Causal Uncertainty and Damages Claims for the Infringement of

Causal Uncertainty and Damages Claims for the Infringement of

Yearbook of European Law Advance Access published November 6, 2015 Yearbook of European Law, (2015), pp. 1–62 doi:10.1093/yel/yev017 Causal Uncertain...

326KB Sizes 0 Downloads 10 Views

Recommend Documents

Causal uncertainty and damages claims for the - UCL Discovery
1 Case C-453/99, Courage Ltd v Bernard Crehan and Bernard Crehan v Courage ... Crehan to the White paper and beyond (Klu

Chapter 20 Damages for the infringement of human rights - jus.unipi.it
regional human rights law we would not speak of constitutionalisation, but .... inviolable rights, i.e. personality righ

Damages for Indirect Patent Infringement - Washington University
Grokster, 545 U.S. at 929–30); Jacob S. Sherkow, Patent Infringement as Criminal Conduct, 19 MICH. TELECOMM. & TECH. L

damages for copyright infringement - Phillips Nizer LLP
Under the statute, the plaintiff will receive an award of between $7,500 and $30,000 per work —as the Court considers

Policy and Procedure for Making Claims of Copyright Infringement
Randy Valentino. [email protected] Wide Open West, 1323 Bond Street, Suite 163, Naperville, IL 60563. Telephone number: (

notice and procedure for making claims of copyright infringement
Service Provider: Jasperson Realty. Name of Agent Designated to Receive Notification of Claimed Infringement: Attorney A

Can Claims for Software Copyright Infringement and - R | McHale Law
Jun 30, 2015 - BancorpSouth Bank (“BCS”) was one of SMI's largest customers. SMI never provided BCS—or any of SMI'

Damages: The Measure of Damages for Anticipatory Repudiation and
Anticipatory Repudiation and Seller's Duty to. Mitigate. Fintan M. Flanagan. Follow this and additional works at: http:/

Claims of Copyright Infringement AMSTED Industries Incorporated
Two Prudential Plaza. 180 N. Stetson, Suite 1800. Chicago, IL 60601. Notification: To be effective, the Notification mus

Causal Uncertainty and Social Rejection by Harmony Driver - QSpace
To my supervisor, Dr. Jill A. Jacobson, thank you for undertaking this path of ..... O-Cam procedure (Goodacre & Zadro,

Yearbook of European Law Advance Access published November 6, 2015 Yearbook of European Law, (2015), pp. 1–62 doi:10.1093/yel/yev017

Causal Uncertainty and Damages Claims for the Infringement of Competition Law in Europe Ioannis Lianos *

Key words: causation, causal link, damages, competition law, antitrust, scientific uncertainty, multiple tortfeasors, comparative law, tort law, joint and several liability, proportionate liability, EU Damages Directive, umbrella customers, indirect purchasers, but-for test

* Professor in Global Competition Law and Public Policy, Faculty of Laws, UCL; Director, Centre for Law, Economics and Society, UCL; Lead Researcher, Skolkovo-HSE Institute on Law & Development, Higher School of Economics. The article draws on research completed for I. Lianos, P. Davis, and P. Nebbia, Damages Claims for the Infringement of Competition Law (forthcoming 2015, OUP). The author would like to thank Peter Davis, David Gilo, Herb Hovenkamp, Wolfgang Kerber, and Sandy Steel for their comments to earlier versions of this paper. The usual disclaimer applies. ß The Author 2015. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Abstract: In a tort law regime established on the basis of corrective justice considerations, causation requirements will tend to play a predominant role in regulating the damages claims brought forward. The requirement of the causal link between the harm suffered and the anticompetitive conduct in damages claims for infringement of EU competition law has nevertheless received remarkably little attention in the recently adopted EU Damages Directive and in academic literature. The Damages Directive and some recent case law of the Court of Justice of the EU proceed to some limited harmonization of evidential presumptions and procedural requirements, as well as the exclusion of national rules that may deny the right of the parties harmed by the competition law infringement to receive compensation. Yet, the contours of the requirement of causal link are left to the interpretative work of national courts, in view of their respective tort law doctrines on causation and the lack of a proper EU tort law. The study first explores the role of the concept of causation in claims for damages for infringement of EU competition law and the different approaches taken by the legal systems of EU Member States in conceptualizing the inquiry of a causal link. It then focuses on the methods used by the tort law systems of the EU Member States, the recent Damages Directive and the case law of the EU Court to engage with situations of causal uncertainty, which may frequently arise in the context of competition law actions for damages, in view of the complexity of the commercial environment and the multiple factors influencing markets.

2

Lianos

I. Introduction

1 Case C-453/99, Courage Ltd v Bernard Crehan and Bernard Crehan v Courage Ltd and Others [2001] ECR I-6297. 2 For an historical perspective, see, among others, V Milutinovic, The Right to Damages under EU Competition Law: From Courage v. Crehan to the White paper and beyond (The Netherlands: Kluwer Law International, 2010). 3 Joined Cases C-295/04 to C-298/04, Manfredi & others [2006] ECR I-6619, paras 61–63. See also, Case C-199/11, Europese Gemeenschap v Otis NV [2013] ECR [not yet published], paras 43–44 (noting the role of the national courts in civil proceedings for damages, to determine the damage and the causal link between the infringement and the damage). 4 Joined Cases C-295/04 to C-298/04, Manfredi & others [2006] ECR I-6619, para. 64. 5 Directive 2014/104/EU of the European Parliament and the Council on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union (hereinafter Damages Directive), [2014] OJ L349/1 (emphasis added).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

In Courage Ltd v Crehan,1 the Court of Justice of the European Union (CJEU) expressly recognized the existence of a right to claim damages on the basis of the direct effect of the provisions of EU competition law. In doing so, the CJEU planted the first seeds of what would constitute one of the most remarkable reforms in the five decades of EU competition law enforcement—introducing damages claims for the infringement of EU and national competition law.2 In Manfredi, the CJEU proclaimed the right for any individual to claim compensation for the harm suffered ‘where there is a causal relationship between that harm and an agreement or practice prohibited’ under Article 101 of the Treaty on the Functioning of the EU (TFEU; and/or Article 102 TFEU).3 The Court did not elaborate further on the requirement of a causal link, but simply referred to ‘the legal system of each Member State to prescribe the detailed rules governing the exercise of [the] right [to damages], including those on the application of the concept of “causal relationship” between an antitrust infringement and the harm suffered’, ‘provided that the principles of equivalence and effectiveness are observed’.4 The recently voted Directive of the European Parliament and the Council on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union (hereinafter the Damages Directive), stipulates in its Article 3 that ‘Member States shall ensure that any natural or legal person who has suffered harm caused by an infringement of competition law is able to claim and to obtain full compensation for that harm’.5 The conferral of standing in EU law to anyone harmed by a competition law infringement theoretically opens the gates to a flood of claims for damages initiated by those having suffered damages. Yet, the analysis of all the published competition law damages cases in the four most important jurisdictions in the European Union—in terms of the size of their economy—from 1999 to 2013 indicates that the overwhelming majority of damages claims leading to court judgments has been initiated by direct purchasers and competitors, with a

Causal Uncertainty and Competition Law Claims for Damages

3

6 For example, Which? brought the first UK representative action for damages against JJB sports in 2007 on behalf of consumers who purchased football shirts at cartelized prices in 2000–01, following an OFT infringement decision in 2003. However, as Which? describe ‘as we had to operate under an opt-in system, the number of consumers opting in was very low considering the degree of publicity, the amount of resources we spent and the external legal costs’. See . 7 Law and economics scholars question the central role of causation in tort actions by taking an ex ante perspective based on the concept of efficiency. They advance the view that responsibility should be imposed on the person best placed to avoid the loss most cheaply (cost avoidance theory): see, Guido Calabresi and Jon T Hirchsoff, ‘Towards a test in strict liability in torts’ (1972) 71 Yale L J, 1055; Richard A Posner, ‘Strict Liability: A Comment’ (1973) 2 Journal of Legal Studies, 205. They also consider that the creation of an increased risk of some harm in these circumstances would constitute causation of that harm, thus subjecting the question of actual causation to policy consideration with regard to the ‘efficient’ attribution of responsibility. According to this view, an actor is held liable if he fails to take care when the burden of care (Bc) is less than the outcome of the probability of loss (pL) and the amount of the loss (aL) (eg Bc
Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

handful of cases being initiated by other groups of claimants, such as indirect purchasers, which may seem rather odd for a remedy perceived as aiming to provide full compensation to those affected by a competition law infringement. The reasons for this poor representation of certain categories of claimants may include the relative uncertainty until recently in some jurisdictions of the standing of indirect purchasers. In addition, the absence of proper collective redress systems in most EU Member States has also clearly been a very significant factor.6 The standing issue resolved, for the future, the question is whether we will see a litigation flood. Whether we do may result from the central role played by the aim of compensation (and more broadly corrective justice) in the legal framework in Europe on actions for competition law damages. An approach focusing on compensation and corrective justice requires that causation be proven and this will inevitably become a central issue in actions for damages for competition law infringements, eventually regulating the extent of competition damages litigation in Europe.7 Surprisingly, causation has received remarkably little attention, as neither the preparatory works for the Damages Directive, nor subsequent legal commentary, have paid any attention to it.8 The issue is not, as such, discussed in the Damages Directive (although it is indirectly touched upon with the establishment of a causal presumption for cartel harm) and the CJEU refused to develop a common EU law-based framework on causality in Kone, despite being invited

Lianos

4

to it by its Advocate General,9 thus leaving the task to national legal systems of general tort law. I will first address the role of the concept of causation in claims for damages for infringement of competition law and discuss the different approaches taken by the legal systems of EU Member States in conceptualizing the inquiry of a causal link (II). I will then focus on the methods used by the tort law systems of the EU Member States, the recent Damages Directive and the case law of the EU court to engage with situations of causal uncertainty, which may frequently arise in the context of competition law actions for damages, in view of the commercial environment and the multiple factors influencing markets (III). A conclusion follows (IV).

A. Setting the stage: the role of causation in actions for damages for infringement of competition law The requirement of causation is a common feature of tort law in the EU Member States for damages actions.10 Causation requirements have also been inherent in the development of the substantive provisions of EU competition law, beyond the context of a claim for damages.11 Causation can be part of the liability analysis (in order to establish the infringement of competition law), as well as an inherent part of damages analysis. This paper will focus on the second aspect. Causation was among the issues put forward by the European Commission in its Green paper on damages actions for breach of the EC antitrust rules as a possible factor having an impact on the development of antitrust damages claims in Europe, although the Commission kept away from making any specific ¨ BB-Infrastruktur AG, ECLI:EU:C:2014:1317. Case C-557/12, Kone AG and Others v O See, Christian von Bar, ‘Causation or attribution’, in The Common European Law of Torts: Volume Two (Oxford University Press, 2000), 433–98; Isabel C Durant, ‘Causation’, in Helmut Koziol and Reiner Schulze (eds), Tort Law of the European Community (Vienna: Springer, 2010), 47– 80. 11 The development of the counterfactual method constitutes in that respect an illustration of the importance of causation in EU competition law: Ioannis Lianos, Causation in EU Substantive Competition Law (forthcoming CLES Research paper 5-2015); More generally on the rise of the counterfactual method in EU competition law, Damien Geradin and Ianis Girgenson, ‘The Counterfactual Method in EU Competition Law: The Cornerstone of the Effects-Based Approach’ (11 December 2011). Available at SSRN: ; Antonio Bavasso & Alistair Lindsay, Causation in EC Merger Control, (2007) 3 (2) Journal of Competition Law & Economics, 181–202. Of particular interest, with regard to this issue, is the way EU competition law deals with attempts of anticompetitive practices (see, Ioannis Lianos, Causation in EU Substantive Competition Law (forth. CLES Research paper 5-2015) 9

10

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

II. Unpacking the requirement of a causal link in actions for damages for competition law infringement

Causal Uncertainty and Competition Law Claims for Damages

5

12 European Commission, Green paper—Damages actions for breach of the EC antitrust rules, SEC(2005) 1732, question N. 13 Ibid. 14 See also, Ashurst, Study on the conditions of claims for damages in case of infringement of EC competition rules, Comparative Report, 31 August 2004. 15 Commission Staff Working paper, Annex to the Green paper, Damages for breach of the EC antitrust rules, COM(2005), 672 final, para. 275. 16 See, . 17 See, . 18 Commission Staff Working paper, Annex to the Green papern 15), 672 final, para. 276.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

suggestions on this topic.12 The Commission Staff Working paper, annexed to the Green paper, highlighted three challenges arising out of the application of the concept of causation in actions for damages for infringement of EU competition law. First, proving a causal link in antitrust damages cases ‘will often require complex economic analysis based on a large number of facts and economic data’, thus hinting at the practical and theoretical challenges arising out of the import of the concept of economic causality in a legal and factual setting.13 Secondly, the legal systems of Member States adopt different approaches with regard to the legal concept of causation: an issue that was also highlighted in the European Commission’s commissioned study on the conditions of claims for damages for infringement of EU competition law.14 The Commission, however, observed that the application of these concepts in concrete cases ‘will not lead to widely diverging results and that the concepts derive from the legal culture of the jurisdictions in question more than to actual differences in appreciation’.15 For this reason, and more generally because of the important role played by the case law in this context, the Commission Staff Working paper did not consider that any harmonization action was necessary in this field in order to facilitate damages claims. One should also keep in mind the harmonization initiatives currently considered, in the aftermath of the publication of the Common Frame of Reference16 and the option for a European Civil Code.17 Third, the Commission stressed that the application of a given national law on causation should not lead to the exclusion of groups of victims of anticompetitive behaviour from recovering their losses, hence compromising the objective of effective competition law enforcement. The application of the national requirements for causation is subject to double discipline in the form of the principles of equivalence and effectiveness, and particularly the second principle, which, as the Commission staff paper highlights, may ‘influence’ notions of causation as existing in national civil law and eventually lead to their clarification so as to facilitate damages actions further.18 No concrete proposals as to how this influence will be exercised and in which direction (strong or weak requirements of causality) have been put forward. This is probably because of important ‘cultural’ differences in the national tort law systems and the way they assess causation, and the difficulties arising out

6

Lianos

19 Commission Staff Working Paper accompanying the White paper on Damages actions for breach of the EC antitrust rules, COM(2008) 165 final. 20 Directive of the European Parliament and the Council on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union (hereinafter Damages Directive), [2014] OJ L349/1. 21 Commission Staff Working Document—Practical Guide Quantifying Harm in Actions for Damages Based on Breaches of Article 101 or 102 of the Treaty on the Functioning of the European Union, C(2013) 3440. 22 Providing for the right of ‘anyone who has suffered harm caused by an infringement of Union or national competition law’ to claim ‘full compensation’ for that harm.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

of the need to integrate the economic concept of causality to a common legal core. The concept of causation was not even mentioned once in the White paper on Damages Actions, although it was acknowledged in the annexed Staff Working paper that having to demonstrate in detail the causation and quantification of damages remains a particular difficulty in competition cases, especially for anticompetitive behaviour to which several infringers have contributed.19 The option of lowering the standard of proof so that less evidence or a lesser degree of likelihood would suffice to prove causation or that of shifting the overall burden of proof were not, however, found to be appropriate solutions to mitigating these difficulties. The recent Damages Directive20 also does not deal explicitly with the issue of causation and does not include any harmonized rule concerning the appropriate causation test and its operation in actions for damages for the infringement of EU competition law. Causation is only mentioned en passant in the Practical Guide on quantifying harm as essentially a matter for national law in the absence of rules at the EU level on this matter. It is also noted in this document that national requirements on causality or proximity that link the illegal act and the harm should observe the principles of equivalence and effectiveness.21 However, the exact application of these two principles, in particular the second one, and the nature of the obligations they impose on Member States’ legal systems remain unclear, the only limit so far explicitly mentioned being that victims of anticompetitive practices enjoying standing should not be, as a group, denied the possibility to claim damages, because of a restrictive interpretation by the national court of domestic causation requirements. While standing issues have been thoroughly considered and determined as a matter of EU law by the Court of Justice in Courage, subsequently in Manfredi, and harmonized in Article 2 of the Damages Directive,22 causation remains subject to the requirements of the domestic legal systems of the EU Member States. The subtle differences in the exact meaning of the concept of causation in various legal traditions and the reluctance of the CJEU to proceed to a definition of the contours of the concept in actions for damages for infringements of EU competition law (in view of the possible repercussions to the general system of tort law in each jurisdiction) may have played a role in this decision initially to

Causal Uncertainty and Competition Law Claims for Damages

7

23 Centre for European Policy Studies, Final Report, Making antitrust damages actions more effective in the EU: welfare impact and potential scenarios, 21 December 2007, available at , pp. 36, 145, 415–16. 24 Including direct and indirect purchasers, potential consumers who would have liked to purchase the given goods, but refrained from doing so because the price was set at a supracompetitive level and who ended up either not buying anything or a lower-quality good, competitors foreclosed from the relevant market as a result of the anticompetitive agreement, future consumers harmed by the exclusion of competitors whose presence could lead to future potential output increase, product differentiation, lower prices, new products and higher quality, umbrella customers, etc. 25 Centre for European Policy Studies, Final Report (n 23), 415–16 (noting that if one takes a corrective justice perspective, ‘a broader set of losses would be included in the awardable damages, and standing would be granted to a larger set of victims’). 26 This is particularly so for tort law regimes that depart from a general rule on liability leaving it to the courts to set the limits of its application (eg France) and do not impose extra requirements, such as the duty of care in English tort law or the requirements of unlawfulness (Rechtswidrigket), intention or negligence (Verschulden), and Tatbestand in German tort law: Cees van Dam, European Tort Law (Oxford: Oxford University Press, 2013), §605–2. 27 Centre for European Policy Studies, Final Report (n 23), p. 418.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

defer to national courts for the interpretation and implementation of this concept. From a theoretical viewpoint, questions of standing, the related issue of passing on and causation are intrinsically linked.23 A broad definition of the categories of victims that will be allowed to sue for antitrust damages (standing) as including anyone harmed,24 thus enabling the consideration of passing on (for indirect purchasers), will inevitably raise issues of causation and proximity of the harm to the illegal act. This is particularly so in view of the fact that the principal objective pursued by the claim for antitrust damages for infringement of EU competition law is corrective justice and compensation and not just deterrence. This inevitably influences the legal concept of causation adopted in this context, as it also does for issues of standing, antitrust injury, and recoverable damages.25 The requirement of causation in order to claim antitrust damages for infringements of EU competition law plays a similar role to the requirements of standing and antitrust injury in US antitrust law in limiting possible claims for damages, possibly leading to under-compensation (in the EU context) or under-deterrence (in the US context), should these concepts be interpreted restrictively.26 Because of a restrictive interpretation of the requirements of causation in some Member States, ‘some categories of harmed individuals or firms may not be able to prove the harm suffered or the causation link’, with the result that the social harm or the full amount of the harm inflicted will be higher than the private harm compensated through antitrust damages actions.27 As it is shown by an empirical analysis of the actions for damages for competition law infringement from 1999 to 2013 in the five most important jurisdictions, in terms of the number of cases, in the EU the overwhelming majority of damages actions were brought by direct purchasers or excluded competitors, a

8

Lianos

B. Distinguishing the factual from the normative element: the interplay of causation in fact and causation in law (or scope of liability) In law, causation refers to causal connections between events.29 The principal function of causation in law is to explain the occurrence of particular events, to control events and to attribute moral responsibility to agents whose action has provoked the events. The idea is that among the variety of relationships between events (eg agency and harm), only some will be considered to constitute a legally causal relationship. Which relationships are selected as causal, will depend on the aims of establishing legal causation. 28 I have examined the case law on damages claims for the infringement of EU and national competition in the five most significant (in terms of the size of the economy) jurisdictions in Europe (Germany, France, UK, Italy, Spain) from January 1999 to December 2013. The only jurisdictions in which damages claims were introduced by indirect purchasers were the UK (representing a little less than 1/10th of the total number cases) and Germany (to our knowledge one case). I have not been able to identify any cases initiated by indirect purchasers in France, Italy, and Spain. Neither was I able to identify any cases brought in any of the examined jurisdictions by umbrella customers or counterfactual customers. For a description and analysis of the results, see Table 4.1. at I Lianos, P Davis, and P Nebbia, Damages Claims for the Infringement of Competition Law (Oxford: Oxford University Press, forthcoming 2015). 29 In that sense, it can be distinguished from a logical connection (as cause and effect are independent of one another, which is not the case for logical connections) and from statistical causal connections, as for legal causation the determination of the causal relation is unique (connection between events). Part of the following developments draw from Jan Hellner, ‘Causality and causation in law’ (2000) 40 Scandinavian Studies in Law, 111–34.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

category of litigants for which establishing a direct causal link between the anticompetitive conduct and the harm suffered does not present insurmountable difficulties,28 As we will explain further below, certain categories of consumers, such as indirect purchasers, umbrella customers, and counterfactual (potential) customers, may find it difficult to establish a direct causal link between anticompetitive conduct and the damage suffered, in view of the restrictive approach followed in certain European tort law systems on the causal link required by domestic tort law. Identifying the causation nexus in order to trace an overcharge may also prove impossible in some cases in which the anticompetitive practice affected various successive market levels, sometimes not vertically linked to the infringer. The twin concepts of causation and damage apportionment, in practice, have the potential to play the same filtering or limiting function that other procedural and substantive rules have played in the development of private action for damages in US antitrust law, although causation does also play its part in the USA, particularly in view of the important implications of the finding of liability for antitrust harm, in particular treble damages. The next section briefly introduces the nature of the inquiry into the existence of a causal link in a legal setting.

