By Maximilian Andres, Lisa Bruttel, and Jana Friedrichsen
In most countries, agreements that are designed to distort competition, e.g. in the form of coordinated pricing behavior, are prohibited (e.g., Article 101 of the Treaty on the Functioning of the European Union from 2012). Despite being illegal, however, cartels are regularly formed in various markets all around the world. In attempting to hinder and destabilize cartels, competition authorities partially rely on corporate leniency programs that exempt a firm from paying a fine if it provides evidence about the illegal activity to the authorities that helps in the prosecution of that cartel.
In many recent cartel cases, the leniency rule played an important role and the general perception appears to be that its introduction was successful as it increased the number of prosecuted cartels. As the number of active cartels that are not prosecuted is unknown, the validity of this claim is hard to judge. BCCP Doctoral Student Maximilian Andres, BCCP Fellow Jana Friedrichsen, and their co-author Lisa Bruttel, therefore, conducted a laboratory experiment on the topic, where successful and stable cartels are also observed.
They study the effectiveness of a corporate leniency program in repeated Bertrand competition where the same three firms repeatedly interact and may communicate in free-form chat before they set their prices in a given round. The focus of their study is on the role of communication. Importantly, their study allows for distinguishing innocuous communication from chat content that is related to the establishment or conduct of a cartel. In particular, their experimental design features the human judgment of communication and competitive conduct that takes place during the investigation of the competition authority.
It turns out that, in this setup, a leniency rule does not significantly affect any relevant outcome. The study does not find significant differences in cartelization, pricing, or communication between an experimental treatment with a leniency rule and a control treatment without it. This finding is in stark contrast to most of the published experimental literature on the topic, which relies on designs where a) firms in a market can only communicate if they unanimously voted to do so, implying that their market counts as a cartel independent of the realized market outcome, and b) communication has to follow a highly structured protocol in the form of pre-coded messages. Andres, Friedrichsen, and Bruttel argue that those design features are crucial for the different results.
Taken together with the results from previous studies, this study suggests that a leniency policy might not be as effective as often claimed because it only finds a very low self-reporting rate. Specifically, the authors argue that leniency mostly attracts self-reports of firms in cartels that are inherently unstable. In these cases, of course, self-reports may provide information that improves the prosecution of cartels. With a low reporting rate, however, the high number of cartel cases might be worrying because it indicates that there are many more undetected cartels.
The full paper “The leniency rule revisited: Experiments on cartel formation with open communication” is available as DIW Discussion Paper No. 1926.
This text is jointly published by BCCP News and BSE Insights.