherché si ces commissions, qui constituaient une charge d’exploitation pour les banques remettantes, avaient pu être répercutées sous la forme d’une hausse de prix, portant sur le tarif de la remise de chèques ou sur le prix d’autres services bancaires (via des subventions croisées). Elle a observé que la perte subie par les consommateurs du fait d’une hausse des prix (ou de leur maintien à un niveau artificiellement élevé) était en l’occurrence d’autant plus importante que les remettants sont des clients captifs du monopole bancaire. Il a ainsi été estimé que pour chaque centime de la commission interbancaire répercuté, l’augmentation des prix payés par les clients s’élevait en réalité à environ 220 millions d’euros sur l’ensemble de la période considérée. Autre exemple, dans sa décision 09-D-05, l’Autorité a sanctionné une entente dans le secteur du travail temporaire par laquelle les plus grandes entreprises de ce secteur s’étaient concertées sur leur politique commerciale à l'égard de leurs grands comptes. Elle a constaté que les entreprises de travail intérimaire avaient pu se réserver des marges supérieures à celles qui auraient résulté du libre jeu de la concurrence et que les travailleurs intérimaires avaient également pâti de la pratique en raison du renchérissement du recours à cette forme de travail, qui avait réduit l’emploi. Le surcoût pour les clients et les conséquences sur l'emploi intérimaire ont ainsi été estimés à plusieurs dizaines de millions d'euros.

et hautement spécialisées, disposant de moyens d’investigation indispensables à l’instruction des cas d’entente et d’une expertise économique renforcée. Cette expertise leur permet notamment de mieux examiner, le cas échéant, les arguments d’efficience économique fondés sur l’article

24 Selon une étude de 2007 (Cohasset associates, information governance : a core requirement for the global enterprise), les entreprises américaines dépenseraient entre 2,5 et 4 millions de dollars sur 1 milliard de chiffre d’affaires annuel en « e-discovery » en application de l’article 34(a)(1)(A) des Federal Rules of civil procedure. 25 La divulgation du dossier d’une affaire ayant débouché sur une décision de l’Autorité de la concurrence est, en principe, punie d’un an emprisonnement et de 15 000 euros d’amende en application de l’article L. 463-6 du Code de commerce. Pour autant, l’utilisation par les parties d’informations couvertes par le secret de l’instruction devant l’Autorité de la concurrence est envisageable dans la mesure nécessaire à l’exercice des droits de la défense (Cass.com, 19 janvier 2010, n°08-19.761, P+B).

11

101, paragraphe 3 TFUE, et d’adopter une « approche par les effets » des plaintes pour abus de position dominante, dans les affaires qui s’y prêtent. Toutes les ANC veillent, par surcroît, à assurer le respect des droits de la défense, conformément aux exigences communes découlant de la Convention européenne des droits de l’homme, mais également aux exigences constitutionnelles et législatives nationales26

.

Enfin, les décisions des ANC contiennent des constatations de fait qui ont déjà fait l’objet d’un débat contradictoire et peuvent nourrir utilement le débat ultérieur devant le juge civil, en réduisant le risque qu’il ne soit statué sur la base de dossiers différents si plusieurs juges sont saisis d’une action en réparation, ce qui contribue également à limiter les hypothèses de décisions contradictoires ultérieures sur la question de la faute. 2.1.3. Les avantages pour le juge civil

Le traitement des actions en réparation formées consécutivement à une décision de l’Autorité de la concurrence est facilité pour le juge civil. L’intervention préalable de l’Autorité simplifie en effet la mise en état du dossier (voir sur ce point l’avis précité du 21 septembre 2006, notamment sur la question des actions autonomes, beaucoup plus délicates à traiter). Le juge peut notamment s’aider, pour évaluer le préjudice, des informations figurant dans la décision rendue par l’Autorité, qui peuvent être relatives au surprix, mais aussi à ses autres composantes possibles, telles que le manque à gagner lié aux ventes perdues du fait d’une éviction ou d’une entente. 2.2.Il conviendrait de tirer parti de l’expérience positive de nombreux Etats membres

en renforçant le statut des décisions d’ANC devenues définitives, après que les entreprises concernées ont exercé leur droit de recours

Afin de tirer le meilleur parti des avantages précédemment décrits des actions en réparation consécutives à une décision d’une ANC, le législateur européen pourrait s’inspirer de l’expérience d’un nombre important d’Etats membres.

26 En France, la loi et la jurisprudence ont ainsi défini un très haut degré d’exigence en la matière. L’Autorité de la concurrence donne aux entreprises l’occasion de faire part de leurs observations à trois occasions au cours de la procédure. Elle conduit tout d’abord deux « tours » de débat contradictoire écrits. Le premier est déclenché par la notification des griefs, qui définit le champ des accusations faisant l’objet de la procédure, expose les faits reprochés, les qualifications juridiques envisagées par les services d’instruction à ce stade de la procédure et les responsabilités de chacun. Cette notification des griefs n’est connue que des parties, et, dans sa version confidentielle protégeant les secrets d’affaires, du plaignant, ce qui permet à l’entreprise de faire valoir sa défense sans s’exposer à la divulgation d’informations confidentielles à des tiers et à l’Autorité de recueillir les observations de l’ensemble des parties de façon impartiale. Un second tour de contradictoire est alors organisé sur la base d’un rapport, qui est communiqué selon des modalités identiques aux parties. Ce rapport affine, à l’aune des observations sur la notification de griefs, l’analyse factuelle et juridique des services d’instruction. A l’issue de cette deuxième étape, le collège de l’Autorité, distinct des services d’instruction, convoque systématiquement à une séance orale l’ensemble des parties, en continuant à veiller à la protection des secrets d’affaires. Cette séance permet aux entreprises mises en cause d’être confrontées aux autres parties, dont le saisissant, et de débattre notamment de la réalité des pratiques, de leurs justifications éventuelles, ainsi que de leur imputation.

12

Dans quatorze Etats membres de l’Union européenne, soit une majorité d’entre eux, les constatations factuelles et les qualifications juridiques au regard du droit de la concurrence retenues dans les décisions de l’ANC compétente devenues définitives bénéficient d’un statut particulier. Au Danemark, en Italie, en Pologne et en Roumanie, la jurisprudence prévoit une inversion de la charge de la preuve. Dans dix autres Etats membres, en application de la loi (Allemagne, Bulgarie, Grèce, Hongrie, République tchèque, Royaume-Uni, Slovaquie, Slovénie) ou de la jurisprudence (Belgique, Finlande), les décisions d’ANC constatant une violation des règles de concurrence et devenues définitives27

après exercice, le cas échéant, des voies de recours suffisent par elles-mêmes à établir l’existence d’une faute des destinataires de ces décisions, susceptible d’ouvrir un droit à réparation.

L’office du juge est, de ce fait, organisé de façon différente du schéma de droit commun, dans un objectif de simplification au bénéfice des victimes de pratiques anticoncurrentielles et de spécialisation des juridictions. Ainsi, si le législateur européen s’inspirait des exemples exposés ci-dessus, l’action en se déroulerait en deux phases en France, concomitantes ou consécutives. D’une part, les juridictions de contrôle spécialisées seraient chargées de statuer, le cas échéant, sur la décision de l’Autorité. En l’absence de recours dans le délai d’un mois prévu par l’article L. 464-8 du Code de commerce, cette dernière deviendrait définitive. D’autre part, le juge civil saisi d’une demande en réparation pourrait reconnaître, en application de la jurisprudence fondée sur l’article 1382 du Code civil28

, l’existence d’une faute du défendeur, si celui-ci était partie à la procédure devant l’Autorité, et pourrait apprécier, sur la base des éléments de fait et de droit qui lui seraient soumis, le lien de causalité entre la violation du droit de la concurrence et le préjudice allégué par le demandeur, pour statuer sur le montant de la réparation éventuelle.

Cette évolution supprimerait également les risques de discrimination à rebours. Si les victimes d’une pratique anticoncurrentielle constatée par une décision de la Commission européenne peuvent, en effet, en application de la jurisprudence Masterfoods29

et de l’article 16, paragraphe 1 du règlement (CE) n°1/2003 du Parlement européen et du Conseil, se prévaloir de celle-ci auprès du juge civil, elles ne peuvent bénéficier d’un dispositif équivalent sur la base d’une décision d’ANC que dans quelques Etats membres, et non dans l’ensemble de l’Union européenne. Les victimes, c’est-à-dire les consommateurs et PME, mais aussi de grands groupes, comme ce fut le cas en France dans l’affaire du travail temporaire, se voient, dans ces conditions privés d’un droit effectif à un recours. Il serait donc souhaitable de mettre fin à cette situation qui peut être source de discriminations.

27 Dans un souci de sécurité juridique, les décisions des ANC ne bénéficieraient ainsi pas de la même présomption de validité que les décisions de la Commission européenne. 28 Voir à titre d’exemple, CA Versailles, 24 juin 2004, 12ème chambre, 2ème section, SA Verimédia c/ SA Médiatrie et a. (JCP E 2004, n° 1487) : une « décision définitive du Conseil de la concurrence désignant nommément la société (demanderesse) comme l’une des victimes d’actions dilatoires concertées ayant eu pour objet ou pour effet d’entraver le libre exercice de la concurrence dans le secteur (en cause) (...) établit suffisamment que les comportements (...) sanctionné(s) à ce titre constituent bien des fautes à l’égard de (la victime) ». 29 CJCE, 14 décembre 2000, Masterfoods Ltd et HB Ice Cream Ltd, C-344/98, Rec. p. I-11369, points 49 et suivants

13

2.3.L’efficacité de l’articulation entre les actions publiques et privées dépend de la

protection effective des programmes de clémence

Enfin, l’une des conditions essentielles de l’efficacité des standards européens en matière d’action collective porte sur la protection des programmes de clémence. Ces programmes incitent les entreprises parties à une entente à révéler elles-mêmes ces infractions néfastes pour l’économie en contrepartie d’une réduction de la sanction administrative encourue. Le législateur français a été un pionnier en la matière puisqu’il a doté l’Autorité de cet outil, encadré par l’article L. 464-2, IV du Code de commerce, dès 2001. Ces programmes bénéficient à l’ensemble de la collectivité car, en leur absence, de nombreux cartels n’auraient pu être détectés par les ANC, et par conséquent portés à la connaissance de leurs victimes, qui ne se verraient dans ces conditions offrir aucun droit à réparation. Il convient donc d’assurer un équilibre entre la préservation de l’efficacité des programmes de clémence, qui suppose la garantie de la confidentialité des déclarations faites par les entreprises demanderesses, et l’accès aux preuves que peuvent se voir offrir les victimes, dont les actions en réparation seront considérablement facilitées du fait des standards européens. Ces standards européens devraient donc organiser de façon uniforme au sein du réseau européen de la concurrence la protection des déclarations et preuves annexées aux demandes de clémence, afin que les entreprises concernées ne soient pas placées dans une position moins favorable devant le juge civil que si elles n’avaient pas transmis des informations à la Commission européenne ou à une ANC. Conclusion : La définition de standards européens permettant de favoriser les actions en réparation des dommages nés de pratiques contraires aux articles 101 et/ou 102 TFUE est souhaitable afin de conforter l’effectivité des droits des victimes, qu’il s’agisse de grands groupes, mais aussi, et surtout, de PME et de consommateurs individuels dont la capacité de négocier une transaction permettant d’éviter un litige est moindre. Elle peut également contribuer à réduire les inégalités qui se développent dans l’Union européenne du fait de divergences dans les droits nationaux. Un modèle propre à l’Union européenne, qui s’acclimaterait ainsi plus aisément dans l’ensemble des Etats membres, peut être construit à partir des expériences nationales et des réflexions des parlementaires européens et nationaux. Il doit être efficace, en veillant à une bonne articulation entre action publique et action privée, qui passe essentiellement par la reconnaissance de l’autorité de la chose décidée aux décisions des ANC devenues définitives et par la protection des déclarations de clémence. Ces conditions d’efficacité répondent également à des considérations d’équilibre car elles participent des garanties effectives contre les recours abusifs, qui doivent faire partie intégrante de ce modèle européen.

Annexe 5

EN EN

COMMISSION OF THE EUROPEAN COMMUNITIES

Brussels, 2.4.2008 COM(2008) 165 final

WHITE PAPER on

Damages actions for breach of the EC antitrust rules

{SEC(2008) 404} {SEC(2008) 405} {SEC(2008) 406}

EN 2 EN

WHITE PAPER on

Damages actions for breach of the EC antitrust rules

1. PURPOSE AND SCOPE OF THE WHITE PAPER

1.1. Why a White Paper on damages actions for breaches of the EC antitrust rules?

Any citizen or business who suffers harm as a result of a breach of EC antitrust rules (Articles 81 and 82 of the EC Treaty) must be able to claim reparation from the party who caused the damage. This right of victims to compensation is guaranteed by Community law, as the European Court of Justice recalled in 2001 and 2006.1

Despite the requirement to establish an effective legal framework turning exercising the right to damages into a realistic possibility, and although there have recently been some signs of improvement in certain Member States, to date in practice victims of EC antitrust infringements only rarely obtain reparation of the harm suffered. The amount of compensation that these victims are forgoing is in the range of several billion euros a year.2

In its 2005 Green Paper, the Commission concluded that this failure is largely due to various legal and procedural hurdles in the Member States’ rules governing actions for antitrust damages before national courts. Indeed, such antitrust damages cases display a number of particular characteristics that are often insufficiently addressed by traditional rules on civil liability and procedure. This gives rise to a great deal of legal uncertainty.3 These particularities include the very complex factual and economic analysis required, the frequent inaccessibility and concealment of crucial evidence in the hands of defendants and the often unfavourable risk/reward balance for claimants.

The current ineffectiveness of antitrust damages actions is best addressed by a combination of measures at both Community and national levels, in order to achieve effective minimum protection of the victims’ right to damages under Articles 81 and 82 in every Member State and a more level playing field and greater legal certainty across the EU.

The European Parliament4 concurred with the findings in the Green Paper, as did other stakeholders, and called upon the Commission to prepare a White Paper with detailed proposals to address the obstacles to effective antitrust damages actions.

1.2. Objectives, guiding principles and scope of the White Paper

This White Paper considers and puts forward proposals for policy choices and specific measures that would ensure, more than is the case today, that all victims of infringements of

1 Case C-453/99, Courage and Crehan, [2001] ECR I-6297, and Joined Cases C-295−298/04, Manfredi,

[2006] ECR I-6619. 2 See section 2.2 of the Impact Assessment Report (IAR). 3 See ibid., section 2.3. 4 Resolution of 25 April 2007 (2006/2207(INI)).

EN 3 EN

EC competition law have access to effective redress mechanisms so that they can be fully compensated for the harm they suffered.

This White Paper is to be read in conjunction with two Commission staff working documents: (a) a Commission staff working paper on EC antitrust damages actions (“the SWP”) which explains in greater detail the considerations underlying the White Paper and also provides a concise overview of the already existing acquis communautaire; and (b) an Impact Assessment Report (the “IAR”) analysing the potential benefits and costs of various policy options, and an executive summary of this report.

The primary objective of this White Paper is to improve the legal conditions for victims to exercise their right under the Treaty to reparation of all damage suffered as a result of a breach of the EC antitrust rules. Full compensation is, therefore, the first and foremost guiding principle.

More effective compensation mechanisms mean that the costs of antitrust infringements would be borne by the infringers, and not by the victims and law-abiding businesses. Effective remedies for private parties also increase the likelihood that a greater number of illegal restrictions of competition will be detected and that the infringers will be held liable.5 Improving compensatory justice would therefore inherently also produce beneficial effects in terms of deterrence of future infringements and greater compliance with EC antitrust rules. Safeguarding undistorted competition is an integral part of the internal market and important for implementing the Lisbon strategy. A competition culture contributes to better allocation of resources, greater economic efficiency, increased innovation and lower prices.

The Commission followed the further guiding principle that the legal framework for more effective antitrust damages actions should be based on a genuinely European approach. The policy choices proposed in this White Paper therefore consist of balanced measures that are rooted in European legal culture and traditions.

Another important guiding principle of the Commission’s policy is to preserve strong public enforcement of Articles 81 and 82 by the Commission and the competition authorities of the Member States. Accordingly, the measures put forward in this White Paper are designed to create an effective system of private enforcement by means of damages actions that complements, but does not replace or jeopardise, public enforcement.

In view of the foregoing and in line with the requirement set out by the Court of Justice that any victim of antitrust infringements must be able to exercise his right to compensation effectively, the issues addressed in the White Paper concern, in principle, all categories of victim, all types of breach of Articles 81 and 82 and all sectors of the economy. The Commission also considers it appropriate that the policy should cover both actions for damages which do, and actions which do not, rely on a prior finding of an infringement by a competition authority.

5 See the IAR, section 2.1.

EN 4 EN

2. THE PROPOSED MEASURES AND POLICY CHOICES

2.1. Standing: indirect purchasers and collective redress

In the context of legal standing to bring an action, the Commission welcomes the confirmation by the Court of Justice that “any individual” who has suffered harm caused by an antitrust infringement must be allowed to claim damages before national courts.6 This principle also applies to indirect purchasers, i.e. purchasers who had no direct dealings with the infringer, but who nonetheless may have suffered considerable harm because an illegal overcharge was passed on to them along the distribution chain.

With respect to collective redress, the Commission considers that there is a clear need for mechanisms allowing aggregation of the individual claims of victims of antitrust infringements. Individual consumers, but also small businesses, especially those who have suffered scattered and relatively low-value damage, are often deterred from bringing an individual action for damages by the costs, delays, uncertainties, risks and burdens involved. As a result, many of these victims currently remain uncompensated. At the rare occasions where a multitude of individual actions are brought in relation to the same infringement, procedural inefficiencies arise, for claimants, defendants and the judicial system alike.

The Commission therefore suggests7 a combination of two complementary mechanisms of collective redress to address effectively those issues in the field of antitrust:

• representative actions, which are brought by qualified entities, such as consumer associations, state bodies or trade associations, on behalf of identified or, in rather restricted cases, identifiable victims. These entities are either (i) officially designated in advance or (ii) certified on an ad hoc basis by a Member State for a particular antitrust infringement to bring an action on behalf of some or all of their members; and

• opt-in collective actions, in which victims expressly decide to combine their individual claims for harm they suffered into one single action.

Considering that qualified entities will not be able or willing to pursue every claim, it is necessary that these two types of action complement each other to ensure effective collective redress for victims of antitrust infringements. In addition, it is important that victims are not deprived of their right to bring an individual action for damages if they so wish. However, safeguards should be put in place to avoid that the same harm is compensated more than once.

These suggestions on damages actions in the field of antitrust are part of the Commission’s wider initiative to strengthen collective redress mechanisms in the EU and may develop further within this context.

2.2. Access to evidence: disclosure inter partes

Competition cases are particularly fact-intensive. Much of the key evidence necessary for proving a case for antitrust damages is often concealed and, being held by the defendant or by third parties, is usually not known in sufficient detail to the claimant.

6 Manfredi (see footnote 1), point 61. 7 For the underlying reasons see Chapter 2 of the SWP.

EN 5 EN

Whilst it is essential to overcome this structural information asymmetry and to improve victims’ access to relevant evidence, it is also important to avoid the negative effects of overly broad and burdensome disclosure obligations, including the risk of abuses.

The Commission therefore suggests that across the EU a minimum level of disclosure inter partes for EC antitrust damages cases should be ensured. Building on the approach in the Intellectual Property Directive (Directive 2004/48/EC), access to evidence should be based on fact-pleading and strict judicial control of the plausibility of the claim and the proportionality of the disclosure request. The Commission therefore suggests8 that:

• national courts should, under specific conditions, have the power to order parties to proceedings or third parties to disclose precise categories of relevant evidence;

• conditions for a disclosure order should include that the claimant has:

– presented all the facts and means of evidence that are reasonably available to him, provided that these show plausible grounds to suspect that he suffered harm as a result of an infringement of competition rules by the defendant;

– shown to the satisfaction of the court that he is unable, applying all efforts that can reasonably be expected, otherwise to produce the requested evidence;

– specified sufficiently precise categories of evidence to be disclosed; and

– satisfied the court that the envisaged disclosure measure is both relevant to the case and necessary and proportionate;

• adequate protection should be given to corporate statements by leniency applicants and to the investigations of competition authorities;

• to prevent destruction of relevant evidence or refusal to comply with a disclosure order, courts should have the power to impose sufficiently deterrent sanctions, including the option to draw adverse inferences in the civil proceedings for damages.

2.3. Binding effect of NCA decisions

Whenever the European Commission finds a breach of Article 81 or 82 of the EC Treaty, victims of the infringement can, by virtue of established case law and Article 16(1) of Regulation 1/2003, rely on this decision as binding proof in civil proceedings for damages. For decisions by national competition authorities (NCAs) finding a breach of Article 81 or 82, similar rules currently exist in only some Member States.

The Commission sees no reason why a final decision9 on Article 81 or 82 taken by an NCA in the European Competition Network (ECN), and a final judgment by a review court upholding

8 For the underlying reasons see Chapter 3 of the SWP. 9 In all Member States, NCA decisions are subject to judicial review. NCA decisions are considered final

when they can no longer be reviewed, i.e. decisions that were not appealed within the applicable time limits and thus accepted by their addressees, and those that were confirmed by the competent review courts.

EN 6 EN

the NCA decision or itself finding an infringement, should not be accepted in every Member State as irrebuttable proof of the infringement in subsequent civil antitrust damages cases.

A rule to this effect would ensure a more consistent application of Articles 81 and 82 by different national bodies and increase legal certainty. It would also significantly increase the effectiveness and procedural efficiency of actions for antitrust damages: if defendants can call into question their own breach of Article 81 or 82 established in a decision by an NCA and, possibly, confirmed by a review court, the courts seized with an action for damages are required to re-examine the facts and legal issues already investigated and assessed by a specialised public authority (and a review court). Such duplication of factual and legal analysis leads to considerable extra costs, duration and imponderability for the victim’s action for damages.

The Commission therefore suggests10 the following rule:

• national courts that have to rule in actions for damages on practices under Article 81 or 82 on which an NCA in the ECN has already given a final decision finding an infringement of those articles, or on which a review court has given a final judgment upholding the NCA decision or itself finding an infringement, cannot take decisions running counter to any such decision or ruling.

This obligation should apply without prejudice to the right, and possible obligation, of national courts to seek clarification on the interpretation of Article 81 or 82 under Article 234 of the EC Treaty.

The rule set out above confers binding effect only on decisions that are final, i.e. where the defendant has exhausted all appeal avenues, and relates only to the same practices and same undertaking(s) for which the NCA or the review court found an infringement.

2.4. Fault requirement

If the breach of Article 81 or 82 has been proven, Member States take diverse approaches concerning the requirement of fault to obtain damages.

Some Member States require no fault at all as a condition for an antitrust damages claim, or irrebuttably presume the existence of fault once an infringement has been proven. The Commission sees no policy grounds against such an approach.

As regards the other Member States, the Court’s case law on the conditions of civil liability for breaches of directly applicable Treaty rules, such as Articles 81 and 82, and the principle of effectiveness suggest that any fault requirements under national law would have to be limited. The Commission sees no reasons to relieve infringers from liability on grounds of absence of fault other than in cases where the infringer made an excusable error.

The Commission therefore suggests11 a measure to make it clear, for Member States that require fault to be proven, that:

10 For the underlying reasons see Chapter 4 of the SWP. 11 For the underlying reasons see Chapter 5 of the SWP.

EN 7 EN

• once the victim has shown a breach of Article 81 or 82, the infringer should be liable for damages caused unless he demonstrates that the infringement was the result of a genuinely excusable error;

• an error would be excusable if a reasonable person applying a high standard of care could not have been aware that the conduct restricted competition.

2.5. Damages

The Commission welcomes the confirmation by the Court of Justice of the types of harm for which victims of antitrust infringements should be able to obtain compensation.12 The Court emphasised that victims must, as a minimum, receive full compensation of the real value of the loss suffered. The entitlement to full compensation therefore extends not only to the actual loss due to an anti-competitive price increase, but also to the loss of profit as a result of any reduction in sales and encompasses a right to interest.

For reasons of legal certainty and to raise awareness amongst potential infringers and victims, the Commission suggests codifying in a Community legislative instrument the current acquis communautaire on the scope of damages that victims of antitrust infringements can recover.

Once the scope of damages is clear, the quantum of these damages must be calculated. This calculation, implying a comparison with the economic situation of the victim in the hypothetical scenario of a competitive market, is often a very cumbersome exercise. It can become excessively difficult or even practically impossible, if the idea that the exact amount of the harm suffered must always be precisely calculated is strictly applied. Moreover, far-reaching calculation requirements can be disproportionate to the amount of damage suffered.

To facilitate the calculation of damages, the Commission therefore intends:13

• to draw up a framework with pragmatic, non-binding guidance for quantification of damages in antitrust cases, e.g. by means of approximate methods of calculation or simplified rules on estimating the loss.

2.6. Passing-on overcharges

If the direct customer of the infringer fully or partially passed on the illegal overcharge to his own customers (the indirect purchasers), several legal issues can arise. At present, these create a great degree of legal uncertainty and difficulties in antitrust damages actions.

Problems arise, on the one hand, if the infringer invokes the passing-on of overcharges as a defence against a damages claimant, arguing that the claimant suffered no loss because he passed on the price increase to his customers.

The Commission recalls the Court’s emphasis on the compensatory principle and its premise that damages should be available to any injured person who can show a sufficient causal link with the infringement. Against this background, infringers should be allowed to invoke the possibility that the overcharge might have been passed on. Indeed, to deny this defence

12 Manfredi (see footnote 1), points 95 and 97. 13 For the underlying reasons see Chapter 6 of the SWP.

EN 8 EN

could result in unjust enrichment of purchasers who passed on the overcharge and in undue multiple compensation for the illegal overcharge by the defendant. The Commission therefore suggests14 that:

• defendants should be entitled to invoke the passing-on defence against a claim for compensation of the overcharge. The standard of proof for this defence should be not lower than the standard imposed on the claimant to prove the damage.

Difficulties also arise, on the other hand, if an indirect purchaser invokes the passing-on of overcharges as a basis to show the harm suffered. Purchasers at, or near the end of the distribution chain are often those most harmed by antitrust infringements, but given their distance from the infringement they find it particularly difficult to produce sufficient proof of the existence and extent of passing-on of the illegal overcharge along the distribution chain. If such claimants are unable to produce this proof, they will not be compensated and the infringer, who may have successfully used the passing-on defence against another claimant upstream, would retain an unjust enrichment.

To avoid such scenario, the Commission therefore proposes to lighten the victim’s burden and suggests15 that:

• indirect purchasers should be able to rely on the rebuttable presumption that the illegal overcharge was passed on to them in its entirety.

In the case of joint, parallel or consecutive actions brought by purchasers at different points in the distribution chain, national courts are encouraged to make full use of all mechanisms at their disposal under national, Community and international law in order to avoid under- and over-compensation of the harm caused by an infringement of competition law.

2.7. Limitation periods

While limitation periods play an important role in providing legal certainty, they can also be a considerable obstacle to recovery of damages, both in stand-alone and follow-on cases.

As regards the commencement of limitation periods, victims can face practical difficulties in the event of a continuous or repeated infringement or when they cannot reasonably have been aware of the infringement. The latter occurs frequently in relation to the most serious and harmful competition law infringements, such as cartels, which often remain covert both during and after their lifespan.

The Commission therefore suggests16 that the limitation period should not start to run:

• in the case of a continuous or repeated infringement, before the day on which the infringement ceases;

• before the victim of the infringement can reasonably be expected to have knowledge of the infringement and of the harm it caused him.

14 For the underlying reasons see Chapter 7 of the SWP. 15 For the underlying reasons see ibid. 16 For the underlying reasons see Chapter 8 of the SWP.

EN 9 EN

To keep open the possibility of follow-on actions, measures should be taken to avoid limitation periods expiring while public enforcement of the competition rules by competition authorities (and review courts) is still ongoing. To this end, the Commission prefers the option of a new limitation period, which starts once a competition authority or a review court adopts an infringement decision, over the option of suspending the limitation period during the public proceedings.