Causal Uncertainty and Competition Law Claims for Damages

9

30 Antony Honore ´, ‘Causation in the Law’, The Stanford Encyclopedia of Philosophy (Winter 2010 Edition), Edward N. Zalta (ed.), URL = . 31 X is a factual cause of Y if (i) Y actually occurred, and (ii) X is a necessary condition for Y. 32 Richard W Wright, ‘Proving causation: probability versus belief’, in R. Goldberg (ed.) Perspectives on Causation (Oxford: Hart Publishing, 2011) 195–220. 33 Honore ´ (n 30). 34 Honore ´ (n 30).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Legal causation may serve two main purposes: (i) it is backward looking/ explanatory; and (ii) it is attributive (eg establishing the responsibility of agents for the outcomes that follow their actions).30 When the concept of causation is used for the first, explanatory, purposes, it is usually referred to as causation in fact;31 when it is employed for the second, attributive, purpose, it is usually referred to as causation in law. These two functions are not always pursued when one employs the concept of causation in science. For example, statistical causality adopts an empirical view of causation focusing on regulatory or constant conjunction as a necessary condition for causation. In general, this would not be considered as sufficient to establish legal causation, as causality employed in science usually relies on a causal generalization (that events of a type similar to event A almost always or regularly occur jointly or simultaneously with events of a type similar to event B, without it being possible to substantiate this finding for all the events of types A and B as there might be some instances in which this conjunction cannot be observed). The legal concept of causation would require instead a concrete instantiation of a causal law on the particular occasion, regarding the existence of a causal link between the specific event A and the specific event B.32 This cautious approach followed by causation jurisprudence may be explained by its function to attribute responsibility to agents, and the legal consequences for the individual’s autonomy that may follow from a finding of liability. The use of the concept of causation in science (eg economics) does not lead to similar normative consequences, the search for a causal link being not confined to a specific decisional context, thus enabling its constant revision, according to more recent findings. Furthermore, one of the functions of causation in science may be to assist the scientist in making predictions over the probable consequences of an event or a category of events, in particular specifying ‘what will happen and by what stages if certain conditions are present together’.33 This predictive and ‘forward-looking’ function of causation is generally absent from tort law, the latter focusing only on examining ‘which earlier conditions best account for some later event or state of affairs’.34 Counterfactual reasoning in tort law should therefore be clearly distinguished from prediction involving speculation, as they relate only to past events and possible worlds that are on a par with actual worlds. This section introduces the different concepts of causation in fact that have emerged in tort law scholarship, in view of the various approaches adopted in

10

Lianos

the tort law of EU Member States, before examining the elusive character of the distinction between causation in fact and causation in law (responsibility) and the need to take into account the values represented and the objectives followed by the choice of the instrument of tort law as a method of regulation in a particular setting, such as the enforcement of competition law.

(i) Causation in fact

(a) Individualizing theories: equivalence of conditions and NESS According to the theory of equivalence of conditions, any necessary condition for the occurrence of the damage may be considered as having caused the damage. The theory derives from Mill’s statement that ‘all the conditions’ are ‘equally indispensable to the production of the consequent’, the statement of the cause being ‘incomplete, unless in some shape or other we introduce them all’.36 In view of the attributive function of causation in the law, which is not predominant in the context of philosophical inquiry (emphasizing explanation), legal scholars have attempted to extract from the ‘necessary’ conditions that which has caused the damage. However, each necessary condition is logically equivalent to the other necessary conditions, thus leading to the impossibility of establishing, at this level of the causal inquiry, a graduation between the necessary conditions for the occurrence of the damage. In view of the fact that all conditions, but one, cannot have produced the damage and that all necessary conditions are equivalent, each of the necessary conditions constitutes the cause of the damage. Hence, as long as a specific conduct has set one of the necessary conditions for the occurrence of the damage, it forms part of the causal chain, even if other facts may also have contributed to the occurrence of the damage. The equivalence of conditions theory requires the identification through a but-for test relying on a thought experiment formulating a counterfactual scenario of potentially necessary conditions, in order to identify the conditions sine quibus 35 Herbert L A Hart and Tony Honore ´, Causation in the Law (2nd edn, Oxford: Oxford University Press, 1985), 433 (emphasis added). 36 John Stuart Mill, A system of logic ratiocinative and inductive, being a connected view of the principles of evidence and the methods of scientific investigation (Longmans, Green & Co, 1900), 315.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

A frequently used taxonomy of causation in fact distinguishes between individualizing and generalizing theories of causation, the former insisting ‘that there is a quality of being a cause or being causally efficacious which inheres in or belongs to particular acts or events and perhaps also omissions’, while the latter rely on the view that ‘every particular statement is implicitly general in the sense that its truth is dependent on the truth of some general statement of regularities’, thus deriving the causal quality of a particular action or event from ‘the fact that it is an instance of a kind of event believed to be regularly or generally connected with an event of some other kind’.35

Causal Uncertainty and Competition Law Claims for Damages

11

37 Richard W Wright, ‘Causation, responsibility, risk, probability, naked statistics, and proof: pruning the bramble bush by clarifying the concepts’ (1988) 73 Iowa Law Review, 1001–77, 1021. 38 Ibid, 1021. 39 Alex Broadbent, ‘Fact and Law in the Causal Inquiry’ (2009) 15(3) Legal Theory, 173–91; Wright (n 37), 1737–828. 40 For example, Milk Marketing Boards and other farming collective selling organizations can have large numbers of farmers as members.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

non of the damage, that is, the necessary conditions, absent which, the specific damage would not have occurred. The ‘but-for test’ entails a strong necessity requirement, expressed in the familiar counterfactual inquiry. According to this test, ‘a condition is a cause of some result if and only if, but for the occurrence of the condition, the result would not have occurred, considering the circumstances that existed on the particular occasion’.37 Hence, when applying the but-for test, ‘the condition being tested, Q, is hypothetically eliminated and the world is run forward from that point, leaving all the other actual conditions the same insofar as possible, to see if the result, R, still would have occurred’.38 The test thus enables the equal inclusion as a cause in fact of conditions that just contributed to the result as well as of conditions that were a substantial factor in the occurrence of the result. The test may lead to a finding of no causation (underinclusiveness issue) even though it is clear that the conduct in question contributed to the injury (the so called overdetermined causation problem). For instance, in cases of causal redundancy, when a cause is redundant with regard to an effect, as the effect would have occurred even if the cause had not (pre-emptive causation), or when two conditions are simultaneous or successive and the actual condition would have caused the same damage as the initial one (duplicative cause), the test may not assist the decision maker in identifying the relevant cause, and hence the actor that should be held responsible.39 In this instance, if the courts are to choose ‘the’ relevant cause, they will have to do this either on the basis of another factual causality test than the ‘but-for test’ or in the context of the requirements of the relevant causation in law. To provide an example related to competition law, suppose that the decision maker considers whether the behaviour of an individual member of a cartel has caused damage to customers. In some instances, at least, it may be argued that the price in the industry would have remained far above the competitive level even if one individual firm had not participated in the cartel. Suppose for the purpose of illustration that a particular firm’s decision to join the cartel had not affected the cartel price. A strict application of the ‘but-for’ criterion for that individual firm might lead one to the conclusion that the firm’s actions caused no damage, as the cartel price would have stayed the same, even if this firm had not contributed to the cartel. However, in at least some cases cartels have been alleged to consist of a large number of small firms.40 If one considered whether

12

Lianos

41 For instance, one part of the conduct might be non-price related and the other one price-related for which the European Commission and most recently the courts apply an as efficient as competitor

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

any individual firm’s behaviour caused the damage to customers, one might arrive at the conclusion that no firm should be held responsible. In order to avoid this problem, various amendments to the test have been proposed, the NESS (Necessary Element of a Sufficient Set of conditions) test being one of them. The NESS test enables the inclusion of conditions that contributed to the result or were a factor in the occurrence of the result, thus dealing with the under-inclusiveness of the ‘but-for’ test. This is achieved by introducing into the analysis the concept of a sufficient set of conditions necessary for the occurrence of an event (the damage), which does not require establishing a direct link between a specific condition and a consequent result, as this is the case with the necessity requirement in the ‘but-for’ test. Hence, although each of the individual components of the set would not be sufficient to produce the specific damage, if it forms part of set of conditions that would be sufficient for the damage tp occur, it would be deemed to have a causal character. By introducing the concept of a set of sufficient conditions, the NESS test dissociates necessity from sufficiency. The existence of an alternative necessary condition, part of the specific set of sufficient conditions, does not deny the finding of causation, as all necessary elements of that sufficient set of conditions would be deemed to constitute the cause. Suppose that any two of three colluding firms could have colluded and the result would have been a high price level. A strict application of the ‘but-for’ or ‘strong necessity’ tests might conclude that the behaviour of any one firm, say firm 1, was not necessary for the result that a cartel led to high prices and so causation is not established. But for the actions of firm 1, firms 2 and 3 would have colluded anyway and this would have led to cartel prices. Of course the same could be said for any individual participant in the cartel. NESS, on the other hand, considers a firm’s behaviour to be a cause if it is a necessary element of (or condition contributing to) some set of antecedent actual conditions that was sufficient for the occurrence of the cartel. Here there is a set of antecedent conditions, that involving the behaviour of firms 1 and 2 coordinating on prices, where firm 1’s behaviour is a necessary element of the conditions sufficient to cause customers to pay cartel prices. However, this test may also lead to a problem of over-inclusiveness in the presence of multiple simultaneous sufficient causes that may constitute competing alternatives, rather than cumulative ones, for the occurrence of damage. Assume that an excluded or marginalized undertaking suffered losses of market share and thus profits, because of the conduct/strategy of a dominant undertaking in a relevant market. Some of this conduct was found to constitute an exclusionary abuse under Article 102 TFEU, while some other parts of this strategy were found to be compatible with Article 102 TFEU, because of the superior efficiency of the undertaking.41 It will be important in this case to

Causal Uncertainty and Competition Law Claims for Damages

13

distinguish the harm that was caused by the conduct found illegal, and the losses caused by the conduct that was found legal. For instance the loss of market share may not constitute, as such, harm caused by the illegal conduct, but may be the result of legal conduct. This is impractical, at least at the step of establishing causation (if this is viewed independently from that of the quantification of harm). The application of the NESS test may hence lead to over-inclusiveness as the loss will be imputed entirely to the conduct found illegal, both causes being simultaneously sufficient for the occurrence of the damage. A way out, for the NESS test, would be to consider that the victim has suffered no loss if their lost market share would have occurred in any event, as a result of the conduct found legal, the illegal conduct being not causative of a loss.42

test. Hence, while the exclusion of a less efficient competitor might constitute a competition law infringement if it is non-price related, it will not if it is price-related. 42 I am indebted to Sandy Steel for this remark. 43 Michael S Moore, Causation and Reponsibility: An Essay in Law, Morals and Metaphysics (Oxford: Oxford University Press, 2009), 471. 44 Hart Honore ´ (n 35), 482, referring to Max Ru¨melin, Der Zufall im Recht (J.C.B. Mo¨hr, 1896).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(b) Generalizing or generalist theories of causation: the theory of adequate cause Moore explains that generalist theories of causation ‘seek to reduce that relation between state of affairs tokens, to some law-based relation between state of affairs types’.43 The theory of adequate cause constitutes the archetype of a generalizing theory of causation in fact, as it focuses on the condition or conditions that would objectively produce the type of damage examined, not necessarily the specific damage or event. The theory of adequate cause presents close characteristics to scientific theories of causation, in the sense that it puts the emphasis on the regularity of the occurrence of types of events in order to infer from this a causal generalization implemented in the specific case examined. This involves the description of a class of events whose probability must be shown to have been significantly increased by the condition or conduct in question. The condition may be described either in light of what the actor knew at the time of the act, or in light of the knowledge of a ‘prudent man’, or, finally, by reference to ‘what was or has become known otherwise, for example, circumstances existing at the time of the . . . act which have been discovered through the subsequent course of events’.44 The ‘scope of the rule’ doctrine often complements the analysis of probabilities, providing some limits to the extent to which conditions are considered as the adequate cause. This is an element of causation in law, although it also plays an important role in determining adequate cause in the context of causation in fact. Courts may refer to the purpose of the rule violated (‘Normzweck’), determined according to its scope (‘Schutzbereich’), in order to focus their inquiry only on those conditions that relate directly to the purpose of the rule

14

Lianos

breached.45 This provides judges with some discretion over the interpretation of the scope of the rule and consequently the conditions to take into account. It also enables them to focus the analysis only on the risks that the legal system expects the actors not to bear for themselves. Depending on how important the analysis of the ‘scope of the rule’ doctrine is in the process of determining causation in fact, one may conclude that the causation in fact inquiry may be disposed of in certain circumstances.

(ii) Criticisms to a strict distinction between cause in fact and legal causation (or responsibility)

45

See the discussion in Hart Honore´ (n 35), 472–8. Ibid, 304–7. 47 Jane Stappleton, ‘Law, causation and common sense’ (1988) 8 (1) Oxford Journal of Legal Studies 111–31, 116. 48 Ibid, 122. 49 Richard W Wright, ‘The nightmare and the noble dream: Hart and honore ´ on causation and responsibility’, in Matthew H Kramer, Claire Grant, Ben Colburn, and Antony Hatzistavrou (eds), 46

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

As shown by the above discussion, it is difficult to separate issues of fact from issues of legal policy in the determination of causation in fact. The operation of the principle of equivalence of conditions or sine qua non condition, frequently used to determine causation in fact, does not conclude the causation in fact inquiry. In addition, it becomes important to examine the presence of the causal connection required by the relevant legal rule. This is an issue that is partly influenced by factors derived from common sense notions of causation (thus, a factual matter), and partly by ‘scope rules’, that is, the question of whether, as a matter of policy, the law ought in this case to enlarge or restrict liability.46 Hence, one should distinguish between causal limitations that are common in all areas of law and limitations resulting out of ‘scope rules’, which may vary in each area of law according to the policy issues that are considered relevant.47 It follows from the above that there are three steps to be distinguished in the causal inquiry: the first step will attempt to identify factual causation, for instance the conditions but for which the damage will not have occurred (conditions sine qua non). The second step will attempt to identify further causal connections not identified by the sine qua non test, based on ‘common sense principles’, common to all areas of law. For instance, the intervention of a deliberate voluntary human act or abnormal coincidence will ‘break the chain of the causal connection’, even if the subsequent events would not have occurred but for the examined conduct. The third step will engage with policy reasons and will include scope limitations ‘truncating liability more narrowly’ than that produced by mere satisfaction of a designated causal requirement.48 Although issues relating to cause in fact (the explanatory function of causation) and issues relating to legal responsibility (the attributive function of causation) are clearly separated,49 the ‘practical interests’ of the decision

Causal Uncertainty and Competition Law Claims for Damages

15

The Legacy of H.L.A. Hart: Legal Political and Moral Philosophy (Oxford: Oxford University Press, 2008), 165–80. 50 von Bar (n 10), 440 (observing that the distinction between factual or scientific causation and legal causation is ‘misconceived: one is always concerned with evaluations in respect of what is known as causation. In other words, even “factual” causation is in truth a legal evaluation’). 51 Wright (n 49), 165–80. 52 Alex Broadbent, ‘Fact and Law in the Causal Inquiry’ (2009) 15(3) Legal Theory, 173–91. 53 Ibid. 54 Wright (n 37), 1012–15. 55 Ibid, 1011.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

maker and the context in which causation should be identified exercise some influence, to a certain extent, on the operation of the natural causal inquiry (causation in fact).50 If one is to distinguish causes from ‘mere circumstances or conditions’, the practical interests and the purpose of the person making the causal statement may influence the decision to treat it as the cause of an event or a deliberate human intervention that makes a difference to the normal course of events which accounts for the difference in outcome.51 In the absence of a universally applicable legal test of causation, the law may employ different tests in different circumstances, indicating that cause in fact can never be a purely factual issue.52 Furthermore, issues of relevance are never absent from a cause in fact inquiry, although not as explicitly as with the cause in law inquiry. One has always to identify the causal facts from the non-causal ones (mere conditions). In practice, the starting point of causal inquiries is often reversed: we start from what is thought to constitute the causally relevant events, before asking whether these amount to causes in fact.53 Finally, causation in law also depends on common sense principles, such as the ability of the agent to reasonably foresee the damage, which is also linked to the scope of the duty of care the agent has towards the victim, hence a mixed facts and law/policy issue. Does this mean that causation in fact and causation in law inquiries should be merged? It is important here to note that the ‘practical interests’ of the inquiry, occasionally taken into account in the causation in fact inquiry, remain distinct from issues of justice, efficiency, and, more generally, policy, which enter the picture only when examining the attributive function of legal causation (responsibility), that is, choosing ‘the’ responsible cause. This is not a purely causal inquiry but a policy inquiry dealing with issues of moral and legal responsibility.54 An event may be a cause of the injury (actual causation/causation in fact) but not ‘the’ responsible cause, if the injury would have occurred anyway as a result of non-responsible conditions or other moral, economic, legal arguments weighing against liability.55 The principal aim of causation in law is to limit liability, once a causal relation has been identified. The distinction between natural causation (causation in fact) and responsibility (causation in law) implies that the causal inquiry takes a factual, empirical approach and should almost always be kept separate from policy issues. However, while recognizing the importance of such distinction for conceptual

16

Lianos

III. Causal uncertainty and actions for competition law damages: a comparative perspective Causal uncertainty may take different forms. It may relate to a situation in which the conduct of multiple tortfeasors may have contributed to the realization of the damage, hence making attribution of the damage suffered to the conduct of a specific defendant particularly difficult (situations of preemptive and duplicative cause). For instance, according to the cumulative effect doctrine of the Court of Justice in Delimitis, a combination of tying/exclusive purchasing agreements might infringe Article 101 TFEU if it were difficult for competing suppliers to enter the market because of the cumulative effects of the networks of agreements by different suppliers in the market (the foreclosure effect of all the networks of agreements) and the supplier network to which the agreement under consideration formed part had a significant contribution to the foreclosure effect (contribution of the specific network of agreements to the foreclosure effect).58 The collective dimension of antitrust liability in this context does not avoid the necessary examination at the stage of the attribution of the damage due to the causal relation between a specific conduct by one of the multiple tortfeasors and the damage suffered.

56

Broadbent n 52), 173–91. Although the use of the term ‘ersatz’ to refer to these concepts may not be entirely correct, as some, such as the ‘scope of the rule’ doctrine, do not dispense with a separate analysis, at a first step, of whether the defendant’s conduct was a factual cause of the loss in question. However, this analysis ultimately depends on the meaning provided to the concept of factual causation by the ‘scope of the rule’ doctrine. 58 Case C-234/89, Stergios Delimitis v Henninger Bra ¨ u AG. [1991] ECR I-935. 57

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

clarity purposes, implementing it constitutes a more complex endeavour, as in reality the law retains a central role in determining causation in fact and, in particular, which natural causation test is to be applied, among the many put forward.56 In addition, the form of the causal requirement becomes an issue of policy as long as ‘cause’ may be conceived narrowly, as referring to a voluntary act of doing harm or, more broadly, as covering an unintentional conduct of ‘occasioning harm’, or, in other words, an act providing the opportunity or means necessary for harm to occur. Rules concerning the existence of presumptions or the allocation of the burden of proof may also introduce policy considerations in the assessment of causation in fact, eventually operating as alternatives to it (ersatz to factual causation).57 The analysis of the way tort law systems deal with situations of causal uncertainty, which are quite frequent in damages claims for competition law infringements, shows the strong interplay of causation in fact and causation in law (responsibility) in this context.

Causal Uncertainty and Competition Law Claims for Damages

17

A second form of uncertainty may relate to the lack of clear evidence on the causal link between the conduct of a defendant and the damage suffered by the claimant, or on the lack of scientific consensus on the inferences that can be drawn from the evidence. The relative uncertainty which may characterize a complex causal inquiry, as that usually required in the context of competition law enforcement, has led legal doctrine to develop auxiliary methods of establishing legal causation, without proceeding to an in depth analysis of factual causation, according to the theories mentioned above.

