Experimental evidence on partial cartels: determinants and policy implications
We intend to use a laboratory experiment to investigate the effectiveness of different antitrust policies. Subjects will act as the manager of a firm and play a repeated Bertrand game with fixed capacities in groups of four. By letting subjects the choice to communicate, we will consider that subjects who took part to the communication have formed a cartel. A novelty of our experiment is to allow partial cartels to form i.e. we consider a cartel has formed if at least two members of a group decide to communicate. In reality, competition authorities frequently sanction only a subset of firms active on a market. Considering only cartels encompassing all the firms of a market, as usually done, may be too restrictive. As treatment variations, we will implement different antitrust policies. We are particularly interested in the effectiveness of leniency programs at deterring and detecting cartel formation. Leniency programs give complete immunity or a sentence reduction to firms that participated to an illegal cartel agreement and report it to the competition authority. In the last 20 years, competition authorities have increasingly used such programs. Up to now, there is little experimental evidence on the effectiveness or drawbacks of leniency programs.