In the latter case, claimants (and defendants) will sometimes find it difficult to calculate the remaining period precisely, given that the opening and closure of proceedings by competition authorities are not always publicly known. Moreover, if a suspension were to commence at a very late stage of the limitation period, there may not be enough time left to prepare a claim.

The Commission therefore suggests17 that:

• a new limitation period of at least two years should start once the infringement decision on which a follow-on claimant relies has become final.

2.8. Costs of damages actions

The costs associated with antitrust damages actions, and also the cost allocation rules, can be a decisive disincentive to bringing an antitrust damages claim, given that these actions may be particularly costly and are generally more complex and time-consuming than other kinds of civil action.

The Commission considers that it would be useful for Member States to reflect on their cost rules and to examine the practices existing across the EU, in order to allow meritorious actions where costs would otherwise prevent claims being brought, particularly by claimants whose financial situation is significantly weaker than that of the defendant.

Due consideration should be given to mechanisms fostering early resolution of cases, e.g. by settlements. This could significantly reduce or eliminate litigation costs for the parties and also the costs for the judicial system.

Member States could also consider introducing, where appropriate, limits on the level of court fees applicable to antitrust damages actions.

Finally, Member States are invited to reflect on their cost allocation rules in order to reduce the uncertainty for potential claimants about the costs for which they may be liable. The “loser pays” principle, which prevails in the EU Member States, plays an important function in filtering out unmeritorious cases. However, under certain circumstances, this principle could also discourage victims with meritorious claims. National courts may therefore have to be empowered to derogate from this principle, for example by guaranteeing that an unsuccessful claimant will not have to bear the defendants’ costs that were unreasonably or vexatiously incurred or are otherwise excessive.

The Commission therefore encourages18 Member States:

• to design procedural rules fostering settlements, as a way to reduce costs;

17 For the underlying reasons see ibid. 18 For the underlying reasons see Chapter 9 of the SWP.

EN 10 EN

• to set court fees in an appropriate manner so that they do not become a disproportionate disincentive to antitrust damages claims;

• to give national courts the possibility of issuing cost orders derogating, in certain justified cases, from the normal cost rules, preferably upfront in the proceedings. Such cost orders would guarantee that the claimant, even if unsuccessful, would not have to bear all costs incurred by the other party.

2.9. Interaction between leniency programmes and actions for damages

It is important, for both public and private enforcement, to ensure that leniency programmes are attractive.

Adequate protection against disclosure in private actions for damages must be ensured for corporate statements submitted by a leniency applicant in order to avoid placing the applicant in a less favourable situation than the co-infringers. Otherwise, the threat of disclosure of the confession submitted by a leniency applicant could have a negative influence on the quality of his submissions, or even dissuade an infringer from applying for leniency altogether.

The Commission therefore suggests19 that such protection should apply:

• to all corporate statements submitted by all applicants for leniency in relation to a breach of Article 81 of the EC Treaty (also where national antitrust law is applied in parallel);

• regardless of whether the application for leniency is accepted, is rejected or leads to no decision by the competition authority.

This protection applies where disclosure is ordered by a court, be it before or after adoption of a decision by the competition authority. Voluntary disclosure of corporate statements by applicants for immunity and reduction of fines should be precluded at least until a statement of objections has been issued.

A further measure to ensure that leniency programmes continue to be fully attractive could be to limit the civil liability of successful immunity applicants. The Commission therefore puts forward for further consideration20 the possibility of limiting the civil liability of the immunity recipient to claims by his direct and indirect contractual partners. This would help to make the scope of damages to be paid by immunity recipients more predictable and more limited, without unduly sheltering them from civil liability for their participation in an infringement. The immunity recipient would have to bear the burden of proving the extent to which his liability would be limited. However, consideration should be given, in particular, to the need for such a measure and the impact it would have on the full compensation of victims of cartels and on the position of the co-infringers, especially other leniency applicants.

19 For the underlying reasons see Chapter 10, section B.1 of the SWP. 20 For the underlying reasons see Chapter 10, section B.2 of the SWP.

EN 11 EN

The Commission invites comments on this White Paper. They may be sent, by 15 July 2008, either by e-mail to:

[email protected]

or by post to:

European Commission Directorate-General for Competition, Unit A 5 Damages actions for breach of the EC antitrust rules B-1049 Brussels.

It is standard practice within DG Competition to publish submissions received in response to a public consultation. However, it is possible to request that submissions, or parts thereof, remain confidential. Should this be the case, please indicate clearly on the front page of your submission that it should not be made public and also send a non-confidential version of your submission to DG Competition for publication.

Annexe 6

1

COMMISSION EUROPÉENNE

Secrétariat général SEC(2010) 1192

Bruxelles, le 5 octobre 2010

OJ 1932

RENFORCER LA COHERENCE DE L’APPROCHE EUROPEENNE EN MATIERE DE

RECOURS COLLECTIF : PROCHAINES ETAPES

Note d'information de Mme REDING, M. ALMUNIA et M. DALLI

Cette question est inscrite à l'ordre du jour de la 1932ème réunion de la Commission le mardi 12 octobre 2010.

Destinataires : Membres de la Commission Directeurs généraux et chefs de service

2

Towards a Coherent European Approach to Collective Redress:

Next Steps

Joint information note by Vice-President Viviane Reding, Vice-President Joaquín Almunia

and Commissioner John Dalli

3

Towards a Coherent European Approach to Collective Redress: Next Steps

The European Commission's Legislative and Work Programme 2010 foresees the launching of a public consultation on a European approach to collective redress. We – the EU Commissioners for Justice, Competition and Consumer Policy – have been thinking about this issue together and would like to share with you our reflections on the best way forward. A. Collective redress as an instrument to strengthen the enforcement of EU law

1. Effective enforcement of EU law is of utmost importance for citizens and businesses alike. As the Europe 2020 strategy and the Stockholm Programme emphasize, we need to ensure that citizens and businesses can use in practice the opportunities offered to them by the Single Market and the European area of justice. Rights which cannot be enforced in practice are worthless. Where substantive EU rights are infringed, citizens and businesses must be able to enforce the rights granted to them by EU legislation.

2. An important instrument for ensuring effective enforcement of EU law in such cases is public enforcement by the European Commission (e.g. infringement action or competition proceedings), often based on complaints of citizens or businesses. As guardian of the Treaties, the Commission can ensure that not only individual, but also public interests and, more broadly, the Union interest are taken into account.

3. However, with the enlargement of the European Union the number of cases requiring enforcement has increased substantially because of the larger territorial scope of application of EU law. This has accentuated the need for a more decentralised enforcement of EU law. This has brought the issue on the agenda whether further mechanisms of private enforcement should be added to the current system of EU remedies in order to strengthen the enforcement of EU law.

4. Private enforcement of EU law can be pursued, first of all, by way of individual redress: natural or legal persons could initiate individually legal proceedings to enforce their EU law rights. However, where the same breach of EU law harms a large group of citizens and businesses, individual lawsuits are not always an effective means to stop unlawful practices or to obtain compensation for the harm caused by these practices. Citizens and businesses are also often reluctant to initiate private lawsuits against unlawful practices, in particular if the individual loss is small in comparison to the costs of litigation. As a result, continued illegal practices cause significant aggregate loss to European citizens and businesses.

5. Moreover, where breaches of EU law do trigger a mass of individual lawsuits, the procedural laws of many Member States leave the courts ill-equipped to deal with the case-load efficiently and within reasonable delay.

4

6. Mechanisms of collective redress could be considered because of these reasons to remedy the current shortcomings in the enforcement of EU law.

B. What is meant by "collective redress"?

7. EU citizens and businesses should be able to take action when harmed by a breach of any EU legislation creating substantive rights. When citizens and businesses are victims of the same breach committed by the same company, bundling of their claims in a single collective redress procedure, or allowing such a claim to be brought by a representative entity or a body acting in the public interest, could simplify the process and reduce costs. "Collective redress" is a broad concept encompassing any mechanism that may accomplish the cessation or prevention of unlawful business practices which affect a multitude of claimants or the compensation for the harm caused by such practices. There are two main forms of collective redress: by way of injunctive relief, claimants seek to stop the continuation of illegal behaviour; by way of compensatory relief, they seek damages for the harm caused. Collective redress procedures can take a variety of forms, including out-of-court mechanisms for dispute resolution or the entrustment of public or representative entities with the enforcement of claims.

C. Existing forms of collective redress in the European Union

8. Collective redress is not a novel concept in the European Union. Existing EU legislation and international agreements require Member States to provide for collective injunctive redress in certain areas. As a consequence, all Member States have procedures in place which grant representatives or groups of claimants the possibility to seek an injunction to stop illegal practices. In the area of consumer law, as a result of the Directive on Injunctions, consumer protection authorities and consumer organisations are entitled to put an end to practices that infringe national and EU consumer protection rules in all Member States. In the area of environmental law, the Aarhus Convention requires Member States to ensure access to justice against infringements of environmental standards. Several Member States like Portugal, UK, Ireland, Latvia, Spain, Estonia and Slovenia have implemented this through different forms of collective injunctive relief (actio popularis).

9. Procedures for the collective claim of compensatory relief in certain areas have been introduced in the majority of Member States, most recently in Poland. The existing mechanisms to compensate a group of victims harmed by illegal business practices vary widely throughout the EU. Essentially, every national system of compensatory redress is unique and there are no two national systems that are alike in this area. Some of the procedures only apply in very specific sectors (e.g. the recovery of capital investments losses in Germany or damage caused by anti-competitive practices in the United Kingdom), others have a larger scope (e.g. the Spanish collective redress procedures). A second difference concerns the legal standing in compensatory redress proceedings: some Member States have vested public authorities with the power to institute proceedings in certain areas (e.g. the Ombudsman in Finland), others grant standing to private organisations such as consumer associations (e.g. Bulgaria) or to individuals acting on behalf of a group (e.g. Portugal). Many Member States have a combination of

5

several rules on standing. A further difference concerns the category of victims that can make use of collective compensatory redress. Most of the national systems referred to above allow for compensatory redress for consumers, whereas only a few also allow for compensatory redress for other victims, such as small businesses. Differences also relate to the effect of a judgment on the members of the group concerned: in most Member States, the decision only binds those who have expressly consented to the proceedings ("opt-in", e.g. Sweden, Italy). In a few Member States, the decision becomes binding for all members of the group unless they opted out (Portugal, Denmark, Netherlands). In addition, there are differences between Member States as to the moment at which those entitled to claims are individually identified: in some Member States, the identification must take place when the representative action is brought (e.g. the United Kingdom), whilst in others, it can take place at a later stage (e.g. Poland and Spain). There are also notable differences governing the funding of collective redress actions, the distribution of proceeds and the use of alternative dispute resolution mechanisms. The impact of any future European measure on the national legal systems would vary depending on whether the Member State concerned already has a system of collective redress in place and what the defining features of this system are.

D. A coherent European approach to collective redress

10. Given the diversity of existing national systems and their different levels of effectiveness, a lack of a consistent approach to collective redress at EU level may undermine the enjoyment of rights by citizens and businesses and gives rise to uneven enforcement of those rights. A coherent European framework drawing on the different national traditions could facilitate strengthening collective redress in targeted areas. In any event, such a framework should contain common principles which any potential future EU initiatives on collective redress in any sector would respect. The objective is to ensure from the outset that any future proposal in this field, while serving the purpose of ensuring a more effective enforcement of EU law, fits well into the EU legal tradition and into the set of procedural remedies already available for the enforcement of EU law.

11. Within the European Commission, work has been undertaken for several years to develop European standards of compensatory collective redress in the field of consumer and competition law. The Commission adopted a Green Paper on anti-trust damages actions in 2005 and a White Paper in 2008. In 2008, the Commission also published a Green Paper on consumer collective redress. Stakeholders' positions on the main issues are known: most consumer organisations are in favour of EU-wide judicial compensatory collective redress schemes, whereas many representatives of industry fear the risks of abusive litigation. Stakeholders also warned against an inconsistency between the different Commission initiatives on collective redress, which pleads for the development of a more coherent approach.

E. Common principles to guide EU initiatives on collective redress

12. Based on the outcome of previous consultations, we have identified a first set of core principles which could form part of a European framework for collective redress:

6

13. Any EU initiative on compensatory collective redress should first and foremost ensure that any right of injured parties to compensation can be effectively and efficiently obtained. In a mass claim situation, bundling of individual claims in a single collective redress procedure should allow savings for the parties involved and increase the efficiency of both judicial and out-of-court redress schemes.

14. Parties should have the possibility to resort to a collective consensual resolution of their dispute by either settling among themselves or using an Alternative Dispute Resolution (ADR) mechanism in connection to a collective court case. In some fields it should be explored whether resorting to ADR could become a legal requirement. The effectiveness of ADR and the fairness of its outcome depend, however, significantly on the incentives of the parties to engage in the process. The availability of an effective judicial collective redress system should act as a strong incentive for parties to agree out-of-court which is likely to solve a considerable number of cases thereby avoiding litigation. An initiative on ADR which deals with individual and collective ADR in consumer matters is under preparation.

15. The rules on European civil and procedural law should work efficiently for collective actions and judgements should be enforceable throughout the EU.

16. Adequate means of financing should be available to allow citizens and businesses to have access to justice. Those means should allow for the funding of meritorious claims but avoid any incentive to pursue unmeritorious claims, the risks of which have been obvious in the US system.

F. Avoiding the risk of abusive litigation

17. Any European approach to collective redress would have to avoid from the outset the risk of abusive litigation. Such abuses have occurred in the US with its "class actions" system. This form of collective redress contains strong economic incentives for parties to bring a case to court even if, on the merits, it is not well founded. These incentives are the result of a combination of several factors, in particular, the availability of punitive damages, the absence of limitations as regards standing (virtually anybody can bring an action on behalf of an open class of injured parties) the possibility of contingency fees for attorneys and the wide-ranging discovery procedure for procuring evidence. Because of the increased risk of abusive litigation resulting from these combined incentives, we believe that these features are not compatible with the European legal tradition. We therefore firmly oppose introducing "class actions" along the US model into the EU legal order.

18. A European collective redress scheme should not give any economic incentives to bring abusive claims. In addition, effective safeguards to avoid abusive collective actions should be defined. This should be inspired by the existing national judicial redress systems. For instance, one commonly used safeguard is the 'loser pays' principle which means that the losing party pays the court and lawyers fees of both parties. The judge could also be given a prominent role in the process. This role could consist in assessing

7

the admissibility of the collective actions or verifying if a representative entity bringing the action respects a number of strict criteria.

G. Next Steps: Launching a public consultation

19. As set out in the Commission Legislative and Work Programme 2010, we will launch a public consultation on a European approach to collective redress in order to identify which forms of collective redress could fit into the EU legal system and into the legal orders of the 27 EU Member States. The purpose of the consultation will be to elaborate the common legal principles laid out above and to identify other common principles which should guide any future proposals for collective redress in EU legislation. The public consultation will also explore in which fields different forms of collective redress could have an added-value for improving the enforcement of EU legislation or for better protecting the rights of victims.

20. Our departments (DG JUSTICE, DG COMP and DG SANCO) will consult all Commission departments concerned on the draft consultation paper. We intend to launch the public consultation in November and run it until the end of February 2011. In addition, in January 2011, we will hold a public hearing. Upon conclusion of the consultation, we will present to you the results and our suggestions on the way forward. It goes without saying that any legislative initiative in this field would have to be preceded by an impact assessment.

Annexe 7

P7_TA(2012)0021

Towards a coherent European approach to collective redress

European Parliament resolution of 2 February 2012 on ‘Towards a Coherent European

Approach to Collective Redress’ (2011/2089(INI))

The European Parliament,

– having regard to the Commission staff working document of 4 February 2011 entitled

‘Public Consultation: Towards a Coherent European Approach to Collective Redress’

(SEC(2011)0173),

– having regard to the draft guidance paper entitled ‘Quantifying harm in actions for damages

based on breaches of Article 101 or 102 of the Treaty on the Functioning of the European

Union’, published by the Commission in June 2011,

– having regard to Directive 2009/22/EC on injunctions for the protection of consumers'

interests1,

– having regard to the consultation paper for discussion on the follow-up to the Green Paper

on consumer collective redress, published by the Commission in 2009,

– having regard to its resolution of 26 March 2009 on the White Paper on damages actions for

breach of the EC antitrust rules2,

– having regard to the Commission Green Paper of 27 November 2008 on consumer

collective redress (COM(2008)0794),

– having regard to its resolution of 20 January 2011 on the Report on Competition Policy

20093,

– having regard to the Commission White Paper of 2 April 2008 on damages actions for

breach of the EC antitrust rules (COM(2008)0165),

– having regard to the Monti report of 9 May 2010 on a new strategy for the single market,

– having regard to the Commission communication of 13 March 2007 entitled ‘EU Consumer

Policy Strategy 2007-2013: Empowering Consumers, Enhancing their Welfare, Effectively

Protecting Them’ (COM(2007)0099),

– having regard to its resolution of 25 October 2011 on alternative dispute resolution in civil,

commercial and family matters4,

– having regard to its resolution of 13 September 2011 on the implementation of the directive

1 OJ L 110, 1.5.2009, p.30. 2 OJ C 117 E, 6.5.2010, p.161. 3 Texts adopted, P7_TA(2011)0023. 4 Texts adopted, P7_TA(2011)0449.

on mediation in the Member States, its impact on mediation and its take-up by the courts1,

– having regard to Rule 48 of its Rules of Procedure,

– having regard to the report of the Committee on Legal Affairs and the opinions of the

Committee on Economic and Monetary Affairs, the Committee on Industry, Research and

Energy and the Committee on the Internal Market and Consumer Protection (A7-

0012/2012),

A. whereas in the European area of justice, citizens and companies must not only enjoy rights

but must also be able to enforce those rights effectively and efficiently;

B. whereas recently adopted EU legislation is designed to enable parties in cross-border

situations either to enforce their rights effectively2 or to seek out-of-court settlement by way

of mediation3;

C. whereas the benefits of the alternative dispute resolution method are undisputed and fair

access to justice should remain available to all EU citizens;

D. whereas, according to the Flash Eurobarometer on ’Consumer attitudes towards cross-

border trade and consumer protection’ published in March 2011, 79% of European

consumers agree that they would be more willing to defend their rights in court if they could

join other consumers complaining about the same issue;

E. whereas consumers affected by a legal infringement who wish to pursue a court case in

order to obtain redress on an individual basis often face significant barriers in terms of

accessibility, effectiveness and affordability owing to sometimes high litigation costs,

potential psychological costs, complex and lengthy procedures, and lack of information on

the available means of redress;

F. whereas, when a group of citizens are victims of the same infringement, individual lawsuits

may not constitute an effective means of stopping unlawful practices or obtaining

compensation, in particular if the individual loss is small in comparison with the litigation

costs;

G. whereas in some Member States the overall performance of the existing consumer redress

and enforcement tools designed at EU level is not deemed satisfactory, or such mechanisms

are not sufficiently well known, which results in their limited use;

H. whereas the integration of European markets and the consequent increase in cross-border

activities highlight the need for a coherent EU-wide approach to address cases where

consumers are left empty-handed as the procedures for the collective claim of compensatory

relief which have been introduced in a number of Member States do not provide for cross-

border solutions;

1 Texts adopted, P7_TA(2011)0361. 2 Regulation (EC) No 861/2007 establishing a European Small Claims Procedure (OJ L 199 ,

31.7.2007, p. 1); Regulation (EC) No1896/2006 creating a European order for payment procedure

(OJ L 399 , 30.12.2006, p. 1); Regulation (EC) No 805/2004 creating a European Enforcement

Order for uncontested claims (OJ L 143, 30.4.2004, p. 15). 3 Directive 2008/52/EC on certain aspects of mediation in civil and commercial matters (OJ L 136,

24.5.2008, p. 3).

I. whereas national and European authorities play a pivotal role in the enforcement of EU law,

and private enforcement should only supplement, but not replace, public enforcement;

J. whereas public enforcement to stop infringements and impose fines does not of itself enable

consumers to be compensated for damage suffered;

K. whereas bundling claims in a single collective redress procedure, or allowing such a claim

to be brought by a representative entity or body acting in the public interest, could simplify

the process and reduce costs for the parties involved;

L. whereas a system based on collective legal actions can usefully supplement, but is no

substitute for, individual legal protection;

M. whereas the Commission must respect the principles of subsidiarity and proportionality with

regard to any proposal that does not fall within the exclusive competence of the Union;

1. Welcomes the abovementioned horizontal consultation and stresses that victims of unlawful

practices – citizens and companies alike – must be able to claim compensation for their

individual loss or damage suffered, in particular in the case of scattered and dispersed

damages, where the cost risk might not be proportionate to the damages suffered;

2. Notes the efforts made by the US Supreme Court to limit frivolous litigation and abuse of

the US class action system1, and stresses that Europe must refrain from introducing a US-

style class action system or any system which does not respect European legal traditions;

3. Welcomes the efforts of Member States to strengthen the rights of victims of unlawful

behaviour by introducing or planning to introduce legislation aimed at facilitating redress

while avoiding an abusive litigation culture, but also recognises that national collective

redress mechanisms are widely divergent, in particular in terms of scope and procedural

characteristics, which may undermine the enjoyment of rights by citizens;

4. Welcomes the Commission's work towards a coherent European approach to collective

redress and asks the Commission to demonstrate in its impact assessment that, pursuant to

the principle of subsidiarity, action is needed at EU level in order to improve the current EU

regulatory framework so as to allow victims of infringements of EU law to be compensated

for the damage they sustain and thus contribute to consumer confidence and smoother

functioning of the internal market;

5. Underlines the possible benefits of collective judicial actions in terms of lower costs and

greater legal certainty for claimants, defendants and the judicial system alike by avoiding

parallel litigation of similar claims;

6. Believes, as regards the competition sector, that public enforcement is essential to

implement the provisions of the Treaties, to fully achieve the goals of the EU and to ensure

the enforcement of EU competition law by the Commission and national competition

authorities;

7. Recalls that, currently, only Member States legislate on national rules for quantifying the

amount of compensation that can be awarded; notes, furthermore, that the enforcement of

1 Wal-Mart Stores Inc. v. Dukes et al. 564 U. S. xxx (2011).

national law must not prevent the uniform application of European law;

8. Calls on the Commission to examine thoroughly the appropriate (legal basis for any

measures in the field of collective redress;

9. Notes that the information available to date, in particular a study carried out for DG

SANCO in 2008 entitled ‘Evaluation of the effectiveness and efficiency of collective

redress mechanisms in the EU’, indicates that collective redress mechanisms available

within the EU have not generated disproportionate economic consequences;

Existing EU legislation and injunctive relief

10. Notes that some enforcement mechanisms for individual cases, such as Directive

2008/52/EC on certain aspects of mediation in civil and commercial matters and Regulation

(EC) No 805/2004 creating a European Enforcement Order for uncontested claims, already

exist at EU level and believes that, in particular, Regulation (EC) No 861/2007 establishing

a European Small Claims Procedure provides access to justice by simplifying cross-border

litigation and reducing costs in cases involving claims for a sum of less than EUR 2 000, but

notes that this legislation is not designed to provide effective access to justice in cases

where a large number of victims suffer similar damage;

11. Takes the view that injunctive relief also plays an important role in safeguarding rights

which citizens and companies enjoy under EU law and believes that the mechanisms

introduced under Regulation (EC) No 2006/2004 on consumer protection cooperation1, as

well as Directive 2009/22/EC on injunctions for the protection of consumer interests can be

significantly improved so as to foster cooperation and injunctive relief in cross-border

situations;

12. Takes the view that the need to improve injunctive relief remedies is particularly great in

the environmental sector; calls on the Commission to explore ways of extending relief to

that sector;

13. Considers that injunctive relief should focus on the protection of both the individual interest

and the public interest, and calls for caution to be exercised when widening access to justice

for organisations, since organisations should not enjoy easier access to justice than

individuals;

14. Calls therefore on the Commission to strengthen and increase the effectiveness of existing

instruments such as Directive 98/27/EC on injunctions for the protection of consumers’

interests and Regulation (EC) No 2006/2004 on cooperation between national authorities

responsible for the enforcement of consumer protection laws, in order to ensure appropriate

public enforcement of consumers’ rights in the EU; emphasises, nonetheless, the fact that

neither Directive 98/27/EC nor Regulation (EC) No 2006/2004 allows consumers to be

compensated for the damage suffered;

Legally binding horizontal framework and safeguards

15. Takes the view that access to justice by means of collective redress comes within the sphere

of procedural law and is concerned that uncoordinated EU initiatives in the field of

1 OJ L 364, 9.12.2004, p. 1.

collective redress will result in a fragmentation of national procedural and damages laws,

which will weaken and not strengthen access to justice within the EU; calls, in the event

that it is decided after detailed consideration that a Union scheme of collective redress is

needed and desirable, for any proposal in the field of collective redress to take the form of a

horizontal framework including a common set of principles providing uniform access to

justice via collective redress within the EU and specifically but not exclusively dealing with

the infringement of consumers' rights;

16. Stresses the need to take due account of the legal traditions and legal orders of the

individual Member States and enhance the coordination of good practices between Member

States, and believes that the work on an EU scheme fostering effective relief for both

consumers and SMEs should not cause delays in the adoption of the horizontal framework;

17. Stresses that any legally binding horizontal framework must cover the core aspects of

obtaining damages collectively; further stresses that, in particular, procedural and

international private law issues must apply to collective actions in general irrespective of

the sector concerned, while a limited number of rules relevant to consumer protection or

competition law, dealing with matters such as the potential binding effect of decisions

adopted by national competition authorities, could be laid down, for instance, in separate

articles or chapters of the horizontal instrument itself or in separate legal instruments in

parallel or subsequent to the adoption of the horizontal instrument;

18. Believes that the individual damage or loss suffered plays a pivotal role when deciding to

file an action as these are inevitably compared with the potential costs of proceeding with

an action; reminds the Commission, therefore, of the need for a horizontal framework on

collective redress to be an efficient and cost-effective tool for all parties and takes the view

that national procedural rules in Member States could use Regulation (EC) No 861/2007

establishing a European Small Claims Procedure as a reference for the purposes of

collective redress in cases where the value of the claim does not exceed that regulation’s

scope;

19. Considers that collective action under a horizontal framework would deliver the most

benefit in cases where the defendant and victims represented are not domiciled in the same

Member State (cross-border dimension) and where the rights alleged to have been infringed

are granted by EU legislation (infringement of EU law); calls for further examination of

how to improve redress in cases of infringements of national law which may have large,

cross-border implications;

20. Reiterates that safeguards must be put in place within the horizontal instrument in order to

avoid unmeritorious claims and misuse of collective redress, so as to guarantee fair court

proceedings, and stresses that such safeguards must cover, inter alia, the following points:

Standing

– for a representative action to be admissible there must be a clearly identified group, and

identification of the group members must have taken place before the claim is brought;

– the European approach to collective redress must be founded on the opt-in principle,

whereby victims are clearly identified and take part in the procedure only if they have

expressly indicated their wish to do so, in order to avoid potential abuses; underlines the

need to respect existing national systems, in accordance with the principle of subsidiarity;

calls on the Commission to consider a system which will provide relevant information to all

potential victims involved, increase the representativeness of collective actions, allow for

the largest number of victims to seek compensation and ensure simple, affordable and

effective access to justice for EU citizens, thereby avoiding excessive litigation and

subsequent unnecessary individual or collective actions concerning the same infringement;

calls on the Member States to put in place efficient mechanisms ensuring that as many

victims as possible are informed and made aware of their rights and obligations, in

particular when they are domiciled in several Member States, whilst avoiding unduly

harming the reputation of the party concerned, in order to respect the principle of the

presumption of innocence;

– a collective redress system where the victims are not identified before the judgment is

delivered must be rejected on the grounds that it is contrary to many Member States’ legal

orders and violates the rights of any victims who might participate in the procedure

unknowingly and yet be bound by the court’s decision;

– Member States should ensure that a judge or similar body continues to have discretionary

powers taking the form of a preliminary admissibility check of any potential collective

action in order to confirm that the qualifying criteria have been met and that the action is fit

to proceed;

– Member States should designate organisations qualified to bring representative actions, and