A. Five ways to deal with situations of causal uncertainty

59

A Porat and A Stein, Tort Liability Under Uncertainty (Oxford: Oxford University Press, 2001). I Gilead, M D Green, and B A Koch, ‘General Report—causal uncertainty and proportional liability: analytical and comparative report’, in I Gilead, M D Green, and B A Koch (eds), Proportional Liability: Analytical and Comparative Perspectives (Berlin: De Gruyter, 2013), 1–73, 5. 61 I. Gilead et al. 2. 60

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Causal uncertainty constitutes a feature of tort litigation, in particular following the development of mass tort litigation and the increasing reliance on scientific evidence in damages cases.59 The application of the traditional but-for or sine qua non causal test in these circumstances, coupled with the prevailing evidentiary rules on the standard of proof may lead to a perceived failure of tort law to provide compensation to victims of mass torts (under-compensation), where the court determines that the required standard of proof was not met, or to overcompensation, should the court decide, despite causal uncertainty, that the required standard of proof is met and that causation has been established.60 In view of the negative effects of such an all-or-nothing approach to corrective justice and deterrence (at least in most circumstances), legal systems have opted for different strategies in order to respond to situations of causal uncertainty. The sources of causal uncertainty may be multiple: broadly defined, it may be evidential, when there is uncertainty over the facts, or scientific, when there is uncertainty over the inferences that should be drawn out of the facts on the basis of the scientific method (as opposed to just common sense). Legal systems may choose to address causal uncertainty as an issue (i) of factual causation, (ii) relating to causation in law (responsibility), (iii) that requires specific procedural rules, (iv) that requires a public law solution, or (v) to be ignored altogether. For instance, with regard to the first approach, one may decide to attribute responsibility for a specific harm, according to the probability that the specific tortious activity was a factual cause of it.61 This may lead the legal system to accept ‘causal proportional liability’, which is defined as ‘the tort liability imposed on D for harm suffered by P, for part of it, or for harm that P may suffer, according to the causal probability that D’s tortious conduct may have

18

Lianos

62

Ibid. Ibid, 3 (emphasis added). As this is referred in Art 9:101 of the Principles of European Tort law (PETL). European Group on Tort Law, Principles of European Tort Law, Text and Commentary (Berlin: Springer, 2005). 65 Gilead et al. (n 60), 5. 66 It is fairer that the infringer incurs the risks of the causal uncertainty, rather than the claimant. 67 Gilead et al. (n 60), 29. 68 Ibid. Although one may rightly hold the view that class actions do not have a ‘public law’ nature, as in some situations they may constitute the best possible mechanism in order to achieve individual corrective justice. 63 64

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

caused the harm or caused part of it or may cause harm in the future’.62 Systems of causal proportional liability may take different forms, one of which constitutes the well-known doctrine of ‘loss of a chance’ (or ‘loss of opportunity’ in the jargon of the newly adopted Damages Directive). With regard to the second approach, which is often used in situations of multiple tortfeasors, the legal system may opt, causal uncertainty notwithstanding, to recognize the existence of the causal link and hence the liability of the multiple tortfeasors and then to provide rules that apportion liability among the multiple tortfeasors (contribution or recourse claims) or between the defendants and the claimants (contributory negligence). As is rightly explained in a recent comparative study by the European Centre of Tort and Insurance Law, these rules differ from causal proportional liability as ‘they apportion harm after it has been established by the required burden and standard of proof that the multiple tortfeasors or [the claimant] were a factual cause of the same share (or all) of the [claimant’s] harm’.63 For this reason, there is joint (or solidary64) and several liability of multiple tortfeasors in competition law infringements. The third approach maintains the traditional all or nothing approach, but shifts the burden of proof from the claimant to the defendant, and is accompanied by solidary liability,65 the objective being to ease the burden of proof for the party that was the alleged victim of the infringement, essentially for fairness reasons.66 The fourth approach recognizes that the usual mechanisms of tort law may not provide adequate compensation to those harmed in situations of causal uncertainty and advances an administrative governance solution that would substitute ‘public law sanctions (taxes, fines, or fund schemes) for tort liability’, the collected money being used ‘to repair [the claimants’] harm, on a collective or individual basis, and to help prevent such harms in the future’.67 A functional equivalent consists in bypassing ‘the causal relation problem’ and allowing ‘class or representative claims in which the claimants are not required to establish the individual harm caused by each [defendant] to each claimant but rather the overall harm caused to the [claimants] as a group’, the awarded amount being distributed among the claimants.68 For instance, this will involve class actions or representative actions, or even a public compensation scheme to compensate those injured by the public enforcement activities of the competition authority.

Causal Uncertainty and Competition Law Claims for Damages

19

69 K Oliphant, ‘Causal uncertainty and proportional liability in England and Wales’, in I Gilead, M D Green, and& B A Koch (eds), Proportional Liability: Analytical and Comparative Perspectives (Berlin: De Gruyter, 2013), 121–39, 139 (noting that ‘a powerful reason why proportional liability has so far remained exceptional in English law is the common law’s traditional standard of proof: the balance of probabilities. . . . The same difficulty does not arise in systems where the standard of proof is a certainty or substantial certainty, as there exists in such systems a wide area of uncertainty between (substantially) certain causation and (substantially) certain non-causation, and it is here that proportional liability can play a role without undermining the established approach to proof’). 70 For a comparative analysis, see Thomas Kadner Graziano, ‘Loss of a chance in European private law’ (2008) 6 European Review of Private Law, 1009–42 (noting that ‘in at least 12 European legal orders, the concept of loss of chance is still either unknown or has been rejected’, giving as examples, Germany, Austria, Switzerland, and Greece. The loss of a chance approach is however well established in France, Belgium, and partly in the Netherlands. Finally, some jurisdictions have adopted the concept of loss of chance with slight modifications (Spain, Italy, and in certain categories of cases, England and Wales, referring to Gregg v Scott [2005] UKHL 2 (Lord Nicholls and Lord Hope dissenting). 71 At least since Cour Cass. Re. 17.7.1889, S 1891, 1, 1399. For a more recent definition, see Article 1346 of the Civil Code as proposed to be amended by the committee chaired by Pierre Catala, according to which ‘la perte d’une chance constitue un prejudice re´parable distinct de l’avantage qu’aurait procure cette chance si elle s’e´tait re´alise´e’: Avant-projet de re´forme du droit des obligations (Articles 1101–1386 du Code Civil), available at . 72 Articles 1382 and 1383 of the French Civil Code. Italian tort law also seems to adopt a similar approach, with the additional complexity that damage is understood as an offence to a legally protected interest and thus the judge should decide’ which types of interests or legal situations qualify for compensation when breached’, thus limiting the scope of compensation. See, Rui Cardona Ferreira, ‘The loss of chance in civil law countries: A comparative and critical analysis’ (2013) 20 Maastricht Journal 56–74, 62 (noting that it may be difficult to export the theory of lost chance, ‘at least to the extent in which it is recognized in France, to civil liability systems that tend to limit tortious civil liability to the breach of absolute rights or legal provisions designed to protect third parties’, such as the German tort system).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Ignoring the problem will lead to over-compensation or under-compensation, depending on the domestic evidence rules on the standard of proof. The interaction of the standard of proof and the strategy adopted by each legal system should not be under-estimated and constitutes in most cases the main reason justifying the choice made.69 We will focus here on the different expressions of causal proportional liability, such as the doctrine of the loss of chance, market share liability, and liability according to the contribution of the tortious conduct to the risk. The primary purpose of the doctrine of the loss of chance, as is the case for other forms of causal proportional liability, is to strengthen the victim’s position in situations of causal uncertainty, in which it is difficult for the claimant to prove the existence of a causal link between the claimant’s conduct and the damage. In some civil legal systems,70 in particular in France, the doctrine of loss of chance constitutes a specific and autonomous type of damage,71 with reference to the final damage, in view of the open and general nature of tortious liability.72 The concept is employed when the occurrence of the damage is certain, and had a particular condition or hypothetical outcome been realized. One may

20

Lianos

73

Cardona Ferreira (n 72), 59. Ibid, 59. Christophe Que´zel-Ambrunaz, Essai sur la causalite´ en droit de la responsabilite´ civile (Dalloz, Nouvelle bibliothe`que de the`ses, vol. 99, 2010), 164–6. 76 Stathis Banakas, ‘Causalite ´ juridique et imputation: re´flexions sur quelques de´veloppements re´cents en droit anglais’ (2007) 40 Revue Lamy Droit Civil, 93–9. 77 On market shares liability, see Mark A Geistfeld, ‘The doctrinal unity of alternative liability and market share liability’ (2006) 155 U. Pa. Law Review, 447–500; Susan Rose-Ackerman, ‘Marketshare allocations in tort law: Strengths and weaknesses’ (1990) 19 Journal of Legal Studies, 739–46. 78 Gilead (n 60), 2. 74 75

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

distinguish this ‘second degree’ or ‘relative’ certainty as to the loss of a chance, from the situation of hypothetical damage, which it is not clear will be compensated under French law, in view of the requirement of certainty and the high standard of proof the claimant needs to achieve. It is, however, necessary that the lost chance is ‘deemed as real and serious’.73 This requires the claimant to establish ‘the mere probability of materialization of the result that the chance offered’.74 The compensation of the victim may also be partial, which contrasts with the all-or-nothing approach of the usual causation in fact requirement with regard to the award for the damage. This may potentially dispose of the requirement of causal link altogether as the plaintiff may be granted partial compensation equivalent to the fraction of the value or the advantage denied as a result of the loss of chance. The compensation is awarded by calculating the advantage expected by the victim, if the condition or hypothetical outcome had been realized, had the probability that this hypothetical outcome been realized.75 The concept thus offers the judge the opportunity to engage in an assessment of probabilities for the chance or opportunity to be realized, based on factors or causal presumptions that are not linked to the specific individual causal link between the harmful conduct and the damage. Compensation is thus not related to evidence of an individual causal link between the harm and the conduct, but requires partial compensation of all parties affected by the loss of the opportunity/chance, once the harmful conduct has been identified, on the basis of how likely/probable the realization of this opportunity/chance was).76 One may also refer to ‘market share liability’ as another situation of causal uncertainty dealt with by causal proportional liability. In US law, the judge may award compensation by reference to the market shares of the entities that have engaged in the harmful conduct, thus linking the market share of each entity to the risk the harmful conduct caused to the affected party and consequently to the amount of the damage to be awarded.77 Alternatively, causal proportional liability may be based ‘on the ex ante risk created by [a defendant], rather than an ex post assessment of the probability of having caused a specific claimant’s harm’.78 Consequently, a duty of reparation may be imposed on those responsible for the loss of chance, in accordance with the risk that their conduct or activities contributed to the loss of that opportunity.

Causal Uncertainty and Competition Law Claims for Damages

21

B. Towards the development of EU causation rules for competition law based actions for damages? The Ashurst study on the conditions of claims for damages acknowledged that ‘the test of causation is approached in very different ways in the Member States’.79 Indeed, a cursory view of the different general tort law regimes of the most significant, in terms of the number of damages actions, EU Member States show important differences in the choice of causation in fact tests and the interaction between causation in fact and the scope of liability rules (causation in law) (subsection (i)). More importantly, the implementation so far of causation principles in damages actions for the infringement of competition law in various EU Member States illustrates that strict causation requirements, or more accurately the perception that causation requirements constitute an important hurdle, may limit the opportunities of certain categories of claimants to bring forward their claims, thus undermining the effectiveness of the EU remedy of damages for competition law infringements (subsection (ii)). This has led to some effort at the EU level to facilitate the position of certain categories of claimants (subsection (iii)).

79 Ashurst, Study on the conditions of claims for damages in case of infringement of EC competition rules, Comparative Report, 31 August 2004, pp. 73–6.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Causal presumptions or evidential short-cuts constitute an alternative tool to dispose of difficult factual causation issues. The reversal of the burden of proof may also constitute an additional ersatz of causation in fact. Inevitably, principles of public policy exercise significant influence in the design of these causal presumptions, which may lead to inferences of a causal link or a reversal of the burden of proof to the defendant, relieving the claimants of any real obligation to show a factual causal connection between specific conduct and a harm or loss suffered. In particular, causal presumptions avoid the hurdle of establishing a direct causal link between the conduct of the defendant and a hypothetical but probable loss. Causal presumptions may ultimately force the defendant to act as an insurer providing compensation to all members of a class of claimants that has suffered losses, even in the presence of only a tenuous relationship with its conduct. The limited institutional competence of judges argues against such presumptions being established by the judiciary, although significant experience with a certain type of cases and evidence may lead the judiciary, in specific circumstances, to develop inferences of causation. The legislator may nevertheless take the initiative to establish causal presumptions in order to facilitate proof for certain categories of claimants. Inevitably, policy considerations and distributive justice concerns, in particular, may explain the choice of such an approach.

22

Lianos

(i) A comparative analysis of causation rules in the EU: England and Wales, Germany, and France A comparative analysis of causation rules in the various tort law legal systems of the EU Member States illustrates the great variety of approaches followed, first with regard to the distinction and subsequently the interaction between causation in fact and the scope of liability (or causation in law) and, secondly, with regard to the approach followed in situations of causal uncertainty.

80 For a recent analysis, see David Hamer, ‘“Factual causation” and “scope of liability”: What’s the difference?’ (2014) 77(2) Modern Law Review, 155–88. 81 Simon F. Deakin, Angus Johnston, and Basil Markesinis, Markesinis and Deakin Tort Law (7th edn, Oxford: University Press, 2013), 218. 82 Ibid, citing Lord Hoffmann in SAAMCO [1997] AC 191.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(a) England and Wales The law applying to England and Wales divides the assessment of causation into two steps: first determining causation in fact and, second, determining causation in law or legal cause.80 With regard to causation in law, English law proceeds with a number of specific enactments that impose a duty of care, if that was the legislative intention, or more generally under the general tort of negligence. The court determines whether, in the particular circumstances of the case, the defendant owed a duty of care to the claimant, in view of the scope of protection afforded by the statutory duty breached. A duty of care is generally found to exist if the economic injury to the claimant was foreseeable, there is a relation of proximity between the claimant and the defendant and it would be ‘fair, just and reasonable to impose a duty of care upon the defendant’.81 The concept of causation in law aims to distinguish those causes to which the law assigns responsibility. Under common law, cause in law is found when the event is not too ‘remote’ from the defendant’s wrongful conduct (eg a free, deliberate, informed action normally breaks the chain of causation between the defendant’s conduct and the victim’s injury). Remoteness involves the likelihood of the harm (hence a test implicitly based on probabilities, excluding from compensation an improbable loss, that is, damage that a reasonable person could not have foreseen), its gravity, or broader policy questions, such as fairness, the insurability of the harm, or ensuring efficient loss shifting or spreading through the implementation of the least cost avoidance principle or the development of standards reflecting social values and concerns with regard to the allocation of risks and responsibilities. Policy or normative considerations may also be taken into account in the course of establishing if there is a duty of care. For instance, with regard to the tort of breach of statutory duty, English courts examine whether the harm suffered by the claimant relates to the statutory duty breached.82 However, this does not go as far as incorporating in English and Welsh law

Causal Uncertainty and Competition Law Claims for Damages

23

the ‘scope of the rule’ doctrine that would have required an elaborate inquiry on the boundaries of the scope of protection and the interests safeguarded by the specific statutory duty breached. With regard to causation in fact, the English courts employ the ‘but-for test’.83 The claimant must show that it is more likely than not that the damage would not have occurred ‘but for’ the defendant’s breach of duty. As it is put by Clerk and Lindsell on Torts, referred to by the Competition Appeal Tribunal (CAT) in 2 Travel Group:

However, the but-for test was modified, or some would argue ignored, in certain cases of causal uncertainty, either because the claimant’s injury was provoked by the conduct of multiple indeterminate defendants, or because of a genuine scientific uncertainty as to the inferences that may be drawn from the factual record. English law has provided a choice of a variety of instruments from those listed above with regard to situations of causal uncertainty. In situations of ‘indeterminate cause’, when the individual responsible for the harm cannot be identified, the courts have opted for an ersatz of causation in fact, the reversal of the burden of proof, so that each of the defendants has to show that his conduct did not produce the harm.85 In the absence of such proof, both tortfeasors may be held liable. This solution, however, ‘only works in English law if the probabilities are equally split between two defendants’, but does not do so ‘if there are more than two defendants, and the probability that each has caused the harm is less than even’.86 In these cases, maintaining the but-for test, coupled with a reversal of the burden of proof, would not be sufficient to establish causation. The but-for test was also found inoperable in situations where the application of the test would have led to the conclusion that each of the defendants’ conduct could have caused the harm, each of the defendants being equally to blame for the claimant’s loss. English and Welsh courts have found multiple tortfeasors to 83 84 85 86

For a reminder, see 2 Travel Group Plc & Cardiff City Transport Services [2012] CAT 19, para. 77. 2 Travel Group Plc & Cardiff City Transport Services [2012] CAT 19, para. 77. Deakin et al. (n 81), 224. Oliphant (n 69), 123.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(t)he first step in establishing causation is to eliminate irrelevant causes, and this is the purpose of the ‘but for’ test. The courts are concerned, not to identify all of the possible causes of a particular incident, but with the effective cause of the resulting damage in order to assign responsibility for that damage. The ‘but for’ test asks: would the damage of which the claimant complains have occurred ‘but for’ the negligence (or other wrongdoing) of the defendant? Or to put it more accurately, can the claimant adduce evidence to show that it is more likely than not, more than 50 per cent probable, that ‘but for’ the defendant’s wrongdoing the relevant damage would not have occurred. In other words, if the damage would have occurred in any event the defendant’s conduct is not a ‘but for’ cause.84

24

Lianos

87

Deakin et al. (81), 224. Fairchild v Glenhaven Funeral Services Ltd [2002] UKHL 22; [2003] 1 AC 32, although indications for this change of approach were already visible in McGhee v National Coal Board, [1972] 3 All E.R. 1008 (H.L). 89 Oliphant (n 69), 124–5. 90 McGhee v National Coal Board, [1972] 3 All E.R. 1008 (H.L.), 1010–1011 Lord Reid. 91 According to the broader interpretation of the rule by Lord Wilberforce in McGhee v National Coal Board, [1972] 3 All E.R. 1008 (H.L.), 1010–11, where the defendant created a risk of harm and the injury occurred within that ambit of risk, then an inference of causation should be made, hence giving rise to the possibility of conflating the issue of the contribution to the risk of harm and contribution to harm itself. Such an approach was allegedly followed in Fairchild v Glenhaven Funeral Services Ltd [2002] UKHL 22; [2003] 1 AC 32 and Barker v Corus UK Ltd [2006] UKHL 20; [2006] 2 AC 572, in which the courts developed a ‘special rule’ for the contraction of mesothelioma finding that each person who has, in breach of duty, been responsible for exposing the victim to a significant quantity of asbestos and thus creating a material increase in risk should be held liable in respect of the disease. This special rule provides an exception to the usual rule in negligence cases that the claimant must establish on the balance of probabilities that the injury was caused by the defendant’s negligence. According to Lord Hoffman in Barker v Corus UK Ltd [2006] 2 AC 572, para. 17, this special rule only applies where ‘a defendant has materially increased the risk that the claimant will suffer damage and may have caused that damage, but cannot be proved to have done so because it is impossible to show, on a balance of probability, that some other exposure to the same risk may not have caused it instead’. 92 See, for instance, Barker v Corus UK Ltd [2006] 2 AC 572; See also, the Supreme Court’s jurisprudence in Sienkiewicz v Greif (UK) Ltd [2011] UKSC 10, [2011] 2 AC 229. 88

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

be liable in solidum (jointly and severally liable), each of them being potentially responsible for the full amount of the loss, even if such solution cannot be reconciled with a strict application of the but-for test, thus choosing to address the issue of proportional liability when deciding the apportionment/ contribution of the harm or the extent of contributory negligence (hence at the redress stage), rather than at the level of causation in fact (which is presumed to have been established).87 Situations of genuine scientific uncertainty were dealt by the introduction of an exception to the application of but-for test in Fairchild,88 an exception ‘to the ordinary application of principles of causation’, and narrowly confined to ‘situations of scientific, as opposed to merely evidential, uncertainty, as to the cause of the condition, and the fact that the condition resulted from the same (or at least similar risks) to which the claimant was exposed’.89 This mesotheliomarelated diseases tort case law has accepted the development of ‘weaker’ or broader causal link tests, for cases presenting an important evidential uncertainty, for instance because of uncertainty in the available scientific evidence on the existence of a causal link when multiple causal factors are at play. Causation in fact was therefore found where the defendant in breach of the statutory duty materially increased the risk of the claimant suffering a particular kind of loss90 or simply because the defendant increased the probability of harm which that statutory duty aimed to avoid.91 Hence, it is possible to establish causation in fact, by virtue of a material increase in a risk of the presence of an evidential impossibility to prove the causal link on the balance of probabilities.92

