Market concentration and incentives to collude in Cournot oligopoly experiments

Multiple Cournot oligopoly experiments found more collusive behavior in markets with fewer firms (Huck et al., 2004; Horstmann et al., 2018). This result could be explained by a higher difficulty to coordinate or by lower incentives to collude in markets with more firms. We show that the Quantal Response Equilibrium can explain how the change in incentives alone could result in more collusive output in smaller markets. We propose a new method to manipulate the group size while keeping constant the locations of key outcomes, payoffs at these outcomes and the incentives to collude. Experiments using this normalized payoff function find that the number of firms has no direct effect on the average output or profit. We conclude that higher rates of aggregate collusion in markets with fewer firms are driven by the changes in incentives or focality rather than purely the number of firms. These findings imply that antitrust policies aimed at preventing collusion should focus on incentives rather than on the market concentration.