European criteria would be useful in order to clearly define these qualified entities; these

criteria could be based on Article 3 of Directive 2009/22/EC on injunctions for the

protection of consumer interests, but need to be further specified in order to ensure both that

abusive litigation is avoided and that access to justice is granted; such criteria should cover,

inter alia, the financial and human resources of qualifying organisations;

– victims must in any case be free to seek the alternative of individual compensatory redress

before a competent court;

Full compensation for actual damage

– the horizontal framework should cover compensation only for the actual damage caused,

and punitive damages must be prohibited; by virtue of the concept of compensation the

damages awarded must be distributed to individual victims in proportion to the harm they

sustained individually; by and large, contingency fees are unknown in Europe and should

not form part of the mandatory horizontal framework;

Access to evidence

– collective claimants must not be in a better position than individual claimants with regard to

access to evidence from the defendant, and each claimant must provide evidence for his

claim; an obligation to disclose documents to the claimants (‘discovery’) is mostly unknown

in Europe and must not form part of the horizontal framework;

Loser pays principle

– there can be no action without financial risk, and Member States are to determine their own

rules on the allocation of costs, under which the unsuccessful party must bear the costs of

the other party in order to avoid the proliferation of unmeritorious claims in an EU-wide

collective redress mechanism;

No third-party funding

– the Commission must not set out any conditions or guidelines on the funding of damages

claims, as recourse to third-party funding is unknown in most Member States’ legal

systems, for instance, by offering a share of the damages awarded; this does not preclude

Member States setting out conditions or guidelines on the funding of damages claims;

21. Suggests that, should the Commission submit a proposal for a horizontal framework

governing collective redress, a principle of follow-on action should be adopted where

appropriate, whereby private enforcement under collective redress may be implemented if

there has been a prior infringement decision by the Commission or a national competition

authority; notes that establishing the principle of follow-on action does not generally

preclude the possibility of providing for both stand-alone and follow-on actions;

22. Calls on the Commission to explore ways of raising consumer awareness of the availability

of collective redress mechanisms and facilitating cooperation between the entities qualified

to bring collective actions; stresses the key role that consumer organisations and the

European Consumer Centres Network (ECC-Net) can play in passing on the information to

as many victims of infringements of EU law as possible;

23. Stresses that many of the infringements of Union law identified by the Commission in the

field of EU consumer protection measures call for the strengthening of injunctive relief1,

while acknowledging that injunctive relief is not sufficient when victims have suffered

damage and have the right to compensation; asks the Commission to identify the EU

legislation in respect of which it is difficult to obtain compensatory redress;

24. Considers that this should be done in order to pinpoint the areas where the horizontal

framework could provide for collective compensatory redress for breach of such legislation,

as well as for breach of EU antitrust law; calls for the relevant EU legislation to be listed in

an annex to the horizontal instrument;

Alternative dispute resolution (ADR)

25. Notes that ADR mechanisms often depend on the trader's willingness to cooperate, and

believes that the availability of an effective judicial redress system would act as a strong

incentive for parties to agree an out-of-court settlement, which is likely to avoid a

considerable amount of litigation; encourages the setting-up of ADR schemes at European

level so as to allow fast and cheap settlement of disputes as a more attractive option than

court proceedings, and suggests that judges performing the preliminary admissibility check

for a collective action should also have the power to order the parties involved to first seek a

collective consensual resolution of the claim before launching collective court proceedings;

believes that the criteria developed by the Court2 should be the starting point for the

establishment of this power; stresses, however, that these mechanisms should remain, as the

1 Study regarding the problems faced by consumers in obtaining redress for infringements of

consumer protection legislation, and the economic consequences of such problems, 26 August 2008,

Part I: Main Report, p. 21 et seq. 2 Judgment of 18 March 2010 in Joined Cases C-317/08, C-318/08, C-319/08 and C-320/08, Alassini,

not yet reported in the ECR.

name indicates, merely an alternative to judicial redress, not a precondition therefor;

Jurisdiction and applicable law

26. Stresses that a horizontal framework should itself lay down rules to prevent a rush to the

courts (‘forum shopping’) whilst not jeopardising access to justice, and that Brussels I

should be taken as a starting point for determining which courts have jurisdiction;

27. Calls for further examination of how the conflict-of-law rules might be amended; believes

that one solution could be to apply the law of the place where the majority of the victims are

domiciled, bearing in mind that individual victims should remain free not to pursue the opt-

in collective action but instead to seek redress individually in accordance with the general

rules of private international law laid down in the Brussels I, Rome I and Rome II

regulations;

28. Emphasises that following the judgment of the Court in Case C-360/09, Pfleiderer, the

Commission must ensure that collective redress does not compromise the effectiveness of

the competition law leniency system and the settlement procedure;

Ordinary legislative procedure

29. Insists that the European Parliament must be involved, within the framework of the ordinary

legislative procedure, in any legislative initiative in the field of collective redress and that

any proposal must be based on a detailed impact assessment;

o

o o

30. Instructs its President to forward this resolution to the Council, the Commission, the

governments and parliaments of the Member States, and the social partners at EU level.

Annexe 8

EN 1 EN

EUROPEAN COMMISSION DG Competition

Brussels, June 2011

DRAFT GUIDANCE PAPER

QUANTIFYING HARM IN ACTIONS FOR DAMAGES BASED ON BREACHES OF

ARTICLE 101 OR 102 OF THE TREATY ON THE FUNCTIONING OF THE EUROPEAN UNION

- Public Consultation -

EN 2 EN

− Public Consultation − The full effectiveness of the EU antitrust rules (Articles 101 and 102 TFEU) requires that any individual can effectively claim compensation for the harm caused by an infringement of these rules. Damages actions based on an infringement of these rules complement public enforcement by allowing those who have been harmed to receive compensation for their harm. While the right to compensation is recognised by EU law, a range of obstacles currently stand in the way of injured parties effectively receiving the compensation to which they are entitled.

In its 2005 Green Paper on damages actions for breach of the EC antitrust rules, the Commission identified difficulties in quantifying the harm suffered by injured parties as one of the key issues in antitrust damages actions. In its 2008 White Paper, Commission announced its intention to draw up a framework with pragmatic, non-binding guidance on quantifying the harm suffered in such actions.1

In preparation of such guidance and to make it suited to the practical needs of judges and parties in antitrust damages cases, DG Competition sought expertise, inter alia, from legal and economic practitioners and academics, and analysed current practices in the quantification of antitrust harm suffered by injured parties in a range of cases before national courts.2

The aim of the present draft Guidance Paper is to offer assistance to courts and parties involved in actions for damages by making more widely available information relevant for quantifying harm caused by infringements of the EU antitrust rules. The Guidance Paper therefore provides insights into the harm caused by infringements of these rules to different categories of injured parties and, in particular, presents the main methods and techniques currently available to quantify such harm.

DG Competition invites all interested parties to make comments on this draft Guidance Paper. We encourage submissions from stakeholders who are directly involved in antitrust damages actions, be it as members of the judiciary, as parties, or as their legal or other advisors. DG Competition would particularly welcome comments on the practical usefulness and linguistic accessibility of the Guidance Paper and possible suggestions for improvements in this respect, as well as on any recent developments in the judicial practice of the courts of the Member States with regard to the quantification of harm.

Comments on this draft Guidance Paper should be sent to DG Competition not later than 30 September 2011. Received contributions will be published on the Internet. It is important to read the specific privacy statement attached to this consultation for information on how your personal data and contribution will be dealt with (see at the end of this document). It is possible to request that submissions or parts thereof remain confidential. Should this be the case, please indicate clearly on the front page of the submission that you request it should not be made publicly available. In this case a non-confidential version of the submission should also be

1 The Commission’s 2005 Green Paper (COM(2005) 672 final) and 2008 White Paper (COM(2008) 165

final) are available at http://ec.europa.eu/competition/antitrust/actionsdamages/index.html. 2 For the studies commissioned by the DG Competition and the submissions received from external

experts see the website indicated in previous footnote.

http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2005:0672:FIN:EN:PDF

http://ec.europa.eu/competition/antitrust/actionsdamages/economist_workshop.html

http://ec.europa.eu/competition/antitrust/actionsdamages/economist_workshop.html

http://ec.europa.eu/competition/antitrust/actionsdamages/quantification_study.pdf

EN 3 EN

forwarded to DG Competition for publication. Comments should be sent either by e-mail or by regular mail to:

European Commission Directorate-General for Competition, Unit A 1 Quantifying harm in antitrust damages actions − HT.2290

B-1049 Brussels, Belgium

[email protected]

This document has been drafted by the services of the European Commission for consultation purposes. It does not purport to represent or pre-judge any formal position of the Commission on the issues set out therein.


EN 4 EN

TABLE OF CONTENTS

PART 1 — CONTEXT AND GENERAL APPROACH TO QUANTIFYING HARM IN COMPETITION CASES 2

I. LEGAL CONTEXT ........................................................................................................... 2

A. The right to compensation............................................................................................ 2

B. National rules on quantification and this Guidance Paper ........................................... 2

II. GENERAL APPROACH TO QUANTIFYING HARM IN COMPETITION CASES.......................... 2

III. STRUCTURE OF THE GUIDANCE PAPER .......................................................................... 2

Part 2 — Methods and Techniques ............................................................................................ 2

I. OVERVIEW .................................................................................................................... 2

II. COMPARATOR-BASED METHODS ................................................................................... 2

A. Methods for establishing a non-infringement scenario................................................ 2

(1) Comparison over time on the same market.................................................................. 2

(2) Comparison with data from other geographic markets ................................................ 2

(3) Comparison with data from other product markets...................................................... 2

(4) Combining comparisons over time and across markets............................................... 2

B. Implementing the method in practice: techniques for estimating the price or other economic variable in the non-infringement scenario ................................................... 2

(1) Simple techniques: individual data observations, averages, interpolation and simple adjustments................................................................................................................... 2

(2) Regression analysis ...................................................................................................... 2

a. Concept and purpose of regression analysis ................................................................ 2

b. Examples and illustrations ........................................................................................... 2

c. Requirements for applying regression analysis............................................................ 2

(3) Choice of techniques .................................................................................................... 2

III. SIMULATION MODELS, COST-BASED ANALYSIS AND OTHER METHODS........................... 2

A. Simulation models........................................................................................................ 2

B. Cost-based method ....................................................................................................... 2

C. Other methods .............................................................................................................. 2

IV. CHOICE OF METHODS .................................................................................................... 2

Part 3 — Quantifying harm caused by a rise in prices............................................................... 2

I. EFFECTS OF INFRINGEMENTS LEADING TO A RISE IN PRICES........................................... 2

EN 5 EN

II. QUANTIFYING THE OVERCHARGE .................................................................................. 2

A. Quantifying overcharges caused by cartels.................................................................. 2

(1) Effects of cartels........................................................................................................... 2

(2) The initial overcharge paid by the direct customer...................................................... 2

a. Comparison over time .................................................................................................. 2

b. Other comparator-based methods ................................................................................ 2

(3) The pass-on of overcharges.......................................................................................... 2

B. Quantifying overcharges caused by other types of infringements leading to overcharge harm........................................................................................................... 2

III. QUANTIFYING THE HARM CAUSED BY THE VOLUME EFFECT.......................................... 2

PART 4 — Quantifying harm from exclusionary practices....................................................... 2

I. EFFECTS OF EXCLUSIONARY PRACTICES ....................................................................... 2

II. QUANTIFYING HARM TO COMPETITORS ......................................................................... 2

A. The time dimension of exclusionary practices............................................................. 2

B. General approach to the quantification of lost profits.................................................. 2

C. Existing competitors..................................................................................................... 2

(1) Comparison over time .................................................................................................. 2

(2) Other comparator-based methods ................................................................................ 2

D. Prevented entry of competitors .................................................................................... 2

E. Compensation for future loss ....................................................................................... 2

III. QUANTIFYING HARM TO CUSTOMERS ............................................................................ 2

A. Recoupment.................................................................................................................. 2

B. Harm to competitors as customers of the infringers .................................................... 2

Table of cases cited .................................................................................................................... 2

EN 6 EN

PART 1 — CONTEXT AND GENERAL APPROACH TO QUANTIFYING HARM IN COMPETITION CASES

I. LEGAL CONTEXT

A. The right to compensation

1. Everyone who has suffered harm because of an infringement of Article 101 or 102 of the Treaty on the Functioning of the European Union (TFEU) has a right to be compensated for that harm. The Court of Justice of the EU held that this right is guaranteed by primary EU law.3 Compensation means placing the injured party in the position it would have been in had there been no infringement. Therefore, compensation includes reparation not only for actual loss suffered (damnum emergens), but also for loss of profit (lucrum cessans) and the payment of interest.4 Actual loss means a reduction in a person’s assets; loss of profit means that an increase in those assets, which would have occurred without the infringement, did not happen.5

2. In so far as there are no EU rules governing the matter, it is for the domestic legal system of each Member State to lay down the detailed rules governing the exercise of this right to compensation guaranteed by EU law. Such rules, however, must not render excessively difficult or practically impossible the exercise of rights conferred on individuals by EU law (principle of effectiveness), and they must not be less favourable than those governing damages actions for breaches of similar rights conferred by domestic law (principle of equivalence).6

B. National rules on quantification and this Guidance Paper

3. In an action for compensation of harm suffered because of an infringement of Article 101 or 102 TFEU, national courts have to determine the amount to be awarded to the claimant in the event that the claim is well-founded.7 Assessing and proving the quantum of damages in actions for damages is often difficult. This is particularly true in competition law cases.8

4. National law – which has to be laid down and applied in accordance with the rules and principles of EU law referred to in paragraphs 1 and 2 above – determines the legal framework in which courts fulfil their function of adjudicating disputes between parties. With regard to quantifying damages, this legal framework will, for instance, include rules on:

3 See Case C-453/99 Courage [2001] ECR I-6297, paragraph 26; Joined Cases C-295/04 to C-298/04

Manfredi [2006] ECR I-6619, paragraph 60. These cases directly concern Article 101 TFEU (ex Article 81 EC Treaty); the same principles apply however also to Article 102 TFEU (ex Article 82 EC Treaty).

4 See Joined Cases C-295/04 to C-298/04 Manfredi [2006] ECR I-6619, paragraph 95. 5 See Opinion of Advocate General Capotorti in Case 238/78 Ireks-Arkady GmbH v Council and

Commission [1979] ECR 2955, paragraph 9. 6 See Case C-453/99 Courage [2001] ECR I-6297, paragraph 29; Joined Cases C-295/04 to C-298/04

Manfredi [2006] ECR I-6619, paragraph 62. 7 This Guidance Paper is only concerned with the assessment of harm in the context of claims for

financial compensation. While the present paper does not specifically cover the determination of the award in other civil law remedies, its insights may also be used in making such determination, in particular with regard to actions (under applicable national law) for restitution.

8 See, in more detail, paragraphs 10 et seq. in Section II below.

EN 7 EN

• the heads of damages to be compensated and general rules of liability governing claims for compensation;

• requirements such as causality or proximity that define a link between the illegal act and the harm which justifies that the infringer is held liable for reparation of that harm;9

• the procedural framework in which claims for damages are adjudicated. National rules typically provide for an allocation of the burden of proof and of the respective responsibilities of the parties to make factual submissions to the court;10

• the appropriate standard of proof, which may vary between different stages of the proceedings, and may also be different for questions of liability for damages and those of the quantum of damages;

• to what extent and how courts are empowered to quantify the harm suffered on the basis of approximate best estimates or equitable considerations; and

• the admissibility and the role of evidence in civil litigation and its evaluation (and in particular of expert evidence).

5. Within their respective legal frameworks, legislators and courts have often adopted pragmatic approaches in determining the amount of damages to be awarded. For instance, they have established presumptions and allowed for the burden of proof to shift, e.g. once a party has provided a certain amount of facts and evidence. Also, the law of the Member States may provide that the illicit profit made by the infringing undertaking(s) plays a role — either directly or indirectly — in estimating the harm suffered by injured parties.11

6. The purpose of this Guidance Paper is to place at the disposal of courts and parties to damages actions economic and practical insights that may be of use when national rules and practices are applied. To this end, the Guidance Paper outlines insights into the harm caused through anticompetitive practices prohibited by the Treaty and information on the main methods and techniques available to quantify such harm.12 Such guidance may also help parties in finding a consensual resolution of their disputes, be it within or outside the context of judicial proceedings or alternative dispute resolution mechanisms.

9 The Court of Justice has clarified in this respect that whilst any individual must be able to claim

compensation for the harm suffered where there is a causal relationship between that harm and an agreement or practice prohibited under EU competition law, it is for national law to prescribe the rules governing the exercise of that right, including those on the application of the concept of ‘causal relationship’, provided that the principles of equivalence and effectiveness are observed; see Joined Cases C-295/04 to C-298/04 Manfredi [2006] ECR I-6619, paragraphs 61, 64; Case C-453/99 Courage [2001] ECR I- 6297, paragraph 29.

10 See, for an example of such distribution of this burden in competition cases, Kammergericht Berlin (Higher Regional Court, Berlin), decision of 1 October 2009, case No 2 U 10/03 Kart (Vitaminpreise).

11 See paragraph 126 in Part 3 below. 12 The Commission has found useful assistance in preparing this Guidance Paper in various studies it

commissioned as well as in the comments received from external experts; see http://ec.europa.eu/ competition/antitrust/actionsdamages/index.html.

EN 8 EN

7. This Guidance Paper is purely informative, does not bind national courts and does not alter the legal rules applicable in the Member States to damages actions based on infringements of Article 101 or 102 TFEU.13 In particular, whether any and, if so, which of the methods and techniques described in this Guidance Paper are considered appropriate to use in a given case before the courts of the Member States depends on national law applied in accordance with the above-mentioned EU law principles of effectiveness and equivalence. Relevant considerations in this respect are likely to include whether a certain method or technique meets the standard required under national law, whether sufficient data are available to the party charged with the burden of proof to apply the method or technique and whether the burden and costs involved are proportionate to the value of the damages claim at stake. Excessive difficulties in exercising the right to damages guaranteed by EU law and therefore concerns in view of the principle of effectiveness could arise, for instance, through disproportionate costs or through overly demanding requirements regarding the degree of certainty and precision of a quantification of the harm suffered.14 It may be that national courts, in a particular case, can use pieces of direct evidence relevant for the quantification of harm, such as documents produced by an infringing undertaking in the course of business regarding agreed price increases and their implementation or assessing the development of its market position. The availability of such evidence may play an important role when a court decides whether any, and if so which, of the methods and techniques set out below are necessary to be used by a party to meet the required standard of proof under applicable law.15

8. Nothing in this Guidance Paper should be understood as arguing against the use of such direct evidence or against the use of more pragmatic approaches, or as raising or lowering the standard of proof or the level of detail of the factual submissions required from the parties in the legal systems of the Member States. Under applicable rules, and in accordance with the principles of effectiveness and equivalence, it may well be sufficient (including in cases where courts are called upon to make a finding based on approximate estimates or equitable considerations) for the parties to provide facts and evidence on the quantum of damages that are less detailed than the methods and techniques discussed in this Guidance Paper.

9. It should also be noted that the economic insights into the harm caused by antitrust infringements and methods and techniques to quantify such harm can evolve over time along with theoretical and empirical research and the judicial practice in this area.

II. GENERAL APPROACH TO QUANTIFYING HARM IN COMPETITION CASES

10. Compensation for harm suffered aims to place the injured party in the position in which it would have been had the infringement of Article 101 or 102 TFEU not occurred. Quantification of harm suffered therefore requires the actual position of the

13 Neither does it affect the rights and obligations of Member States and natural or legal persons under EU

law. 14 See also paragraphs 13 and 14 below. 15 Which substantive law is applicable in a given individual case will often be determined by EU

Regulations, in particular Article 6 of Regulation 864/2007 on the law applicable to non-contractual obligations, OJ 2007 L 199/40. The applicable procedural rules will usually be those in force in the country of the court hearing the action (lex fori).

EN 9 EN

injured party to be compared with the position in which this party would have been but for the infringement. This assessment is sometimes called ‘but-for analysis’.

11. The central question in antitrust damages quantification is hence to determine what is likely to have happened without the infringement. This hypothetical situation cannot, however, be observed and some form of estimation is therefore necessary to construct a reference scenario with which the actual situation can be compared. This reference scenario is referred to as the ‘non-infringement scenario’ or the ‘counterfactual scenario’.

12. The type of harm for which the claimant seeks compensation determines which kind of economic variables (such as, for instance, prices, sales volumes, profits, costs or market shares) need to be considered. For example, in a cartel leading to higher prices for customers of the cartelists, a non-infringement price will need to be estimated to have a reference point for comparing it with the price actually paid by these customers. In an abuse of dominance case leading to the market foreclosure of competitors, the profits lost by these competitors may be measured by comparing their actual turnover and profit margins with the turnover and profit margins they were likely to have generated without the infringement.

13. It is impossible to know with certainty how a market would have exactly evolved in the absence of the infringement of Article 101 or 102 TFEU. Prices, sales volumes, and profit margins depend on a range of factors and complex interactions between market participants that are not easily estimated. Estimation of the hypothetical non-infringement scenario will thus by definition rely on a number of assumptions.16 In practice, the unavailability or inaccessibility of data will often add to this intrinsic limitation.

14. For these reasons, quantification of harm in competition cases is, by its very nature, subject to considerable limits as to the degree of certainty and precision that can be expected. There cannot be a single ‘true’ value of the harm suffered that could be determined, but only best estimates relying on assumptions and approximations. Applicable national legal rules and their interpretation should reflect these inherent limits in the quantification of harm in damages actions for breaches of Articles 101 and 102 TFEU in accordance with the EU law principle of effectiveness so that the exercise of the right to damages guaranteed by the Treaty is not made practically impossible or excessively difficult.

15. This Guidance Paper outlines a number of methods and techniques that have been developed in economics and legal practice to establish a suitable reference scenario and to estimate the value of the economic variable of interest (for example, in a price cartel the likely price that would have been charged for the product had the

16 The limits and implications of such assessment of a hypothetical situation have been recognised by the

Court of Justice (in the context of quantifying loss of earnings in an action for damages against the European Community in the agricultural sector): ‘the loss of earnings is the result not of a simple mathematical calculation but of an evaluation and assessment of complex economic data. The Court is thus called upon to evaluate economic activities which are of a largely hypothetical nature. Like a national court, it therefore has a broad discretion as to both the figures and the statistical data to be chosen and also, above all, as to the way in which they are to be used to calculate and evaluate the damage’, see Joined Cases C-104/89 and C-37/90 Mulder and others v Council and Commission [2000] ECR I-203, at paragraph 79.

EN 10 EN

infringement not occurred).17 The methods and techniques are based on different approaches and vary in terms of the underlying assumptions and the variety and detail of data needed. They also differ in the extent to which they control for factors other than the infringement that may have affected the situation of the claimant. As a result, these methods and techniques may be more or less difficult, time-consuming and cost-intensive to apply.

16. Once a value for the relevant economic variable (such as, price or, for instance, profit margins, or sales volumes) in the hypothetical non-infringement scenario has been estimated, a comparison with the actual circumstances (e.g. the price actually paid by the injured party) is necessary to quantify the harm caused by the infringement of Article 101 or 102 TFEU.

17. Addition of interest will also need to be considered. The award of interest, pursuant to the applicable national rules, is an essential component of compensation for harm suffered through infringements of rights conferred by the Treaty. As the Court of Justice has emphasised, full compensation for the harm sustained must include the reparation of the adverse effects resulting from the lapse of time since the occurrence of the harm caused by the infringement.18 These effects are monetary devaluation19 and the lost opportunity for the injured party to have the capital at its disposal.20 National law may account for these effects in the form of statutory interest or other forms of interest, as long as they are in accordance with the above-mentioned principles of effectiveness and equivalence.

III. STRUCTURE OF THE GUIDANCE PAPER

18. The basis of a claim for damages is the submission that an infringement of Article 101 or 102 TFEU adversely affected the situation of the claimant. Broadly speaking, two principal categories of harmful effects of such infringements can be distinguished:

(a) Infringers can exploit their market power by raising the prices their direct customers pay.21 Among the infringements having such effect are exploitative abuses within the meaning of Article 102 TFEU. Undertakings can also raise prices to their customers by engaging in the kind of practices forbidden by Article 101 TFEU. Typical examples are price fixing, market sharing or output limitation cartels.

17 See Part 2 below. 18 See Case C-271/91 Marshall [1993] ECR I-4367, paragraph 31; Joined Cases C-295/04 to C-298/04

Manfredi [2006] ECR I-6619, paragraph 97; see also European Commission, White Paper on damages actions for breach of the EC antitrust rules (COM(2008) 165), section 2.5 and the accompanying Commission Staff Working Paper (SEC(2008) 404), paragraph 187.

19 See Case C-308/87 Grifoni II [1994] ECR I-341, paragraph 40 and Opinion of Advocate General Tesauro in Case C-308/87 Grifoni II [1994] ECR I-341, paragraph 25; Joined Cases C-104/89 and C-37/90 Mulder and others v Council and Commission [2000] ECR I-203, paragraph 51. In the context of loss of purchasing power, see Joined Cases T-17/89, T-21/89 and T-25/89 Brazzelli Lualdi [1992] ECR II-293, paragraph 40.

20 See opinion of Advocate General Saggio in Joined Cases C-104/89 and C-37/90 Mulder and others v Council and Commission [2000] ECR I-203, paragraph 105.

21 Where the infringement affects the buying activity of the infringing undertakings, the corresponding effect will be the decrease in the purchase prices that these undertakings have to pay to their suppliers. See paragraph 115 in Part 3, Section 1 for more details.

EN 11 EN

Raised prices mean that the customers who purchase the affected product or service22 pay an overcharge. Moreover, a rise in prices may also lead to less demand and may entail a loss of profits for customers who use the product for their own commercial activities.23

(b) Undertakings can also infringe Articles 101 and 102 TFEU by excluding competitors from a market or reducing their market share. Typical examples are abuses of a dominant position through margin squeeze, predatory pricing or tying, or certain vertical exclusivity agreements between suppliers and distributors that infringe competition law.24 Such practices have a significant effect on competitors, who suffer harm as they forego business opportunities and profit in this market. Where foreclosure of competitors is successful and competitive pressure in a market diminishes, customers will be harmed too, typically by a rise in prices.

19. Infringements of Articles 101 and 102 TFEU can also have further harmful effects, for example adverse impacts on product quality and innovation. The Guidance Paper focuses on the two principal categories of harm and the categories of injured parties25 described in paragraph 18. The methods and techniques described in the Guidance Paper may, nonetheless, also be relevant in damages actions concerning other types of harm and other injured parties.

20. Part 3 of the Guidance Paper addresses specifically the quantification of the kind of harm referred to in paragraph 18(a). This part includes a description of the basic effects on the market of price increases resulting from an infringement and illustrates how these types of harm (in particular the harm resulting from the payment of an overcharge and the harm associated with a reduction in demand) can be quantified.

21. Part 4 of the Guidance Paper addresses specifically the quantification of the kind of harm referred to in paragraph 18(b). This part includes a description of the possible effects of the exclusion of competitors from a market and illustrates through examples how these types of harm (namely the loss of profit of the excluded competitor and the harm to customers) can be quantified.

22. The main methods and techniques available to quantify the harm resulting from infringements of Article 101 or 102 TFEU are common to all kinds of harm caused by such infringements. Part 2 of the Guidance Paper therefore provides a general overview of these methods and techniques. In particular, it gives more information on the basic assumptions on which these methods rely and explains their application in practice.

22 For ease of presentation, in the following reference will only be made to ‘products’ affected by an

infringement, which should however be understood as also referring to the ‘services’ affected. 23 See paragraphs 109 et seq. in Part 3, Section I for more details. 24 Vertical agreements are those concluded between undertakings from different levels of the supply

chain. 25 The Guidance Paper does not specifically address the situation of persons other than those mentioned in

points (a) and (b) of paragraph 18, although other persons (such as suppliers of the infringers or customers of law-abiding competitors of the infringers) may also be harmed by infringements leading to price overcharges or the exclusion of competitors; see also footnote 112.