Causal Uncertainty and Competition Law Claims for Damages

25

93

Oliphant (n 69), 125. In essence, the risk of causal uncertainty is burdening the defendants, if the first (solidary liability) approach is chosen, while it burdens the claimant if the second (proportionate liability) is chosen, as the claimant would have to incur more expenses and bring actions for damages against all the defendants. The results may be unfair for certain categories of claimants. While the House of Lords in Fairchild v Glenhaven Funeral Services Ltd and others; Fox v Spousal (Midlands) Ltd; Matthews v Associated Portland Cement Manufacturers (1978) Ltd and others [2002] UKHL 22, seemed to leave open the possibility to a worker who had contracted mesothelioma after being wrongfully exposed to significant quantities of asbestos dust at different times by more than one employer or occupier of premises, to sue any of them for the whole damage, notwithstanding that he could not prove that the exposure engendered by the defendant’s conduct had caused the whole damage in Barker v Corus UK Ltd [2006] 2 AC 572, the House of Lords introduced the concept of ‘proportionate’ or ‘aliquot’ liability as the defendant should only be held liable for the proportion of the harm for which he materially increased the risk. Hence, ‘liability should be divided according to the probability that one or other caused the harm’ (Lord Hoffman, para. 43). In view of the burdens that this aliquot liability imposed to victims of mesothelioma-related diseases seeking compensation, the Compensation Act 2006 reversed the Barker v Corus UK Ltd solution, but only for damages claims in various mesothelioma claims arising from unlawful exposure to asbestos, making all parties jointly and severally liable in full for the harm. The decision in Barker v Corus UK Ltd, continues however to be valid for any non-mesothelioma cases involving evidentiary difficulties for the claimant due to scientific uncertainty (‘rock of uncertainty’) or where there is an ‘indeterminate wrongdoer situation’, in the case of multiple defendants in breach of a duty, and it is impossible for the claimant to demonstrate on the balance of probabilities which of these defendants has caused the damage. 95 Trigger BAI (Run off ) v Durham [2012] UKSC 14, Lord Mance noting that ‘there is no magic about concepts such as causation or causal requirements, wherever they appear. They have the meanings assigned to them and understood in ordinary usage in their context’. He further concluded (paras 73–74) that ‘it is not in any event accurate to treat the liability as being either solely or strictly for the risk’ and that ‘(t)he risk is no more than an element or condition necessary to establish liability’. 94

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

The finding of such liability leads to a question of the extent of the liability of the defendants (and/or the contributory negligence of the claimant), the choice being between a rule that would have found them joint and severally liable for the full amount of the damage (thus choosing to address the issue at the stage of the apportionment of damages to the harm/redress stage, the claimant being able to seek compensation in full from any of the solidary tortfeasors) and a rule that would have found them liable for only a proportionate share reflecting their contribution to the total risk (hence the claimant should seek compensation from each party in relation to the harm that party caused).93 Although these two doctrines are considered in some contexts interchangeable as methods to resolve causal uncertainty, as we have explained above, they constitute two genuinely different approaches to situations of scientific uncertainty and impose different burdens on the claimant and defendants. The English and Welsh courts finally chose the route of proportionate liability.94 They also prefer to risk unfairness to the claimants by reversing the burden of proof to the defendant if the claimant can show a material increase in risk, rather than taking the view that the increased risk constitutes the damage itself, which may have changed the substantive law’s notion of damage, simply because of an evidentiary difficulty.95 For all these cases where there is ‘inherent uncertainty’ in the proof that the defendant’s wrongful conduct caused the claimant’s loss, because of the above evidentiary

26

Lianos

96 Sandy Steel, ‘Rationalising loss of a chance in tort’, in Stephen G A Pitel, Jason W Neyers, and Erika Chamberlain (eds) Tort Law: Challenging Orthodoxy (Oxford: Hart Publishing, 2013), 242. 97 Allied Maples Group Ltd v Simmons & Simmons [1995] 1 WLR 1602. Oliphant (n 69), 131, notes that ‘(a)s the chance has economic value, it is clear that its loss itself constitutes damage’ but cautions that ‘it is only exceptional that pure economic loss is recoverable’. The author noted the ‘evident tension’ of rejecting the loss of a chance doctrine, while accepting the weaker causation in fact standard of material contribution to risk. 98 Oliphant (n 69), 128–9. 99 Ibid, 126. 100 Ibid. 101 Albion Water Ltd v Dwr Cymru Cyfyngedig [2011] CAT 1.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

difficulties, but where it is possible to estimate the chance that the defendant caused the loss, damages are calculated by multiplying the magnitude of the chance lost, which can be estimated, by the total loss suffered. As Steel explains, the doctrine ‘is better thought as creating proportionate liability in proportion to the chance that the defendant has caused’ the harm: ‘(d)amages are for the chance of causation, not causation of a lost chance’.96 Although the doctrine of loss of chance is not properly adopted in English and Welsh law with regard to cases of personal injury, loss of chance awards have been accepted in the context of compensation actions for pure economic loss.97 In terms of the ‘market share liability’ option, this has not yet been accepted by English courts.98 With regard to causal uncertainty as to future harm, English law distinguishes between ‘cases in which it is uncertain whether the [claimant] will suffer any harm at all and cases in which harm has already been caused but the scope of the harm (or the development of a new harm) in the future is unknown’.99 While the first category does not give rise to causation and thus damages are not awarded, the second is a matter usually addressed at the stage of the quantification of damages.100 Moving to concrete competition law examples, in Albion Water, the UK Competition Appeal Tribunal (CAT) examined a follow on claim for damages brought by a water company, Albion Water, against the defendant utility comˆ r Cymru, relying on the finding of the CAT that the defendant had pany, Dw engaged in an abuse of a dominant position by adopting the practices of margin squeeze and charging excessive prices.101 The CAT examined the issues of causation and quantum. Albion was asserting that if it had been offered a reasonable ˆ r Cymru it would have accepted that offer and price for common carriage by Dw thus would have been able to make more money on its supply of water to some of its existing customers and could also have bid to supply others. The CAT thus proceeded by comparing the money that Albion would have made in the counterfactual world (but for the infringement) with the money it in fact made from the supply to existing customers in order to see if it has suffered any loss. The CAT constructed the counterfactual cautiously by assuming that the defendant would not have engaged in illegal conduct and would have negotiated a reasonable agreement, in view of the possibility of the regulator, OFWAT, to

Causal Uncertainty and Competition Law Claims for Damages

27

What has to be proved to establish a causal link between the negligence of the defendants and the loss sustained by the plaintiffs depends in the first instance on whether the negligence consists in some positive act or misfeasance, or an omission or non-feasance. In the former case, the question of causation is one of historical fact. . . . Once established on the balance of probability, that fact is taken as true and the plaintiff recovers his damage in full. There is no discount because the judge considers that the balance is only just tipped in favour of the plaintiff; and the plaintiff gets nothing if he fails to establish that it is more likely than not that the accident resulted in the injury. Questions of quantification of the plaintiff’s loss, however, may depend upon future uncertain events. . . . It is trite law that these questions are not decided on a balance of probability, but rather on the court’s assessment, often expressed in percentage terms, of the risk eventuating or the prospect of promotion, which, it should be noted, depends in part at least on the hypothetical acts of a third party.104 The Court of appeal also noted that if the plaintiff’s loss depends on the hypothetical action of a third party, either in addition to action by the plaintiff or independently, the plaintiff may succeed provided that he shows on the balance of probabilities there is a substantial chance, rather than a speculative one. The evaluation of the substantial chance is an issue relating to the quantification of damages, the quantum being in this 102 103 104

Albion Water Ltd v Dwr Cymru Cyfyngedig [2011] CAT 1 Albion Water Ltd v Dwr Cymru Cyfyngedig [2011] CAT 1, para. 202. Allied Maples Group Ltd v Simmons & Simmons [1995] 1 WLR 1602, para. 191.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

intervene and apply concurrently competition law as well as the Water Industry Act. Although the CAT recognized that there is a range of lawful access prices ˆ r Cymru could have offered, the CAT took the middle of that range, on that Dw the assumption that the defendant would have offered a reasonable access price rather than the highest price it could lawfully have charged.102 It also held that on the balance of probabilities Albion would have agreed to this rate. When examining Albion’s other costs of supply in the counterfactual world, the CAT took care to evaluate the commercial reality of the parties’ relative positions and their bargaining power, as well as overall industry practice. The case is particularly interesting in view of the analysis by the CAT of the loss of opportunity that Albion would have had in the counterfactual world to bid for a contract to supply water to a new customer, which had invited Albion to tender. Albion was alleging that the defendant’s abusive behaviour effectively deprived it of that opportunity, as Albion was obliged to focus its very limited resources on the competition case and simply did not have the resources to pursue other business opportunities at the same time.103 The CAT analysed this as a loss of chance claim. The CAT referred to the Allied Maples Group Ltd v Simmons & Simmons case law of the Court of Appeal in which Stuart-Smith LJ highlighted the distinction between causation and quantification of damages by the different function probability judgments played in each context:

28

Lianos

case the relevant percentage of the total profit that would have been made if Albion had indeed won the business. The assessment of the chance in the context of quantification is done in a ‘range lying somewhere between something that just qualifies as real or substantial on the one hand and near certainty on the other.105

105 Albion Water Ltd v Dwr Cymru Cyfyngedig [2011] CAT 1, para. 218, citing Allied Maples Group Ltd v Simmons & Simmons [1995] 1 WLR 1602, 1614D. 106 Enron Coal Services Ltd & English Welsh & Scottish Railway Ltd [2011] EWCA Civ 2; Enron Coal Services Ltd & English Welsh & Scottish Railway Ltd [2009] CAT 36.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

The CAT found in this case that it would have been highly likely for Albion to win this contract, although not as high as certain or near certain and consequently reduced Albion’s damages for its lost opportunity by one third (33 per cent). In Enron Coal Services Ltd & English Welsh & Scottish Railway Ltd, the Court of appeal (Civil Division) and the CAT examined the claim by Enron (ECSL) that it had been deprived, following the anticompetitive conduct of EWS, of a real or substantial chance of winning a contract for the supply of coal to one of Edison’s (EME’s) power stations.106 EWS was found by the UK Office of Rail Regulation (ORR) to have infringed the Chapter II of the UK Competition Act 1998 prohibition by having abused of its dominant position in the market for coal haulage by rail, by setting higher prices to Enron for coal haulage than what it did to other customers, thus putting Enron at a competitive disadvantage, in particular with regard to coal haulage supply to two power stations owned by Edison. Enron brought a follow on damages claim arguing that EWS’s conduct caused it to lose a real or substantial chance of gaining a profitable conduct with Edison. The CAT held that to succeed in this claim Enron had to show on the balance of probabilities what it would have done, but for the infringement, thus requiring Enron to build a counterfactual (but for) world in which it would have submitted a bid to EME achieving similar terms than those awarded by Edison to Enron’s competitor, EWS, and that it would have used this opportunity in order to negotiate with Edison a contract to supply coal to one of its power stations. Furthermore, Enron had to establish that “there was a real or substantial chance” that negotiations would have led EME to award the contract of coal supply to Enron. The CAT, confirmed by the Court of appeal, dismissed Enron’s claim taking the view that it would have failed to prove what was needed on either of these issues. The competitive disadvantage that Enron was suffering following the discriminatory conduct of EWS was not in itself a finding that it had suffered loss, although this fact was relevant to the question of causation. It was not however determinative of causation as the award of the contract by EME was not only a question of price, but also depended on other considerations, such as the quality and reliability of the proposed service. It is noteworthy that the CAT and the Court of appeal constructed the but for scenario on the basis of circumstantial evidence in the “real world”, for instance

Causal Uncertainty and Competition Law Claims for Damages

29

107 108

see Section (c). 2 Travel Group Plc (In Liquidation) v Cardiff City Transport Services Ltd [2012] CAT 19.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

the difficult previous commercial relations between the parties, the lack of enthusiasm of Enron’s senior staff about winning the contract and Edison’s general unwillingness to enter into a long term coal supply contract, as this transpired from the evidential findings in the ORR decision. Hence, even if EWS’s anticompetitive conduct and the subsequent competitive disadvantage suffered by Enron was a factor preventing Enron from contracting successfully with Edison for coal supply, this contribution to Enron’s loss of a chance was not considered sufficient to constitute evidence of causation. Hence, the allor-nothing approach followed by the application of the “but-for” test in this context led the CAT and the Court of appeal to reject Enron’s claim, without being possible to compensate Enron’s loss of a chance/opportunity to win the contract as a specific head of damage, as it could have been possible under French tort law.107 Causation was also thoroughly examined in Cardiff Bus, a case involving a follow on claim for damages, relying on an Office of Fair Ttrading (OFT) decision finding that a bus company, Cardiff Bus, had abused its dominant position by engaging in exclusionary practices (predatory conduct through the launching of loss-making bus services, the so called ‘White Services’ on top of their regular liveried service) against a rival bus company, 2 Travel, and leading to it going into liquidation.108 The damages claimed by 2 Travel were substantial and covered lost profits, loss of a capital asset, loss of a commercial opportunity, wasted staff and management time, and the costs relating to 2 Travel’s liquidation. The defendant, Cardiff Bus, raised the issue of factual causation and advanced a strict but-for test that would have compared the losses in the real world with those in the but-for world, but only to the extent that these losses would not have occurred in the but-for world scenario. Cardiff Bus’ argument was that, in any case, 2 Travel would have gone bankrupt in view of its weak financial position and its mismanagement, hence the costs involved by 2 Travel’s liquidation would also have been incurred in the ‘but-for’ world scenario. 2 Travel was advancing a broader causation test, arguing that it should focus instead on the factors that contributed to the damage, in this case 2 Travel’s exit from the market. Hence, if Cardiff Bus’s predatory conduct was a contributing factor to its exit from the market, that would be a sufficient finding upon which to infer causation. The nature of the behaviour (its predatory character) and the intent of Cardiff Bus to cause the ‘type’ of damage actually incurred, would be sufficient elements, according to 2 Travel, for the finding of causation. Without explicitly employing such terminology, 2 Travel was reasoning, in terms of probabilities, that Cardiff Bus’s predatory conduct had caused the damage. The CAT rejected the broader view of causation advanced by 2 Travel, finding it ‘vague’. It also held that intention alone cannot be said to be causative,

30

Lianos

(b) Germany In German tort law, the requirement of causation derives from the general part of the law of obligations relating to the general theory of the reparation of damage.110 As in English law, the causal inquiry is bifurcated between a factual phase (‘Kausalita¨t im natu¨rlichen Sinne’ or natural causation) and a more normative/policy oriented one (‘Kausalita¨t im rechtlichen Sinne’ or causation in the legal sense).111 With regard to the normative dimension, a distinction is made between causation giving rise to the violation of a protected interest or ‘causation as a foundation of liability’ (‘Haftungsbegru¨dende Kausalita¨t’), which links the conduct of the defendant and the infringement of a protected interest, and ‘causation as a determinant of the scope of liability’ (Haftungsausfu¨llende Kausalita¨t), which relates to the issue of the extent of damage.112 While the first must be 109 2 Travel Group Plc (In Liquidation) v Cardiff City Transport Services Ltd [2012] CAT 19, para. 332. 110 For instance, §§ 823 BGB [(1)A person who, intentionally or negligently, unlawfully injures the life, body, health, freedom, property, or another right of another person is liable to make compensation to the other party for the damage arising from this. (2) The same duty is held by a person who commits a breach of a statute that is intended to protect another person. If, according to the contents of the statute, it may also be breached without fault, then liability to compensation only exists in the case of fault]. 111 Walter van Gerven, Pierre Larouche, and Jeremy Lever, Cases, Materials and Text on National, Supranational and International Tort Law (Oxford, Hart Publishing, 2000), 453. 112 Basil S. Markesinis and Hannes Unberath, The German Law of Torts (4th edn, Oxford: Hart Publishing, 2002), 106.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

particularly if the defendant has not taken any effective steps to cause such harm. The CAT seems to have relied on a clear distinction between factual causation and causation in law, with normative aspects, such as the protective scope of competition law, exercising no influence on the assessment of causation in fact. The anticompetitive intent of Cardiff Bus, which could have been a factor considered in the analysis of the scope of the rule, was therefore found to be irrelevant in the analysis of the factual causation step. That said, the CAT engaged in a thorough but cautious analysis of the relevant counter-factual scenario, in particular exploring what 2 Travel’s position would have been, had the ‘White Services’ never operated. For example, with relation to the claim of a causative link between the decline of 2 Travel’s bus services and the infringement, the CAT found, on the balance of probabilities, that this falling off was due to 2 Travel’s ‘basic operational failings with its too hasty expansion of services’ and not the infringement.109 Also, with regard to loss of passengers and diversion of passenger revenue, the CAT engaged in a considerate analysis of the other options that passengers of ‘White Services’ disposed and did not assume that all these passengers would have made the choice of 2 Travel’s bus service. The Tribunal also opted for a frequency-based allocation of passengers, in view of the importance the frequency of service had for consumers, adequately adjusted with a number of other factors.

Causal Uncertainty and Competition Law Claims for Damages

31

113

van Gerven et al. (n 111), 397. Which may employ different methods to construct a counterfactual: either an ‘elimination test’, according to which, the defendant’s conduct is eliminated from the scene, all other conditions remaining exactly the same, or a ‘substitution’ test, which replaces the unlawful conduct with a counterfactual in which the defendant acted lawfully, all other conditions remaining the same. 115 Ludwig von Bar, Die Lehre vom Causalzusammenhange im Rechte, besonders im Strafrechte, (Leipzig: Bernhard Tauchnitz, 1871). 116 The concept is of particular interest for our purposes as it seems that it has also exercised some influence on the development of the logical concept of probability in the work of the Vienna circle: Michael Heidelberger, ‘Origins of the logical theory of probability: von Kries, Wittgenstein, Waismann’ (2001) 15 International Studies in the Philosophy of Science, 177–88; and hence the development of the concept of economic causality: Guido Fioretti, ‘John Maynard Keynes and Johannes von Kries’ (1998) VI(3) History of Economic Ideas, 51–80. 117 On the question of the evaluation of the significant increase of probabilities, see our discussion above on the different positions between those advancing an ‘individual foresight approach’, such as Johannes von Kries, those arguing for an ‘objective handsight’ theory, such as Max Ru¨melin, or the intermediary approach defended by Ludwig Traeger. 118 van Gerven et al. (n 111), 453. 119 Notes to VI.-4.101 of the Common Frame of Reference referring to German tort law, available at , p. 3432. 120 van Gerven et al. (n 111), 453. 114

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

proved by the plaintiff, the second can be assessed by the court on the basis of available evidence.113 The factual dimension of the causal inquiry in German tort law requires claimants to demonstrate the existence of an adequate causal nexus between the conduct (an act or an omission) and the damage. This is undertaken under the equivalence of conditions test (the condition-sine-qua-non or but-for test).114 This is also complemented with the more generalizing theory of adequate cause (‘Ada¨quanztheorie’). The theory has been interpreted differently by German legal scholarship, some authors insisting on the idea that a cause of an ultimate event is the factor that departs from the ordinary course of events,115 while others have relied on a probabilistic view of causation arguing that a fact should be considered as the adequate cause of the damage, if and only if, it has significantly increased the objective probability116 of the occurrence of the damage.117 According to the formulations usually adopted by German courts, an event is the adequate cause of a particular result, ‘if [it] general[ly] increases, in a significant way, the objective probability of a result of the kind which occurred118’ or, alternatively, the event would have been considered as an adequate cause of the damage, ‘if it could engender the damage in question only under particularly unique, improbable circumstances which would have been disregarded had events followed their usual course’,119 ‘these probabilities being assessed from the point of view of an “optimal observer” which knows all the circumstances surrounding the injurious event which could be known at the time of that event and is furthermore equipped with the general experience of mankind’.120 Hence, German law adopts the generalizing theory of adequate cause, essentially based on probabilities.

