نقش مشارکت در بین متهمان در شکایت های قانونی ضد تراست خصوصی

The incidence of private actions for damages in antitrust cases has varied markedly across jurisdictions. The procedural rules surrounding private litigation may account for some of these differences. This paper explores the effect of rules concerning contribution among multiple defendants who are jointly and severally liable for a cartel infringement. The no-contribution rule is shown to lead to higher levels of aggregate damages and more information revelation to the private plaintiff. However the no-contribution rule also has the potential to neutralise any public leniency programme, thereby possibly reducing the number of cartels detected.

The last few years have seen a strong push by the European Commission for an increased level of private enforcement of competition law.1 This has generated renewed interest in understanding how private enforcement works. With hard core cartels singled out by the majority of enforcement agencies as the “worst” form of infringement, it is of particular interest to understand the potential role of private enforcement in these cases. Anti-competitive agreements to collude on prices or quantities involve multiple defendants, all of whom have engaged in a violation of competition law. To ease the burden on the private plaintiffs, such co-defendants have typically been made jointly and severally liable, ensuring that the plaintiff needs only sue one of the defendants. However joint and several liability raises the question of whether the defendant, who has been ordered to pay the damages caused collectively by all the defendants, can ask for contributions from the co-defendants. The answer to this question appears to differ between the US and the majority of EU Member States. In the former jurisdiction, the rule is very explicitly one of no contribution2 so that a defendant who is taken to court faces the whole liability for the losses incurred by the plaintiff, minus any amounts already settled, and cannot recover any of this from co-defendants. In the latter jurisdictions there is scope for at least some level of contribution3 so that the defendant who is found liable in court can, if it is faced with a disproportionate liability, demand contributions from co-defendants. The aim of this paper is to provide a better understanding of the impact of contribution rules when defendants are jointly and severally liable by demonstrating how a “no-contribution” rule has the effect of introducing a type of leniency programme for co-defendants in the area of private enforcement, leading to more information being revealed. This information revelation effect is in addition to the already identified ability of the no-contribution rule to increase the expected total settlement. This provides two avenues through which the no-contribution rule delivers greater deterrence of cartel behaviour, better information about the violation and increased liability. Most jurisdictions with a competition law also have a public leniency programme aimed at undermining cartels. Although the details of how these programmes work differ across jurisdictions, they all involve a reduction in the penalty imposed by the competition authority on a firm in return for information and in some cases an admission of guilt. These public leniency programmes have become a corner-stone in the fight against hard core cartels. With two parallel leniency programmes, one explicitly designed for the purpose by the competition authority and another arising as a side-effect of a rule about liability, comes the question of how they may interact with each other. In particular we consider whether there are any negative effects on the public leniency programmes arising from this aspect of private enforcement. Such concern about private enforcement undermining the public leniency programme are not new. The US Antitrust Criminal Penalty Enhancement and Reform Act of 2004 removed joint and several liability from a firm granted immunity under the leniency programme and in addition reduced the private liability of that firm to single damages. The literature on contribution in antitrust is relatively sparse, with early contributions by Easterbrook, Landes, and Posner (1980) and Polinsky and Shavell (1981) who offer a theoretical analysis of the merits of contribution and claims reduction.4 Both papers put considerable emphasis on risk-aversion, something which is not considered in our paper. Their models have subsequently been generalised. Cavanagh (1987) cites the corrugated container litigation in the US as an example of the ability of the no-contribution rule to increase the recovery by the plaintiff.5 He argues that the no-contribution rule is unfair on firms with only a marginal participation in a price-fixing conspiracy: “these companies, precisely because they are small players with marginal culpability, are rarely if ever offered settlement terms comparable to those provided to the ringleaders” Cavanagh (1987, p. 1920). Cirace (1980) focuses on how contribution rules affect the incentives of defendants to settle early or late, showing how no-contribution leads to the former and contribution to the latter. Stanley (1994) assumes asymmetric information with the defendants having better information than the plaintiff. He assumes that the settlement bargaining between the plaintiff and all of the defendants occurs simultaneously and that any private information revealed in a settlement increases the probability that the plaintiff prevails in court. Stanley (1994) is the only paper which clearly identifies the scope of early settlements to lead to information revelation, though the focus of the analysis is on the effect of the different contribution rules on deterrence and the likelihood of settlements rather than information revelation. However, partly due to the assumptions made, the full effect of information revelation is not explored. Finally, the recent paper by Goetz, Higgins, and McChesney (2006) summarises and builds on most of the existing literature in the US regarding contribution. Unlike Cirace (1980) and Stanley (1994), they assume sequential settlement. However, as they do not ensure that the equilibria they find are subgame perfect, there are some doubts about the predictive power of their analysis. The focus of this paper is on an industry with N firms who may decide to form a cartel, depending on the substantive law and the procedural rules. Where a cartel is uncovered, members are assumed to be jointly and severally liable and we contrast the two cases of no-contribution and contribution. 6 We will assume that those harmed by the cartel are sufficiently alike that any private case they would pursue can be consolidated into a single class action. While a class action is not generally available in the EU, many of the jurisdictions have other means for claims to be consolidated. 7 We assume sequential bargaining, which we feel is closer to reality, and that private information enhances the level of damages awarded. The remainder of the paper is organised as follows. Section 2 provides an overview of the basic model of joint and several liability with and without contribution and demonstrates that the no-contribution rule leads to a larger overall settlement for the plaintiff. Section 3 extends this to the case where defendants have private information about the case. We consider the situation where revelation of this private information to the plaintiff will increase the total level of damages which can be established in court. The main insight of this section is that the no-contribution rule gives an incentive to reveal information in return for a reduced level of damages, in effect introducing a private leniency programme. The section also highlights how a firm applying for public immunity may be at a disadvantage in the private settlement game. Up to that point we simply look at the aftermath of an infringement with multiple defendants having been detected. The classic case which fits this scenario is collusion. In sections four and five we take the analysis further by focusing on collusion and considering first (in Section 4) how a no-contribution rule for private damages interacts with the stability of a collusive arrangement and secondly (in Section 5) the interplay between a public leniency programme and the private leniency programme arising from the no-contribution rule. Section 6 concludes. All proofs are collected in Appendix A.