EN 12 EN

PART 2 — METHODS AND TECHNIQUES

I. OVERVIEW

23. Various methods are available to construct a non-infringement scenario for the purposes of quantifying the harm in damages actions in competition cases. The methods most widely used by parties and courts estimate what would have happened without the infringement by looking at the time periods before or after the infringement or at other markets that have not been affected by the infringement. Such comparator-based methods take the data (prices, sales volumes, profit margins or other economic variables) observed in the unaffected period or on the unaffected markets as an indication of the hypothetical scenario without the infringement. The implementation of these methods is sometimes refined by the use of econometric techniques.26 Various comparator-based methods and techniques to implement these methods are described in Section II below.

24. Methods other than comparator-based are addressed in Section III below. One of these methods uses economic models fitted to the actual market to simulate the likely market outcome that would have occurred without the infringement. These models draw on economic theory to explain the likely functioning of a market in view of its main features (e.g. the number of competitors, the way these compete with each other, the degree of product differentiation, entry barriers). Further methods include the cost-based method, which uses production costs for the affected product and a mark-up for a ‘reasonable’ profit margin to estimate the hypothetical non-infringement scenario.

25. Each of these methods and techniques has particular features, strengths and weaknesses that may make them more or less suitable to estimate the harm suffered in a given set of circumstances. In particular, they differ in the degree to which they rely on data that are the outcome of actual market interactions or on assumptions based on economic theory and in the extent to which they control for factors other than the infringement that may have affected the claimant for damages. Moreover, the methods and techniques differ in the degree to which they are simple to use and in the amount of data required.

26. While these methods seek to construct how the market in question would have evolved absent the infringement, more direct evidence available to the parties and to the court (for instance, internal documents of the infringing undertakings on agreed price increases) may also provide, under applicable national legal rules, useful information for assessing quantum of damages in a given case.27

26 Econometrics combines economic theory with statistical or quantitative methods to identify and

measure economic relationships between variables. 27 See for an example of such an approach Oberlandesgericht Karlsruhe (Higher Regional Court,

Karlsruhe), decision of 11 June 2010, case No 6 U 118/05 (appeal pending), where specifically agreed price increases of the infringing undertakings of a cartel were used, under applicable legal rules on the distribution of fact pleading and the establishment of prima facie evidence, to determine the damages award.

EN 13 EN

27. Section IV below sets out considerations on the choice of method, which will usually depend on the specific features of that case and on the requirements under applicable law.28

II. COMPARATOR-BASED METHODS

28. To appreciate how comparator-based methods work in practice, it is useful to consider a (entirely fictitious) example of a hypothetical damages action based on a cartel infringing Article 101 TFEU.29

The flour cartel Assume that all of the milling companies in a particular Member State have been found, by the national competition authority, to have fixed among themselves the prices for the grinding of cereals and the production of flour.

A bakery that regularly purchased flour in recent years brings a damages claim against one of the milling companies. The bakery submits that the infringement has led to an illegal rise in prices for the flour it purchased from that milling company. The bakery asks for compensation for this price overcharge it paid over the past years.

29. The key question regarding the quantification of harm in the aforementioned example is to find out what price the claimant bakery would have paid for flour had there been no infringement. If a comparator-based method is used to do so, these methods compare the price in the infringement scenario with a non-infringement scenario that is established on the basis of price data observed either

• on the same market at a time before and/or after the infringement (1); or

• on a different but similar geographic market (2); or

• on a different but similar product market (3).

It is also possible to combine a comparison over time with a comparison across different geographic or product markets (4).

30. In the example of the flour cartel, the application of the methods focuses on price. It is, however, likewise possible to use these methods to estimate other economic variables such as market shares, profit margins, rate of return on capital, value of assets, or the level of costs of an undertaking. Which economic variable can be usefully considered for the purposes of damages quantification depends on the circumstances of the case at hand.

31. The data used in such a comparison across markets or over time can be data that relate to the entire market (i.e. the average of the price for flour charged to all bakeries operating in a neighbouring geographic market) or data that relate to certain specific market participants only (i.e. the price charged for flour to certain customer groups such as wholesale purchasers operating in a neighbouring market). It could

28 See, in this context, paragraph 7 above. 29 This example is further developed at paragraph 127.

EN 14 EN

also be appropriate, in particular in cases concerning exclusionary practices, to compare data relating to only one market participant. An example for such a comparison between individual companies, i.e. the injured party and a sufficiently similar comparator firm, may be the comparison between the profits achieved by a company trying to enter a new market where it faced exclusionary practices in breach of the EU competition rules and the profits that a comparable new entrant achieved on a different but similar geographic market without being affected by anticompetitive practices. Sections A.1 to 4 below cover the comparison with aggregated market data and firm-level data alike.30

32. The strength of all comparator-based methods lies in the fact that they use real-life data that are observed on the same or a similar market.31 The comparator-based methods rely on the premise that the comparator scenario can be considered representative of the likely non-infringement scenario and that the difference between the infringement data and the data chosen as a comparator is due to the infringement. Whether the level of similarity between infringement and comparator markets or time periods is considered sufficient in order to perform a comparison depends on national legal systems. Where significant differences exist between the time periods or markets considered, various techniques are available to account for such differences32.

A. Methods for establishing a non-infringement scenario

(1) Comparison over time on the same market

33. One frequently used method consists in comparing the actual situation during the period when the infringement produced effects with the situation on the same market before the infringement produced effects or after they ceased.33 For instance, where an undertaking abused its dominant position by foreclosing a competitor from the market during 2004 and 2005, the method could look at e.g. the competitor’s profits during the infringement period and its profits in 2002 and 2003 when there was not yet an infringement.34 Another example would be a price fixing cartel (such as the flour cartel example mentioned above) that lasted from 2005 to 2007 where the method could compare the price paid by the cartel customers during the infringement

30 The comparison with firm-level data of another company could, theoretically, be made not only for

companies that operate in another geographic or product market as discussed in Sections 2-4 below, but also for data of companies operating in the same product and geographic market as the injured party. In practice, such intra-market comparisons do not play a significant role, possibly because within the same market it can be difficult to find a sufficiently comparable other company that was not affected by the infringement. The following sections therefore do not further discuss such comparisons within a market.

31 This aspect is emphasised, for instance, by the Bundesgerichtshof (Federal Court of Justice, Germany), decision of 19 June 2007, case No KRB 12/07 (Paper Wholesale Cartel).

32 See for more detail paragraphs 53-84 in Section B below. 33 See, for example, Corte d’Appello di Milano (Court of Appeal, Milan), decision of 11 July 2003,

(Bluvacanze) and Corte d’Appello di Milano (Court of Appeal, Milan), decision of 3 February 2000, case No I, 308 (Inaz Paghe v Associazione Nazionale Consulenti del Lavoro) (in both cases, comparison before, during and after); Landgericht Dortmund (Regional Court, Dortmund), decision of 1 April 2004, case No 13 O 55/02 Kart (Vitaminpreise) (during and after comparison); Landesgericht für Zivilrechtssachen Graz (Regional Civil Court of Graz), decision of 17 August 2007, case No 17 R 91/07 p (Driving school) (accepting a comparison during and after).

34 For more detailed examples of the method’s application in cases of exclusionary practices, see Part 4 below.

EN 15 EN

period with the price paid by customers in a period after the infringement, e.g. in 2008 and 2009.35

34. There are, in principle, three different points of reference that can be used for the comparison over time:36

• an unaffected pre-infringement period (comparison ‘before and during’ — in the flour cartel example: comparison of the prices paid for flour in the same market before the infringement had effects with those affected by the infringement);

• an unaffected post-infringement period (comparison ‘during and after’ — in the flour cartel example: comparison of the prices affected by the infringement with prices paid in the same market after the infringement ended); and

• both an unaffected pre- and post-infringement period (comparison ‘before, during and after’).

35. An advantage of all methods comparing, over time, data from the same geographic and product market is that market characteristics such as the degree of competition, market structure, costs and demand characteristics may be more comparable than in a comparison with different product or geographic markets. However, also in comparisons over time it happens that some differences between the two data sets are not only due to the infringement. In such cases, it may be appropriate to make adjustments to the data observed in the comparator period to account for differences with the infringement period37 or to choose a different comparator period or market. For instance, in the case of a long-lasting infringement, the assumption that e.g. prices of 10 years ago would have remained unchanged over time absent the infringement is probably overly strong and may lead to opting e.g. for a comparison with the pre-infringement period and the post-infringement period.38

36. In a comparison over time, choosing among the reference periods described in paragraph 34 above will depend on the specifics of the case, including market characteristics and availability of data. For example, in the case of long-lasting infringements, data from the pre-infringement period may simply no longer be available.

37. Where data are available, the choice between a comparison ‘before and during’, ‘during and after’ or ‘before, during and after’ can be determined by a range of factors. It is highly unlikely to find any reference period where market circumstances exactly represent what would have happened in the infringement period had the infringement not occurred. It is only possible to identify a sufficiently similar time period that allows a likely non-infringement scenario to be reasonably approximated.

35 For more detailed examples of the method’s application in cases of infringements that lead to a price

overcharge, see Part 3 below. 36 The comparison over time method is also referred to as the ‘before-after method’ or ‘benchmark

method’. 37 On such adjustments and, in particular, the possibility to use regression analysis, see paragraphs 53-84

in Section B below. 38 In cases of long infringement periods, it may also be appropriate to address practical issues of

comparability of data that result from changes in the way that data have been recorded by companies (e.g. changes in accounting practices or changes in the data organisation software).

EN 16 EN

Factors to be considered in this context may include uncertainties as to which time periods were actually not affected by the infringement. Some infringements start, or cease, gradually; and often doubts exist regarding the exact beginning of an infringement and, in particular, the effects it produces. Indeed, decisions of competition authorities regularly mention evidence suggesting that the infringement may have started earlier than the period established as the infringement period for the purposes of the decision.39 Econometric analysis of observed data can be a way to identify when the infringement’s effects started or ceased.

38. The ending of an infringement and its effects may be more easily established than its beginning, but here too uncertainties could arise as to whether the period immediately after the infringement’s end is unaffected by the anticompetitive behaviour.40 For example, when there is some delay until market conditions return to a non-infringement level, using data from the period immediately after the infringement could lead to an underestimation of the effect of the infringement.41 The pre-infringement period may be a more suitable reference point where central market characteristics changed radically towards the end of the infringement period due to exogenous factors (e.g. a steep increase in raw material costs or an increase in demand for the product).42

39. Nonetheless, even when there are doubts as to whether or not a certain period before or after the infringement was affected by the infringement, this period could, in principle, still serve as a reference period in order to obtain a safe estimate of the harm that will at least have been suffered (“lower-bound” estimate or “minimum damage”).43

40. In certain circumstances, the non-infringement scenario may be appropriately estimated on the basis of two reference periods (before and after the infringement), for example, by using the average from these periods or by using other techniques to reflect a trend in the development of market circumstances during the infringement.44

39 It is possible that a competition authority limits the finding of an infringement to a certain period, while

in fact the infringement may have had a longer duration. 40 See the decision of the Oberlandesgericht Karlsruhe (Higher Regional Court, Karlsruhe) of 11 June

2010 in case No 6 U 118/05 (appeal pending), for an example where a national court ruled that the prices charged in the five months after the infringement ended were still influenced by the cartel.

41 It may also occur that prices are, for a short period after the end of a cartel, particularly low as companies might temporarily engage in aggressive pricing strategies until the ‘normal’, i.e. non-infringement, equilibrium on the market is reached. Specifically in oligopolistic markets another issue may arise, namely that the participants in a cartel can use the knowledge gained through the cartel to coordinate their behaviour afterwards without infringing Article 101. In such a situation, post-infringement prices are likely to be higher than without the infringement and can only serve to make a lower-bound estimate of the harm suffered; see also paragraphs 134 et seq. in Part 3, Section II below.

42 For the short period of the infringement after such a change, post-infringement data can be the more appropriate comparator as they may better reflect the market characteristics after the change. However, where the change in market characteristics was caused by the infringement itself (e.g. where due to anticompetitive foreclosure several competitors exited the market), the post-infringement period is obviously not a suitable comparator to estimate the situation that would have existed without the infringement.

43 If during the infringement exogenous factors lead to a decrease in prices (e.g. a sharp fall in input costs of the infringer), the inference of a lower bound could be rebutted.

44 For example interpolation or regression analysis. For these different techniques to implement comparator-based methods, see paragraphs 53-84 in Section B below.

EN 17 EN

Pre-infringement data could also be used as the reference period up to a certain point during the infringement when a significant change in market circumstances occurred, and post-infringement data as the reference period for the time thereafter.

41. Also the choice of data can contribute to building a sufficiently similar basis for the comparison: there can be situations where aggregated data such as industry price averages (or averages for certain groups of firms) are sufficiently representative,45 whilst in other situations it would be more appropriate to use only data from pre- or post-infringement transactions by the injured company or average data that relate to similar companies. For example, where the injured party or the infringer belongs to a specific group of market players such as wholesale customers (as opposed to end customers), pre- or post-infringement prices charged to wholesale customers may be an appropriate reference point.

42. Making an informed choice of reference period and type of data will usually require good knowledge of the industry in question. The choice will also be influenced by the availability of data and the requirements of applicable rules regarding the standard and burden of proof.

(2) Comparison with data from other geographic markets

43. Another comparator-based method consists in looking at data observed in a different geographic market46 for the purpose of estimating a non-infringement scenario.47 These may be data observed across the entire geographic comparator market or data observed in relation to certain market participants only. For instance, in the example of a flour cartel mentioned above at paragraph 28, the prices paid by the claimant bakery during the infringement period could be compared with the prices paid on average by similar bakeries, in a different geographic market untouched by the infringement. The same type of comparison can, in principle, be undertaken with regard to any other economic variable, e.g. the market shares, profit margins, rate of return on capital, value of assets, or level of costs of an undertaking. A comparison with the commercial performance of firms active on another geographic market that is unaffected by the infringement48 will be particularly relevant in cases of exclusionary behaviour.

45 For further detail on the use of averages in implementing comparator-based approaches, see

paragraph 64 in Part 2, Section II below. 46 For the concepts of relevant (geographic and product) market, see Commission Notice on the definition

of the relevant market for the purposes of Community competition law, OJ C 372, 9.12.1997, p. 5. 47 This method is also referred to as ‘yardstick method’ or ‘cross-sectional method’. These terms are also

used to refer to the comparator-based method that looks at data observed in different but similar product markets, see paragraphs 48-49 in Section 3 below.

For examples of the use of the comparator-based method looking at different geographic markets see, for instance, Cour d’Appel de Paris (Court of Appeal, Paris), decision of 23 June 2003 (Lescarcelle-De Memoris v OGF); Juzgado Mercantil numero 5 de Madrid (Commercial Court, Madrid), decision of 11 November 2005, case No 85/2005 (Conduit-Europe, S.A. v Telefónica de España S.A.), confirmed by Audiencia Provincial de Madrid (Court of Appeal, Madrid), decision of 25 May 2006, case No 73/2006; Bundesgerichtshof (Federal Court of Justice, Germany), decision of 19 June 2007, case No KBR 12/07 (Paper Wholesale Cartel) (in the context of assessing the illicit gain by cartelists for the purpose of calculating a fine).

48 The comparator firm might, in principle, also be a firm active on the infringement market provided that its performance was not significantly influenced by the exclusionary behaviour. Even if the comparator firm was not directly affected by the infringement, it may still have been indirectly affected, e.g. by

http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:31997Y1209(01):EN:NOT


EN 18 EN

44. The more a geographic market is similar (except for the infringement effects) to the market affected by the infringement, the more it is likely to be suitable as a comparator market. This means that the products traded in the two geographic markets compared should be the same or, at least, sufficiently similar. Also the competitive characteristics of the geographic comparator market should be similar to the characteristics of the affected market except for the infringement. This may well be a market that is not perfectly competitive.

45. The method of using geographic comparator markets for deriving a non-infringement scenario is, in practice, mainly used when the infringement concerns geographic markets that are local, regional or national in scope.49 Where the infringement market and the geographic comparator market are neighbouring areas, possibly within one country, there may be an increased likelihood that they are sufficiently similar for the purpose of a comparison.50

46. The comparator market does not always need to be sufficiently similar in its entirety. Where, for instance, the prices paid by one customer group (e.g. wholesalers) or the profits earned by one competitor company (e.g. a new entrant) in the comparator market are used as a reference, it is important that the market position of this customer group or this competitor is sufficiently similar to that of the injured party on the infringement market.51

47. The choice of a geographic comparator market may also be influenced by uncertainties about the geographic scope of an infringement. Geographic markets on which the same or a similar infringement occurred are, in principle, not good candidates for being used as comparator markets. Also neighbouring markets on which no similar infringement occurred may still have been influenced by the anticompetitive practices on the infringement market (e.g. because prices on the neighbouring market were raised in view of the increased prices on the infringement market and lesser competitive pressure emanating from this market). A comparison with such markets will not show the full extent of the harm suffered, but they may, nonetheless, constitute a useful basis to establish a lower-bound estimate of the harm caused on the infringement market. This means that a party to an action for damages could, in principle, safely choose to rely on the comparison with a geographic market that was influenced by the same or a similar infringement, in particular where such influence is likely to have been rather small.

gaining market shares from a foreclosed competitor. The risk of being directly or indirectly influenced by the infringement is lower if the comparison is carried out in relation to a similar firm active on another geographic market. Characteristics that could be relevant when considering the sufficient similarity of firms include their size, cost structure, customers and features of the product they sell.

49 It might, however, also be used when the relevant market is wider than national provided that a sufficiently similar comparator market can be identified.

50 See, however, paragraph 47 below. 51 This may, however, not easily be the case if the overall characteristics on the comparator market are all

too different from the infringement market.

EN 19 EN

(3) Comparison with data from other product markets

48. Similar to the comparison across geographic markets is the approach to look at a different product market52 with similar market characteristics.53 For example, in a case of exclusionary behaviour partially foreclosing a company selling one product, the profit margin earned by that company in the infringement market could be compared with the profit margin for another product that is traded (by a similar or the same company) in a distinct but similar product market.

49. The considerations discussed in the context of geographic comparator markets are, mutatis mutandis, also likely to be relevant for the choice of a suitable comparator product market. They will often relate to the degree of similarity between the two product markets. In particular, the comparator product should be carefully chosen with a view to the nature of the products compared, the way they are traded and the characteristics of the market e.g. in terms of number of competitors, their cost structure and the buying power of customers.54 Uncertainties as to whether a potential comparator product market was affected by the infringement or a similar infringement of Article 101 or 102 TFEU can also play a role.

(4) Combining comparisons over time and across markets

50. Where sufficient data are available, it may be possible to combine comparisons over time and comparisons across markets. This approach is sometimes called the ‘difference in differences’ method because it looks at the development of the relevant economic variable (e.g. the price for flour) in the infringement market during a certain period (difference over time on the infringement market) and compares it to the development of the same variable during the same time period on an unaffected comparator market (difference over time on the non-infringement market).55 The comparison shows the difference between these two differences over time. This gives an estimate of the change in the variable produced by the infringement and excludes all those factors that affected both the infringement and the comparator market in the same way. The method is thus a way to isolate the effects of the infringement from other influences on the relevant variable.

51. A simple example derived from the flour cartel mentioned above may illustrate the method: assume that a before, during and after comparison reveals an increase in price of € 40 per 100 kg bag of flour in the Member State where the cartel occurred between 2005 and 2008. Looking at an unaffected geographic market over the same period may show that prices for flour rose by € 10 per 100 kg bag due to increased costs for an input product (cereals). Assuming that the increased input costs also concerned the infringement market, a comparison of the different development of

52 For the concepts of the relevant (geographic and product) market, see Commission Notice on the

definition of the relevant market for the purposes of Community competition law, OJ C 372, 9.12.1997, p. 5.

53 This method is sometimes also referred to as ‘yardstick method’ or ‘cross-sectional method’ (as is the the comparator-based method looking at different geographic markets).

54 Similarity of market characteristics may be more likely if the two products compared are traded in the same geographic market. Although the circumstances may also be sufficiently similar where the same or similar products from different geographic markets are compared.

55 This can be a geographic or a product comparator market.



EN 20 EN

prices on the infringement and the comparator market would indicate the price difference caused by the flour cartel. In the example, this would be € 30 per unit.

52. The strength of the “difference in differences” method is therefore that it can subtract out changes unrelated to the infringement that occurred during the same time period as the infringement.56 It rests, however, to a large extent on the assumption that these other changes affected both markets similarly.57 The considerations regarding the application of the comparison over time and across market methods, in particular the need for sufficient similarity, are also relevant for the difference in differences method. From a practical point of view, this method usually requires a range of data from different markets and periods of time that may not always be easy to obtain; lesser amounts of data may, however, still allow lower-bound or approximate estimates to be derived.58

B. Implementing the method in practice: techniques for estimating the price or other economic variable in the non-infringement scenario

53. Once a suitable comparator-based method for establishing a non-infringement scenario has been chosen, various techniques are available to implement this method in practice. These techniques differ mainly in the degree to which they rely on individual or average data (e.g. price observations), and in the degree to which the data observed in the comparator market59 or period are subject to further adjustment. As a consequence, these techniques differ in the amount of data they require in order to be carried out.

54. One possibility in implementing comparator-based methods is to use comparator data directly in the form they are observed and to estimate on this basis a value for the economic variable under consideration in the non-infringement scenario (e.g., in the above example, the price of flour). Where more than one data observation is available (e.g. the price of flour in a range of transactions on a geographic comparator market), they can be combined through a calculation of averages into one or more values for the non-infringement scenario. Such average value(s) for the non-infringement scenarios could then be compared to the average value(s) actually observed during the infringement, e.g. the prices really paid for flour (see in more detail in Section (1) below).

55. Where certain factors (such as an increase in raw material prices) have influenced only the comparator or only the infringement market or period, it should be considered, depending on the standard of proof required and depending on applicable

56 Compared to a simple comparison across markets, the difference in differences method also has the

advantage of filtering out fixed differences between markets (such as differences due to constantly lower input costs in one of the markets).

57 If, for example, price increases unrelated to the infringement were higher in the affected market than in the comparator market during the infringement period, application of the difference in differences method using simple averages would overestimate the amount of damages. An econometric implementation of the difference in differences technique may help control for such factors.

58 See, for an example of a national court establishing a lower bound in the course of estimating the quantum of damages (although not using the difference-in-differences method, but the comparison over time method), Kammergericht Berlin (Higher Regional Court, Berlin), decision of 1 October 2009, case No 2 U 10/03 Kart.

59 As mentioned in paragraph 31 above, the data used in such comparison across markets or over time can be data that relate to the entire market or data that relate to certain specific market participants only.

EN 21 EN

rules regarding causality, whether adjustments need to be made to the observed data in order to account for such influences. These could be simple adjustments to the data in cases where the influencing factor and the magnitude of its effects can be relatively easily ascertained and accounted for (see Section (1) below). More sophisticated adjustments of observed comparator data can be obtained on the basis of econometric techniques, in particular through the use of regression analysis, which is described in Section (2) below. Whether it is for the defendant or the claimant to plead, substantiate and prove such adjustments is a matter of applicable law.60

56. In a given case, the choice between these different techniques depends on the specific circumstances of the case and applicable legal rules, taking account of the different advantages and disadvantages of these techniques, for instance with regard to their accuracy and precision and the data requirements they entail (see Section (3) below).

(1) Simple techniques: individual data observations, averages, interpolation and simple adjustments

57. Depending on the requirements under applicable national law and on the circumstances of the case, especially the degree of similarity between the infringement market and the comparator market or period, the data observed may be compared directly, i.e. without further adjustments, with the data observed in the infringement market.61

58. The amount of data observed for the variable of interest (e.g., in the flour cartel example, the price for flour) in the comparator markets or comparator time periods may range from only one or very few data observations (i.e. the price observed in a small number of transactions) to a large number of data observations. In bidding markets, for example, auctions may occur very infrequently and at the time of the damages estimation only the price observed in the one tender after the infringement may be available. A similar situation could occur in industries where long-term contracts are common. It may be appropriate to use damages estimations based on single data observations where these are sufficiently representative for the period of interest.

59. Where looking at comparator markets or time periods produces a greater number of data observations, e.g. the prices paid by the injured party in a series of post-infringement transactions, or the prices paid by a number of customers in another geographic market, these data observations can be used either individually or in the form of averages.62

60 See, for instance, Kammergericht Berlin (Higher Regional Court, Berlin), decision of 1 October 2009,

case No 2 U 10/03 Kart., as an example of the distribution of fact pleading obligations in the quantification of harm.

61 For instance, time-based comparison could be based on the simple observation of prices before and during the infringement. For an example of the legal implications of such method see Corte Suprema di Cassazione (Supreme Court of Cassation, Italy), decision of 2 February 2007, case No 2305 (Fondiaria SAI SpA v Nigriello).

62 For the purposes of this Guidance Paper, the term ‘average’ is used as referring to the mean, i.e. the average calculated by dividing the sum of observations by the number of observations. There may, however, be situations where it may be more appropriate to use other descriptive statistics (i.e. the median or the mode). For example, where in a market of 25 companies, 21 charge a price of € 50 and

EN 22 EN

60. The use of various forms of averages or other forms of data aggregation can be appropriate, provided that like with like is compared. For example, where a wholesaler claims damages for having purchased a product in January, May, July and October 2009 from the participants in a price cartel and where the chosen method is comparison with another geographic market, the monthly average prices paid in that market by the same type of customer (wholesaler) during the same months may be the appropriate reference point (i.e. comparing January data with January data, May data with May data, and so forth). Comparing data from the same months will, for instance, account for seasonal differences over a year and thus make the comparison more reliable. If, however, little monthly price variation exists, the average price on the comparator market for the entire year of 2009 may be considered an appropriate indicator. It may also be the case that yearly data or other average data (e.g. aggregated industry data) are simply the only information available. Legal systems in the Member States may generally allow parties to rely on average data whilst granting the defendant the opportunity to show that significant differences exist, and they may require the use of more disaggregated data where available.

61. Another simple technique for deriving a comparator value from a range of data observations is linear interpolation. Where a comparison over time has produced price series from before and after the infringement, the “non-infringement” or “counterfactual” price during the infringement period can be estimated by drawing a line between the pre-infringement price and the post-infringement price, as shown in the illustration below. From this line, a comparator value can be read for each relevant point in time during the infringement period. Compared with the calculation of a single average value for price during the entire infringement period, interpolation therefore allows to some degree to account for trends in price developments over time that are not due to the infringement. Reading comparator data from the interpolated line will, therefore, produce more accurate results than using an average value for the period, e.g. in cases where damages are claimed that result from transactions (or other events) which occurred only towards the beginning or the end of the infringement period.63 The following illustration gives a simple example of linear interpolation (the dotted line shows the interpolated non-infringement price, the full line the actually observed prices):

four a price of € 75, the modal price of € 50 (the price most observed in the sample) may be the more meaningful representation of the market price than the mean of € 54 (in this example, the modal price equals the median price, which is the price charged by the middle-ranked company).

63 Interpolation likewise has advantages over using averages where the number of transactions (or other events) is unevenly distributed during the infringement period.

EN 23 EN

05

1015202530354045

1994 1999 2004 2009

Price Actual priceEstimated price

Infringementeffects start

Infringement effects end

Linear extrapolation works similarly to interpolation except that the line is continued from either only pre- or only post-infringement data.64

62. There may be situations where it is quite straightforward to identify a differentiating factor between an infringement market (or period) and a comparator market (or period) and to make the corresponding adjustment to the value of the observed comparator data. For example, certain seasonal effects occurring on a market may have a pattern and a magnitude that can in some cases be rather easily understood from internal business records of a party or from other sources, such as expert statements. In these cases, for example, the straight line obtained in a simple linear interpolation should be adjusted to reflect such patterns.65

(2) Regression analysis

a. Concept and purpose of regression analysis

63. Regression analysis uses statistical techniques to investigate patterns in the relationship between economic variables and to measure to what extent a certain variable of interest66 (e.g., in the flour cartel example, the price for flour67) is influenced by other variables that are not affected by the infringement68 (e.g. raw material costs, variations in customer demand, product characteristics, the level of market concentration69). Regression analysis therefore makes it possible to assess

64 Extrapolation thus extends a trend existing in a time series either before or after the infringement. For

example, if in the three years before a cartel prices were € 12, € 13.20, and € 14.52 respectively (reflecting a 10 % increase each year), a simple technique would be to estimate that prices during the two-year duration of a cartel were € 15.97 and € 17.57 respectively; a more accurate estimation of the underlying trend could be obtained through using regression analysis.