32

Lianos

121

Notes to VI.-4.101 of the Common Frame of Reference (n 119), p. 3432. U. Magnus, ‘Causal uncertainty and proportional liability in Germany’, in I Gilead, M D Green, and B A Koch (eds), Proportional Liability: Analytical and Comparative Perspectives (Berlin: De Gruyter, 2013), 153–69, 154. 123 Ibid, 161 (noting that ‘(u)nless otherwise agreed upon, order by law or indicated by the circumstances, all tortfeasors bear an equal share towards each other’, this resulting ‘in a final partial liability of all tortfeasors’). 124 Ibid, 153. 125 Ibid, 155. Of course, an eventual contribution of the victim (contributory negligence) to the realization of the damage should be taken into account at the level of redress/quantification of the damage. 122

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

German legal scholarship and the courts, have also developed the concept of the ‘scope of the rule’ (‘Schutzzweck der Norm’) in order to delimit the damages that can be recovered. The rule postulates that ‘an obligation to make reparation will only arise, if the damage claimed, according to its type and its origin, stems from a sphere of danger which the infringed norm was enacted to protect against’.121 The theory inevitably introduces normative and policy considerations in the assessment of causation in fact, thus conflating with ‘causation as a foundation of liability’ (‘Haftungsbegru¨dende Kausalita¨t’), which, as we have previously highlighted, forms part of the normative phase of the causal inquiry. The concept offers the necessary degree of flexibility to take into account the interests protected by the norm and the sphere of risks each of them should be expected to bear. With regard to causal uncertainty arising out of several tortfeasors causing the same damage, German law prefers the solution of the solidary liability of all defendants, as ‘it is thought that it should not be the victim’s risk to prove who caused which part of the damage’.122 Where one of the several liable tortfeasors compensates the claimant for the full damage, the claimant may benefit from a redress/contribution claim towards the other tortfeasors (§426 BGB).123 With regard to genuine causal uncertainty, the application of the generalizing theory of adequate cause is limited by its narrow confines and also the strict standards of proof applied in German law with regard to Haftungsbegru¨dende Kausalita¨t. In relation to the first limit, German tort law is satisfied not only with evidence that the defendant created a risk of harm and that the plaintiff suffered or may have suffered a risk-related loss to establish causation in fact, but further requires ‘a strict limit—in time and space—between the defendant’s act and the damage, and that ‘(t)he particular risk which the defendant created and the particular situation must have been likely and apt to cause the concrete damage that ensued’.124 If the conditions of liability are met and the damage is certain, then the defendant must compensate the damage in full, regardless of how minimal the defendant’s contribution in regard to causation or fault: ‘the slightest negligence leads to the establishment of full liability’ (an all or nothing approach).125 In relation to the second limit, German law requires that the judge, who is allowed a free assessment of the evidence, is convinced (has a

Causal Uncertainty and Competition Law Claims for Damages

33

126

Ibid, 159. U Ibid, 160. A lower standard of conviction is required with respect to causation concerning the extent of the damage (Haftungsausfu¨llende Kausalita¨t), where ‘the belief of preponderant probability [+50,1%] suffices’. 128 On the distinction between ‘belief probability’ and ‘fact probability’, see S Gold, ‘Causation in toxic torts: Burdens of proof, standards of persuasion, and statistical evidence’ (1986) 96 Yale Law Journal, 376–402; Richard W Wright, ‘Proving causation: Probability versus belief’, in R Goldberg (ed.) Perspectives on Causation (Oxford: Hart Publishing 2011) 195–220. 129 Magnus (n 122), 159. On the proof proximity principle as an EU evidence principle, see C Volpin, ‘The ball is in your court: Evidential burden of proof and the proof-proximity principle in EU competition law’ (2014) 51(4) Common Market Law Review, 1159–85. 130 Magnus (n 122), 160 & 163. 131 Ibid, 164 132 P von Dietze and H Janssen, Kartellrecht in der anwaltlichen Praxis (4th edn, Munchen: C.H. Beck, 2011), at 836; G. Ma¨sch, Praxiskommentar zum deutschen und europa¨ischen Kartellrecht (Mu¨nster: Lexis Nexis, ZAP, 2010). 127

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

firm belief without any reasonable doubts remaining) that the factual allegation is or is not true and is obliged to state the reasons that led him to this convinction.126 To be convincing, a belief of probability should be close to certainty, thus a balance or probabilities standard will not be sufficient to carry conviction. According to some interpretations of the German civil procedure rules,127 even a belief of a 99 per cent probability will not be considered sufficient, if for instance it relates to a fact probability of 51 per cent that the defendant’s contribution caused the damage.128 Hence, in these cases, in the absence of certainty, liability is unjustified, unless a specific causal presumption applies for fairness purposes, in which case the burden of proof shifts to the defendant. The general risk of uncertainty is thus borne by the claimant, although occasionally some procedural rules may facilitate the claimant’s burden (rules on prima facie evidence, proof proximity or the ‘so called secondary burden of evidence’ for the defendant to establish the facts where the claimant has done what he reasonably could and the defendant has the means to adduce further evidence).129 German tort law does not accept ‘market share’ liability, nor the ‘loss of chance’ doctrine.130 With regard to future harm that remains uncertain, an award will also be rejected, although if harm is certain but its extent (in the future) is uncertain, the German courts will assess this issue at the level of the quantification of damages (Haftungsausfu¨llende Kausalita¨t), the standard of proof required in these cases being lower.131 Damage claims for competition law infringements under section 33(1) of the German Competition Act (§33 Gesetz gegen Wettbewerbsbeschra¨nkungen— GWB) require the establishment of a causal connection between the cartel agreement and the damage suffered. Relevant in this respect is the so called theory of adequate causation (Ada¨quanztheorie),132 as it has been developed in the context of German tort law: In principle, the cartel agreement must be sufficient to cause the specific financial loss suffered. Furthermore, it is required that the damage claimed (as defined by reference to the type, the level, and the

34

Lianos

133 T. Lu ¨ bbig, in T. Lu¨bbig, K. Markert, and K. Stockman (eds.), Mu¨nchener Kommentar zum Europa¨ischen und Deutschen Wettbewerbsrecht, Vol II (2nd edn, Munchen: C.H. Beck, 2014), §33, para. 36 et seq; W.H. Roth, in W. Jaeger; H. Achenbach (eds.), Frankfurter Kommentar zum Kartellrecht (3rd edn, Ko¨ln: Verlag Dr. Otto Schmidt, 2011), §33, para. 111. 134 Cp. Court of Appeal of Berlin (Kammergericht Berlin), decision of 1 October 2010—2 U 17/03 Kart, BeckRS 2009, 88509 and Court of Appeal of Karlsruhe (Oberlandesgericht Karlsruhe), decision of 31 July 2013—6 U51/12 (Kart), BWGZ 2013, 1011; see also J. Topel, in D. Schroeder, G. Wiedemann and T. Lu¨bbig (eds.), Handbuch des Kartellrechts (2nd edn, Munchen: C.H. Beck, 2008), §50, para. 130 with reference to Monopolkommission, Das allgemeine Wettbewerbsrecht in der Siebten GWB-Novelle, Sondergutachten der Monopolkommission gema¨ß §44 Abs. 1 Satz 4 GWB, para. 61. 135 J. Haucap and T. Stu ¨ hmeier, Wie hoch sind durch Kartelle verursachte Scha¨den: Antworten aus Sicht der Wirtschaftstheorie, Wirtschaft und Wettbewerb (WuW), 2008(4): 413–424, at 420 et seq. 136 See Topel (n 134), §50, para. 132; in favour of this approach also Federal Supreme Court (BGH Kartellsenat), decision of 28 June 2011—KZR 75/10 (ORWI), BGHZ 190, 145–72. 137 J. Bornkamm, in E. Langen and H.J. Bunte (eds.), Kommentar zum deutschen und europa ¨ ischen Kartellrecht (12th edn, Berlin: Hermann Luchterhand Verlag, 2014), §33, para. 152 et seq. 138 F W Bulst, Schadensersatzanspru ¨ cheder Marktgegenseite im Kartellrecht (Baden Baden: Nomos, Berlin: Beck, 2006), at 133. 139 The German Federal Supreme Court (Bundesgerichtshof) in its recent ruling in ORWI shifted the burden of proving the causal link between the breach of cartel law and the damage to the claimant, BGH Kartellsenat, decision of 28 June 2011—KZR 75/10 (ORWI), BGHZ 190, 145–72; see also J

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

origin of the damage) be consistent with the protective purpose of the (allegedly) infringed provision.133 Generally, the necessary causal link between the prohibited agreement and the damage claimed is not to be presumed. However, it seems that in practice the direct purchaser can easily prove a causal connection between the cartel agreement and the damage suffered.134 The situation is different with respect to the damage (ie the increased price) suffered by indirect purchasers at subsequent levels of the market, where pricing depends upon several factors related to the market structure and the prevailing commercial strategies. In fact, an increased price on a retail market might not be attributable to the cartel agreement on the upstream market, but rather to the (direct) purchaser’s pricing strategy, in light of her particular market position and/or the conditions in the next market level (as for instance, the price elasticity of demand and supply, the intensity of competition etc.).135 Accordingly, the causation requirement has to be assessed on a case-by-case basis.136 The damage is, thus, causally linked to the prohibited agreement where most of the providers (direct purchasers) in the aftermarket have inflated their prices due to the cartel agreement on the upstream market. On the contrary, no causal connection can be asserted where the price increase results from special commercial services or from the provider’s dominant position.137 In any event, the (direct) purchaser’s independent price fixing strategy does not interrupt the chain of causation between the initial cartel agreement and the price increase in the aftermarket.138 Regard must also be had to the principle that it is up to the claimant bringing antitrust damages actions under Section 33(1) GWB to establish the causal connection between the cartel agreement and the damage suffered.139 As regards

Causal Uncertainty and Competition Law Claims for Damages

35

Bornkamm, Wem stehen zivilrechtliche Anspru¨che bei Kartellversto¨ßen zu? in P Behrens, M HartmannRuu¨ppel, and J Herrlinger (eds), Schadensersatzklagen gegen Kartellmitglieder (2010), 51, at 58; Lu¨bbig (n 133), §33, para. 106 et seq. 140 K. Schmidt in U. Immenga and E.J. Mestma ¨cker (eds), EU Wettbewerbsrecht (5th edn, Mu¨nchen: C.H. Beck, 2012) §33 GWB, para. 28 and G Meesen, Der Anspruch auf Schadensersatz bei Versto¨en gegen EU-Kartellrecht (Tu¨bingen: Mohr Siebeck, 2011), at 414. 141 ORWI, BGH, 28 June 2011—KZR 75/10, BGHZ 190, 145 = BB 2012, 75 = WRP 2012, 209 = ZIP 2012, 390 = GRUR 2012, 291 = NJW 2012, 928 = JZ 2012, 789. 142 ORWI, OLG Karlsruhe, 11 June 2010—6 U 118/05 (Kart), BeckRS 2011, 26582. 143 ORWI, BGH, NJW 2012, 928, ô 45.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

the applicable standard of proof, it is sufficient for the claimant to provide prima facie evidence of the causal link.140 The ORWI case of the German Federal Court of Justice (BGH) constitutes the only example so far of a case in which the claim for damages was introduced by an indirect purchaser.141 The plaintiff was a printer business which had purchased carbonless paper from a wholesaler, a subsidiary of one of the paper producers fined by the Commission for participating in a price fixing cartel in the carbonless paper industry from 1992 until 1995. The printer business became insolvent and assigned its damages claims against the cartelists to the plaintiff, a German savings bank, which brought an action against the supplier. The action was dismissed by the Regional Court of Mannheim, finding that the customers of a subsidiary constitute indirect purchasers who may not claim damages. Although the decision was partially confirmed by the Higher Regional Court (OLG) of Karlsruhe, it nevertheless accepted the claim, by invoking that the wholesaler was a 100 per cent subsidiary of the cartelists, and thus not providing the possibility for an action for damages in this case would have simply enabled cartel members to avoid liability by strategically using its subsidiary to supply customers.142 The BGH nevertheless rejected the decision of the OLG Karlsruhe, thus allowing indirect purchasers to raise damages claims against the members of a cartel, at the same time accepting the consequent availability for the defendant of the passing-on defense. The BGH highlighted that it is the indirect purchaser alleging that the overcharge was passed on to the next market level who bears the burden of proving it, considering the theory of adequate causation, as has been established in the framework of German tort law in accordance with EU Law, which is of relevance in this respect. The Court also stated that in view of the economic complexity inherent in pricing and competitive pressure in the downstream markets, a causal link between an overcharge in the aftermarket, which coincides in time with the prohibited agreement or practice under Article 101(1) TFEU, and the relevant cartel agreement may not be presumed, thus requiring a substantiation on a case by case basis of the causal link between the cartel agreement and the pricing in the next market level.143 This causal link should be established by reference to the hypothetical price level in the absence of the cartel, in particular that the price increase is attributed to the cartel

36

Lianos

agreement, rather than other factors having an impact on the pricing strategy. Indeed, as the Court explains, ‘the purchaser’s pricing strategy on the upstream market level is possibly not based on the conditions in the market where the cartel agreement is implemented, but rather on its particular market position and/or the specific conditions in the aftermarket’.144

144

ORWI, BGH, NJW 2012, 928, ô 46. Article 1151 on contractual liability exemplifies this requirement of direct and certain damage, similar principles animating non-contractual liability, the requirement of direct causal link being considered as the most important requirement. 146 Duncan Fairgrieve and Florence G’Sell-Macrez, ‘Causation in French law: Pragmatism and policy’, in Richard Goldberg (ed.), Perspectives on Causation (Oxford: Hart Publishing, 2011), 111. 147 For a more extensive analysis of the concept of causation in French law see, Florence G’SellMacrez, Recherches sur la notion de causalite´ (th. Paris I, 2005); Ce´dric Beadeux, La causalite´, fondement pour une the´orie ge´ne´rale de la responsabilite´ civile (Editions universitaires europe´ennes, 2010); Christophe Que´zel-Ambrunaz, Essai sur la causalite´ en droit de la responsabilite´ civile (Paris: Dalloz : Nouvelle bibliothe`que de the`ses, vol. 99, 2010). 148 The recent reforms of French tort law initiated by the publication of the preliminary draft reform prepared by a committee chaired by Pierre Catala in September 2005 noted that any attempt to define causation would be futile, hence it just referred to the requirement of causality in Article 1347 145

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(c) France In contrast to English and German law, article 1382 of the French Civil Code, employed as the legal basis for damages claims, puts in place a single heading of civil liability, espousing a general open-ended principle of liability without any a priori limits on the boundaries of the potential liability of the wrongdoer. In order to compensate for such a wide-ranging approach, French law incorporates strict causation in law (proximate cause) requirements, which also inevitably influence the choice of the causation test. Indeed, in order to be recoverable the damage should be certain, actual, and direct145 and civil courts have occasionally referred to the notions of direct link, even when examining causation in fact.146 With regard to the factual/non normative aspect of causation, French courts seem to take a variety of approaches. Some civil courts apply the individualizing ‘theory of equivalence of conditions’ or sine qua non condition, according to which every factor contributing to the occurrence of the effect should be considered as a cause.147 In other words, all the conditions without which the effect would not have occurred are necessarily factual causes. Other courts adopt the generalizing ‘theory of adequate causation’, which involves a selection from the various conditions of the factor that departs from the ordinary course of events. However, French courts do not adopt a probability-based version of the adequate cause theory, preferring a version that emphasizes a departure from the ordinary course of events, for instance by the interference of a third party or the plaintiff, or a natural event, following the occurrence of the condition generated by the defendant’s conduct, which is unforeseeable (‘impre´visible’) and unavoidable (‘irresistible’), without however making any explicit or implicit reference to probabilities.148

Causal Uncertainty and Competition Law Claims for Damages

37

without any effort to provide a specific content to the concept. The preliminary draft also advanced the solution of liability in solidum (article 1348 of the preliminary draft) in circumstances of causal uncertainty because of multiple sufficient causes (the intervention of multiple defendants each of which would have been sufficient to produce the result). The more recent reform draft, currently under discussion in the French Parliament, adopts explicitly (in Article 10) the adequate cause theory for the definition of causation, although it makes reference only to its version emphasizing the ordinary course of events. It also includes a limitation to the theory by requiring that the person liable for the damage should only be held responsible for the immediate and direct consequences of it (‘suites imme´diates et directes’). Finally, the government’s draft also referred to the possibility of establishing the existence of a causal link with any type of evidence (‘par tous moyens’), thus opening the possibility to the development of causal presumptions. In a report published in February 2012, the Working Group on the reform of French tort law, chaired by professor Franc¸ois Terre´, suggested to eliminate any reference to a specific ‘philosophical theory’ of causality, such as the theory of adequate cause, and also proposed to eliminate the reference to the ‘immediacy’ of the causal link between the conduct of the defendant and the loss, as this did not add anything to the requirement of direct link. 149 O. More ´teau, ‘Causal uncertainty and proportional liability in France’, in I Gilead, M D Green, and B A Koch (eds), Proportional Liability: Analytical and Comparative Perspectives (Berlin: De Gruyter, 2013), 141–52, 144–7. 150 Ibid, 143. 151 Ibid, 150.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

French tort law has predominately chosen to by-pass situations of causal uncertainty through the concept of loss of chance (‘perte de chance’), which has allowed French courts to express the causal impact of the defendant’s conduct in probabilistic terms. In particular, French courts employ the ‘perte de chance’ doctrine as a response to scientific uncertainty about the existence of a causal link between the condition and the result, responsibility in solidum being mostly used for evidential uncertainty with regard to multiple tortfeasors.149 Loss of chance enables French courts to conceal scientific uncertainty, as loss of a chance is defined as a specific head of damage, which is certain, as far as this concerns the loss of a favourable opportunity for the claimant. The compensation ‘must correspond to a share of the victim’s various heads of damage’ and ‘lower courts must to that effect assess each head of damage and decide the proportion to which it constitutes a reparable loss of a chance’.150 This comes close to the apportionment which may be undertaken under proportional liability, although here the damage is considered as ‘certain’ (in view of the strict requirements of the standard of proof for damages in France). The solution of French tort law is therefore markedly different from that adopted in the context of German tort law and to a certain degree English law, which prefer instead to respond to causal uncertainty with procedural rules, such as reversing the burden of proof to the defendant, rather than substantive principles such as the establishment of a specific type of damages for loss of chance. Although compensation for future hypothetical harm is not awarded, it is possible to compensate future loss when ‘all the conditions for its existence in the future already exist at the time of the facts’.151 The concept of loss of a chance may provide a possible way out in such situations as well as situations where the scope of the future harm is uncertain.

