The impact of leniency and whistle-blowing programs on cartels

Abstract

Antitrust authorities have set up leniency programs for cartel members that denounce their collusive agreements. These programs help prosecute participants and can thereby deter collusion. We compare the impact of reduced fines and positive rewards and argue that rewarding individuals, including firm employees, can deter collusion in a more effective way.

We discuss possible adverse effects of whistle-blowing programs on firms' behavior, and particularly on turnover, incentives to innovate and cooperation. We also explore explanations for the puzzling fact that managers keep incriminating evidence and argue reward programs actually provide additional incentives for keeping such evidence.

Introduction

There is almost universal agreement that price-fixing and market allocation cartels reduce economic efficiency. Yet although most competition laws forbid cartels, cartels continue to form and operate in a significant number of industries (see Connor, 2004a, Connor, 2004b, Levenstein and Suslow, 2004). Evidence from, e.g., the food additive and vitamin cartels, that lasted for 10 years, suggests moreover that a number of cartels succeed in maintaining discipline over a long period of time. The main problem is thus an implementation issue, i.e., how to enforce cartel prohibition (see Rey, 2003).

National competition agencies have nevertheless recently achieved notable successes in prosecuting cartels. Cartel decisions are increasingly numerous and concern all sectors, from vitamins to plasterboard or fine art auctions – with Christie's denouncing its participation in a cartel with Sotheby's. These successes are in great part the result of leniency programs set up in the United States in 1993 and in the European Union in 1996.¹ This paper investigates the effect of leniency programs on the behavior of firms and argues that more generous programs, offering positive rewards to firms but also to individual informants, might be more effective.

Leniency programs reduce the fines for cartel members that bring evidence to the antitrust authority, and their impact can be seen in the recent increase in successfully prosecuted cartels. For example, its Amnesty program allowed the Antitrust Division to secure more fines in 1999 only, than the total sum of fines imposed under the Sherman Act since its adoption, more than a century before. And in 19 months only, over 2002 and 2003, the European Commission took 19 decisions, involving more than 100 companies, for a total amount of fines of almost 3 billion Euro.²

The corporate leniency program set up in the United States in 1978 has been revised in 1993, to grant full amnesty to the first informant, together with amnesty for individuals (‘directors, officials and employees’ of the informant firm). The European leniency program has similarly been reinforced in 2002 so as to grant full amnesty to the first reporting firm.³ The judicial security offered to informants constitutes a key success factor of these revised programs. Confidentiality is guaranteed, in particular with respect to other relevant jurisdictions.⁴ Several differences still exist between the American and the European systems: First, individuals are not liable in Europe under the current legislation, contrary to the US.⁵ Second, the US leniency program does not grant amnesty to cartel ringleaders. A similar requirement existed in the first version of the EU program but has been essentially removed in 2002 in order to foster deterrence.⁶

Encouraging insiders to bring evidence seems sensible, as they are more likely to possess the type of information needed for establishing a violation of the law. Cartels must reach agreements (on total output levels, output shares, sales territories, etc.) and monitor compliance with agreed-upon terms. Such cooperation often leaves traces that can be used as proofs by an insider informant. Yet reduced fines may not be sufficiently attractive to fully deter collusion. One might therefore contemplate giving informants not only favorable treatment (partial or full amnesty) as in a leniency program, but also positive rewards.

In practice, only a few systems offer bounties to informants. One such system⁷ is the U.S. Civil False Claims Act, which has been used extensively to attack fraud involving the U.S. government's role as purchaser of goods and services (e.g., procurement fraud against the Department of Defense) and as insurer (e.g., fraud involving health care programs covered by Medicare). This Act rewards individuals who inform the government of fraud in procurement contracts by a substantial share of the fines collected – an amount sufficient to compensate employees for the stream of foregone future wages.⁸

As in McCutcheon (1997), our working assumption is that communication is necessary for collusion: it brings an agreement that firms are then free to implement or not, as in a standard tacit collusion situation, but it also generates evidence that can be found by the antitrust authority, as well as by firms or individuals. This framework is similar to the one used by Motta and Polo (2001) and Spagnolo (2003), who also study the role of leniency programs (on self-reporting, see also Innes, 1999).

Motta and Polo (2001) analyze the impact of reduced fines for cartel members that inform the antitrust authority, when the probability of antitrust intervention is endogenously determined under a balanced budget constraint. They show that it can be efficient to reduce fines even when the antitrust authority has already started an investigation, but has not yet obtained evidence of misbehavior. While such leniency reduces the cost of collusion for cartel members, it encourages reports that decrease investigation costs. In this model, if the budget of the antitrust authority was high enough, it would be optimal to have no Leniency Program, and to intervene often enough to fully deter collusion: reduced fines are a second best instrument.

Spagnolo (2003) also examines the effect of Leniency Programs on the sustainability of cartels, in a framework that is closer to ours. Contrary to Motta and Polo, he assumes that when a cartel is detected, it is also convicted. This allows to focus on the impact of Leniency Programs on cartels that are not already under investigation. Allowing for generic punishment strategies,⁹ Spagnolo shows that the antitrust authority should not fine firms that deviate from a cartel agreement, should reward only the first party that reports, and should offer a positive reward equal to the sum of the fines paid by the convicted firms (assuming that the antitrust authority is budget constrained). Provided that the maximum fine is high enough, such a reward policy can implement the first-best outcome: full deterrence at no cost. Even if the antitrust authority cannot offer positive rewards, reduced fines for reporting firms can be useful by decreasing the cost of deviating from the cartel agreement. Last, Spagnolo shows that reduced fines always increase the riskiness of an agreement.

Our framework differs from the above papers in that we take the probability of an investigation as given, and focus on the consequences of rewards on decisions taken within the cartel and within firms. We show that positive rewards have a larger deterrence effect than reduced fines, and that rewards for individuals can be more effective than corporate ones. We then turn to the potentially adverse effects of rewards mentioned above, such as preventing efficient cooperation between firms, restricting information flows between employees, or inducing a more rigid employment structure. We show that reward programs can be adapted so as to mitigate these costs.

A major puzzle is why evidence is not destroyed once communication has taken place. We explore explanations based on agency problems, both at the cartel level and at the level of the individual firms. We point out that positive rewards may exacerbate these agency problems and encourage firms (or individuals) to gather and keep evidence. This indicates that, while firms may adapt to the new judicial environment and better hide or destroy all traces of communication with other cartel members,¹⁰ they will probably have to keep some incriminating evidence.

The paper is organized as follows. Section 2 sets up the model. Section 3 analyzes the effect of leniency and reward programs on cartel sustainability. Section 4 considers the effects of rewards on inter-firm communication, intra-firm hiring strategies, and investment incentives. Explanations for the fact that evidence of collusion is often not destroyed are explored in Section 5. They rely on agency problems, within the cartel or within the firm. Section 6 concludes.