65 Such adjustment, could, data permitting, be done in a more sophisticated way by using regression analysis as explained in the following section.

66 Also referred to as an ‘explained variable’ or ‘dependent variable’. 67 Possible other variables of interest for which regression analysis may be applied include, for instance,

sales volumes, market shares or profit margins (e.g. those of an excluded competitor who claims damages for loss of profits through a reduction of sales or a decrease in its margins), costs of production (which may also be relevant in the context of a loss of profits estimation).

68 Also referred to as ‘explanatory variable(s)’ or ‘influencing variable(s)’. 69 Other factors influencing the variable of interest may, for example, include customer and order sizes,

the technology used for production, the size and cost structure of the firms offering the product, or advertising expenditure.

EN 24 EN

whether, and by how much, factors other than the infringement have contributed to the difference between the value of the variable of interest observed on the infringement market during the infringement period and the value observed in a comparator market or during a comparator time period. Regression analysis is thus a way to account for alternative causes for the difference between the compared data sets. All comparator-based methods are, in principle, capable of being implemented through regression analysis provided that sufficient data observations are available.70

64. In a regression analysis, a number of data observations for the variable of interest and the likely influencing variables are examined by means of statistical techniques. The relationship identified is usually described in the form of an equation (referred to as a ‘regression equation’ or ‘regression model’). This equation makes it possible to estimate the effects of influencing variables on the variable of interest and to isolate them from the effects of the infringement. Regression analysis estimates how closely the relevant variables are correlated71 with each other, which may in some instances be suggestive of a causal influence of one variable on the other.72

65. There are two main approaches to carrying out a regression analysis for damages estimation, depending on whether only data from non-infringement periods (markets) are used to build the regression equation or whether, in addition to non-infringement data, also data from within the infringement period (market) are used. If only data from non-infringement periods are used to estimate the regression, the regression equation would be used to ‘forecast’ the effect on the variable of interest during the infringement period on the basis of the pattern identified outside this period (‘forecasting approach’).73 Where, in addition, also data from the infringement period (market) are used to estimate the regression, the effect of the infringement would be accounted for in the regression equation through a separate indicator variable (called ‘dummy variable’).74 Whether it is more appropriate to apply the forecasting or the dummy variable approach will depend on the circumstances of the case.75

70 A sufficient number of data observations is, however, required to apply statistical methods in a

meaningful manner. Such sets of data observations could be obtained (in comparisons over time) from time series of observations, or (in comparisons at one point in time) from a range of comparator markets or from a range of firms or a range of transactions, or a combination of both (observations over time from a range of markets, firms or transactions).

71 In multivariate regression analysis (see in more detail below), the correlation established is a conditional correlation, i.e. one where the effect of other variables is controlled for.

72 Provided this is consistent with a coherent economic framework and with other pieces of qualitative and quantitative evidence.

73 This ‘forecasting approach’ is sometimes also referred to as a ‘residual model approach’. This approach is illustrated in the graph in paragraph 72 below.

74 Such a ‘dummy variable’ measures whether there has been an upward shift in the variable of interest during the infringement period.

75 In particular, while the forecasting method has the advantage of allowing the choice of a regression model that is only based on data observations from the non-infringement period (and hence, untainted by the effects of the infringement), using data from both periods/markets may allow a more precise and accurate estimation of the parameters of interest, in particular if the available non-infringement data are limited or do not allow the dynamics of the industry at hand to be fully captured. In practice, both methods can often be combined, e.g. by selecting the model on the basis of the pre-infringement period and estimating a dummy-variable regression using data from both periods (and allowing, if appropriate, the effects of the other influencing variables to vary in the infringement and non-infringement periods).

EN 25 EN

b. Examples and illustrations

66. A simple example that, for illustrative purposes, looks only at one potential influencing variable may show the basic steps in regression analysis. Assume that, in the above-mentioned example of a flour cartel, the prices paid by bakeries during the cartel period to the milling companies are compared with the prices paid by bakeries to the milling companies in the pre-infringement period, and that this comparison shows a price increase during the infringement period of 20 %. Assume further that there are indications that this increase is not exclusively due to the cartel but that during the infringement period costs for an important input material (e.g. cereals) also increased significantly. It is therefore not clear how much of the increase in price for flour is due to the infringement and how much is due to the increased input costs (the rise in cereals prices).

67. One option to address this uncertainty could be to use data from another period or market where input costs (price for cereals) were more similar and no infringement existed, but there may be situations where this is not possible.76 Regression analysis can offer a tool to account for the variation in input costs, by showing the statistical relationship between input costs and price for flour. To this end, a range of data observations on input costs (cereal prices) and on prices for flour during the period not affected by the infringement could be examined.77 Through applying statistical techniques to these data observations, it is possible to establish a pattern of how the prices for cereals influenced the price for flour in a period where the flour prices were not influenced by the infringement. It is then possible to deduce a statistical relationship between the price for flour and the price for cereals from this period. By applying the insight on this relationship to the prices for flour from the infringement period, it is possible to eliminate the part of the increase of prices for flour not imputable to the infringement, but to the change in input costs. This allows to ‘forecast’ prices for flour without the cartel overcharge but including the price increase caused by higher input costs.

68. The following graph gives a simple illustration of how such a statistical relationship is deduced. The chart shows several data observations of the input costs (cereals prices) and the corresponding price for flour at the same point in time during a non-infringement period. For instance, when at one particular moment the price for cereals was 60, the price for flour was 128. It is possible to calculate the coordinates of the line that best fits all data observations in order to represent the statistical relationship (correlation) between the price for cereals and the price for flour. This relationship is expressed in the graph below as a line and can be, and usually is, also expressed as an equation.78 The steepness of this line shows what increase in the price for flour is associated with a certain increase in the price for cereals. In the

76 For instance, because reliable data from other periods (or markets) are not available or because in such

other periods (or markets) market characteristics differed significantly. 77 On the possibility of whether or not to also consider data from the infringement period (market), see

paragraph 75 below. 78 Estimating a regression of price (as the variable of interest) over input costs (as the influencing variable)

provides the coordinates of this line. In this example, the ordinary least squares (OLS) technique is used to calculate the coordinates of a straight line that is located at a minimal distance (‘least squares’) from the set of data points on the graph. The OLS technique is a common statistical method to estimate the parameters of a linear regression model. Which statistical technique is used and how it is adjusted to the case at hand will depend on what is known about the general functioning of the industry in question.

EN 26 EN

example shown in the graph, the identified relationship indicates that e.g. a rise in the price for cereals from 50 to 60 relates to a rise in price for flour from 120 to 130. As an increase in input costs (cereals) by € 10 is associated with a flour price increase of € 10, the statistical relationship thus shows that an increase in this input cost is fully passed on.

100

110

120

130

140

150

10 20 30 40 50 60 70 80 90 100Input costs

PriceData observations

Regression line

69. Knowing the pattern of how the input cost (cereals prices) influenced prices for flour outside the infringement period makes it possible to estimate (‘forecast’) how much the observed higher values of these costs (cereals prices) during the infringement period influenced prices for flour. Excluding these effects from the price comparison allows the price overcharge caused by the infringement to be estimated on a more reliable basis than without the regression analysis. In the above example, if during the infringement period the price for flour was 140 instead of 120 during the non-infringement period, but the input cost (cereals prices) increased from 50 to 60, the likely price for flour without the cartel would not be 120 but 130.

70. Whilst the example described so far concerned only the influence of a single other variable (cereals price as input cost) on the variable of interest (flour price), regression analysis in competition practice usually has to account for several other factors influencing the variable of interest (multiple regression analysis79). In this situation, data need to be observed for all additional relevant influencing variables and a regression equation needs to be deduced from these data that reflects their relationship to the variable of interest. For instance, in the above-mentioned flour cartel example, it may be the case that during the infringement period the milling companies not only had to pay higher prices for cereals, but were also subject to an increase in energy and labour costs and introduced a more efficient milling and packaging technology, all of which may have had an impact on the price of the flour

79 Also referred to as ‘multivariate regression analysis’ as opposed to ‘single variable (‘univariate’)

regression analysis’ as used in the above example.

EN 27 EN

they sold to bakeries during the cartel period. To identify the statistical pattern of how these factors influenced the flour price, series of data observations for each of these influencing variables need to be analysed.

71. When undertaking a regression analysis, it is important to consider all variables that are relevant in the specific case. Suppose that either the defendant or the claimant uses, in a comparison of the flour prices charged by a mill before and during an infringement, a multiple regression analysis to control for the potential influence on the flour price of the above-mentioned factors (i.e. the cereal prices, energy and labour costs and milling and packaging technology). If, however, a significant demand change took place during the cartel (e.g. higher demand by bakeries for flour due to an increased demand by end customers for bread and cake) and if the influence of this event on the price for flour is not accounted for in the regression equation, the estimate of the infringement effect is likely to be biased, despite the otherwise comprehensive regression analysis.80 It is for the applicable national law to determine, in accordance with the principle of effectiveness, the party on which the burden falls to invoke and prove facts, such as the above-mentioned change in demand or the completeness of the variables considered in a regression analysis.

72. The basis of each damages quantification using regression analysis is thus the statistical relationship between the variable of interest (e.g. price) and the relevant explanatory variable(s) expressed in a regression equation. When the forecasting approach is used,81 the estimation of a regression equation using data from the non-infringement period constitutes the first step. In a second step, using this regression equation and the observed values of these relevant variables during the infringement period, the price injured parties are likely to have paid without the infringement can then be estimated. In a third step, the difference between this likely non-infringement price and the price actually paid by the injured parties gives an estimate of the overcharge resulting from the infringement. The graph below illustrates the second and the third step. When the dummy variable approach is used, the regression analysis combines the three steps described above.82

80 It is, however, important not only to include all relevant factors in the regression model, but also to

refrain from including variables that appear clearly irrelevant (on the basis of industry knowledge). In fact, damages estimates could be wrongly lowered (even down to zero) if irrelevant variables are included in order to explain the price variation in the model.

81 The alternative approach is the dummy variable approach; see paragraph 65 above. Unlike the forecasting approach, the dummy variable approach estimates the effect of the infringement in a single step, by carrying out a regression analysis using data from both the infringement and non-infringement periods. In the case of the example above, this approach would estimate the effect of the cartel as the upward shift in price that is observed during the cartel period (i.e. the coefficient of the dummy variable in the regression equation) and is not explained by changes in other influencing variables, such as raw material costs.

82 In this case, the regression equation is estimated using data from both the infringement and non-infringement periods and directly indicates how much the variable of interest changed during the infringement period after accounting for the effect of other explanatory variables.

EN 28 EN

60

70

80

90

100

110

120

130

1990 1995 2000 2005 2010

PriceActual priceEstimated price

Infringementeffects start

Infringement effects end

73. The regression analysis illustrated in this graph is based on the forecasting approach, in which a regression is carried out on pre- and post-infringement data to establish in an equation the statistical relationship between price and various relevant explanatory variables (input costs and other relevant factors). Using this equation and the observed values of the relevant explanatory variables, an estimated price can be derived that is likely to have prevailed absent the infringement (dotted line). The continuous line is the actually observed price. The difference between the continuous and the dotted line during the infringement period is the estimated overcharge. The dotted line outside the infringement period is also derived from the regression equation and can serve, through comparison with the actually observed non-infringement prices (continuous line), to assess the predictive power of the regression model.

c. Requirements for applying regression analysis

74. Carrying out a regression analysis requires knowledge of various statistical techniques to measure the relationship between variables, to construct an appropriate regression equation and to calculate the precision of the parameters in this equation. In addition, it is necessary to have a good understanding of the industry concerned, in the first place, to formulate the right hypotheses when constructing the regression equation and to make the right choice as to the factors that are likely to have significantly influenced the variable of interest (and which should therefore be included in the analysis). Industry understanding is furthermore necessary to make informed choices about which statistical techniques to use in a given situation, for instance, to account for unusual observations (outliers) or other specific features in data sets. In particular, where the influencing variables were themselves affected by the infringement, biased results may occur if this aspect is not taken into account,

EN 29 EN

e.g. through applying specific statistical techniques83 or through using data observations that lie outside the infringement period or market.84

75. Without a sufficient number of data observations, statistical analysis cannot identify relationships between economic variables. To identify the effect of influencing variables on the variable of interest therefore requires that a sufficient range of data observations is available for all variables considered. Regression analysis therefore typically requires extensive data. However, statistical techniques may help to overcome some gaps in data or biases in their interpretation85 and there can be situations where also the analysis of a smaller number of data observations is meaningful.

76. Data observations can, in principle, be gathered at different levels of aggregation. For example, where the relationship between price and input cost is to be analysed, data series either for the prices charged in individual transactions, for annual industry average prices or — in between — monthly data at firm level could be examined next to data series either for individual input costs per unit or for industry cost averages respectively. Using disaggregated data makes it possible to analyse a greater number of observations and therefore to obtain more precise estimates. Where such disaggregated data do not exist or are not accessible to the party carrying out the regression analysis, the analysis of aggregated data may still produce informative results, in particular if the aggregated data have a high frequency.

77. Where used appropriately and on the basis of sufficient data observations, regression analysis can considerably refine the damages estimation through comparator-based methods. It should be stressed, however, that even very sophisticated regression equations rely on a range of assumptions and will (like any technique to predict a hypothetical situation) only be able to deliver estimates. It is good practice to consider the assumptions underlying a regression equation, because some assumptions may be more appropriate than others in a given situation and may lead to significantly different results.

78. Whether, by which party and at which stage of the proceedings a regression analysis is carried out in a court case will inter alia depend on the existence or accessibility of data and the rules under applicable law regarding fact pleading requirements, disclosure of evidence, the standard of proof and the allocation of the burden of proof between the claimant and the defendant.

79. The different forms of regression analysis mentioned above (paragraphs 65 et seq.) are sometimes referred to as ‘reduced form’ approaches, as they directly estimate parameters of an equation that are themselves derived from other economic relationships (e.g. the interaction of supply and demand), without modelling these explicitly. Alternatively, econometric models can be built to estimate these underlying economic relationships. Although such econometric models, which are usually referred to as ‘structural’, often rely on particularly strong assumptions, they

83 For example, the use of instrumental variables, an econometric technique that may be applied to correct

such bias. 84 In particular, by using the forecasting approach described above, where the value of the influencing

variables included in the model to predict the counterfactual are corrected for the infringement effect on these variables.

85 E.g. where a sample of data observations is not fully representative.

EN 30 EN

may bring a deeper understanding of the market concerned and form an integral part of simulation exercises to estimate damages (as further detailed in section III.A).

(3) Choice of techniques

80. Sections 1 and 2 above have described different techniques whereby comparator-based methods can be implemented in practice. In a given case, the choice of technique will usually depend on a range of aspects, in particular the legal requirements and the factual circumstances of the case. Considerations relating to the standard and burden of proof are likely to be very relevant in practice.

81. Econometric techniques can increase the degree of accuracy of a damages estimate and may thus help in meeting a higher standard of proof if required under applicable rules. Whether regression analysis is required (possibly in addition to other evidence available) to meet such a standard, and on which party the burden of proof falls in this respect are questions of applicable law, including the EU law principle of effectiveness. It should be considered that carrying out an econometric analysis usually requires a significant number of data observations, which may not always be accessible. Moreover, it may also be that in a given procedural situation the applicable standard of proof does not require the party charged with the burden of proof to go further than the techniques mentioned in Section 1 above. This could be because the national legal system concerned considers the markets or periods compared as sufficiently similar and the estimate of damages resulting from the simple comparison as sufficiently accurate for what the party has to show in the given procedural situation. It may also be that the legal system, in view of the damages estimation presented by a claimant and the data that are reasonably accessible to him, provides for a shift of the burden of proof from the claimant to the defendant. In such a situation, the defendant may consider carrying out a regression analysis to rebut the submission of the claimant.

82. Considerations of proportionality may also play an important role, as the gathering of data and their econometric analysis can entail considerable costs (including those of third parties) that may be disproportionate to or even exceed the value of the damages claim at hand. Such considerations may also become relevant with a view to the principle of effectiveness.86

83. Courts in the EU have mainly used straightforward implementations of comparator-based methods without regression analysis, often on the basis of averages.87 They have also accepted simple adjustments to the value of observed data when it is quite straightforward to identify a differentiating factor between an infringement market (or period) and a comparator market (or period). To date, little experience exists with econometric analysis in actions for antitrust damages before courts in the EU,88 although such techniques can, as described above, provide valuable help in quantifying the harm suffered through infringements of Article 101 or 102 TFEU.

86 See above paragraph 2 in Part 1, Section 1. 87 The use of averages was accepted in e.g. Landgericht Dortmund (Regional Court, Dortmund), decision

of 1 April 2004, Case No 13 O 55/02 Kart (Vitaminpreise); WuW/DE-R 1352. 88 For a recent example concerning lost profits in an exclusionary case see Juzgado Mercantil numero 2 de

Barcelona (Commercial Court, Barcelona), decision of 20 January 2011, case No 45/2010 (Céntrica Energìa S.L.U./Endesa Distribuciòn Eléctrica S.A.)

EN 31 EN

84. Courts in the EU sometimes also apply a ‘safety discount’, i.e. they deduct from the observed data values an amount sufficient, under the standards of applicable law, to take account of uncertainties in a damages estimate.89 Regression analysis can also be considered to account for these other possible influencing factors, and to obtain a “lower bound estimate” of the damages incurred.90

III. SIMULATION MODELS, COST-BASED ANALYSIS AND OTHER METHODS

85. Alongside comparator-based methods, other methods exist to establish an estimate for the hypothetical non-infringement situation. Such other methods include, in particular, the simulation of market outcomes on the basis of economic models (A), and the approach to estimate a likely non-infringement scenario on the basis of costs of production and a reasonable profit margin (B).

A. Simulation models

86. Simulation methods draw on economic models of market behaviour. Economic studies on how markets function and how firms compete with each other have shown that markets with certain characteristics may allow the likely outcomes of market interaction to be predicted, for instance the likely price or production levels or profit margins. The branch of economics known as industrial organisation has developed models of competition for various types of markets that can simulate such outcomes. These models range from monopoly models to, at the other end of the spectrum, perfect competition models. Intermediate models designed to reflect firm behaviour in oligopolistic markets are, in particular, those designed originally in the 19th century by the economists Augustin Cournot and Joseph Bertrand and numerous extensions and variations of the Cournot and Bertrand models. These include, in particular, also dynamic oligopoly models based on game theory91 that take into account the repeated interaction between firms in the market.92

87. Prices are likely to be highest (and sales volumes lowest) in a monopoly and prices are likely to be lowest (and sales volumes highest) in a situation of perfect competition. The particular oligopolies described by Bertrand (‘Bertrand oligopoly’)93 in markets with differentiated goods94 and by Cournot (‘Cournot

89 For instance, to exclude the effects on the variable of interest of possible other factors. See e.g.

Kammergericht Berlin (Higher Regional Court, Berlin), decision of 1 October 2009, case No 2 U 10/03 Kart.; Oberlandesgericht Karlsruhe (Higher Regional Court, Karlsruhe) of 11 June 2010 in case No 6 U 118/05 (appeal pending).

90 Indeed, in addition to providing damages estimates that already control for the influence of other factors, regression analysis also measures the precision of these estimates (in the form of standard errors), from which lower (and upper) bounds on the estimated damages can be obtained.

91 Game theory is the study of how people and firms behave in strategic situations in which they must consider how others respond to their action.

92 Taking into account the repeated interaction between firms in the market can be useful to explain, for instance, coordinated behaviour between firms or market entry of a new competitor. Dynamic models can be particularly useful for explaining exclusionary behaviour and to build a counterfactual to estimate the resulting damage.

93 The Bertrand oligopoly model of competition describes a market with a relatively small number of firms (and high barriers to entry) that compete on price, not output quantity. Firms set their price based on their beliefs about the prices their competitors will charge. In this model, prices increase with the degree of product differentiation.

EN 32 EN

oligopoly’)95 will normally lead to prices and volumes somewhere between perfect competition and monopoly levels; the exact outcome depends on the number of firms in the market and barriers to entry, on the degree of differentiation between them and their products and on other characteristics of the market at hand, such as demand characteristics (especially, how sensitive customers are to changes in price), and the capacities and cost structure of producers.

88. Based on such theoretical insights that link the market outcome e.g. in terms of prices to a given set of market characteristics, simulation models can be built to estimate the prices (or other variables) that are likely to have prevailed in the market had an infringement of Article 101 or 102 TFEU not occurred. The simulation model should be constructed in such a way that it replicates (a) the most significant factors influencing supply (in particular, the way competition takes place between firms (‘competitive interactions’)96 and the cost structure of firms) and (b) demand conditions (in particular, the extent to which customers respond to price changes). These factors would be expressed as a set of equations in which a number of parameter values need to be included. These values may be known, estimated econometrically or assumed so that the output of the model matches some observed variables. When using simulation models to generate a non-infringement scenario, the relevant market structure and other characteristics must be those that would have existed without the infringement; these may correspond to the structure and other characteristics of the market observed in the infringement scenario, but they may also differ to some extent.97

89. An example may illustrate the use of simulation modelling to estimate damages. In the example of a cartel on a differentiated product market (e.g. confectionary chocolates), non-infringement prices could be estimated as follows, using data from the non-infringement period. First, one would estimate how the demand for each chocolate product varies with its own price (own-price elasticity) and with the price of competing products (cross-price elasticity).98 Second, one would decide which

94 In a market with homogeneous goods with no capacity constraints, Bertrand price competition will, in

contrast, lead to very competitive outcomes. Homogenous goods are goods that have little differences in terms of quality or features.

95 The Cournot oligopoly model of competition describes a market with a relatively small number of firms (and high barriers to entry) that compete on the amount of output they will produce. Before they choose prices, they set their quantity (or capacity) simultaneously on the basis of how much they each believe the other firms will produce.

96 The term ‘competitive interactions’ is used to indicate how competition between firms takes place, e.g. (but not limited to) Bertrand or Cournot competition, or how firms refrain from competing between each other (in the case of collusive behaviour infringing competition rules). Markets on which price formation occurs through auctions or other bidding processes may also be conducive to modelling as interaction between competitors often follows fixed rules (prices or output quantities likely to result from an auction or other bidding process not affected by the infringement could, in particular, be estimated by oligopoly models that incorporate game theory to simulate the likely bidding behaviour of competitors in a non-infringement scenario).

97 As the infringement may have led to a change in the market structure or may have prevented changes in the market that would otherwise have occurred (e.g. the exit of an inefficient competitor), the (hypothetical) market characteristics in the non-infringement scenario are not necessarily the same as those that could be observed in the infringement scenario. In addition, market shares observed during an infringement may significantly differ from those that would have prevailed in the absence of the infringement as cartel members may allocate markets between themselves.

98 Technically, this would involve estimating a demand system, which is an example of the structural econometric analysis mentioned in paragraph 79.

EN 33 EN

model appropriately reflects the competitive interaction between firms in the non-infringement period (e.g. the Bertrand model of competition in the confectionary chocolates example). On this basis, it can be calculated at which prices the profits of the firms are maximised in view of the cost parameters (e.g. marginal costs) and demand parameters (e.g. the level of demand).99 The value of some of these parameters can then be adapted to reflect the relevant conditions during the infringement period (e.g. supposing the cost of cocoa increases by 10%). With all this information expressed in equations, it can be simulated (under the assumption that firms strive for maximised profits) what prices these firms are likely to have charged during the infringement period. The cartel overcharge can then be estimated by taking the difference between the observed prices and the simulated non-infringement prices.

90. This example is particularly demanding in terms of data requirements and assumptions. Simpler simulation models may be envisaged to estimate damages but they rely even more heavily on crucial assumptions that are difficult to verify. For example, damages following a cartel infringement could be calculated by comparing monopoly prices (aimed at reflecting prices during the cartel) with prices expected under a Cournot model (aimed at reflecting prices in the non-infringement scenario), using data such as market shares, costs, and market price elasticity. However, such a method crucially depends on the assumed competitive interactions in the infringement and non-infringement scenarios and entails the risk that these do not mirror sufficiently closely the way in which the cartel operates during the infringement period and the way in which competition on the market would have operated absent the infringement.

91. Simulation models can be used to estimate market outcomes not only in cartel cases (or other price raising infringements), but also in cases of exclusionary behaviour. For example, an oligopoly model could be used to simulate the sales volume and the market share a foreclosed competitor would have attained had the infringement not taken place.

92. Each model simulating market outcomes is an approximation of reality and relies on theoretical and often also factual assumptions regarding market characteristics and the likely behaviour of producers and customers. Although, by their very nature, models rely on simplification of reality, even simple models may in certain cases provide useful insights regarding the likely damages.100 Building a comprehensive model that replicates a range of specific features of the market in question, if it can be properly solved and evaluated, can increase the likelihood that the result of the simulation is a reasonable estimate for the hypothetical non-infringement scenario. Even very comprehensive models, though, still depend very much on the right assumptions being made, in particular regarding the central questions of what is the likely mode of competition and the likely customer demand in the non-infringement scenario. Moreover, the development of complex simulation models can be technically demanding and may require significant amounts of data that may not

99 The value of these parameters (e.g. the value for marginal costs used in the calculation) in the non-

infringement period can be determined so that the derived prices and volumes match the observed data. 100 Therefore, pointing out that a model relies on seemingly simplifying assumptions should therefore on

its own not be sufficient to dismiss it; rather, one should consider how some of the simplifying assumptions are likely to affect its results.

EN 34 EN

always be accessible to the party concerned or possible to be estimated with sufficient reliability.

93. Nonetheless, both simple and more complex simulation models could provide useful insights when estimating the outcomes that a market would have produced absent an infringement of Article 101 or 102 TFEU. Whether and in which procedural situation legal systems will consider that the use of an economic simulation is appropriate and its results are sufficiently reliable will depend on the specific circumstances of the case in point and the requirements under applicable legal rules.

B. Cost-based method

94. Another approach to estimating the likely prices that would have emerged absent the infringement is provided by the cost-based method.101 This method consists in using some measure of production costs per unit, and adding a mark-up for a profit that would have been ‘reasonable’ in the non-infringement scenario. The resulting estimate for a per unit non-infringement price can be compared to the per unit price actually charged by the infringing undertaking(s) to obtain an estimate of the overcharge.102

95. Different types of production costs may be suitable for implementing the cost-based method, depending on the characteristics of the industry concerned. It is, however, essential to ensure that the treatment of costs and margins is consistent. For example, if variable costs (i.e. costs that vary with the level of production) are considered as the basis of this exercise, a gross margin (i.e. the margin earned once variable costs have been deducted) should be added to calculate the price. It should also be noted that the relevant cost for determining prices may be not only the cost of the infringer, but also the cost of one of its competitors (e.g. if the price in the market is determined by the least efficient producer).

96. The first step of the cost-based method is to determine the production cost per unit. Per unit costs can be estimated by dividing the actual relevant production costs incurred by the infringer(s) for the relevant business activity by the total number of products produced. This approach can be rather straightforward where companies or separate business divisions of companies produce only one main product. Such companies or business divisions sometimes publish their major cost data or file this information as part of their audited accounts with public registries. In other situations, the access to data and the allocation of costs to the product affected by the infringement is more difficult. Where accounting data are available, adjustments may be necessary given that the notions of costs in accounting terms can differ from the notions of costs in economic terms.

101 This method is also referred to as the ‘cost plus method’ or ‘bottom-up costing method’. It is

mentioned, as a subsidiary approach in cases where comparator-based methods are not appropriate, by the Bundesgerichtshof (Federal Court of Justice, Germany), decision of 19 June 2007, case No KBR 12/07 (Paper Wholesale Cartel).