38

Lianos

152 Paris Court of Appeal, SNC Doux Aliments Bretagne etc v SAS Ajinamoto Eurolysine, No. 07/ 10478 (10 June 2009). 153 French Supreme Court, commercial, Ajinomoto Eurolysine v SNC Doux Aliment Bretagne and others, No. 09/15816 (15 June 2010). 154 For a similar approach in the context of another claim for damages in the lysine case, see Paris Court of Appeal, La SCA Le Gouessant v SA CEVA SANTE ANIMALE & SAS Ajinomoto Eurolysine (16 February 2011), finding that the claimants have not proved to the sufficient standard that they did not pass on any of the overcharge to their customers. See also, Tribunal de Commerce de Nanterre, SA Les Laboratoires Pharmaceutiques Arkopharma v Ste Roche etc, No. 02004F02643 (11 May 2006) noting that the fact that the Commission in its vitamins cartel infringement decision

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Causation jurisprudence in competition law cases in France is intrinsically linked with the doctrine of loss of chance (‘perte de chance’), which provides an escape valve for the strict requirements of certainty and directness of the causal link between the damage and the infringement, although this is also not without ts limits. We will comment on some cases that highlight the specificities of competition law causation jurisprudence in France. The decision of the European Commission sanctioning the lysine price-fixing cartel has led to a number of follow-on damages claims in French courts by purchasers of lysine, active in various downstream markets, which saw their commercial margins affected by the cartel overcharge. Yet, despite the existence of a Commission infringement decision, purchasers have faced substantial difficulties in actions for damages. For instance, the Doux group, Europe’s largest producer of poultry claimed, relying on an expert report, that the lysine cartel led to an increase of 30 per cent in the price of lysine, as a result of which it suffered damage, due to reduced margins and a drop in competitiveness, which amounted to 30 per cent of its purchases of lysine during the period of the cartel. Ajinamoto Eurolysine, the defendant in the action and an addressee of the Commission’s infringement decision argued that Doux has passed on the overcharge to its consumers and that in any case the expert report on which Doux’s claim relied was imprecise and did not account for the important fluctuations of the price of lysine during this period. The Court of Appeal of Paris rejected Ajinamoto’s first argument and used the loss of chance doctrine in order to respond to the second, thus finding that Doux had suffered a loss of a chance that had to be compensated to the level of 30 per cent of the amount of damages requested.152 No explanation was provided in the judgment in the choice of the 30 per cent figure, presumably referring to the probability of occurrence of the loss of chance in this case. The French Supreme Court quashed the judgment of the Court of appeal on several grounds.153 First, it held that the Court of Appeal should have considered the possibility that Doux could have passed on the overcharge to its consumers, thus requiring the consideration of an eventual passing on of an overcharge, but more controversially allocating to the claimant the burden of proving that no pass on was effectuated in the circumstances of the case.154

Causal Uncertainty and Competition Law Claims for Damages

39

found that the cartel had effects on the final consumers only, precluded intermediary consumers, such as Arkopharma, which purchased vitamins from Roche, to claim that they did not pass on the eventual overcharges to the final consumers, thus breaking the causal link between the higher prices as a result of the cartel and the damage from which Akropharma allegedly suffered. The Court surprisingly observed that since the cartel was international and covered 80% of the global market, the intermediary consumers could pass on the overcharge to the final consumers without incurring any risk that competitors will not pass on the overcharge. A further example of the strict standard of proof required by the claimants to prove that they have not passed on the overcharge (the burden being on them rather than on the defendants) is provided by Tribunal de Commerce de Paris, Socie´te´ les Laboratoires Juva Production etc v SAS Roche etc, No. RG2003048044 (10 September 2003), also a case brought following a Commission’s decision on the vitamins cartel, which noted that in view of the ‘essential’ nature of the products sold by Juva (food complements), there would be no volume effect following the price increase because of the cartel, and in any case, the cartelized products constitute only a small proportion of the cost of the products sold by Juva, hence it was possible for the latter to pass on the overcharges to consumers by means of relatively modest price increases. 155 Paris Court of Appeal, SNC Doux Aliments Bretagne etc v SAS Ajinamoto Eurolysine, No. 10/ 18285 (27 February 2014).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

This is, of course, incompatible with Article 13 of the Damages Directive, which places the burden of proof on the defendant (passing on is considered as a defence). Secondly, it held that the causal link between the damage and the fault/infringement should not be hypothetical, noting that the ‘erratic’ fluctuations of the lysine price during the cartel period did not enable the court to make an inference of causal link between the damage suffered and the infringement, this appearing to be purely hypothetical in this context. The case was then sent back to a different formation of the Paris Court of Appeal for reconsideration.155 The claimants were arguing on three counts: one principal and two subsidiary. The principal consisted in compensation of the loss suffered during the period of the cartel; the subsidiaries related to the loss of chance and the disgorgement of illicit profits (referring to the cartel overcharge). The Court of Appeal considered the possibility of a preliminary ruling question to the Court of Justice on the issue of who bears the burden of proof for an eventual pass on of an overcharge, yet it found that the issue was clearly settled by the case law of the French Supreme Court and that this interpretation of Article 1382 of the French Civil Code did not conflict with the principles of equivalence and effectiveness. It decided eventually not to send any preliminary question to the CJEU on this issue. The reader is reminded that the recently adopted Damages Directive proceeds to an EU harmonization of the matter by instituting a passing on defence. The Court of Appeal of Paris then proceeded to examine the principal count of damages claimed by the Doux group to find that although the expert report submitted by the claimants in the first instance relied on, in some situation, imprecise methods, these findings were based on the ‘incontestable’ data contained in the Commission’s decision and therefore proved beyond doubt the surcharge suffered by Doux and the subsequent damages engendered by the additional expenses incurred in order to mitigate the results of the price increase.

40

Lianos

(ii) Causation standards and categories of claimants Although the jurisprudence of the CJEU and/or the Damages Directive do not restrict the category of possible claimants in actions for damages and take the widest possible approach with regard to standing,158 in practice, it may be more difficult to establish causation in bringing successful damages actions159 for parties which are not direct purchasers or competitors of the infringers. There

156 Court of Appeal of Versailles, SA Concurrence v. SA Aiwa France, No. 01/08413 (9 December 2003). 157 Paris, Court of Appeal, M. Merhi Bassam v SNC Socie ´te´ Presse Paris Services—SPPS, No. 08/ 21750 (27 April 2011). 158 According to Article 2(1) of the Damages Directive, ‘Member States shall ensure that any natural or legal person who has suffered harm caused by an infringement of competition law is able to claim and to obtain full compensation for that harm’. This essentially codifies the expansive case law of the CJEU in Joined Cases C-295/04 to C-298/04, Manfredi & others [2006] ECR I-6619, para. 64. In contrast, Section 4 of the Sherman Act imposes some limits: ‘(a)ny person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue thereof’ (emphasis added). US courts have relied on the directness test in order to exclude standing for officers of the company, shareholders, creditors of an injured corporation, suppliers (in some circumstances when the anticompetitive conduct concerns the downstream market, rather than upstream) a landlord of an injured lessee, a patent holder licensing the patent to an injured licensee, a franchisor of an injured franchisee, employees (when the competition law infringement does not concern their wages or other conditions of employment), or trade unions: E L Cramer and D C Simons, ‘Parties entitled to pursue a claim’, in A A Foer and R M Stutz, Private Enforcement of Antitrust Law in the United States (Cheltenham: Edward Elgar, 2012), 64–94. 159 OXERA, Quantifying Antitrust Damages—Towards non-binding guidance for courts, December 2009, p. 24.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

The French courts have had recourse to the doctrine of the loss of chance on more than one occasion. For instance, in SA Concurrence v/ SA Aiwa, the Court of Appeal of Versailles had to assess the damage a discounter had suffered as a result of the selective rebates policy adopted by one of its suppliers, which benefited only to some of its competitors.156 The Court analysed the damage as a loss of chance to make additional sales and, taking into account the volume of sales of the discounter, its usual margin, and the purchases it made from the supplier during the period of the infringement, the court granted damages that corresponded to less than one-sixth of the amount initially requested by the claimant. In M. Merhi Bassam v SNC SPPS, the Court of Appeal of Paris also employed the loss of chance doctrine to provide compensation to a potential buyer of a newspaper stall who was excluded from the distribution network of SNC SPPS, the dominant distribution channel for newspapers in the Paris region, for having rejected the presumably abusive clauses of the latter’s standard contract, on the basis of the existence of an agreement with the previous owner of the newspaper stall to transfer the property rights on the stall subject to the approval of SNC SPPS.157 Hence, the damage occurred was not hypothetical, as its occurrence relied on past events (the sale of the stall).

Causal Uncertainty and Competition Law Claims for Damages

41

160 One may also put forward that financial difficulties as a result of exclusionary practices against an undertaking may affect its employees if that undertaking imposes salary cuts or fires some employees (although in this case the causation chain may break because of the intervention of a third party, the management of the excluded undertaking or the employer). Yet, if one adopts the broad view of AG Kokott in Case C-557/12, KONE AG and Others (30 January 2014), and consider that all foreseeable losses by the cartel members or the injurer in general may be recoverable, one may argue that such losses are foreseeable and thus recoverable.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

are several parties that might be affected by a competition law infringement such as a cartel agreement. First, there are direct purchasers/customers that purchase goods or services directly from the infringer. These customers forgo the benefit of additional sales in the form of utility in consumption or profit when resold to the final customer in a competitive environment. An equivalent effect occurs for upstream suppliers. By exercising buyer power, a cartel may enforce lower input prices upstream. Depending on the specific market conditions, input price reduction may be enforced by the cartel through output contraction affecting both existing and potential suppliers. Moreover, upstream suppliers may (partially) pass-on the worsened sales conditions to their own upstream suppliers. Indirect purchasers, purchasing goods or services from a downstream supplier of the infringer may also be affected where the overcharge was passed on to their level. Exclusionary practices may affect (potential) competitors outside the cartel. Competitors in the same relevant market that are not participating in the cartel agreement, or potential competitors in related product or neighbouring regional markets, are potentially affected by exclusionary practices. The opposite can also happen: competitors outside the cartel could benefit by softened competition, enjoying the higher prices that are due to the cartel (the so-called umbrella effect). The direct customers of these competitors may suffer an overcharge because of the umbrella effect. Finally, one should also take account of the effect on potential customers (intermediate customers or end consumers) who would have purchased the products or services directly or indirectly from the infringers in the absence of the infringement (counterfactual customers). These suffer harm in view of the fact that they did not purchase the product or service, although they would have done so if charged the counterfactual price, or they are pushed to purchase lesspreferred alternative products and services. Other injured parties may include employees, where the undertaking for which they worked goes bankrupt and consequently they lose their employment,160 or shareholders, where the shares of the undertaking they have invested in, as victims of the anticompetitive conduct, dropped, following the anticompetitive/exclusionary conduct.

42

Lianos

There is also a significant ‘structural asymmetry in the distribution of information required by claimants to prove causation.161 As the Commission recognized in the Staff Working paper accompanying its 2008 White Paper:

A careful analysis of the damages cases for infringement of competition law brought in the most significant jurisdictions in Europe, in terms of the size of their economy and number of cases, shows that, in the overwhelming majority of cases, damages actions are brought by direct purchasers or competitors, and it is extremely rare they are introduced by indirect purchasers. We have not been able to identify one case in which the damage claim was initiated by counterfactual direct (or indirect) purchasers. A possible explanation for this phenomenon is that it is often difficult to identify these counterfactual customers and measure with certainty the harm incurred. Indeed, it is difficult to establish how much a counterfactual customer would have bought, in the event that he was charged the counterfactual price. Any finding will be subject to uncertainty and assumptions about the probable behaviour of the counterfactual customer and his preferences. Requirements of certainty and foreseeability may curb the ability of those customers to receive compensation for the harm incurred. Similar difficulties persist with indirect purchasers or end customers who are only 161

National Grid Electricity Transmission Plc v ABB Ltd [2012] EWHC 869 (Ch), para. 41. Staff working paper accompanying the White Paper on Damages actions for breach of the EC antitrust rules COM(2008) 165 final, para. 89. 162

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Even where claimants are in a position to describe and prove the factual elements necessary for finding an infringement, having to demonstrate in detail the causation and quantification of their damages remains a particular difficulty in competition cases. To establish their damage, claimants have to compare the anti-competitive situation to a situation which would have existed in the absence of the infringement, i.e. a hypothetical competitive market. In a breach of contract case, a claimant can normally use market prices at the time of the breach of contract as the benchmark for calculating his loss. However, in a typical competition case, the claimant cannot rely on the prices at the time of the infringement and has to establish what the price would have been in the absence of the restriction of competition. For this purpose, he will often depend on information that is in the sphere of the defendant and possibly their partners in the infringement: for example, notes on the price overcharges agreed secretly between cartel members, details on how and when they influenced price and other parameters of competition, or internal documents of the infringer showing his analysis of market conditions and developments. Also the reconstruction of a hypothetical competitive market to quantify the damage caused by the infringer usually presupposes knowledge of facts on the commercial activities of the infringer and other players on the relevant market. The same or similar types of difficulty arise in the context of causation, e.g. when claimants try to identify the precise elements of anticompetitive behaviour by an infringer that have caused the claimants damage, or the extent to which several infringers have individually contributed to the damage caused.162

Causal Uncertainty and Competition Law Claims for Damages

43

considered if there is plausible evidence of passing on.163 In the absence of passing on, the causality link between them and the infringer is broken. It is also clear that counterfactual customers resulting from the volume effect (customers not buying the product from the indirect purchasers as a result of the price increase following the passing on of the overcharge) may not be able to receive compensation. Competitors may also be unable to overcome the strict requirements of causation, with regard to eventual lost profits resulting from the competition law infringement, in view of the limits of the counterfactual inquiry.

In the absence of a clear definition of causation in the different legal systems of the EU Member States,164 any effort to define a uniform test for establishing causation in fact and causation in law remains a Sisyphean task. The Damages Directive does not proceed to the development of harmonized rules on causation, but only includes provisions that relate to causation, by advancing causal presumptions favouring certain categories of defendants (subsection (a)). Hence, with regard to causation, the reference in Manfredi to ‘the legal system of each Member State to prescribe the detailed rules governing the exercise of [the] right [to damages], including those on the application of the concept of ‘causal relationship’, remains unchanged. The Court of Justice judgment in Kone provides a first view of the emergent EU rules on causation, in order to ensure the effectiveness of the right to claim damages for a certain category, umbrella customers, which might find themselves disadvantaged by strict national tort law requirements on causation (subsection (b)). (a) Causation in the Damages Directive Although the declared aim of the Damages Directive is the ‘approximation’ of national rules, with a view to ‘prevent[ing] the emergence of wider differences between the Member States’ rules governing actions for damages in competition cases’,165 the Directive observes that the issue of the causal relationship ‘is not dealt with in this Directive’ and makes explicit reference to national rules and procedures to deal with this issue, under the dual framing of the principles of effectiveness and equivalence, following the well-established case law of the EU 163 See, however, the causal presumption introduced by the Damages Directive in order to deal with this issue examined below Section (iii). 164 See, for instance Notes to VI.-4.101 of the Common Frame of Reference, available at , p. 3422: ‘The term “causation” is used in all European systems of legal liability, but is nowhere defined’. 165 Directive 2014/104/EU of the European Parliament and of the Council of 26 November 2014 on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union [2014] OJ L349/1, Recital 9.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(iii) Different strategies of EU ‘harmonization’ of causation requirements in competition law based actions for damages

44

Lianos

166

Directive, Recital 11. Ibid. 168 Directive, Recital 12; Articles 1(1), 3(1) and 3(2). 169 Another possible limitation on claims for damages may be the principle of no overcompensation of the injured party, which, however, applies ‘without prejudice to compensation for loss of opportunity’. See Directive, Recital 13; Directive, Article 3(3) and 12(2). (Note that the expression ‘loss of opportunity’ used by Recital 13 replaced the expression ‘loss of a chance’ employed in earlier versions of the Directive). Overcompensation is explained in Article 12(2) of the Directive as requiring that ‘compensation for actual loss at any level of the supply chain does not exceed the overcharge harm suffered at that level’.This principle seems to be somehow related to the issue of causality and quantum, or to constitute an expression of the causality principle. First, ‘harm’ is defined as the ‘actual loss’ that can result ‘from the price difference between what was actually paid and what would otherwise have been paid in the absence of the infringement’ (thus indicating that the familiar but-for test in order to find the existence of a causal link will also determine what is harm and thus recoverable): Directive, Recital 39. The principle of overcompensation does not apply to loss of profit claims (which makes sense as loss of profits is an ersatz to causation). Secondly, in evaluating the quantum of the damage, national enforcers should assess ‘how the market in question would have evolved had there been no infringement’ (again introducing the causal ‘but-for’ test albeit this time with a more prospective perspective): recital 46. 170 Directive, Article 4. 171 As a requirement that ‘all national rules and procedures relating to the exercise of claims for damages are designed and applied in such a way that they do not render practically impossible or excessively difficult the exercise of the Union right to full compensation for harm caused by an infringement of competition law’: Directive, Article 4. 172 Articles 1 and 3 of the Directive. 173 Article 2(20) of the Directive. 167

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Courts on this issue.166 Yet, the Directive explicitly mentions that compensation for harm suffered may be claimed from any person, subject to the requirement of a causal relationship between the harm suffered and the infringement of the competition law.167 In contrast to US law, the Directive takes an open approach to standing (as any injured party may claim full compensation of the harm168), while leaving to the requirement of causation the function of potentially circumscribing damages claims.169 These potential limitations to claims for damages should nonetheless be implemented in accordance with the principle of effectiveness, which imposed on Member States the obligation to ensure the effective exercise of the ‘subjective’ right to claim damages for infringements of competition law.170 Although the principle of effectiveness is defined negatively in the directive,171 the ‘substantive effectiveness’ of the exercise of the right might also be perceived positively as the ‘right to full compensation’ for the harm ‘caused’,172 without that of course leading to overcompensation, thus indicating the central function of the causation element in the EU framework and the need to implement this in a relatively liberal way. The Directive contains provisions that illustrate the need to enhance full compensation by facilitating the evidence of a causal link in certain specific circumstances. First, direct or subsequent purchasers may claim compensation for the overcharge, that is, the ‘difference between the price actually paid and the price that would have prevailed in the absence of an infringement of competition law’.173

Causal Uncertainty and Competition Law Claims for Damages

45

it is appropriate to presume that cartel infringements result in harm, in particular via an effect on prices. Depending on the facts of the case, cartels result in a rise in prices,

174 175 176 177 178 179

Directive, recital 39 and Ch IV, in particular Art 13.. Directive, Recital 36. Directive, Article 16(1). Directive, Recital 36. Directive, Article 16(2). Directive, Recital 42.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

This may be more than the harm incurred, had the purchasers passed on the overcharge or part of it to their consumers (their harm is the overchargepass on). The infringer may of course raise a passing on defence, by proving the existence and the extent of the pass on of the overcharge.174 Yet, the practical implication of such a theoretical construction is that it leads to a reversal of the burden of proof to the infringer, who needs to establish the existence of the pass on (as pass on constitutes a defence), instead of requiring the claimants to bring only evidence of the specific harm caused to them by the infringer (thus taking into account in defining their harm the possible pass on of the overcharge). Secondly, the Directive puts in place a presumption of causality for the benefit of indirect purchasers only, in view of the difficulties of ‘consumers or undertakings that did not themselves make any purchase from the infringer to prove the scope of that harm’.175 Hence, as is explained by the Directive, ‘taking into account the commercial practice that price increases are passed on down the supply chain’,176, it is ‘appropriate to provide that, where the existence of a claim for damages or the amount to be awarded depends on whether or to what degree an overcharge paid by the direct purchaser of the infringer has been passed on to the indirect purchaser, the latter is regarded as having brought the proof that an overcharge paid by that direct purchaser has been passed on to his level, where he is able to show prima facie that such passing-on has occurred, unless the infringer can credibly demonstrate to the satisfaction of the court that the actual loss has not or not entirely been passed on to the indirect purchaser’.177 Hence, the indirect purchaser may carry more easily the burden of proof of the existence and scope of pass on by simply providing prima facie evidence that ‘(a) the defendant has committed an infringement of competition law; (b) the infringement of competition law resulted in an overcharge for the direct purchaser of the defendant; and (c) the indirect purchaser has purchased the goods or services that were the object of the infringement of competition law, or has purchased goods or services derived from or containing them’.178 Third, the Damages Directive sets up a causal presumption for cartels in order to ‘remedy the information asymmetry and some of the difficulties associated with quantifying antitrust harm, and to ensure the effectiveness of claims for damages’.179 As is explained in the Directive,

46

Lianos

or prevent a lowering of prices which would otherwise have occurred but for the cartel. This presumption should not cover the concrete amount of harm.180

180

Ibid. See, Practical Guide on quantifying harm in actions for damages based on breaches of Article 101 or 102 of the Treaty on the Functioning of the European Union, SWD(2013) 205, available at , 44, referring to the results of an External study prepared for the Commission ‘Quantifying antitrust damages’ (2009), pages 88 ff., available at (finding that in 93% of all cartel cases considered, cartels do lead to an overcharge). 182 Directive, Article 17(2). 183 Directive, Recital 33 and Article 11. See also the analysis in section (d) of this study. 184 Directive, Article 11(5) and recital 52. 185 Directive, Recital 37. 181

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

Such presumption may result from the Commission’s reliance on studies indicating that only 7 per cent of cartels do not lead to overcharging and, more generally, enforcement priorities.181 Accordingly, the Directive requires Member States to establish a presumption that cartel infringements cause harm, also recognizing the right to the infringer to rebut this presumption.182 Of particular interest is the strategy followed by the Directive in relation to conduct by multiple defendants causing harm. The Directive holds undertakings jointly infringing competition law (eg cartels but also other forms of competition law infringement, resulting from other horizontal or vertical agreements, the abuse of a collective dominant position) jointly and severally liable for the ‘entire harm caused by the infringement’.183 The Directive thus adopts the solution of responsibility in solidum, instead of engaging in the application of the equivalence of condition test (‘but-for test’), which would have led to a finding of no liability where the decision maker is faced with multiple sufficient causes or considers the problem to be a causation issue by applying the generalizing theory of adequate causation or the individualizing efficiency theory of causation in order to determine the relevant cause. Yet, the causation issue is not entirely irrelevant as it is indirectly applicable in the context of the subsequent right recognized by the Directive of the infringing undertaking to recover the overpaid contribution from any other infringing undertaking, the amount of which is determined ‘in light of their relative responsibility for the harm caused by the infringement of competition law’.184 Hence, liability is divided according to the probability that one or other caused the harm. According to the Directive, ‘(t)he determination of that share as the relative responsibility of a given infringer and the relevant criteria, such as turnover, market share, or role in the cartel, is a matter for the applicable national law, while respecting the principles of effectiveness and equivalence’.185 The Directive thus opens the door to a possible implementation of the market shares rule or any other proportional apportionment principle developed by

Causal Uncertainty and Competition Law Claims for Damages

47

186

Case C-234/89, Stergios Delimitis v Henninger Bra¨u AG. [1991] ECR I-935. In application of the second step of the Delimitis test: Case C-234/89, Stergios Delimitis v Henninger Bra¨u AG. [1991] ECR I-935 (contribution of the specific network of agreements to the cumulative effect). The first step assesses the existence of a cumulative effect of all the networks of similar agreements contributing to it. 188 Directive, Recitals 39, 43. 187

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

each Member State, which, as we have previously explained, constitutes an alternative to the determination of causation in fact. In the case of an infringement of Article 101 TFEU following the cumulative foreclosure effect of parallel networks of distribution agreements, such as in Delimitis,186 the suppliers whose networks of distribution agreements would have been found to contribute to the cumulative foreclosure effect and hence to infringe Article 101 TFEU would be found jointly and severally liable for the entire harm caused by the infringement. This would be so even if the cumulative effect would have resulted without the intervention of their network, if the contribution of their network to the cumulative effect is not substantial, but is of a level to be considered an infringement of Article 101 TFEU.187 The application of the but-for test would have resulted in no liability being found in such circumstances, as the condition (the network of distribution agreements) would not have been necessary for the occurrence of the harm. However, the contribution of each network of similar agreements to the cumulative effect would be of relevance when determining the apportionment of the damages between the different infringers found jointly and severally liable, should one of them request the recovery from any other infringing undertaking of the damages paid in excess of its relative responsibility. In this case, reference may be made to the respective market shares of the undertaking and/or its share of the cumulative effect, as determined in the context of the second step of the Delimitis test. The Damages Directive does not provide for equivalent presumptions of causality for injured parties other than indirect purchasers, such as counterfactual customers, and umbrella customers, who may find it difficult to prove the causal link between the anticompetitive conduct of the defendant and the harm they suffered. Harm is also defined narrowly as only referring to the higher prices paid as a result of the anticompetitive conduct or reduced sales leading to a loss of profit, without the Directive making any reference to other types of harm that may result from an anticompetitive exercise of market power, such as a reduction of consumer choice, or reduction of innovation or quality, which do not benefit from equivalent causal presumptions.188 It follows that the EU Courts will inevitably be called, not only to interpret the provisions of the Directive, but also to fill in the numerous areas of the law on which the Directive remains silent. One of these is the issue of causation. A recent preliminary ruling procedure at the CJEU has examined precisely this question.