This paper has focused on the rules surrounding contribution from co-defendants when damages have been awarded against one defendant who is one among many who are jointly and severally liable. While some may consider this a somewhat obscure procedural rule, we have shown that the design affects incentives, demonstrating that this is an example of a much more general claim that the choice of procedural rules within a legal system can have quite subtle but substantial effects on the ability of the law to deter anti-competitive behaviour. The no-contribution rule, where the defendant against whom damages has been awarded has to meet this liability in full, has a number of benefits over the contribution rule, where liability is shared among co-defendants. One well understood effect of the no-contribution rule is that there is a strong incentive to settle early and because of the competition to do so, total settlement is higher. This in itself helps deter cartel activity by increasing the penalty on the cartel members if caught. To this we have added the insight that a no-contribution rule makes defendants keen to settle early in return for revealing compromising information. Because contribution is ruled out, such liability-increasing revelation does not rebound on the defendant who provided the information through later claims by other defendants. The direct effect of the resulting increase in the total expected liability of the cartel (Proposition 1) is to increase deterrence (Corollary 2). Having shown that the no-contribution rule in effect leads to the creation of a private leniency programme where earlier settlements are reduced in return for information revelation (Proposition 2) we also analyse how this interacts with any public leniency programme. In particular we show in Section 3.2.1 that if obtaining public leniency singles out this defendant as the most likely to be taken to court, there is a strong disincentive to apply for leniency. We find cases where a cartel is undermined not by the public leniency programme but by the private leniency effects of the no-contribution rule (Proposition 4). These results further justify existing concerns about the adverse effect of private damages on public leniency programmes through increasing the potential liability of each of the cartel members so that they have little incentive to reveal a cartel even where an investigation has been initiated. It also adds a second concern, namely that the firm seeking public leniency is in particular danger of having to face most of the private damages claims. This concern is heightened where a ringleader is able to threaten credibly to reveal all to a private plaintiff. Thus no-contribution rules have potentially additional negative effects on public leniency programmes. In policy terms, it may be the case that joint and several liability and no-contribution are incompatible with a successful public leniency programme. Since the latter is seen as one of the great antitrust policy success stories, it would appear that if something has to give, it would be either joint and several liability or no-contribution. In the US we have seen that, through the Antitrust Criminal Penalty Enhancement and Reform Act of 2004, it was the former which was abolished so that the firm obtaining immunity from public fines also was protected from joint and several liability. The results in this paper lend support to the aims of the Act, which will be revisited 2010 due to a sunset clause. The possible adverse effects of private enforcement on leniency programmes do not appear to be fully appreciated by the European Commission. While neither the 2005 EU Green Paper nor the associated country reports mentions contribution as an issue, the extensive report30 commissioned by the European Commission does so in Section 6.3, although it does not identify the information revelation issues raised in this paper. From the White Paper and the accompanying discussion paper, the proposal appears to be that contribution should continue to be the rule within the EU but that, in a bid to protect its leniency programme, the Commission might consider steps to limit the civil liability of successful immunity applicants.31 A recent paper by Leslie (2009)32 approaches the problems discussed in this paper from another, purely US focused, direction. Although the US has a no-contribution rule, it is legal for defendants who might be held jointly and severally liable to enter a so-called Judgements-Sharing Agreement [JSA for short]. A JSA, if enforceable, has the same effects as replacing the no-contribution rule with a contribution rule. In discussing the merits of the JSA, Leslie (2009, p. 772) identifies that the JSA would have a negative effect on information flows along the lines demonstrated formally in Proposition 2 above. While providing no formal analysis, Leslie (2009, pp. 782–784) also offers a discussion of the impact of JSAs on cartel stability and on public policy, focusing narrowly on the current rules surrounding the US leniency programme. This paper thus also contributes to the discussion initiated by Leslie (2009) about whether or not JSAs should be enforceable. Our analysis rests on a number of simplifying assumptions. One of the more critical assumptions which simplified the analysis considerably is that there is a single plaintiff. We conjecture that the qualitative results would continue to hold if either one plaintiff accounts for the majority of the harm caused or that any information revelation is plaintiff specific and that other plaintiffs can be prevented from acquiring the information. Where private enforcement is possible without consolidating the claims in some way, defendants might be reluctant to provide information to one plaintiff for fear that this information would get disseminated to other plaintiffs with whom they had not yet reached a settlement. In such a case we would still get the higher damages arising from the “whipsaw effect” but not necessarily those associated with information revelation. The existence of several plaintiffs may be due to lack of information among the plaintiffs about each others existence and one might expect that this is a problem more associated with stand alone cases than follow-on actions where the class of those harmed would be more obvious. If information revelation is seen as desirable, making the class as large as possible, possibly by having an opt-out rule rather than an opt-in rule, may be appropriate. This provides further support for the idea that the mix of procedural rules matters to deliver maximal or even predictable effects.