102 Usually, the cost-based method is considered for quantifying price overcharges. The method, or elements of it, may, however, also be used for quantifying other forms of harm such as the profits lost by foreclosed competitors. For instance, the Oberlandesgericht Düsseldorf (Higher Regional Court, Düsseldorf), decision of 16 April 2008, case No VI-2 U (kart) 8/06, 2 U 8/06 (Stadtwerke Düsseldorf), estimated the lost profits of a foreclosed competitor by considering the costs of the competitor and the likely profit margin expressed as a proportion of these costs.

EN 35 EN

97. It may occur that the observed production costs during the infringement are not representative of the production costs that would have been likely without the infringement. This could mainly be for two reasons: first, in the event of infringements of Article 101, companies which due to their collusive behaviour are not subject to the competitive pressure that would exist in the non-infringement scenario may operate less efficiently and therefore generate higher production costs than under competitive pressure. Second, infringers may restrict output and may therefore, during the infringement, forego economies of scale that would have led to lower production costs. Where indications for such situations exist, adjustments to the observed costs data of the infringer(s) may be appropriate. Where such adjustments are not made, the observed costs may still contribute, under the cost-based method, to a lower-bound estimate of the possible price overcharge.

98. The second step of the cost-based method requires a ‘reasonable’ profit margin to be estimated and added to the per unit production costs. Various approaches exist to estimate a ‘reasonable’ profit margin. They are based either on a comparison over time or across markets, or on economic models, and thus have commonalities with the methods described in the preceding Sections. For instance, an estimate for the profit margin that could reasonably be expected in a non-infringement scenario may be derived from the profit margins made by similar undertakings in a comparable geographic market not affected by the infringement or in comparable product markets.103 Similarly, the profit margins of the infringing (or a similar) undertaking during the pre- or post infringement periods could be used as a basis for the estimate. Both these comparator-based methods rest on the assumption that the reference period, market or firm are sufficiently similar,104 in particular with respect to market characteristics that are relevant for profit margins such as the level of competition in the market,105 the cost structure of producers (including costs of innovation), capacity utilisation and capacity constraints. These assumptions are not always easily verified, as a large number of factors and strategic decisions are likely to determine a firm’s price and margin setting.

99. Another approach to estimating a ‘reasonable’ profit margin is to consider the nature of competition and the characteristics of the market absent the infringement and to derive a likely profit margin from the insights from industrial organisation models.106 For instance, absent the infringement, prices may be likely to tend towards costs due to relative homogeneity of goods and overcapacities in the market; in such cases, the likely profit margin of producers would be relatively low.107

103 See Bundesgerichtshof (Federal Court of Justice, Germany), decision of 19 June 2007, case No KBR

12/07 (Paper Wholesale Cartel), referring to the profit margins generated in ‘comparable industries’. 104 For relevant considerations regarding sufficient similarity see above paragraphs 33-52 in Part 2,

Section II. 105 E.g. whether competition would have been so strong as to drive the price downwards towards marginal

costs (as assumed in the model of perfect competition) or whether profit margins, due to an oligopolistic structure, would have been higher even without the infringement.

106 See above paragraphs 86 et seq. in Part 2, Section III. 107 The cost of capital (i.e. the cost at which a firm can obtain capital on the market) is sometimes

considered as an approximation of a ‘reasonable’ profit margin in such cases. However, margins in the absence of an infringement may significantly differ from the cost of capital, for example in the absence of perfect competition or in the presence of firm-specific cost advantages for certain firms, or demand and supply shocks.

EN 36 EN

100. It is clear from the above that both the estimation of likely non-infringement costs and the estimation of a ‘reasonable’ profit margin can, in practice, require a range of difficult issues to be considered. In addition, the cost-based method supposes access to data that may be in the possession of the opposing party or a third party. Nonetheless, depending on the circumstances of the particular case and on the requirements under applicable legal rules, it may provide useful insights to support an estimation of the harm suffered through an antitrust infringement.

C. Other methods

101. The methods described in this Guidance Paper are those that have received most consideration so far in legal practice and academic scholarship. They should, however, not be seen as an exhaustive list, firstly, as the methods described could further evolve or others could be developed in practice.

102. Secondly, there are methods not discussed in this Guidance Paper could nonetheless prove useful, in particular, in order to establish an upper- or lower-bound108 or approximate estimate109 for the harm suffered. Especially where the legal systems provide for the possibility of an approximate estimation, national courts have opted for pragmatic techniques rather than a sophisticated implementation of the methods set out in Sections A and B above to establish the amount of damages to be awarded to injured parties. For instance, in cases where a new entrant has been foreclosed in breach of Article 101 or 102 TFEU, business plans have sometimes been used110 as a source of information on the likely profits of a business, albeit in some instances adjusted depending on the market circumstances or through the use of data from a comparator market or undertaking.

103. It is for national courts to establish whether, under the applicable rules, a method can be accepted for the quantification of harm in a given case, provided that the principles of effectiveness and equivalence of EU law are observed.

IV. CHOICE OF METHODS

104. Each of the methods described in Sections II and III above can, in principle, provide useful insights in relation to all infringements of Article 101 or 102 TFEU and the different types of harm such infringements tend to produce. In particular, they make it possible to estimate not only the amount of illegal price overcharge in a price fixing cartel but also, for example, the sales volume or the profit lost by a company suffering harm through an exclusionary abuse by a dominant competitor.

105. It should be stressed that it is only possible to estimate, not to measure with certainty and precision, what the hypothetical non-infringement scenario is likely to have looked like. There is no method that could be singled out as the one that would in all

108 For example, an upper-bound estimation could be obtained through critical loss analysis. This technique

assesses for a price increase what loss in quantities would make that price increase unprofitable. 109 For instance, counterfactual profits could be prima facie identified by taking as a benchmark the cost of

capital, on the assumption that, absent the infringement, the undertaking would have earned the cost of capital, which represents the minimum return required by providers of capital to an undertaking. On the limitations of this approach, see footnote 107.

110 See for instance Højesteret (Danish Supreme Court), judgment of 20 April 2005, case UFR 2005.217H (GT Linien A/S v De Danske Statsbaner DSB and Scandlines A/S).

EN 37 EN

cases be more appropriate than others. Each of the methods described above has particular features, strengths and weaknesses that may make it more or less suitable to estimate the harm suffered in a given set of circumstances. In particular, the methods differ in the degree to which they are simple to apply, in the degree to which they rely on data that are the outcome of actual market interactions or on assumptions based on economic theory and in the extent to which they take into account factors other than the infringement that may have affected the situation of the parties.

106. In the specific circumstances of any given case, the appropriate approach to quantification must be determined under the applicable rules of law. Relevant considerations may include, alongside the standard and burden of proof under applicable legal rules, the availability of data, the costs and time involved and their proportionality in relation to the value of the damages claim at stake. The costs to be considered in this context may not only be those incurred when the party bearing the burden of proof applies the method, but also include the costs for the other party to rebut its submissions and the costs to the judicial system when the court assesses the results produced by the method, possibly with the help of a court-appointed expert. The costs and burden for an injured party and their proportionality may become particularly relevant with a view to the principle of effectiveness.111 Moreover, the decision under applicable law as to whether and, if so, which of the methods and techniques described in this Guidance Paper should be used may also depend on the availability of other evidence, for instance documentary evidence produced by the undertakings on the course of business showing that an illegally agreed price increase was actually implemented at a certain amount.

107. It may be that in a given case the application of several methods (e.g. comparison over time and comparison across geographic markets) is envisaged, either alternatively or cumulatively. Where two different methods yield results that are similar, such findings may lead a legal system to attribute stronger evidentiary value to the damages estimate, possibly a lower bound, based on these methods. Where, however, the application of two methods produces apparently contradictory results (especially when two opposing parties each rely on a different method), it is normally not appropriate either to simply take the average of the two results nor would it be appropriate to consider that the contradictory results cancel each other out in the sense that both methods should be disregarded. In such a scenario it would rather be appropriate to examine the reasons for the diverging results and to carefully consider the strengths and weaknesses of each method and its implementation in the case at hand.

111 See above paragraph 2 in Part 1, Section 1.

EN 38 EN

PART 3 — QUANTIFYING HARM CAUSED BY A RISE IN PRICES

I. EFFECTS OF INFRINGEMENTS LEADING TO A RISE IN PRICES

108. Anticompetitive practices can have the effect of raising the prices that direct and often also indirect customers112 of the infringing undertakings pay for the product concerned. Typical examples of infringements leading to such increases are price cartels, or excessive pricing by a dominant undertaking. Customers can also be affected by practices that limit output or allocate customers or markets — distortions of competition which in turn normally lead to a rise in prices. A different type of harm is caused where infringements adversely affect the market position of competitors; the quantification of such harm and its consequences for customers is discussed in Part 4 below.

109. In so far as infringements lead to a rise in prices for the products concerned, two main kinds113 of harm caused by such infringement can be distinguished:

(a) the harm resulting from the fact that direct and indirect customers of the infringing undertakings have to pay more for each product they purchase than without the infringement (the ‘overcharge’). This type of harm is further discussed in Section II; and

(b) the harm resulting from the so-called ‘volume effect’, which is caused by the fact that fewer of the products in question are bought due to the rise in prices. This type of harm is further discussed in Section III.

112 The direct customers of the infringing undertakings are those who purchase a product directly from one

of the infringing undertakings; indirect customers are those who purchase a product affected by the infringement from such direct customers or from other indirect customers. In some instances, undertakings that do not infringe the competition rules themselves can raise their prices, as market prices are higher because of the infringement. Customers who purchase from these undertakings are sometimes referred to as ‘umbrella customers’. To what extent such customers can claim compensation for the harm from the infringing undertakings depends on the applicable legal rules.

113 For other kinds of harm, see above paragraph 19 in Part 1, Section III.

EN 39 EN

The following figure represents in a stylised way these two main effects:

P1

P2

B

Price

Demand

Quantity0

A

Q2 Q1

110. P1 is the price charged if no infringement of Article 101 or 102 TFEU affects the market. In a perfectly competitive market, this price will equal the supplier’s marginal cost114 of providing that product. Many markets are in fact not perfectly competitive and non-infringement prices on these markets will be above the level of marginal costs. At price P1, Q1 is the quantity of the product bought by customers.

111. P2 is the higher price resulting from an infringement having an effect on price. This in turn leads to lower demand (Q2) because some customers will consider that the higher price they have to pay exceeds the value of owning the product or of benefiting from the service. This effect is referred to as the ‘volume effect’ or the ‘quantity effect’. The degree to which a rise in prices affects demand depends on demand elasticity.115

112. Rectangle A represents the value transferred from the customers to the infringers due to the infringement: the customers who buy at the higher price P2 have to transfer more money to the infringing undertaking(s) in order to obtain the product. They can demand compensation for having had to pay more and Section II below will explain how to quantify this harm.

113. Triangle B represents the volume effect and thus the value foregone by those who would have bought the product for price P1, but refrain from doing so when the price rises to P2.116

114. Some customers use the product in question for their own commercial activities — for example to sell it on or to manufacture other goods. When they do not buy at

114 These are the costs for suppliers to produce one more unit. 115 Demand elasticity measures by what percentage the quantity sold of a product in a given market varies

in response to a one percent price change for a particular demand level, and provides a useful indication of the magnitude of the volume effect for small price changes.

116 For the economy as a whole, this triangle therefore represents the loss in value for customers due to a reduction in output: while the overcharge affects the distribution of assets within the economy, triangle B means welfare not created because of the infringement. This is referred to in economics as ‘deadweight loss’.

EN 40 EN

price P2 (or buy less), they forego the profit they would have made had they been able to purchase at price P1. They can claim reparation for this loss of profit and Section III below will illustrate how to quantify this harm. Other customers are end-consumers. If these do not purchase at price P2 this means that they fail to enjoy the utility of these products or services, for which they would have been prepared to pay price P1.117 Applicable legal rules may provide that some or all of such harm should be compensated for such failure to enjoy the usefulness of the product. At a minimum, end-consumers who have to bear higher costs (for example for the purchase of a substitute good) and who therefore have suffered an actual loss118 must be able to obtain compensation.

115. The foregoing summarises the basic effects on the market of infringements that lead to a higher selling price. Infringements of Article 101 or 102 TFEU can also affect the demand side and lead to lower purchasing prices paid by infringers in their own supply with products, for example in the case of a buyers’ cartel or in the abuse of market power exercised by a dominant buyer vis-à-vis its suppliers. In such a case, the price effects would consist in an ‘undercharge’ for the supplier of the infringer, and often also an overcharge on the downstream markets, i.e. for the direct and indirect customers of the infringer.119 The same methods used to quantify an overcharge can, in principle, also be used to quantify the undercharge, e.g. the lower prices paid by the members of a buyers’ cartel vis-à-vis their suppliers.

116. The same methods can, in principle, also be used120 where at first sight no overcharge is visible, because the infringement served to artificially stabilise prices over a certain period of time in which prices would under normal market circumstances (i.e. without infringement) have declined. In the following, the term “overcharge” designates also these situations.

II. QUANTIFYING THE OVERCHARGE

117. Different types of infringements lead directly or indirectly to overcharges. Antitrust damages actions often deal with overcharges caused by cartels, which will be addressed in Section A below. The quantification of overcharges caused by other types of infringements will be addressed in Section B below.

A. Quantifying overcharges caused by cartels

118. In an action for compensation, it will be necessary — within the framework of applicable legal rules — to quantify the overcharge paid by the claimant(s). Economic and legal studies have analysed the effects of cartels; some insights from

117 It is also possible that customers would have been prepared to pay a price higher than P1, but lower than

P2. 118 See, for this legal term, Joined Cases C-295/04 to C-298/04 Manfredi [2006] ECR I-6619, paragraph

95. 119 In order to drive down input prices, the cartel members/dominant buyers with downstream market

power are likely to restrict their input purchases, hence also reducing output sales and increasing downstream prices.

120 Only the method based on comparison between time periods in the variant of ‘before and during’ comparison (i.e. comparing the infringement prices with pre-infringement prices) would obviously be unsuitable, unless regression analysis or simple adjustments are applied to account for the factors that would lead to a price decrease under normal market circumstances (e.g. decreased raw material costs).

EN 41 EN

these studies are set out below in Section 1. In actions for damages, it is useful to distinguish between the initial overcharge paid by the direct customer of the infringing undertaking (see below Section 2) and the possible harm that such overcharge causes to indirect customers at different levels of the supply chain (Section 3).

(1) Effects of cartels

119. Cartels are agreements and concerted practices between two or more undertakings aimed at influencing the parameters of competition through practices such as fixing the purchase or selling price or other trading conditions, allocating production or sales quotas or sharing markets (including bid-rigging). For the purpose of finding whether such practices infringe Article 101 TFEU there is no need to quantify the concrete effects of such a practice, because the object of the cartel agreement is the prevention, restriction or distortion of competition.121

120. Infringing the competition rules exposes the cartel members to the risk of being discovered and thus subject to a decision finding an infringement and imposing fines. The fact alone that undertakings nonetheless engage in such illegal activity suggests that they expect to reap substantial benefits from their actions, i.e. that they expect the cartel to have effects on the market and, hence, on their customers.122

121. A study undertaken for the Commission examined the empirical evidence on the existence of overcharge effects and on their magnitude.123 This study draws on a range of existing empirical studies on the effects of cartels. In particular, it refines the sample of cartels examined in the most comprehensive existing study by considering only cartels (a) that started after 1960 (thus taking into account only more recent cartels), (b) for which an estimate of the average overcharge was available (rather than only an estimate of the highest or lowest overcharge), (c) for which the relevant background study explicitly explained the method for calculating the average overcharge estimate, and (d) which were discussed in peer-reviewed academic articles or chapters in books.124 While some care is required in interpreting the results of this exercise,125 the study undertaken for the Commission contains some useful information as to the effects of cartels.

121 See judgments of the General Court in Joined Cases T-25/95 etc. Cimenteries CBR SA v Commission

[2000] ECR II-491, paragraphs 837, 1531, 2589; Case T-202/98 Tate & Lyle v Commission [2001] ECR II-2035, paragraphs 72–74; see also Communication from the Commission: Guidelines on the application of Article 81(3) of the Treaty, OJ C 101, 27.4.2004, p. 97, paragraphs 20–23.

122 See also the decision of the Kammergericht Berlin (Higher Regional Court, Berlin) of 1 October 2009 in case No 2 U 10/03, where the court referred to a similar argument.

123 See the external study prepared for the Commission ‘Quantifying antitrust damages’ (2009), pages 88 et seq., available at http://ec.europa.eu/competition/antitrust/actionsdamages/index.html.

124 In all, the study considers 114 cartels based on different types of collusion, including bid-rigging. The sample includes international and national cartels that affected a wide range of different industries. The geographic spread of the sample extends to the US and Canada as well as cartels from Europe and other regions.

125 In particular, it seems possible that cartels that do have an effect on the market receive more attention in empirical studies than those that have no effects, which may lead to a certain bias in the findings; see the study ‘Quantifying antitrust damages’, page 89 (ref. in footnote 123), for further details about the interpretation of the data used in the study.

EN 42 EN

122. On the basis of the data observed, this study found that in 93 % of all cartel cases considered, cartels do lead to an overcharge. As to the magnitude of the cartel overcharge, this study made the following findings:126

0

5

10

15

20

25

30

35

40

No overcharge

>0 -10 >10 -20 >20--30 >30 -40 >40 -50 >50- 60 >60 -70

% of obser vations

Overcharge (%)

123. According to this study, there is thus a considerable spread of the overcharges observed (with some cartels even having an overcharge of more than 50 %). About 70 % of all cartels considered in this study have an overcharge of between 10 % and 40 %. The average overcharge observed in these cartels is around 20 %.

124. The insights of this study concord with those of other available empirical studies, namely that (a) the vast majority of cartels do in fact lead to an overcharge, and (b) there is considerable variance in the overcharges observed. Also, all of these other empirical studies come largely to a similar estimate of the magnitude of the average overcharges as described above.127

125. These insights into the effects of cartels do not replace the quantification of the specific harm suffered by claimants in a particular case. However, national courts have, on the basis of such empirical knowledge, asserted that it is likely that cartels normally do lead to an overcharge and that the longer and more sustainable a cartel was, the more difficult it would be for a defendant to argue that no adverse impact on price did take place in a concrete case.128 Such inferences, however, are a matter for national legal systems.

126 Study ‘Quantifying antitrust damages’, page 91 (ref. in footnote 123). That magnitude is expressed as a

percentage of the actual price. This means that if the actual price (meaning the price paid as influenced by the infringement) is € 100 and the overcharge is said to be 10 %, the price absent the infringement is deemed to be € 90.

127 For details and further references see the study ‘Quantifying antitrust damages’, pages 89 et seq. (ref. in footnote 123).

128 See for example Bundesgerichtshof (Federal Court of Justice, Germany), decision of 28 June 2005, case No KRB 2/05 (Transportable concrete) (in the context of assessing the illicit gain by cartelists for the purpose of calculating a fine).

EN 43 EN

(2) The initial overcharge paid by the direct customer

126. All of the methods and techniques described above in Part 2 can, in principle, be used to quantify the initial overcharge paid by the direct customers of the infringing undertakings.129 Also, other types of evidence (such as, for instance, a specific agreement on the rise in prices as shown by internal documents) may also provide valuable insights into the scope of the overcharge. As the initial overcharge is a transfer of money from the direct customer to the infringing undertaking(s), any information that may exist on the illicit profits made by infringers can also serve to quantify this overcharge.130

127. In order to illustrate how methods and techniques can be used to estimate prices in a non-infringement scenario and, based on this estimate, to determine the overcharge paid by the customers of infringing undertakings, it is useful to consider the stylised example of a flour cartel already mentioned in Part 2.131

The flour cartel In this example, all the flour in a certain Member State is produced by four milling companies (Mill A, Mill B, Mill C and Mill D). These mills purchase cereals from various farmers, grind the cereals and apply the appropriate treatments, package the flour and sell it on to bakers. These bakers use the flour to bake bread, which they sell on to consumers as well as to supermarkets.

The national competition authority investigates the market on suspicion of price-fixing and in January 2008 carries out unannounced inspections on the premises of the milling companies. In July 2010 the competition authority adopts a decision in which it establishes that all four milling companies infringed Article 101 TFEU by participating, during the period from 1 January 2005 till 31 December 2007, in a single and continuous infringement regarding the production of flour, covering the whole Member State, which consisted of fixing prices.

A bakery company having purchased flour from one of the milling companies (Mill A) sues this company for compensation of the harm suffered because of the infringement of Article 101 TFEU.132 The bakery claims that the infringement has led to a rise in prices for the flour and demands compensation for the payment of this overcharge for all purchases made in 2005, 2006 and 2007.

128. The bakery is a direct customer of one of the infringing undertakings. If the infringement caused higher prices, the bakery paid an overcharge for each of the units of flour purchased while price was affected. Application of the methods and techniques described will yield an estimate of the price which the bakery would have

129 These methods and techniques can also be used to quantify the overcharge paid by the umbrella

customers (see footnote 112 above), provided that applicable legal rules provide that they are entitled to receive compensation.

130 See also Section 33(3)(3) of the German Act against restraints on competition (Gesetz gegen Wettbewerbsbeschränkungen), which states that the proportion of the profit which the infringing undertaking made from the infringement may be taken into account when estimating damages.

131 Any resemblance of this fictitious example to real events would be purely coincidental; the example cannot be seen as reflecting the Commission’s views regarding any specific undertaking or sector or the market definition in such a sector.

132 National law might well provide that all members of a cartel are jointly and severally liable for the entire harm caused by the cartel. The present example has no implications for these rules.

EN 44 EN

paid for the flour had there been no infringement. By subtracting that non-infringement price from the price actually paid by the bakery, the cartel overcharge per unit purchased can be determined. That figure has to be multiplied by the number of units bought by the bakery in order to determine the actual direct overcharge loss (assuming that there were no significant changes in the overcharge during the infringement period).

129. As discussed above in Part 2, several methods and techniques are generally suitable to quantify the non-infringement price. It should be stressed that this non-infringement price is necessarily hypothetical and can therefore only be an estimate.133 The choice of a method and technique in a given litigation scenario will depend on a wide range of factors, in particular the standard of proof under applicable legal rules and the availability of data.134 For the estimation of the overcharge paid by the bakery in the present example, the use of comparator-based methods will be illustrated as these are most often used in practice and will often yield helpful results in quantifying the initial overcharge.

a. Comparison over time

130. In the present example, the claimant bakery company bought flour from Mill A before, during and after the time for which the national competition authority found an infringement. As described above, using the prices actually paid before or after the infringement to reconstruct the prices as they would have been without the infringement makes it necessary, first, to determine which prices were affected by the infringement and which were not. This means finding out at which point the cartel infringement began to have an effect on the flour market and at which point that effect ended.

131. In the present case, the national competition authority has determined the duration of the infringement. In fact, the decision details the evidence the authority had, which indicates that the milling companies met in January 2005 to discuss prices and thereafter continued to meet on a monthly basis, adjusting their pricing arrangements. The last meeting was held in December 2007. The authority found no evidence of meetings after it inspected the companies in January 2008. In a first step, therefore, the prices before January 2005 and after December 2007 appear to be suitable material for a time-based comparison. However, as described in Part 2, further consideration should be given to the extent to which these figures are useful to serve as comparators.

132. As mentioned above, the decision by a competition authority might limit the finding of an infringement to a certain period for which solid evidence is available to the authority, while indicating that the infringement might have had a longer duration.135 It may then be appropriate not to use the relevant price data for the period that might have been affected by the infringement (and thus include an overcharge), although such data may nonetheless be used to determine a lower bound for the damages estimation, i.e. a safe estimate of what the harm suffered has been at least.

133 See above paragraphs 10-15 in Part 1, Section II. 134 See above paragraphs 101-103 in Part 2, Section III. 135 See paragraph 37 in Part 2, Section II.

EN 45 EN

133. Also, the timing of the cartel infringement may be different from the timing of the effects of the infringement: the milling companies infringed Article 101 TFEU by entering into an anticompetitive agreement. For the purpose of determining which prices observed could be regarded as unaffected by the infringement, it is necessary to look at the timing of the effects of that agreement, not its conclusion. If it can be shown that the companies met in January 2005 for the first time, but that their agreement was implemented from March 2005 onwards, prices before March 2005 would not be tainted by the infringement.

134. As regards the suitability of using post-infringement price observations, it is possible that the cartel produced effects on the market even after the cartel members had ceased to engage in the kind of cooperation forbidden by Article 101 TFEU.136 This may, in particular, be the case in oligopoly markets, where the information gathered because of the cartel might allow cartel members to adopt on a sustainable basis — after the cartel infringement has ended — a course of action aimed at selling at a price higher than the price likely associated with absence of the cartel infringement, without engaging in the sort of practices forbidden by Article 101 TFEU.137 It is also possible that, after the end of the cartel, former cartel members resort to another type of infringement of the competition rules that raises prices for their customers. In these cases any time comparison based on the prices observed after the infringement ceased might lead to an underestimation of the overcharge paid by the customers of the infringers, as the post-infringement prices might still be influenced by an infringement. Where in the present example, the claimant bakery has reasons to believe that this might be the case for the prices paid in 2008 and thereafter, it could only use these prices in its submission to the court to estimate a lower bound of the overcharge harm suffered.

135. In the present example, the claimant bakery finds that the prices paid before the infringement are well suited to estimate the likely hypothetical price. If the bakery compares infringement and non-infringement prices as they are observed, it implicitly assumes that the entire difference between the prices paid in the non-infringement years 2003 and 2004 and the prices paid in the infringement years 2005, 2006 and 2007 is due to the infringement. It is possible, however, that causes other than the infringement had a significant influence on the development of prices during the infringement period. Changes in grain prices, for instance, might be an alternative cause that influenced price developments, and they may be accounted for by using the techniques set out in Part 2, Section II B above. In so far as significant other influences can be identified and the price data are adjusted for their effects, the submission that the remaining difference between the prices in the non-infringement and the infringement periods is due to the infringement gains additional strength.138

136 See also paragraph 38 in Part 2, Section II. 137 For further insights into the workings of such ‘coordinated effects’, see Commission, Guidelines on the

assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C 31, 5.2.2004, p. 5, paragraph 39.

138 This is without implications for the application of national rules allowing the claimant to use the basic, unadjusted comparisons between prices charged in infringement and non-infringement periods to make an initial pleading, or to fulfil the duties incumbent upon him under national legal rules with regard to fact-pleading (in particular where national law allows a court to determine the damages award by way of approximate estimation or determination on an ex-aequo-et-bono basis). Also, rules on the standard and the burden of proof remain unaffected.

EN 46 EN

The circumstances in which such adjustments would be required from claimants or defendants will depend on the rules of applicable law.

b. Other comparator-based methods

136. Besides comparisons over time, other comparator-based methods as described above in Part 2 may also be useful in quantifying the amount of the initial overcharge paid by the direct customer. In the example of the flour cartel, the claimant bakery could alternatively use a comparison with prices from another geographic market or another product market to show what the prices in its own market are likely to have been without the infringement.

137. One possibility would be the comparison with price data observed on a different geographic market for flour. On the assumption that the flour cartel as described above covered a national market, price data from another Member State could be used to construct the non-infringement price. In the case of markets with a sub-national regional scope, sales prices for flour from a different regional market could be a suitable reference point.

138. In order to be a suitable indicator for the prices as they would have been absent the infringement, the comparator prices should themselves not be influenced by the same or a similar infringement of the competition rules. If in the example of the flour cartel price data from a neighbouring geographic market are used and there is evidence that the anticompetitive agreement also covered that neighbouring market, prices from that market would lead to an underestimation of the overcharge. Also, in the case of neighbouring markets, the infringement in one market may have had an influence on that neighbouring market (for example through a rise in demand in the market without infringement), which might therefore not reflect non-infringement prices either.

139. Where the comparator market has different market characteristics, price data from that market might likewise not be sufficiently indicative of the prices as they would have been had there been no infringement. In the present example, the market concerned by the infringement is supplied by four milling companies. For instance, if it can be shown that prior to entering into the infringing practices, vigorous competition existed, price data from a neighbouring market characterised by the presence of a dominant milling company might not adequately reflect the prices as they would have been had there been no cartel and may only serve as a basis for a lower-bound estimate.