48

Lianos

189 On this expression See with see, Frank Maier Rigaud, Umbrella effects and the ubiquity of damage resulting from competition violations, Journal of European Competition Law & Practice, (2014) 5: 247–251. 190 As the European Commission explained in its Proposal for a Directive of the European Parliament and of the Council on certain rules governing actions for damages under national law for infringements of the competition law provisions of the member States and the European Union, COM(2013) 404 final, para. 30, ‘(w)hile indirect purchasers are entitled to claim compensation, national rules of causality (including rules on foreseeability and remoteness), applied in accordance with principles of Union law, may entail that certain persons (for instance at a level of the supply chain which is remote from the infringement) are legally unable to claim compensation in a given case’. 191 European Commission, Proposal for a Directive of the European Parliament and of the Council on certain rules governing actions for damages under national law for infringements of the competition law provisions of the member States and the European Union, COM(2013) 404 final, para. 30. 192 Emphasis added.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(b) Causation in the CJEU: umbrella customers and the emergence of the approximation by jurisprudence method Inasmuch as both the CJEU and the EU Legislator have emphasized that full compensation of anyone harmed is the cornerstone of the EU law on damages claims for competition law infringements, the case law and the legislative proposals remain silent as to the possibility of injured parties that are remote in the ‘vertical value chain189’ of the infringer190, or customers who are not part of the vertical value chain, such as counterfactual or umbrella customers. With regard to the first type of customers, the Commission’s proposal for the Damages Directive acknowledged that remoteness or foreseeability requirements may make it legally ‘impossible’ for certain categories of indirect purchasers in a remote position of the vertical value chain to claim full compensation. The solution advanced by the Commission provided that ‘in a situation where the overcharge was passed on to persons who are legally unable to claim compensation, it is not appropriate to allow the infringing undertaking to invoke the passing-on defence, as this would render it free of liability for the harm which it has caused.’191 No equivalent provision exists in the final version of the Directive, thus indicating that if such an issue emerges, it will have to be dealt by the national courts and eventually the CJEU. With regard to the second category of customers, the recent reference for a preliminary ruling from the Austrian Supreme Court to the CJEU in KONE AG and Others arising against the background of the elevator cartel case, illustrates some of the questions to which the CJEU will inevitably be asked to elucidate. The Austrian Supreme Court’s question was formulated by the national court as referring to the issue of standing, as it questioned the CJEU whether ‘any person may claim from members of a cartel damages also for the loss which he has been caused by a person not party to the cartel who, benefiting from the protection of the increased market prices, raises his own prices for his products more than he would have done without the cartel (umbrella pricing)’,192 but implicitly it essentially raised an issue of causation, as it concerned a question, ‘not yet settled

Causal Uncertainty and Competition Law Claims for Damages

49

at European Union level, of whether the civil liability in damages of the members of a cartel also extends to “umbrella effects” or “umbrella pricing”’.193 AG Kokott reformulated the question in order to address two issues (standing and causation) explaining that

Yet, as it is further explained, the issue of causation was the most crucial one as (f )rom a legal point of view, the issue of whether members of cartels can also be held civilly liable for umbrella pricing hangs on the existence or otherwise of a causal link. The question is whether there is a sufficiently close connection between the cartel and the losses resulting from umbrella pricing caused by a cartel, or whether these are excessively remote losses for which damages cannot reasonably be awarded against the members of the cartel.195

The injured party claiming damages had indeed purchased elevators from a manufacturer not involved in the cartel, paying a price which in its opinion was set under the protection of the elevator cartel and was thus higher than would otherwise have been expected under competitive conditions. However, the implementation of Austrian tort law dismissed ‘from the outset’ such claim for compensation, because according to the Austrian courts the loss on account of which the injured party had brought its action could not be attributed to the parties to the cartel on legal grounds. First, the adequate causal link required under Austrian law was not present, and secondly, the loss alleged was not deemed covered by the protective purpose of the competition rules.196 The Austrian court had doubts as to the application of the restrictive national tort law rules in view of the principle of effectiveness. Advocate General Kokott developed a threefold strategy in order to provide a reply to the referring court and bring causation issues within the scope of the institutional competence of the EU court to provide an interpretation of EU law. Although the CJEU did not follow that strategy in its final judgment, and 193

Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 1. Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 2. Emphasis added. 195 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 19. 196 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 14. 194

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(t)here is said to be umbrella pricing when undertakings that are not themselves party to a cartel, benefiting from the protection of the cartel’s practices (operating ‘under the cartel’s umbrella’, so to speak), knowingly or unknowingly set their own prices higher than they would otherwise have been able to under competitive conditions. Does European Union law require that customers of undertakings not party to the cartel should be able to claim compensation for the inflated prices charged by those undertakings from the members of the cartel before the national courts? [the standing issue] Or, conversely, may such an obligation to award compensation be excluded in national civil law on the ground that the loss suffered is indirect and too remote? [the causation issue].194

50

Lianos

197 198 199

Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 28. See, for instance, Courage/Crehan, Manfredi, above. Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 31.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

preferred a more ‘traditional’ approach, relying on the principles of effectiveness and equivalence, Advocate General Kokott’s Opinion provides an interesting blueprint on what an EU rule on causation would look like, if the CJEU had decided to proceed to some harmonized EU rule on the existence of a causal link. The first part consisted in linking the issue of causation, which the CJEU in Manfredi addressed as an issue to be dealt with by the domestic legal system of each Member State, within the ambit of EU law. To do so, AG Kokott introduced a distinction between the existence of the constitutive rules of the right to claim damages for competition law infringement (‘the question of whether compensation is to be granted’), which is an issue of EU law, and ‘the details of application of such claims and the rules for their actual enforcement’ (‘the question of how compensation is to be granted’), which are left to domestic legal systems. The AG then mentioned the ‘direct anchoring’ of the principle that any individual is entitled to claim compensation for loss sustained in the presence of a causal relationship between a loss and an infringement of EU law. She based her interpretation on a variety of grounds: the need for consistency within the EU system of the liability of Member States for their infringements of EU law, the direct effect of Articles 101 and 102 TFEU, and the principle of the ‘effet utile’ of EU law. Having established that constitutive principles emanate from and are determined by EU law, the AG classified the question of the recoverability of losses engendered by umbrella pricing as relating to the ‘fundamental question of whether (cartel members) can be sued by persons who are not their direct or indirect customers’,197 thus formulating the issue as being conceptually close to the questions of standing and the types of losses that are recoverable under EU law, which have already been determined by EU law.198 Consequently, the AG felt that EU law also regulated the issue of the establishment of a causal link. Surprisingly, she did not make any reference to the principle of the supremacy of EU law, which should have entered into play in the presence of a conflicting national rule, had this issue, as she advocated, been one regulated by EU law. The second part of the AG’s Opinion considered ‘the specific conditions’ that ‘may be attached under European law to the establishment of a causal link between a cartel and umbrella pricing’.199 One may consider that this issue relates more to the question of ‘how compensation is to be granted’, the modalities of compensation, rather than the constitutive rules of ‘whether compensation should be granted’, yet the a priori practical exclusion of claims introduced by umbrella customers under Austrian law enabled the AG to conflate the two issues in order to provide a useful answer to the national court. The

Causal Uncertainty and Competition Law Claims for Damages

51

200

Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), paras 32–35. Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 36. Emphasis in the original. 202 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 37. 203 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 38. 204 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 39. 201

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

AG proceeded to elaborate on some broad principles of EU causation law, such as a reference to the conditio sine qua non or ‘but for causal link’, the requirement of a ‘sufficiently direct causal nexus’ in the implementation of Article 340 TFEU on the liability of Member States for infringement of EU law or some EU merger cases, and the importance of a ‘normative examination’ in establishing legal causation leading to the identification in all European legal systems, despite the different terminology employed (‘legal causation, adaquat Kausalita¨t and the like’), of normative principles that ‘inform the concept of a sufficiently direct causal link’.200 By separating the causation in fact from the causation in law or ‘scope of the liability’ element, AG Kokott seems to have aimed to bring the latter within the scope of the constitutive rules of the right to claim damages, and thus under the ambit of EU law, while leaving causation in fact to be determined according to the techniques developed in each domestic legal system. Of course, as we have previously shown, the distinction is somehow artificial as ultimately the two issues are profoundly interlinked. Yet, this approach enabled AG Kokott to comment freely on the way the requirement of causal link should be interpreted as a matter of EU law. She stressed that the requirement of ‘a direct causal link must not be regarded as being the same as a single causal link’ and that ‘there is sufficient support for the assumption of a direct causal link if the cartel was at least a contributory cause of the umbrella pricing’.201 The aim was of course to deal with situations of multiple causes, which are frequent, if not always the case, in the implementation of EU competition law, because of the intervention of the conduct of a third party in the realization of the harm. This is intrinsically linked to the normal operation of the market system and the fact that undertakings are ‘(also) guided by the relevant trading conditions’, which may occasionally break the causal chain.202 According to AG Kokott, such issues are also raised with regard to indirect purchasers who are not denied under EU law the possibility of claiming damages even if the incurrence of their loss ‘is ultimately dictated by the freelymade corporate decision of a third party (the intermediary trader)’.203 In any case, ‘(p)rices rarely have only one cause’, hence ‘it would be unreasonable to make the civil liability of cartel members subject to the condition of their being the single cause’.204 That led the AG to espouse a relatively broad concept of cause, by accepting that a simple contribution of the conduct ‘towards a distortion of the price formation mechanisms that normally apply on the market in question’ may constitute relevant cause. With that potentially very broad

52

Lianos

205

Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para.40. Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para 42. Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), paras 45–52. 208 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 30, 2014), paras 54–70. 209 This is the interpretation that one may have of the following excerpt of AG Kokott’s Opinion, above, para. 66: ‘the objectives of European competition law cannot be reduced to allowing undertakings active on the internal market to operate as cost-efficiently as possible. In a European Union based on the rule of law which has set itself the objective of achieving a highly competitive social market economy (Article 3(3) TEU), functioning markets characterised by undistorted competition are in themselves an asset beyond all cost–benefit consideration’, reflecting her long held view that Articles 101 and 102 do not protect specific interests, consumers or producers, but aim to protect ‘the structure of the market and competition as such’, providing a moral/deontological (as opposed to consequentialist) argument for competition law enforcement. 206 207

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

liability principle announced, the AG then addressed the issue of the necessary limitation of its scope. This constitutes the third step in the strategy of AG Kokott to develop broader EU causation rules. The AG clearly indicated that ‘the criterion of a sufficiently direct causal link is in substance intended, on the one hand, to ensure that a person who has acted unlawfully is liable only for such loss as he could reasonably have foreseen’ and ‘(o)n the other hand, a person is liable only for loss the compensation of which is consistent with the objectives of the provision of law which he has infringed’.205 With regard to the first issue, the AG distinguished the situation of ‘umbrella pricing’ which is foreseeable by cartel members from ‘loss which results from an entirely extraordinary train of events, and therefore, ensues via an atypical causal chain’.206 Her main argument was that essentially that was ‘very much in the normal way of things’ and that cartel members should have expected that non cartel members would have raised their prices as a result of the cartel, in comparison to the prices in the absence of the cartel, such behaviour being ‘far from unusual’ and ‘economically rational’, and thus foreseeable by the members of the cartel but also in the interest of the cartel members, in order to ensure the success of their cartel.207 But more important for the reasoning of the AG was to link such an interpretation of the causation in law requirement with the protective scope of Articles 101 and 102 TFEU, transposing in EU law the ‘scope of the rule’ doctrine of German tort law.208 Summarizing a highly interesting discussion, AG Kokott referred to the compensatory aim of EU competition law enforcement and to deterrence, perceived not as an efficiency objective, but an expression of corrective justice.209 Such interpretation of the scope of protection of EU competition law opens the gate to enforcement by anyone incurring some form of harm (economic and moral?) to which a competition law infringement may have contributed. This is not a narrow definition of the scope of protection, and AG Kokott understands that well, as she mentions that no overloading of national courts with claims from injured parties by umbrella pricing is to be expected, in view of ‘the relatively high hurdles in terms of burden of

Causal Uncertainty and Competition Law Claims for Damages

53

210 Burden of proof or standard of proof? Is the AG making a comparison with the situation of indirect purchasers, for whom the Damages Directive recognizes the benefit of a causal presumption with regard to passing on, so that in this case it would be a burden of proof issue? 211 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 69. 212 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), paras 74–75. 213 Case C-557/12, KONE AG and Others, ECLI:EU:C:2014:1317 para. 32. 214 Case C-557/12, KONE AG and Others, ECLI:EU:C:2014:1317 para. 32. 215 Case C-557/12, KONE AG and Others, ECLI:EU:C:2014:1317 para. 33. 216 Case C-557/12, KONE AG and Others, ECLI:EU:C:2014:1317 para. 33. Emphasis added.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

proof’210 that await such claimants in the civil courts, suggesting that ‘any potential “umbrella plaintiff” “weigh up carefully the pros and cons of taking out a civil action against cartel members”’.211 Finally, no evidence of fault or intent is required, the existence of a causal link being based on ‘purely objective criteria’, hence, mere negligence is sufficient.212 The only limit to the extent of the losses that umbrella customers may request compensation from cartel members is the principle of ‘overcompensation’, which is defined by AG Kokott as being conceptually linked to the issue of causation, as it ‘requires cartel members only to make good the loss which they have caused (or to which they have contributed) on the market in question by their anti-competitive practices’, which is actually somewhat tautological in view of the weak standard of causation required by the AG (a simple ‘contribution’). The CJEU did not follow the full strategy proposed by AG Kokott and reaffirmed the preference for national law to deal with causation issues, under the double discipline of the principles of effectiveness and equivalence. The Court confirmed once again that ‘it is, in principle, for the domestic legal system of each Member State to lay down the detailed rules governing the application of the concept of the “causal link”’, and that ‘national legislation must ensure that European Union competition law is fully effective’.213 Effectiveness in this context was interpreted as requiring national rules on causation to ‘specifically take into account the objective pursued by Article 101 TFEU, which aims to guarantee effective and undistorted competition in the internal market, and, accordingly, prices set on the basis of free competition’ and the need to ‘recognise the right of any individual to claim compensation for loss sustained’.214 Effectiveness would have been ‘put at risk if the right of any individual to claim compensation for harm suffered were subjected by national law, categorically and regardless of the particular circumstances of the case, to the existence of a direct causal link while excluding that right because the individual concerned had no contractual links with a member of the cartel, but with an undertaking not party thereto’.215 According to the Court, this causal link cannot be excluded in principle as the pricing policy of the undertaking inflicting damage to umbrella customers, ‘is a result of the cartel that contributed to the distortion of price formation mechanisms governing competitive markets’.216 By emphasizing ‘contribution’ as a sufficient element for establishing a causal link in these circumstances, the CJEU seems to frame the minimum core

54

Lianos

required in the definition of the causal link for that to ensure the effectiveness of EU competition law prohibitions. The exclusion of a category of victims, umbrella customers, is not compatible with this minimum EU law defined core.

217 See, for instance, Frank Maier Rigaud, Umbrella effects and the ubiquity of damage resulting from competition violations, Journal of European Competition Law & Practice, (2014) 5: 247–251. 218 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 84. 219 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 85. 220 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 86. 221 Opinion of AG Kokott, Case C-557/12, KONE AG and Others (30 January 2014), para. 87.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(c) Towards a broad and instrumental conception of causation in EU competition law on damages actions As has rightly been observed by some commentators,217 the opinion of AG Kokott relied greatly on the evidence put forward by the parties, being relatively agnostic as to the ‘merits’ of the case beforehand, based on some categorical analysis of the ‘type’ of claim or loss brought. As the AG herself puts it, ‘it will always be necessary to carry out a comprehensive assessment of all the relevant circumstances in order to determine whether the cartel in the case in question has given rise to umbrella pricing’.218 Furthermore, ‘(s)hifting the umbrella pricing issue from the level of pure theory to that of the production of evidence seems to me to be the best way of contributing to the effective enforcement of the European competition rules, while taking due account of the interests of all economic operators’.219 Indeed, harm resulting from umbrella pricing is neither ‘speculative’, nor ‘uncertain’.220 Looking at the facts, a ‘categorical exclusion’ of such harm (through standing or causation rules), would not make much sense.221 The CJEU also agreed that ‘a loss being suffered by the customer of an undertaking not party to a cartel, but benefiting from the economic conditions of umbrella pricing, because of an offer price higher than it would have been but for the existence of that cartel is one of the possible effects of the cartel, that the members thereof cannot disregard’, although it did not get into a detailed examination of the topic, which in any case is not expected by the CJEU in a preliminary ruling procedure. The possibility that a cartel contributed to high prices for umbrella customers ‘cannot be ruled out’, which implies that a categorical rule excluding umbrella customers would affect the legitimate right of certain victims of competition law infringements to receive compensation. The third party intervention in this context does not break the causal link between the cartel and the damage. However, the CJEU did not provide guidance as to the causal test that would apply in order to assess the individual circumstances of the case and conclude, or not, on the existence of a causal link. The only guidance provided consists in affirming that ‘contribution’ may be a sufficient factor in establishing causation.