140. If the claimant bakery uses price data from a different geographic market in the form in which they are observed, it makes the implicit assumption that the remaining differences between the prices actually paid to the infringers and the prices prevailing on that comparator market are due to the infringement. Depending on the circumstances of the case and requirements under applicable law, the techniques described in Part 2 Section II B above may be used to identify and account for possible alternative influences on prices. It is a matter of applicable law to determine which party bears the burden of proof for such alternative influences or other counterarguments against the claimant’s estimation of a quantum of damages.

EN 47 EN

141. A further possibility to estimate the non-infringement price is comparison with price data observed on other product markets. In the case of flour, it may, however, be difficult to find a sufficiently similar product market not affected by the same or a similar infringement.

(3) The pass-on of overcharges

142. Direct customers of the infringing undertakings who pay an overcharge caused by the cartel may themselves sell on the affected products (or use them as input for their own production of other goods or services). In the example of the flour cartel discussed above, the bakeries are the direct customers of the infringing undertakings and they use the purchased flour to bake bread, which they then sell on either directly to final customers or to supermarkets. These direct customers (bakeries), in reaction to the price increase they face, may raise the prices for their own goods or services (the bread they sell on), thereby passing on some or the entire initial overcharge to their own customers (the consumers or supermarkets). The same effect exists where it is indirect customers (such as, for instance, the supermarkets in the present example) who themselves raise their own selling prices in their business deals with their customers, thereby passing on an overcharge which was first passed on to them.

143. Such pass-on of overcharges normally entails a volume effect: as described above in paragraphs 109 et seq., a rise in prices normally leads to a decrease in demand. In the example of the flour cartel, in so far as the bakery passes on the overcharge by raising the prices it charges for the bread to the supermarkets and end customers, it may reduce the adverse financial impact of the overcharge on itself, but it will suffer decreased demand.139 This decrease in demand means, for the bakery, less sales and a loss of profit — harm that is also caused by the infringement and should be compensated (see Section III below).

144. The price increase through pass-on and the reduction in sales are thus intrinsically connected. In fact, both pass-on and volume effects are determined by the same factors, in particular, the elasticity of demand from downstream customers. This is because the market conditions regarding downstream demand affect both the sales price and the corresponding sales volumes at which the bakery would maximise its profits.

145. In the context of a claim for compensation of overcharges in an antitrust damages action, the pass-on of overcharges can become relevant in two different types of situations:

(a) In an action brought by the direct customer claiming reparation for the initial overcharge paid by him (in the present example: the claim by the bakery against Mill A), the defendant cartel infringer might argue that the direct customer should not, in fact, be compensated for the overcharge harm to the extent that he raised his own prices and thus passed on the overcharge. This is commonly referred to as the

139 This connection between a company passing on an overcharge and its own sales volume has, in a

different context, also been emphasised by the Court of Justice in Case C-147/01 Weber’s Wine World [2003] ECR I-11365, at paragraphs 98-99: ‘even where it is established that the … charge … has been passed on in whole or in part to third parties … the person may suffer as a result of a fall in the volume of his sales’.

EN 48 EN

‘passing-on defence’. Pass-on by the purchaser may, as mentioned above, lead to a loss of sales and therefore a loss of profit for him.

(b) An action brought by an indirect customer against the infringer (for example, a supermarket or a consumer who purchased bread from the bakery and who brings a claim against the milling companies) will also depend on a pass-on argument. Indeed, the indirect purchaser can claim compensation for an overcharge only where the initial overcharge paid by the direct customer has been passed on partially or entirely to him. This can be of relevance for claimants situated at different levels of the supply chain, including end customers.

146. Different legal rules exist concerning the availability of the passing-on defence and the burden of proof in this context.140 The economic insights into the quantification of pass-on set out in paragraphs 148 et seq. below can be of use no matter how these rules are designed.

147. In both situations considered above, claimants and defendants could rely on two different approaches to substantiate their claim that the overcharge was passed on to the indirect customer: they could either (a) quantify the initial overcharge and determine the pass-on rate to the indirect customer, possibly at several levels of the supply chain and using the econometric techniques outlined above, or (b) use the methods and techniques outlined above to determine whether the indirect customer concerned paid an overcharge. For instance, where an indirect customer brings a claim for compensation of an overcharge caused by a cartel, that indirect customer can either show that there was an initial overcharge and that this overcharge was passed on to him141 or he may quantify the overcharge passed on to his level in the same manner as a direct customer would quantify an initial overcharge, namely by comparing the actual price he paid with the likely price in a non-infringement scenario: comparator-based methods can provide useful insights into the amount of overcharge paid by indirect customers, without it being necessary to identify the degree of pass-on. By using a time comparison, for instance, for the prices paid by the indirect customer before and during the infringement, it can be possible to ascertain how much those prices rose because of the infringement, without having to make a finding concerning the pass-on rate.

148. It is not possible to establish a typical pass-on rate that would apply in most situations. Rather, careful examination of all the characteristics of the market in question will be necessary to assess pass-on rates. In a specific case, the existence and degree of pass-on is determined by a range of different criteria and can therefore only be assessed having regard to the conditions of the market in question.

149. Where the direct customer of the infringing undertakings uses the cartelised goods to compete in a downstream market, it is likely that the direct customer will normally not be able to pass on this increase in cost (or only to a very limited degree) if his own competitors in that downstream market are not subject to the same or a similar

140 See Commission White Paper on damages actions for breach of the EC antitrust rules (COM(2008) 165

final, 2.4.2008) for policy proposals concerning the treatment of pass-on in antitrust damages actions. 141 Where the indirect customer substantiates his claim with reference to a pass-on rate and the

infringement concerns a cost factor which is small compared to the entire cost of the product, the pass-on rates of other, more important cost factors that may be more easily estimated might serve as a useful indicator.

EN 49 EN

overcharge (for example, where they receive their input from a market that is not subject to the cartel). In the example of the flour cartel, the claimant bakery is in competition with other bakeries for the production and supply of bread. In so far as these other bakeries do not obtain their flour from the cartel members, but are able to buy it at a lower price elsewhere, the bakery having to buy from the cartel is placed at a competitive disadvantage vis-à-vis its own competitors that prevents it from passing on the extra cost of the overcharge.

150. Where all the undertakings in that downstream market are hit by the cartel and are thus similarly exposed to the payment of the direct overcharge, it is likely that the direct customer will be able to pass on at least part of that overcharge. The degree of such pass-on is influenced by the intensity of competition in the downstream market: if the downstream market is perfectly competitive, the pass-on rate in this case will be virtually 100 %. This reflects the fact that in perfectly competitive markets, price equals marginal costs and a rise in prices for the input will therefore directly lead to an equal rise in cost/output price. For less than perfectly competitive markets, it is likely that affected firms will pass on at least part of the overcharge, though not necessarily 100 %. For example, if the direct customer is a monopolist on the downstream market, he will choose a pass-on rate that reflects — for him — a profit-maximising price in view of the decrease in demand that the pass-on of the overcharge is likely to generate.142

151. The other characteristics that may also have an influence on the degree of pass-on in such situations (everything else being constant) include:

• The price elasticity of demand and the question whether customers become more or less sensitive to price as prices rise. In particular, pass-on is generally more likely if customers do not easily switch to other products following a price increase (inelastic demand) and if customers become less sensitive to price increases when prices are higher.

• The variation of marginal cost with output changes. For instance, a substantial pass-on is less likely if marginal cost significantly decreases following a reduction in output, because the lower output would become less costly to produce (e.g. in the presence of capacity constraints). Conversely, a substantial pass-on is more likely if marginal cost does not significantly decrease following a reduction in output (e.g. due to the absence of capacity constraints).

• The impact of the infringement on different types of costs. Where the infringement impacts on variable costs, this renders pass-on more likely than if the impact is on fixed costs.

• The duration of the infringement and the frequency of business exchanges. Where infringements last for a long time, it is more likely that some level of pass-on occurs; the same applies to sectors where business exchanges and price adjustments are frequent.

142 The exact extent of this pass-on will depend on the demand the direct customer faces and his cost

structure. For example, in the simple case of a monopolist facing linear demand (meaning that the relationship between the quantity and price can be represented by a straight line) and constant marginal costs, the pass-on will be 50 % of the direct overcharge.

EN 50 EN

B. Quantifying overcharges caused by other types of infringements leading to overcharge harm

152. Cartels are but one of the infringements leading to a rise in prices for customers of the infringing undertakings and thus to overcharge harm (or, in the case of infringements pertaining to the supply to the infringing undertakings, to an ‘undercharge’). Other examples of behaviour that can lead to overcharge harm include infringements of Article 101 TFEU by way of certain anti-competitive joint ventures and the abusive charging of excessive prices by a dominant undertaking within the meaning of Article 102 TFEU.

153. A common feature of these infringements is the fact that they may directly or indirectly allow the infringing undertaking(s) to raise the prices for their customers.143 The payment of such overcharge in turn leads to a decrease in demand and thus to a volume effect as described above.

154. The methods and techniques whose application to the case of cartel overcharge has been described above144 can be used to quantify the overcharge harm caused by other infringements.

III. QUANTIFYING THE HARM CAUSED BY THE VOLUME EFFECT

155. A rise in prices for a particular product leads to less demand. The degrees to which both prices rise and quantities decrease following an infringement depends on the same cost and demand parameters, and are determined jointly. Hence, the overcharge and volume effects are intrinsically linked.

156. For an overcharge to an intermediate customer (as discussed above in paragraphs 142 et seq.), the volume effect is also closely linked to the pass-on of overcharges along the supply chain to the final customer: where a customer of the infringing undertakings does not pass on the overcharge and thus absorbs it entirely, his own sales will not decrease because of the infringement as his customers will not experience a rise in prices due to the infringement. Where, however, the overcharge is passed on partly or entirely to the final customer, that customer will be subject to the rise in prices described in paragraph 109 and will reduce his demand. This in turn will reduce demand upstream in the supply chain.

157. As explained above, for those direct or indirect customers of the infringing undertakings who use the product in question for their own commercial activities, this decrease in demand (‘volume effect’) means that they sell less because of the infringement and therefore forego the profit they would have made on the units they failed to sell because of this effect. This loss of profit is harm for which compensation may be awarded145 and, in principle, the methods and techniques described above in Part 2 could be used to quantify it.146

143 Or, if the infringement relates to the supply to the infringing undertakings, to lower the price these

suppliers obtain from their customers. 144 See paragraphs 130 and following, and 136 and following. 145 Joined Cases C-295/04 to C-298/04 Manfredi [2006] ECR I-6619, paragraph 95. 146 Except for the cost-based method.

EN 51 EN

158. In particular, the comparator-based methods and techniques, whose application to the quantification of the initial overcharge paid by the direct customer is discussed above, can provide the claimant with useful insights in determining the decrease in his turnover and profits. For instance, a comparison over time or across markets can be used to reconstruct the sales volume in the non-infringement scenario, i.e. how many units the claimant would have been able to sell had there been no infringement. Likewise, the application of these methods and techniques can be used to arrive at the hypothetical profit margin in a non-infringement scenario. In some instances, a court may also agree to these methods being used in a simplified fashion, for instance by determining an average profit margin per transaction and then multiplying it by the units that were not sold because of the infringement.147

159. Loss of profit is a form of harm often associated with infringements that have the effect of excluding competitors from the market. Part 4 of the Guidance Paper discusses the quantification of such harm in more detail. The insights presented in that part can also be relevant when it comes to quantifying the loss of profit caused by a rise in prices.

147 See also below, paragraph 171.

EN 52 EN

PART 4 — QUANTIFYING HARM FROM EXCLUSIONARY PRACTICES

I. EFFECTS OF EXCLUSIONARY PRACTICES

160. Infringements of Article 101 or 102 TFEU can have the effect of completely excluding competitors from a market or of reducing their market shares. Such effects of infringements on competitors are commonly referred to as ‘foreclosure’. Examples of these practices are abuses of a dominant position prohibited by Article 102 TFEU through, for instance, predation, exclusive dealing, refusal to supply, tying, bundling, or margin squeeze.148 Such abuses are called “exclusionary abuses”. Foreclosure of a competitor can also be the object or effect of a practice prohibited by Article 101 TFEU. It is therefore possible to refer to “exclusionary pratices”, covering both infringements of Article 101 and of Article 102 TFEU.

161. Through exclusionary practices prohibited by the Treaty’s competition rules, infringers distort competition in order to improve or artificially maintain their position on the market. This immediately affects their competitors by deteriorating their position in a market, driving them out of a market or preventing them from entering a market. Exclusionary practices can affect the costs borne by a competitor, the price it is able to charge for its products, or the quantities it is capable of producing and selling. They typically lead to a loss of profit for the competitors concerned.

162. Moreover, by illegally affecting the market position of competitors and thereby the level of competition in the market, such practices lead to harm to customers in the form of higher prices or reduced choice, quality or innovation. However, the detrimental effects of exclusionary practices on customers may not always manifest themselves immediately, as these practices target competitors in the first place, thereby reducing the competitive constraints exerted by them on the infringer(s). Whereas infringements of the kind described in Part 3 normally produce an immediate illegal profit for the infringers and immediate harm for their customers, exclusionary practices could result in an initial disadvantage for the infringers and in better prices for customers in the short run, as typically occurs in predatory pricing. The following sections will separately approach the issues of quantifying harm suffered by competitors (Section II) and harm suffered by customers (Section III).

163. The Treaty guarantees consumers and undertakings that have suffered harm caused by an exclusionary practice a right to compensation regardless of whether they are customers or competitors of the infringers. As already stated, the Court of Justice has specified that such compensation encompasses the actual loss suffered (damnum emergens), compensation for the profit they have lost due to the infringement (lucrum cessans), and the payment of interest.149 For the purposes of quantifying harm from exclusionary practices, the following Sections will primarily refer to the concept of ‘loss of profit’, in line with the case-law of the Court of Justice. The concept of ‘loss of profit’ will be used in a broad sense, as meaning any difference

148 For a description of these practices see also Communication from the Commission — Guidance on the

Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings, OJ C 45, 24.2.2009, p. 7.

149 Joined Cases C-295/04 to 298/04 Manfredi [2006] ECR I-6619, paragraph 95.

EN 53 EN

between the actual profits generated by an undertaking and the profits it would have generated in the absence of the infringement. The approaches to quantifying such loss of profit described in the following are without prejudice to the possibility of injured parties to claim compensation under other heads of damage under national law. Indeed, some elements of lost profits in a broad sense may be classified under different legal concepts under the law of Member States (such as loss of chance150 or loss of reputation) and there may also be heads of damage caused by exclusionary behaviour that go beyond the notion of lost profits.

II. QUANTIFYING HARM TO COMPETITORS

164. Loss of profit to competitors can be caused by reduced revenues (e.g. through the reduction in the quantity that such competitors can sell) or increased costs (e.g. when the infringement affects the price of an input). The overall situation can be reflected in a decrease in the competitor’s market share. In the following Sections, after a short description of how exclusionary practices affect competitors over time (A), and an outline of the general approach to the quantification of lost profits (B), some typical situations in the quantification of exclusionary practices will be addressed, namely in cases where they affect existing competitors (C) and new entrants (D) and when the harm they produce extends also to the future (E).

A. The time dimension of exclusionary practices

165. Depending on the period considered, exclusionary practices can affect competitors in different ways. When an exclusionary practice starts, competitors typically face difficulties in selling their products or (where the practice concerns the upstream market) obtaining supplies. This translates into a deterioration of their profit through higher costs or reduced revenues. Competitors may typically suffer a drop in their market shares, or a lower market share than they could have expected absent the infringement (for instance where their expansion is prevented). This phase may coincide with an increase in profits for infringers. This is, however, not necessarily so, since infringers may have to bear costs due to the implementation of the exclusionary practice (e.g. by lowering their price, by not supplying a competitor and thus reducing their own sales, or by offering rebates or other advantages to customers that could lower profits in the short term). Competitors may eventually be forced out of the market.

166. Once competitors have been successfully prevented from entering a market, or once their market presence has been reduced or eliminated, infringers usually recoup and benefit from increased profits to the detriment of customers and foreclosed competitors. When this occurs (either very soon after the infringement started or after a certain period of time), customers may have to pay a higher price and suffer a loss of quality or choice. The full exclusion of a competitor from a market is not a prerequisite for these effects on customers. Such effects may occur also from the very beginning of the exclusionary practice, and even if competitors are still on the market, provided the competitive pressure they exercise is weakened.

150 Loss of a chance identifies the business opportunities forgone by an undertaking due to the illegal

exclusionary practice.

EN 54 EN

167. When the exclusionary practice is detected by public enforcers or brought to an end as a result of private actions, competitive conditions could be progressively restored. It is important to stress that the restoration of market conditions as if the infringement had not occurred is factually impossible in many cases. This depends mainly on structural effects of the infringement that may be difficult and lengthy to undo (existing contractual obligations, network effects, or other barriers to the re-entry of a foreclosed competitor). Therefore, in some instances full convergence between the non-infringement scenario and the actual market development cannot take place.

B. General approach to the quantification of lost profits

168. In order to determine whether and to what extent competitors have suffered a loss of profits, it is necessary to compare the profit obtained by competitors during the infringement in the market affected by it with the profit they would have obtained from those products in a non-infringement scenario (i.e. the counterfactual scenario).151 Whenever it can be shown that the foreclosed competitor would have earned higher profits in a non-infringement scenario, and that the difference is caused by the infringement, the competitor has suffered harm, even if its market share is unchanged or profits increased due to other factors.152

169. The actual profits earned by the undertaking in question are normally determined by deducting the actual costs incurred from the actual revenues earned. Similarly, profits that would have been obtained in a non-infringement scenario (counterfactual profits) can be determined by deducting the estimated costs in a non-infringement scenario (counterfactual costs)153 from the revenues expected in the absence of the infringement (counterfactual revenues).154 The amount of profits lost is the difference between counterfactual and actual profits. In the case of prevented entry, the actual profits are normally zero, or can even be a negative number if the foreclosed competitor incurred costs (e.g. investment to enter the market) that did not return any revenue.

170. This basic approach to calculating lost profits can be put into practice in different ways. For instance, it is possible to compare the revenues of the foreclosed competitor in the non-infringement scenario with actual revenues from the market as affected by the infringement. Once the lost revenues have been established, it is possible to deduct the costs that the undertaking has avoided due to the lower volumes produced, in order to obtain a value of lost profits. This approach to

151 This does not concern claims aimed at recovering only part of that loss, e.g. only the additional costs

incurred. Such claims arise in practice also because of the availability of more straightforward approaches to quantifying the harm suffered. See below, paragraph 172.

152 For instance, a new entrant with high potential for growth may maintain its profit levels but would have increased them absent the infringement.

153 When estimating the profit lost by the undertaking in question, it is necessary to take into account the additional costs it would have naturally faced to increase production. In this respect, the cost per unit incurred by the undertaking does not necessarily correspond to its cost per unit in the counterfactual scenario. For instance, in the case of increasing returns to scale, the cost per unit in the counterfactual scenario would be lower than the observed cost as the undertaking's production would be higher in the counterfactual scenario (i.e. had it not been affected by the infringement).

154 E.g. Stockholms tingsrätt (Stockholm District Court), judgment of 20 November 2008, joined cases T 32799-05 and T 34227-05 (Europe Investor Direct AB and others v VPC Aktiebolag), appeal pending.

EN 55 EN

assessing lost profits does not make it necessary to quantify the entire costs that would have been incurred by the company, but only an estimate of those costs that have not been incurred because of the infringement.

171. There are also some further pragmatic approaches to assessing lost profits that may be suitable in certain specific cases. For instance, an average profit margin per unit of the product traded in the non-infringement scenario could be estimated and then multiplied by the number of units that have not been sold due to the infringement.155 Such an estimate of the average per unit profit may be based on one or more transactions that can be considered as sufficiently representative of the claimant’s business for the product concerned. It is worth noting that in this calculation the avoided costs would implicitly be included156.

172. Practice of antitrust damages actions shows that foreclosed competitors sometimes choose to claim damages only for part of the harm, for instance the costs incurred in order to respond to an exclusionary practice,157 the non recoverable costs ('sunk costs') incurred with a view to entering a market from which they have been foreclosed158 or the amount judged excessive in cases of margin squeeze or of discriminatory pricing159 that infringe EU competition law. The quantification methods and techniques illustrated in the present Guidance Paper do not prejudice the possibility to bring claims only for part of the harm suffered. This choice is sometimes prompted by the consideration that quantifying such heads of damage is more straightforward or may require less data, and that evidence is more easily available. Also when claimants seek compensation for loss of profits, quantification of harm on the basis of additional costs incurred (sunk and non-sunk) will generally constitute a lower bound when estimating the full loss of profit.

173. Whichever the method or technique chosen, quantifying lost profits may entail evaluating complex data referring to a hypothetical non-infringement situation against which the actual position of the foreclosed competitor needs to be assessed, often with a view at likely future developments. Assessing the profits that a company would have made, including future profits, may depend on such a number of factors that it could be appropriate to provide for less demanding requirements when it comes to quantification. Therefore, legal systems may allow courts to exercise some

155 For an example of a pragmatic approach based on real data on costs and revenues implemented through

regression techniques, see Juzgado Mercantil numero 2 de Barcelona (Commercial Court, Barcelona), decision of 20 January 2011, case No 45/2010 (Céntrica Energìa S.L.U./Endesa Distribuciòn Eléctrica S.A.)

156 In order to estimate the average profit margin, it could be still appropriate to consider how costs and revenues in the counterfactual scenario would have evolved without the infringement. For example, profit margins observed in a pre-infringement period could have been reduced during the infringement period for reasons unrelated to the infringement, due to a reduction in demand or an increase in input costs that are caused by other factors. In addition, the reduction in the output of the excluded competitor could affect its unit cost, hence also affecting the margin on the units it continues to sell.

157 E.g. additional marketing expenses necessary to retain the market position. 158 E.g. the costs of building a new factory on that market. 159 See e.g. Lietuvos apeliacinis teismas (Lithuanian Court of Appeal), decision of 26 May 2006, case No

2A-41/2006 (Stumbras); Højesteret (Supreme Court, Denmark), decision of 20 April 2005, case No 387/2002 (GT Linien A/S v DSB).

EN 56 EN

discretion as to the figures and statistical method to be chosen, and the way in which they are to be used to evaluate the damage.160

C. Existing competitors

174. In order to quantify the harm they suffered because of an exclusionary practice, competitors may choose to rely on the methods or techniques described in Part 2. The non-infringement scenario could be reconstructed by comparison with the performance of the same undertaking in a time period that was not affected by an infringement, a similar undertaking on the same market, aggregated industry profits161 or the performance of the same or a similar undertaking in a market other than the one in which the exclusionary practice occurred. Alternatively, methods based on simulations may provide an estimate of the non-infringement scenario, i.e. simulating on the basis of a number of assumptions (regarding e.g. the type of competitive interactions among firms) what the likely situation would have been if the excluded competitor could have been active on the market and unaffected by the exclusionary practice. The use of other methods is also possible, e.g. financial data from the undertakings involved could provide useful insights on the likely returns of companies had they not been affected by an infringement.

Refusal to supply an essential input for commercial solvents Worldco is a leading international producer of raw materials that are an essential input in the manufacturing of commercial solvents. Eusolv is a company that has been active on the market for commercial solvents since 1995, and most of its turnover is made from sales of Betanol. In order to produce Betanol, Eusolv purchases Rawbeta from Worldco. Worldco is dominant in the production of Rawbeta, which is the only raw material suitable for producing Betanol on an industrial scale and at prices that enable Betanol to be marketed. Worldco also supplies Rawbeta to its subsidiary Subco, which since 2004 has been producing Betanol and competes with Eusolv.

In 2006, Worldco decides to stop supplying Rawbeta to companies selling Betanol in the European Union, with the exception of its own subsidiary Subco. Eusolv initially tries to acquire sufficient Rawbeta from alternative suppliers or to replace its Rawbeta input with other raw materials produced through experimental processes, which are significantly more costly and produce sharp rises in the sales price of Betanol, together with a decrease in its quality and suitability for commercial purposes. As a consequence, Eusolv suffers a progressive decline in its sales and finally discontinues the production of Betanol in 2010. In the same year, Eusolv brings a damages action against Worldco and its subsidiary Subco in order to recover the profits it lost due to the refusal to supply. The court holds that Worldco’s practice amounted to an abuse of a dominant position prohibited by Article 102 TFEU.

(1) Comparison over time

175. When an exclusionary practice affects existing competitors, it is likely that data from the same undertaking in an unaffected period are available. In such cases, the profits

160 See for instance Joined Cases C-104/89 and C-37/90 Mulder and others v Council and Commission

[2000] ECR I-203, paragraph 79. 161 See above at paragraphs 31, 41 and 60.

EN 57 EN

lost by the harmed competitor could be estimated by means of a comparison over time. The non-infringement scenario could, for example, be constructed by reference to data on revenues and costs of the harmed undertaking before the exclusionary infringement produced effects.162 In many exclusionary practices cases, data from after the infringement may not be available or would not be equally suitable, particularly if the infringement produced effects that may alter the structure of a market and are unlikely to disappear in the short term, for instance when the competitor is excluded from the market and there are barriers to re-entry in the short term, or when the competitor has lost market shares that could be difficult to regain because of network effects.163

In the Betanol example, reliable data from after the infringement are not available, since Eusolv, the harmed undertaking, is no longer active on the market, and its effective re-entry into the market may not occur promptly after the termination of the infringement. Eusolv thus decides to construct a likely non-infringement scenario by using data from before 2006, when the exclusionary practice was initiated.

176. Under some circumstances, the pre-infringement revenue and cost data used for the comparison could be refined. For instance, and depending on applicable national rules on evidence and on the burden of proof, a defendant may challenge the amount estimated by the claimant by indicating other elements that may have adversely influenced the performance of an undertaking and are not related to the infringement, such as a drop in marketing investment, a loss of competitiveness of the product, or an increase in the cost of inputs that is specific to the competitor claiming damages. Conversely, it could be shown that the harmed competitor's situation in the non-infringement scenario would have been better than it was before the infringement, for instance because it had a potential for growth. Generally, the reference to an earlier unaffected time period on the same market is likely to be more reliable the longer the competitor has been on that market and the more stable its market position has been. In other words, the reference to a pre-infringement scenario could benefit more from adjustments164 if the harmed competitor was a recent entrant on the market, since its market share could have been more likely subject to fluctuations.

162 For an example of the application of a before and during comparison to estimate the harm from an

exclusionary practice prohibited by Article 101 TFEU see Corte d’Appello di Milano (Court of Appeal, Milan), decision of 3 February 2000, case No I, 308 (Inaz Paghe v Associazione Nazionale Consulenti del Lavoro).

163 A product is subject to network effects if its value for each user increases as the number of users increases.

164 Such adjustments could be performed through the techniques described above at paragraphs 53 et seq.

EN 58 EN

In the example, Eusolv provides data on its overall actual revenues and costs from the production and sale of Betanol, as set out in the following chart:

0

20

40

60

80

100

120

140

160

180

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

CostsRevenues

Refusal to supply starts

In order to establish a reliable non-infringement scenario, data from before 2004 are not taken into account because Subco, the most significant competitor of Eusolv, was not yet active on the market, whereas after 2004 and until 2006 Eusolv held a stable share of the market.

Eusolv, in accordance with national rules on the burden and the standard of proof, provides figures on the ‘counterfactual’ quantities, revenues and costs that would have occurred in the absence of the infringement.

Due to increasing industrial applications of Betanol, it is observed that the total demand for this product (thus, the size of the market) has grown steadily. The stability of Eusolv’s market share after Subco’s entry into the Betanol market is used by Eusolv to rely on the assumption that, absent the infringement, it would have maintained a similar market share. On this assumption Eusolv provides figures on its ‘counterfactual’ revenues for the years 2006–2010, calculated on the basis of the total value of the market and Eusolv’s share of it. From its internal accounts, Eusolv provides figures on its unit costs for the years 2004 to 2006.165 It is shown that costs closely followed the prices of the inputs for the production of Betanol, i.e. that, for instance, a rise in the input prices directly leads to a corresponding increase in costs. Using available industry data on input prices, Eusolv’s experts estimate ‘counterfactual’ unit costs and, e.g. through regression analysis, account for the evolution in input prices and efficiencies related to the production of higher volumes. The figure for overall ‘counterfactual’ costs in the years 2006–2010 is then obtained by multiplying the estimated ‘counterfactual’ unit cost by the number of units it would have sold in the absence of the infringement.

165 These include sunk costs, distributed over time.

EN 59 EN

0

20

40

60

80

100

120

140

160

180

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Costs (mln €)Revenues (mln €)Counterfactual costsCounterfactual revenues

Refusal to supply starts

The figures obtained are compared with the actual revenues and costs faced by Eusolv as follows: the actual profits (actual revenues minus actual costs) are deducted from counterfactual profits (counterfactual revenues minus counterfactual costs). This constitutes the final estimate of the damages claimed by Eusolv.

However, Worldco and Subco argue that in order to be able to supply the expected increasing number of units in 2006-2010, Eusolv would have needed to expand its capacity, facing extra sunk costs that have not been included in the calculation. The defence is accepted by the court, and the compensation for lost profits is reduced accordingly (by deducting the expected extra sunk costs for the years in question, on a pro-rata basis, from the figure submitted by Eusolv).

177. In exclusionary practices cases, market shares can play an important role as an indicator in the calculation of lost profits through comparator-based methods such as time comparisons. For instance, a comparator-based method could be used to obtain the likely market share of the foreclosed competitor absent the infringement. Lost profits could then be quantified by multiplying the observed data on actual per-unit costs and revenues (or the actual average profit margin) by the extra quantities corresponding to the higher ‘counterfactual’ market share expected in the absence of the infringement. This relies on the assumption that costs and revenues per unit would not have significantly changed in the non-infringement scenario, and could be accepted by a legal system as an estimate of the harm suffered, possibly as prima facie evidence or as sufficient to shift the burden of proof.166 A more refined estimate would assess the evolution of costs and revenues in the non-infringement scenario, provided that sufficient data are available.

166 For an example of a court estimation based on multiplying the total number of contracts concluded by

the infringer by the market share held by claimants before the exclusionary practice started, see Corte d’Appello di Roma (Court of Appeal, Rome), decision of 20 January 2003, case No I, 2474 (Albacom S.p.A. v Telecom Italia S.p.A.).

EN 60 EN

178. When the market share is taken as an indicator in the estimation of lost profits, consideration should be given to the fact that it may be subject to fluctuations due to factors other than the infringement, such as the 2004 fall in Eusolv’s market share in the Betanol example due to the entry of Subco as a competitor.167 It may also be the case that if the infringement shrank the total size of the market, revenues for the excluded competitor estimated on the basis of actual market shares would result in an underestimate.

(2) Other comparator-based methods

179. Other geographic or product markets may also be used as a comparator in order to construct the non-infringement scenario.168 Thus, costs and revenues of the same or a similar undertaking on a different market could be taken as a reference to estimate the costs and revenues that would have been yielded by the harmed competitor had the infringement not occurred. These methods can also be used as a means to assess the reliability of an estimation obtained by a comparison over time or other methods. For instance, if the pre-infringement performance of the sole competitor of a historically monopolistic undertaking indicates that it would have held a certain market share absent the infringement, the estimation could be comforted by the finding that the same or a similar undertaking which competes with the formerly monopolistic incumbent on a comparable geographic market actually holds a similar market share, taking into account possible differences between the undertakings or the markets concerned.

D. Prevented entry of competitors

180. Exclusionary practices can not only lead to the deterioration of the market position of an existing competitor, but also prevent the entry of a potential competitor that was not already active on the market. The foreclosure of new entrants can cause them a very significant harm for which they are entitled to compensation. Legal systems should take account of the inherent difficulties of quantifying such harm and should ensure that damages actions by prevented market entrants are not made practically impossible or excessively difficult.169

181. The situation of prevented entry presents some peculiar circumstances that can be taken into account when quantifying the harm. In particular, if the harmed undertaking was willing to enter a market where it was not active before, there is an inherent lack of observable data on its performance on that market.

167 For this reason, in the example the market share considered for the quantification is the stable market

share held by Eusolv after 2004. 168 See Juzgado Mercantil numero 5 de Madrid (Commercial Court, Madrid), decision of 11 November

2005, case No 85/2005 (Conduit-Europe, S.A. v Telefónica de España S.A.), confirmed by Audiencia Provincial de Madrid (Court of Appeal, Madrid), decision of 25 May 2006, case No 73/2006.

169 In some cases it is possible under applicable legal rules to quantify this harm through pragmatic approaches, such as calculation of the total value of the lost market in terms of profits, multiplied by a percentage expressing the share of the market that the foreclosed undertaking would have been likely to acquire. For instance if the total profits generated by undertakings active on the relevant market after the infringement amount to 200 million euros, and it is estimated that, in the absence of the infringement, the foreclosed competitor would have held a market share of 30 per cent, the lost profit could be estimated, under this approach, at 60 million euros.

EN 61 EN

182. The general approach to quantifying the profits lost by competitors in such situations is not essentially different from the situation of foreclosure of competitors that see their existing market position deteriorate, as it also involves an assessment of the profits that could have been yielded by the excluded competitor absent the infringement. These can then be compared with the actual situation. In cases of prevented entry, it is likely that the excluded competitor made no profits or even sustained losses (for instance where the competitor had to bear costs it did not recover through not being able to enter the market).

183. As mentioned above, foreclosed competitors may decide to seek damages only in relation to the costs borne in order to enter the market rather than the whole of the profits foregone. This approach can be more straightforward than claiming compensation for loss of profits as it only involves quantifying the sunk costs incurred by the claimant.

The medical equipment case Newco is an undertaking that was committed to entering the market for a particular type of medical device in a Member State where Medco has a dominant position. In order to be profitable, Newco would have needed to achieve a minimum size on the market to take advantage of economies of scale.

Fearing to lose substantial sales to Newco, Medco concluded exclusive purchasing agreements with a number of customers in order to prevent Newco from achieving this minimum scale. As a result, Newco could not compete with Medco for these customers and was unable to profitably enter the market, which led to higher average prices for consumers than if Newco had entered the market. As Medco’s conduct was considered to infringe Article 102 TFEU, Newco would be entitled to claim compensation for the profits it lost as a result of the infringement. However, in order to avoid carrying out a full loss of profit analysis, Newco only claimed compensation for the sunk costs it had already incurred to set up a new plant and enter the market (including e.g. financial costs and non-recoverable losses on purchased input material).

184. In cases where entry of competitors is prevented, there are no pre-infringement revenue and cost data for the market concerned, while post-infringement data could equally not lend themselves to be a reference for a time comparison because of the effects of the infringement. In such instances, reference to a comparable geographic or product market where the same or a comparable undertaking is active could prove a better means to construct a non-infringement scenario. Product or geographic markets concerned should offer a sufficient degree of similarity, although it may be possible to adjust for some differences between the markets.170

170 This could be done, for instance, through regression analysis, provided that sufficient data are available.

See above, paragraph 63 et seq. For an example of an exclusionary practice where the use of a different geographic market was, in principle, accepted as a comparator see Juzgado Mercantil numero 5 de Madrid (Commercial Court, Madrid), decision of 11 November 2005, case No 85/2005 (Conduit-Europe, S.A. v Telefónica de España S.A.), confirmed by Audiencia Provincial de Madrid (Court of Appeal, Madrid), decision of 25 May 2006, case No 73/2006.

EN 62 EN

185. In some cases, assessment of the competitor’s financial performance may suffice to find data in order to estimate the profits in the non-infringement scenario.171

In the situation referred in the example above, assume that Newco is willing to supply the three biggest private health centres in a Member State with an innovative type of films for X-ray machines. Assume that normally the market for this type of medical equipment for private health centres is a bidding market. Thanks to a technological improvement, Newco is capable of offering its products at a lower price than Medco. However, Medco, which holds a dominant position in the market for X-ray machines, ties the products by applying a higher price for X-ray machines to centres that do not purchase films from it. As a result, Newco does not obtain any contract. In such circumstances, Newco showed that it was actually capable of supplying the quantities demanded by the centres for the price offered, and provided detailed data on its own costs. On the basis of these data, and on the assumption that Newco would have been chosen as a contractor in those instances where it offered the lowest price, expected profit margins could be estimated without resorting to a comparison in time or with other geographic or product markets.

E. Compensation for future loss

186. When foreclosed competitors claim compensation, they may seek compensation not only for the profits lost during the infringement period, but also for the profits foregone after its termination.172 This is relevant, in particular, where they could not re-enter the market or fully recover their market share because of lasting effects of the terminated infringement. Compensation would then be asked for future profits, i.e. profits that are likely to be lost after the claim for compensation is brought and adjudicated.

187. The challenges for quantifying such loss not only lie in the techniques to be deployed, but also have to do with the time frame during which a lost profit can still be identified and compensated. National law plays an important role in this context, for instance by determining under which circumstances a future loss can be recovered, or by establishing pragmatic rules to address this issue on a case-by-case basis.173

188. Factors likely to affect the choice of the relevant limit in time for claiming loss of future profit may encompass, for instance, the likely time needed to re-enter the market in question. In other cases, this assessment could be easier because of the circumstances of the case. For instance, in the X-ray machine example above, the duration of the contracts Newco was bidding for could constitute a reasonable lapse of time over which loss of future profits should be compensated under applicable national rules. In other cases, the time over which the undertaking could reasonably

171 For an illustration of the quantification of harm to a foreclosed new entrant in a bidding market see

Oberlandesgericht Düsseldorf (Higher Regional Court, Düsseldorf), decision of 16 April 2008, case No VI-2 U (kart) 8/06, 2 U 8/06 (Stadtwerke Düsseldorf).

172 For an example of a damages award also for the period subsequent to the end of an infringement see ∅stre landsrets (Eastern High Court, Denmark), decision of 20 May 2009, case No B-3355-06 (Forbruger-Kontakt a-s v Post Danmark A/S).

173 When future profits are estimated, it is normally appropriate to discount their value in order to reflect the loss in the value of money over time.

EN 63 EN

have continued producing goods or providing services in the absence of new investments could also be considered.

In the Betanol example, Eusolv may claim compensation also for the profits it could have obtained after 2010, when it was driven out of the market and brought an action for damages. In such a case, it would be possible to use the same techniques employed to reconstruct the non-infringement scenario in the years 2006–2010 and project it further into the future. Of course, lost profits for the future cannot be claimed for an indefinite duration. Eusolv decided to take as a benchmark the likely lapse of time that would be needed for Eusolv to re-enter the market once the infringement was brought to an end.

0

20

40

60

80

100

120

140

160

180

200

2001

2003

2005

2007

2009

2011

2013

Costs (mln €)

Revenues (mln €)

CounterfactualcostsCounterfactualrevenuesCosts new entrant

Revenues newentrant

Refusal tosupply

starts

Claim brought

III. QUANTIFYING HARM TO CUSTOMERS

189. Undertakings that collude or abuse their dominant position in order to foreclose a competitor might face costs or a temporary reduction in their profits in order to implement the infringement. This sacrifice is borne in order to achieve a distortion of the competitive process that will eventually place the infringers in a position where they gain higher profits thanks to the distorted market conditions achieved, thus allowing them to recoup, at the expense of their customers, the temporary loss or reduction in profits borne in order to attain that position. The following sections will address two typical situations of harm to customers caused by exclusionary practices. For the purposes of quantification, the harm caused to customers by exclusionary practices can be analogous to that caused by infringements leading to a rise in prices, which is discussed in more detail in Part 3 of the Guidance Paper.

A. Recoupment

190. The most straightforward example of the harm caused to customers in the recoupment phase of exclusionary practices is price predation, where an undertaking

EN 64 EN

abuses its dominant position by setting its prices at an artificially low level that cannot be matched by its competitors, who will eventually leave the market or suffer a reduction in their market share. Once the competitors have been excluded from the market, or once a higher market share has been achieved, infringers can enjoy higher profits due to the weaker competitive constraints.

191. Recoupment can be seen as a complementary phase of the infringement that can result in overcharge effects for the customers of the infringers. These overcharge effects constitute harm caused by the exclusionary practice, and compensation for them can be sought by customers.

Recoupment in a predatory pricing case Consider, for example, the market for flights on a particular route between two cities. Operating on this market in a dominant position is Titan Airlines, an established undertaking which offers high quality in-flight service for a standard fare of 1 000 euros. Another player on this specific market is the smaller Bluesky Airlines, which recently started operating on the same route with prices of 800 euros.

Titan Airlines engages in predatory pricing by strategically lowering its fares to a standard price of 500 euros. Bluesky Airlines experiences difficulties in meeting these predatory fares, as a result of which it fails to remain profitable, and is eventually driven out of the market. The dominant Titan Airlines will in that case take advantage of the reduction of competition and increase its profits by raising fares to a level beyond pre-predation fares, i.e. exceeding its initial standard price of 1 000 euros. If Titan Airlines, until re-entry of a competitor, were to charge a price of 1 100 euros, its customers would, due to the infringement, pay an overcharge of 100 euros.

192. When overcharges resulting from recoupment are to be quantified, the conceptual framework that applies is in principle not different from that discussed in Part 3, namely regarding infringements leading more directly to a rise in prices. Since the harm caused by an exclusionary practice is not confined to competitors of the infringer but extends to all customers in a specific market, the issues discussed in the framework of overcharge harm are thus relevant also in this scenario.

193. The position achieved by an undertaking on the market due to an exclusionary infringement does not lead in all cases to a rise in price for customers of the infringing undertaking. However, also in such cases customers may still be harmed by the infringement, for instance if it results in reduced quality. In the example, it could happen that the dominant undertaking Titan Airlines reinstates the same standard price of 1 000 euros, not exceeding the fares it charged prior to the exclusion of Bluesky Airlines. Passengers travelling on this particular route are nevertheless adversely affected, for instance, if Titan Airlines seizes the opportunity of less competitive constraints to lower the standard of its in-flight service.

194. Customers of the foreclosed competitor could be in a different situation than customers of the infringers, because they may have to switch to the products sold by the infringing undertaking as the competitor is driven out of the market. Apart from the possibility of reduced quality, they may also have to pay to the infringing undertaking prices that are higher than the prices paid for the products sold by the foreclosed undertaking. Depending on applicable legal rules, they could be allowed

EN 65 EN

to show that, in the absence of the infringement, they would have purchased from the foreclosed competitor at a lower price. In such case, the effect to be considered is, in principle, similar to an overcharge. The overcharge can be calculated by comparing the price of the product sold by the infringing undertaking in the actual scenario with that charged by the foreclosed undertaking in the non-infringement scenario.

For instance, passengers travelling with Bluesky Airlines prior to its foreclosure may face an overcharge when, due to Bluesky Airlines’ exclusion from the market, they are forced to fly at more expensive fares with Titan Airlines. The overcharge could be estimated as the difference between the actual price of 1 000 euros paid to Titan Airlines and the price of 800 euros which Bluesky Airlines would have charged, had it not been driven out of the market. In such case, the overcharge suffered by passengers constrained to switch from Bluesky Airlines to Titan Airline could be estimated at 200 euros.

B. Harm to competitors as customers of the infringers

195. In cases where a competitor is also a customer of the infringer, the exclusionary practice could damage the competitor in so far as it purchases from the infringer. In these situations, the foreclosed competitor can not only claim compensation for the increase in costs produced by the infringement, but also choose to claim compensation for the profits lost because the resulting volumes produced or sold are lower than if the infringement had not occurred.174

196. It can be observed that for the purposes of quantification, competitors that suffer an overcharge are in a position analogous to that of customers of the members of a cartel or another infringement leading to an overcharge. In order to explain this, it is possible to take the example of Betanol, and assume that rather than refusing to supply Rawbeta to Eusolv, the dominant firm Worldco decides to increase the price of Rawbeta charged to Eusolv so as to squeeze its profit margins. In such a situation, similar considerations arise as in the case of an increase in price generated by other types of infringements. In the example Eusolv would claim compensation for the overcharge represented by the difference between the price it paid as a result of the exclusionary practice and the price it would have paid in the absence of the infringement. If the overcharge has been passed on, claims for damages could be also brought by Eusolv’s own customers, and Eusolv itself could claim compensation for the volumes lost because of the price increase.

174 For an example of the estimation of damages in a discriminatory pricing affecting a competitor as a

customer of the infringer, see Højesteret (Supreme Court, Denmark), decision of 20 April 2005, case No 387/2002 (GT Linien A/S v DSB).

EN 66 EN

TABLE OF CASES CITED

Court of Justice of the EU Case 238/78 Ireks-Arkady GmbH v Council and Commission [1979] ECR 2955. Case C-271/91 Marshall [1993] ECR I-4367. Case C-308/87 Grifoni II [1994] ECR I-341. Joined Cases C-104/89 and C-37/90 Mulder and others v Council and Commission [2000] ECR I-203 Case C-453/99 Courage [2001] ECR I- 6297. Case C-147/01 Weber’s Wine World [2003] ECR I-11365 Joined Cases C-295/04 to 298/04 Manfredi [2006] ECR I-6619.

General Court Case T-202/98 Tate & Lyle v Commission [2001] ECR II-2035 Joined Cases T-25/95 etc. Cimenteries CBR SA v Commission [2000] ECR II-491

Courts of the Member States Corte d’Appello di Milano (Court of Appeal, Milan), decision of 3 February 2000, case No I, 308 (Inaz Paghe v

Associazione Nazionale Consulenti del Lavoro). Corte d’Appello di Roma (Court of Appeal, Rome), decision of 20 January 2003, case No I, 2474 (Albacom

S.p.A. v Telecom Italia S.p.A.). Corte d’Appello di Milano (Court of Appeal, Milan), decision of 11 July 2003, (Bluvacanze) Cour d’Appel de Paris (Court of Appeal, Paris), decision of 23 June 2003 (Lescarcelle-De Memoris v OGF) Landgericht Dortmund (Regional Court, Dortmund), decision of 1 April 2004, Case No 13 O 55/02 Kart

(Vitaminpreise). Højesteret (Supreme Court, Denmark), decision of 20 April 2005, case No 387/2002 (GT Linien A/S v DSB). Bundesgerichtshof (Federal Court of Justice, Germany), decision of 28 June 2005, case No KRB 2/05

(Transportable concrete) Juzgado Mercantil numero 5 de Madrid (Commercial Court, Madrid), decision of 11 November 2005, case No

85/2005 (Conduit-Europe, S.A. v Telefónica de España S.A.), Audiencia Provincial de Madrid (Court of Appeal, Madrid), decision of 25 May 2006, case No 73/2006. Lietuvos apeliacinis teismas (Lithuanian Court of Appeal), decision of 26 May 2006, case No 2A-41/2006

(Stumbras) Corte Suprema di Cassazione (Supreme Court of Cassation, Italy), decision of 2 February 2007, case No 2305

(Fondiaria SAI SpA v Nigriello). Bundesgerichtshof (Federal Court of Justice, Germany), decision of 19 June 2007, case No KBR 12/07 (Paper

Wholesale Cartel) Landesgericht für Zivilrechtssachen Graz (Regional Civil Court of Graz), decision of 17 August 2007, case No

17 R 91/07 p (Driving school). Oberlandesgericht Düsseldorf (Higher Regional Court, Düsseldorf), decision of 16 April 2008, case No VI-2 U

(kart) 8/06, 2 U 8/06 (Stadtwerke Düsseldorf). Stockholms tingsrätt (Stockholm District Court), judgment of 20 November 2008, joined cases T 32799-05 and

T 34227-05 (Europe Investor Direct AB and others v VPC Aktiebolag), appeal pending. ∅stre landsrets (Eastern High Court, Denmark), decision of 20 May 2009, case No B-3355-06 (Forbruger-

Kontakt a-s v Post Danmark A/S). Kammergericht Berlin (Higher Regional Court, Berlin), decision of 1 October 2009, case No 2 U 10/03 Kart. Oberlandesgericht Karlsruhe (Higher Regional Court, Karlsruhe) of 11 June 2010 in case No 6 U 118/05

(appeal pending)

EN 67 EN

Juzgado Mercantil numero 2 de Barcelona (Commercial Court, Barcelona), decision of 20 January 2011, case No 45/2010 (Céntrica Energìa S.L.U./Endesa Distribuciòn Eléctrica S.A.)

EN 68 EN

Privacy Statement

Public consultation "Quantification of harm in actions for damages based on breaches of Articles 101 or 102 of the Treaty" (referred as "consultation" in the text)

1. Objective The objective of this consultation is to receive the views of stakeholders or people concerned by the topic of the consultation and potentially to publish them on the Internet, under the responsibility of the Head of the Unit “COMP.A1", Directorate-General Competition, acting as the Controller. As this consultation collects and further processes personal data, Regulation (EC) 45/2001, of the European Parliament and of the Council of 18 December 2000 on the protection of individuals with regard to the processing of personal data by the Community institutions and bodies and on the free movement of such data, is applicable.

2. What personal information do we collect and through which technical means?

Identification Data The personal data collected and further processed are data necessary for the participation in the consultation, such as name/surname/profession/ postal & e-mail addresses/phone number/fax number..., of the contributors, including their views on the topics concerned. The processing operations on personal data linked to the organisation and management of this consultation are necessary for the management and functioning of the Commission, as mandated by the Treaties, and more specifically articles 5, 7 and 211 - 219 of the EC Treaty.

Technical information Your reply and personal data will be collected through post or e-mail. The e-mail system of the European Commission abides by the Commission's security decisions and provisions established by the Directorate of Security.

3. Who has access to your information and to whom is it disclosed? Received contributions, together with the identity of the contributor, will be published on the Internet, unless the contributor objects to publication of the personal data on the grounds that such publication would harm his or her legitimate interests. In this case the contribution may be published in anonymous form. Otherwise the contribution will not be published nor will, in principle, its content be taken into account. Any objections concerning publication of personal data should be sent to the service responsible for the consultation (see Contact information below).

4. How do we protect and safeguard your information?

Your replies, together with your chosen language used for drafting the reply, are recorded in a secured and protected database hosted by the Data Centre of the European Commission, the operations of which abide by the Commission's security decisions and provisions established by the Directorate of Security for this kind of servers and services. The database is not accessible from outside the Commission. Inside the Commission the database can be accessed using a UserId/Password.

5. How can you verify, modify or delete your information?

EN 69 EN

In case you want to verify which personal data is stored on your behalf by the responsible controller, have it modified, corrected or deleted, please contact the Controller by using the Contact Information below and by explicitly specifying your request.

6. How long do we keep your data? All replies to the consultation will remain in the database until the results have been completely analysed, and they will be archived for reference purposes thereafter. Your personal data will be part of a list of contact details shared internally amongst the staff under the responsibility of the Controller for the purpose of contacting you in the future in the context of further activities related to the consultation. If you do not agree with this, please indicate this in your reply. If you have already sent your reply, please contact the Controller by using the Contact Information below specifying your request.

7. Contact Information In case you wish to verify which personal data is stored on your behalf by the responsible controller, have it modified, corrected, or deleted, or if you have questions regarding the consultation, or concerning any information processed in the context of the consultation, or on your rights, feel free to contact the support team, operating under the responsibility of the Controller, using the following contact information:

European Commission Directorate-General for Competition Unit A1 – Private Enforcement Reference: HT 2290 [email protected]

8. Recourse Complaints, in case of conflict, can be addressed to the European Data Protection Supervisor.

http://edps.europa.eu/EDPSWEB/

Annexe 9

1

Resolution of the Meeting of Heads of the European Competition Authorities of 23 May 2012

Protection of leniency material in the context of civil damages actions

The purpose of this Resolution is to explain the joint position of all ECN Competition

Authorities (CAs) on the importance of appropriate protection of leniency material1 in

the context of civil damages actions. Such protection is fundamental for the

effectiveness of anti-cartel enforcement. The CAs consider that it is necessary to

reiterate the pertinence of such protection following the Pfleiderer judgment2.

The application of leniency programmes3 by the CAs has successfully uncovered

numerous secret cartels throughout the European Union. These programmes are

among the most effective tools for the detection, investigation and punishment of

cartels as well as for providing effective deterrence against cartelisation. Anti-cartel

enforcement constitutes one of the core competences and obligations of CAs in the

general - European and national - interest. The CAs also welcome the developing

private enforcement of competition rules, in particular through damages actions, as a

complementary tool to enforce competition rules.

1 The term “leniency material” refers to statements by leniency applicants submitted under the leniency programme and witness statements made by employees and directors of undertakings cooperating under the leniency programme (whether oral or written). Depending on the protection in the relevant jurisdictions, other information submitted to the CA by a leniency applicant (including pre-existing documents) will also qualify for protection as leniency material. 2 Judgment of the European Court of Justice in Case C-360/09, Pfleiderer AG v Bundeskartellamt. 3 The term “leniency” refers to immunity as well as a reduction of any penalty which would otherwise have been imposed on a participant in a cartel, in exchange for the voluntary disclosure of information regarding the cartel which satisfies specific criteria prior to or during the investigative stage of the case. A “leniency programme” sets out the framework for leniency in a particular jurisdiction.

2

Secret cartels are difficult to detect and investigate. Therefore, cooperation by cartel

participants is often crucial to uncover and punish these highly detrimental illegal

practices. Leniency programmes destabilise cartels as cartelists face the risk that any

cartel member may report the cartel to CAs in the light of incentives offered. Through

leniency programmes, cartels are uncovered and put to an end, preventing further

damage being inflicted on businesses and consumers and helping cartel victims to

bring forward their claims for damages. At the same time, the possibility that any

cartel member may apply for leniency makes it more risky to form new cartels.

The CAs are determined to defend the effectiveness of leniency programmes in order

to ensure a high level of anti-cartel enforcement. This effectiveness heavily depends

on the incentives which the leniency programmes offer to potential leniency

applicants to come forward and cooperate with CAs. The most important of these

incentives is the immunity from penalties (or the reduction of penalties) which would

otherwise have been imposed. The experience of the CAs shows that when deciding

whether or not to cooperate with CAs under a leniency programme, potential leniency

applicants consider as an important factor the impact of such cooperation on their

position in civil proceedings as compared with the situation where they decide not to

cooperate with CAs.

The CAs recognise the importance of effective recovery of damages caused by

cartels. The existence of such a right also strengthens the working of the EU

competition rules and discourages illegal practices4. From that point of view, civil

damage actions can make a very significant contribution to the maintenance of

effective competition in the European Union.

Civil damage claims and leniency programmes are complementary tools to enforce

competition law and deter further infringements. However, at present, civil damage

claims in cartel cases in the European Union mostly rely on public enforcement

(follow-on actions) and public anti-cartel enforcement is nourished by leniency

programmes. Thus, appropriate protection of leniency material is necessary to ensure

the effectiveness of leniency programmes and to enable the authorities to uncover and

4 See Case C-360/09, Pfleiderer AG v Bundeskartellamt.

3

terminate cartels as well as to document and establish their existence and the

participation therein of the companies up to the requisite legal standard. It is often this

action by the authorities that in turn constitutes the basis for civil damages claims for

harm caused by cartels. Therefore, if the incentives to cooperate under the leniency

programmes are not preserved, the victims of currently hidden and future cartels are

unlikely to learn about those cartels in the first place and would be deprived of

exercising their rights to an effective remedy. It is thus through the public

enforcement of competition law, complemented by the emerging private enforcement,

that fierce competition, strong innovation, economic growth and consumer welfare

can be fostered in the European Union.

Moreover, an equivalent standard of appropriate protection of leniency material

across the European Union facilitates opportunities for constructive cooperation

among CAs, as well as the effective allocation of cases and resources and contributes

to the effectiveness of their enforcement tools. Competition enforcement is closely

interlinked across Europe and even world-wide.

In conclusion, as far as possible under the applicable laws in their respective

jurisdictions and without unduly restricting the right to civil damages, CAs take the

joint position that leniency materials should be protected against disclosure to the

extent necessary to ensure the effectiveness of leniency programmes.

Documents

Séminaire Concurrence « Class action & damage claims ......12. En 2005, le directeur général de la concurrence, de la consommation et de la répression des fraudes et le directeur