Causal Uncertainty and Competition Law Claims for Damages

55

222

Wright (n 37), 1041. Ibid. Ibid. 225 Ibid. 226 Jane Stapleton, ‘Choosing what we mean by “Causation” in the Law’ (2008) 73 Missouri Law Review 433–480, 434. 223 224

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

The introduction by AG Kokott in her Opinion in KONE AG of the ‘scope of the rule’ doctrine of German law as a principle of EU law in the assessment of causation could have had important implications as to the development of the EU law on damages claims for competition law infringements. First, her conception of the causal link as ‘contribution’, which was accepted by the CJEU, may lead to a broad interpretation of the causation in law and causation in fact requirements. Secondly, her broad interpretation of the scope of the protection of competition law provides the basis for a development by the case law of the EU Courts of flexible causation rules for damages actions brought for EU competition law infringements that will take into account the specificities of antitrust damages claims, in comparison with torts for other kinds of injuries. This may indeed provide for flexible rules in order to assess the merely economic damage suffered and the use of economic and econometric evidence. This part of her strategy was not followed by the CJEU. It is important here to note the close-ended character of the counterfactual inquiry in law. The objective is not to speculate on any imaginable turn of events and hypothetical scenarios or the most likely alternative cause of action for the actors involved, which would be ‘hopelessly indeterminate’ and ‘inevitably influenced by policy-based conventions, expectations, or goals’, but, as is explained by Wright, instead to determine ‘the causal processes at work, in the one scenario that did happen’.222 Hence, we usually proceed by eliminating the condition being tested from the sufficient set of actual antecedent conditions in order to determine ‘by matching the remaining conditions in the set against the applicable causal generalization’, ‘whether the set still would be sufficient for the occurrence of the result’.223 Resolving this causal inquiry depends, according to Wright, on ‘our empirical knowledge of the conditions that existed and the possibly applicable causal generalizations, not on policy considerations.’224 Hence, his conclusion that any indeterminacy results not from the openended character of the counterfactual test but from ‘defects in our empirical knowledge’.225 At the other end of the spectrum, Stapleton advances the concept of ‘involvement’, instead of the narrower one of ‘necessity’, the information conveyed by the notion of ‘involvement’ being ‘context-specific’ and depending on the specification of the particular factor of interest to the legal system and consequently of the underlying interrogation that the concept of causation aims to unravel.226 According to Stapleton, ‘lawyers should choose an interrogation underlying causal usage in the law that captures all ways in which the (specified) factor

Lianos

56

227 228 229 230

Ibid, 441. Ibid, 443. Ibid, 448–9. Ibid, 445.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

might be involved in the existence of the particular phenomenon in issue’.227 This may extend beyond necessity and sufficiency and may identify simple ‘contribution’ as a cause, ‘even where the factor is neither necessary nor sufficient for the existence of the particular phenomenon.’228 Stapleton refutes the argument that the broader concept of ‘involvement’ could lead to ‘overinclusiveness (that is too many causes)’, because of the ‘doctrinal filters’ offered by the concept of causation in law and the requirements of proximate cause,229 or one could also add the ‘scope of the rule’, as it is adopted in some continental European systems and suggested by AG Kokott as an EU law rule in KONE AG, although this was not accepted by the CJEU. From this perspective, causation in fact should provide the requisite ‘width of coverage’ that is needed to accommodate the regulatory strategies followed by the legal system in this context. If the aim of law’s intervention is to prevent harm, such a project requires the legal system ‘to address all involved factors— even those that some might describe as mere conditions’.230 Were the law to select a narrower interrogation underlying the causal language, this might have jeopardized the specific regulatory project. The ‘scope of the rule’ doctrine would be a particularly salient factor in determining the ‘width of coverage’ of the causation enquiry in law. The existence of ‘filtering devices’, such as proximate cause or the ‘scope of the rule’ doctrine, enabling ‘the specification of a small finite number of factors whose possible involvement in the existence of a particular phenomenon in the actual world is being examined’, distinguishes the causation inquiry in law from that undertaken in the social sciences and/or metaphysics. By choosing to rely on a broad concept of the causal link, defined as a ‘contribution’, the CJEU is able to assist the claimant(s) in putting forward a causation narrative. This broad conception may potentially also favour defendants, when these argue a pass on defence. While not being open-ended, this concept of causation remains extremely broad and offers considerable discretion to national judges to frame it according to the circumstances of the case and the requirements of their legal system. Potentially, the ‘closure’ of the system through the adoption of a ‘scope of the rule’ doctrine (as suggested by AG Kokott) or other alternatives is something that may be envisaged in the future, should a broad concept of causation lead to a flood of competition damages cases and difficulties as to the definition of the conceptual contours of the causal link. The experience gained in the meantime would be valuable, in particular in view of the complex economic evidence usually relied upon by litigants in this area of law.

Causal Uncertainty and Competition Law Claims for Damages

57

Although the Damages Directive and the EU courts have been reticent to engage in an extensive harmonization of the causation requirement, a more extensive harmonization approach was nevertheless followed with regard to situations of causal uncertainty that are generated by the existence of multiple tortfeasors. In this case, the Damages Directive provides for more harmonized rules.

231 Art. 11(2), introduced when the Commission’s proposal was examined by the European Parliament, provides an exception to this rule for infringers that are small or medium sized enterprises, who shall be liable only to their own direct and indirect purchasers [note there is no reference to providers as in Article 11(3)] under certain conditions, in particular if their ‘market share in the relevant market was below 5% at any time during the infringement’ or if the application of the general rule on joint and several liability would ‘irretrievably jeopardize its economic viability and cause its assets to lose all their value’. According to Article 11(2), the small or medium sized enterprise will not however benefit from such exception if it led the infringement or coerced other undertakings to participate in the infringement or if it was a recidivist. 232 Case C-234/89, Stergios Delimitis v Henninger Bra ¨ u AG. [1991] ECR I-935: a combination of tying/exclusive purchasing agreements might infringe Article 101 TFEU if it would be difficult for competing suppliers to enter the market because of the cumulative effects of the networks of agreements by different suppliers in the market (the foreclosure effect of all the networks of agreements) and the supplier network to which the agreement under consideration formed part had a significant contribution to the foreclosure effect (contribution of the specific network of agreements to the foreclosure effect).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

(d) Beyond causation: The instrument of joint (solidary) and several liability Article 11 of the Damages Directive establishes the rule of joint and several liability of multiple tortfeasors for competition law damages, representing the choice of the EU legislator for a regime of several and joint liability instead of choosing to channel liability to one tortfeasor for the whole damage and then resolve the attribution issue by resorting to some sort of causal proportional liability rule. According to Article 11(1), ‘Member States shall ensure that undertakings which have infringed competition law through joint behaviour are jointly and severally liable for the harm caused by the infringement of competition law: each of those undertakings is bound to compensate for the harm in full, and the injured party has the right to require full compensation from any of them until he has been fully compensated’.231 Hence, in a fact pattern where the damage was caused by the cumulative foreclosure effect created by the parallel networks of agreements with similar clauses, the so called Delimitis cumulative effects doctrine, all the undertakings whose network of contracts participated to this cumulative effect will be found liable, under the joint and several liability doctrine.232 Hence, each of the undertakings is bound to compensate for the harm in full, and the injured party may require full compensation from any of them. Article 11(4) requires Member States to ensure that an infringing undertaking may recover a contribution from any other infringing undertaking, the amount of which shall be determined in the light of their relative responsibility for the harm caused by the infringement of competition law. This right of contribution

58

Lianos

233 Case C-106/89, Marleasing SA v La Comercial Internacional de Alimentacion SA [1990] ECR I4135. 234 A different approach has prevailed in the USA, where the Supreme Court in its judgment in Texas Industries v Radcliff Materials, 451 U.S. 630 (1981) affirmed the common law rule against contribution among competition law defendants and found that there was no indication that Congress considered contribution among joint violators under the US antitrust laws, leaving congressional action as the only avenue for those seeking to allow contribution. A tortfeasor who paid damages to a plaintiff is not entitled to a contribution from other competition law tortfeasors, since each tortfeasor remains responsible for the entire loss and actions for contribution among co-defendants are barred in antitrust cases. Furthermore, where there have been settlements between the parties injured and some of the tortfeasors, only the actual amount of the settlements is deducted from the treble damages awards against the tortfeasors who have not settled who are also responsible for the part of the damage inflicted by the settling defendant which was not covered by the settlement. The no-contribution rule was originally adopted by the English courts for joint tortfeasors in Merryweather v Nixan (1799) 101 Eng. Rep. 1337 (K.B.), on the basis of the principle of ex turpi causa non oritur actio: out of a base (illegal or immoral) consideration, no action arises. At that time the term ‘joint tortfeasors’ referred only to those defendants that have acted intentionally and in concert. It was later extended to all joint tortfeasors, even those that concurrently caused indivisible harm by negligence, although in this case the extent of the rule was limited by the courts. For a discussion, see J S Dillbary, ‘Apportioning Liability Behind a Veil of Uncertainty’ (2011) 62 Hastings Law Journal, 1729–92. In view of the harshness of the rule, the UK Parliament enacted the Law Reform (Married Women and Tortfeasors) Act in 1935, providing for a contribution rule among tortfeasors, followed by the Law Reform (Contributory Negligence) Act, 1945. Similarly in the USA, many States have fashioned rules of contribution in one form or another in the first decades of the twentiethcentury. Yet, in Texas Industries v Radcliff Materials, the Supreme Court refused to follow that trend and adopt contribution for antitrust violators. For an analysis of the respective merits of the no contribution or contribution rules in the US context, see generally W M Landes and R Posner, ‘Joint and multiple tortfeasors: An economic analysis’ (1980) 9 Journal of Legal Studies, 517–55 (noting that from an efficiency perspective, it is impossible to pronounce that no

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

extends only to ‘infringing undertakings’ and thus does not seem to cover the situation of a competitor not participating in the cartel, and thus not committing a competition law infringement, who may have benefited from the higher cartel prices by charging higher prices to his consumers (umbrella customers). Hence, it is possible that, following KONE, umbrella customers may claim full compensation for the harm caused by the infringing undertakings (eg members of the cartel or found liable for having abused their collective dominant position), the latter not being able to recover a contribution, on the basis of Article 11(4) of the Directive, from a non-cartelist competitor benefiting from the umbrella price. This windfall profit (from the competition law infringement) may, however, be recovered on the basis of national rules on restitution (eg unjust enrichment), the Directive being silent on this issue. One may still claim that awarding recovery in this context will limit deterrence and might be indirectly incompatible, if not with the letter, at least with the spirit of the Directive, assuming that national courts have an obligation to interpret national law in conformity with an EU directive, even more so if the Damages Directive is interpreted as also covering actions for damages based on restitution.233 Deterrence would favour a no contribution rule in this context, as it would also have been the case in the context of an action from an infringing undertaking against another infringing undertaking.234 Yet, as it is clear from the

Causal Uncertainty and Competition Law Claims for Damages

59

recovery rule included in Article 11(4), fairness concerns, rather than deterrence, animate the EU rules applying in this context.235

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

contribution is a more efficient rule than contribution, as the result ‘depends on the trade-off between the administrative cost savings under the former and the informational and insurance benefits under the latter’, although they also expressed the feeling that ‘a broad contribution rule is almost certainly less efficient than no contribution’); F Easterbrook, W Landes, and R Posner, ‘Contribution among antitrust defendants: A legal and economic analysis’ (1980) 23 Journal of Law and Economics, 331–70 (finding that even if the no contribution rule leads to more deterrence, its efficiency ‘from an overall social standpoint’ remains a difficult question as one should consider issues of risk aversion and legal error); M Polinsky and S Shavell, ‘Contribution and claim reduction among antitrust defendants: An economic analysis’ (1981) 33 Stanford Law Review, 447–71 (noting, among others, that a no contribution rule might be less efficient as it achieves deterrence by imposing greater risks on innocent parties and allocates liability imprecisely, thus discouraging some firms from engaging in socially beneficial activities); L A.Kornhauser and R L Revesz, ‘Joint and several liability’, in M. Faure (ed.), Tort Law and Economics (Cheltenham: Edward Elgar, 2009) 109–34 (remarking that when it comes to fairness, joint and several liability with no contribution may place a disproportionate burden on the defendant with the smaller share of the liability); American Bar Association (ABA), Antitrust Section, Contribution and Claim Reduction in Antitrust Litigation (Monograph 11) (ABA: 1986) (highlighting, among other issues, the possible over-deterrence or under-deterrence effects that a contribution or a no contribution rule will have, depending on the nature of the antitrust violation and assumptions on the risk averse or risk seeking character of competition law infringers). In response to these harsh rules, firms engaged in or accused of price fixing often enter into judgment-sharing agreements, in which the firms agree to allocate financial responsibility for any private liability among themselves: for a discussion, see C R Leslie, ‘Judgment-sharing agreements’ (2009) 58(5) Duke Law Journal, 747–825. Recently the US Antitrust Modernization Commission discussed a recommendation to US Congress to enact legislation allowing non-settling defendants to seek contribution from other non-settling defendants and that would also permit nonsettling defendants to obtain a reduction of the plaintiffs’ claims by the amount of the settlement(s) or the allocated share(s) of liability of the settling defendant(s), whichever amount is greater: See, Antitrust Modernization Commission AMC), Report & Recommendation (April 2007), 18 (proposal 46), 243–4, 251–5. The AMC indeed noted (at 244) that ‘(t)he existing rules of joint and several liability without a right of contribution and only limited claim reduction have given rise to substantial criticism regarding fairness. These rules permit plaintiffs to settle with some defendants at an early stage for a relatively small amount of damages, leaving remaining, non-settling defendants potentially liable for nearly the entire damages caused by the joint conduct, trebled. As a result, these rules can cause a ‘race’ to settle, potentially leaving defendants that had a small or no role in the overall anticompetitive scheme with disproportionately large potential liability’. Yet, no legislative initiative has so far been enacted by US Congress on this issue, thus indicating a divergence between the deterrence-oriented US no contribution rule and the fairness-oriented European contribution rule. For a discussion with regard to the European context, see M Hviid and A Medvedev, The role of contribution among defendants in private antitrust litigation, CCP Working Paper 08–3, available at (noting that although the no contribution rule may lead to higher levels of aggregate damages and more information available to the private plaintiff, it also has the potential to ‘neuter any public leniency programme, thereby possibly reducing the number of cartels detected’). 235 There are various options on offer: apart from the basic choice of a no contribution rule or a contribution rule; if the contribution rule option is selected, there are various options regarding the dividing and apportioning of damages among a cluster of antitrust co-conspirators. According to the Supreme Court in Texas Industries v Radcliff Materials, 451 U.S. 630 (1981), ‘various formulae are suggested: damages may be allocated according to market shares, relative profits, sales to the particular plaintiff, the role in the organization and operation of the conspiracy, or simply pro rata, assessing an equal amount against each participant on the theory that each one is equally liable for the injury caused by collective action’.

60

Lianos

IV. Conclusion The emergence of actions for damages as one of the main forms of competition law enforcement in Europe is a recent phenomenon. The issue of the standing of indirect purchasers has been a primary focus for legal commentators, as is attested by the number of studies published on this topic. The most recent book-length studies published on competition law damages typically dedicate 236 Section 1(1) of the UK Civil Liability (Contribution) Act 1978. According to Section 2(2) of the Act, ‘the amount of the contribution recoverable from any person shall be such as may be found by the court to be just and equitable having regard to the extent of that person’s responsibility for the damage in question’. It is also mentioned in Section 1(3) that ‘the person from whom the contribution is sought shall not by virtue of any contribution awarded . . . be required to pay in respect of the damage a greater amount than the amount of those damages as so limited or reduced’, for instance by ‘any limit imposed by or under any enactment or by any agreement made before the damage occurred’ or by ‘any corresponding limit or reduction under the law of a country outside England and Wales’. For a discussion of the English law of contribution see, C Mitchel, The Law of Contribution and Reimbursement (Oxford: Oxford University Press, 2003). 237 Section 1(4) of the UK Civil Liability (Contribution) Act 1978. 238 As it is remarked in the notes to Art. III.-4:106, ‘(t)he sharing of liability for damages between those responsible for the same damage is done in different ways in different legal systems. Sometimes it is done according to the degree of seriousness of the wrongdoing of those involved [see for instance, in French, Greek, Danish, Polish jurisprudence]. . . . Elsewhere causal participation is what matters [see for instance in German, Austrian, Italian, Spanish jurisprudence]’. Finally, under English law (Civil Liability (Contribution) Act 1978) the matter is left to the discretion of the judge.

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

The contribution rule indeed constitutes a feature of tort law regimes in Europe. In the UK, the Civil Liability (Contribution) Act 1978 provides that ‘(a)ny person liable in respect of any damage suffered by another person may recover contribution from any other person liable in respect of the same damage (whether jointly with him or otherwise)’.236 The above provision applies not only where there is a judgment against the defendant, but also where the defendant has settled the claim.237 At the European level, the Draft Common Frame of Reference employs the terminology ‘solidary liability’ (instead of joint and several liability) and provides for a default rule of equal sharing for solidary debtors [Art. III.-4:106(1)], ‘unless different shares of liability are more appropriate having regard to all the circumstances of the case and in particular to fault or to the extent to which a source of danger for which one of them was responsible contributed to the occurrence or extent of the damage’ [Art. III.-4:106(2)], thus stipulating a special rule for cases of solidary liability resulting from causing the same damage.238 It is also provided that ‘(a) solidary debtor who has performed more than that debtor’s share has a right to recover the excess from any of the other debtors to the extent of each debtor’s unperformed share, together with a share of any costs reasonably incurred [Art. III.-4:107]. Although these rules are not binding, they provide a reference point for the multiple national rules on this issue.

Causal Uncertainty and Competition Law Claims for Damages

61

239 See, for instance, D Mc Fadden, The Private Enforcement of Competition Law in Ireland (Oxford: Hart Publishing, 2013); D Ashton and D Henry, Competition Damages Actions in the EU—Law and Practice (Cheltenham: Edward Elgar, 2013). 240 Illinois Brick Co. v Illinois, 431 U.S. 720 (1977). Indirect purchasers are not however barred from seeking injunctive relief under Clayton Act § 16, provided they meet the other requirements set in this context: Cargill, Inc. v Monfort of Colorade Inc., 479 U.S. 104 (1986). Some States have nevertheless provided standing to indirect purchasers by passing ‘Illinois Brick repealer’ statutes, a practice accepted by the Supreme Court. These statutes cover approximately 70% of the US population: E L Cramer and D C Simons, ‘Parties entitled to pursue a claim’, in A A Foer and R M Stutz (eds), Private Enforcement of Antitrust Law in the United States (Cheltenham: Edward Elgar, 2012) 64–94, 80–3. Passage of the Class Action Fairness Act (CAFA) of 2005 may direct some of this litigation, when it concerns small claims multi-state, to the federal courts, if it was introduced after the implementation of CAFA. CAFA proceeds to an amendment of 28 USC §1332(d) and establishes federal jurisdiction for class actions in which there is minimal diversity between the parties and the matter in controversy exceeds $5 million. 241 C Hodges, S Vogenauer, and M Tulibacka (eds), The Costs and Funding of Civil Litigation: A Comparative Perspective (Oxford: Hart Publishing. 2010) 3–186. 242 Report for the European Commission, Making antitrust damages actions more effective in the EU: welfare impact and potential scenarios (21 December 2007) available at , pp. 562–6 (scenario 1: double damages for all competition cases) and pp. 579–94 (Scenario 2: double damages only for cartel cases).

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

a chapter to standing issues.239 This choice may have been influenced by the debate in the United States, which has focused predominately on this issue, since the Supreme Court barred indirect purchasers from pursuing treble damage claims under the Clayton Act §4 in Illinois Brick, certain limited exceptions set aside.240 The US solution and the debate over indirect purchasers may be explained by the specific function played by private enforcement in the USA, its overall architecture aiming to ‘provide a mechanism for wider public regulatory and observance goals’, in particular deterrence, thus going beyond compensation for loss, a feature that also explains the various specificities of the US system of private enforcement (eg cost rules), in comparison to the role of private enforcement in other parts of the world.241 In contrast, in the European civil justice systems, the role of private enforcement has principally, but not exclusively, been that of providing compensation and operating as an instrument of corrective justice. Hence, issues of standing, as with any other dimension of a tort law system cannot be determined on the sole basis of deterrence considerations. The competition law field is no exception and the EU Courts, followed by the European Commission, have underlined the goal of compensation as the principal, but again non-exclusive, aim of private enforcement in EU competition law. The choice for a system of single, and not multiple, damages attest to that normative choice, even if there was evidence that double damages might have been a more efficient option, from a deterrence perspective but also from a mixed deterrence–corrective justice perspective (eg double damages only for cartel cases).242 In a regime based on corrective justice considerations, causation requirements will tend to play a predominant role in regulating the damages claims brought

62

Lianos

Downloaded from http://yel.oxfordjournals.org/ by guest on November 24, 2015

forward. By relying on the tort law systems of the various Member States and avoiding any harmonization of these rules at the EU level, the EU legislator and the Court of Justice chose to promote inter-jurisdictional competition, under the framework of the course of the principles of equivalence and effectiveness, in particular the second one, which is interpreted as aiming to facilitate damages actions. Some limited harmonization of evidential presumptions and procedural requirements, as well as the exclusion of national rules that may deny the right of the parties harmed by the competition law infringement to receive compensation also ensures the effectiveness of the remedy. Yet, the contours of the requirement of causal link are left to the interpretative work of the national courts, in view of their respective tort law doctrines on causation and the lack of a proper EU tort law. The study examined the different solutions that have emerged in these various national tort law systems in order to cope with situations of causal uncertainty that inevitably arise in competition law cases, in view of the complexity of the commercial environment and the multiple factors influencing markets. It transpired from the discussion that it is only in circumstances in which national tort law systems may not effectively guarantee the right to compensation for the victims of competition law infringements that EU law intervenes, for instance dealing with damages caused by multiple tortfeasors through provisions on joint (solidary) and several liability or by instituting causal presumptions. The promotion of the inter-jurisdictional competition that lies behind the decision not to harmonize the causation requirements should therefore be understood as aiming to guarantee more effective enforcement of the rights of the victims of competition law infringements, thus leading to a need to adopt an instrumental approach when implementing causation requirements in competition law related damages cases.