Social identity and implicit collusion in Cournot interactions

Abstract

This research applies and extends the standard industrial organization models of repeated interaction between firms by incorporating group identity to evaluate the ability of group identity, thereby summarizing the theories of observed collusion. The model is used to outline circumstances under which collusion is easier to happen in a single market, and it will break down.

A general overview of literature based on laboratory experiments is presented to study the effects of social identity and study oligopoly markets. We construct lab experiments to test the effects of a single factor on collusion, i.e. whether the two players share the same group identity. University students were enrolled as research subjects in the laboratory experiments to test the validity of behaviour predictions. All experiments serve to answers two questions: a) How far is the market outcome away from the Standard Nash equilibrium? b) How good is the Nash prediction?

Study 1 investigates the effects of group identity on randomly rematches one-shot Cournot interactions. Study 2 describes the results of finitely repeated Cournot interactions that behaviour is more collusive when the players were from the same group than those from different groups or nogroup players. Study 3 concentrates on the indefinitely repeated interactions, finding that outgroup favouritism could be reflected in average quantity choices and collusion. Therefore, we determine that the effect of group identity on collusion is greater in repeated Cournot interactions than one-shot Cournot interactions, and that the repeated interaction devices enhance the difference between the players without group identity and players with primed group identity. The inspecting of individual behaviour indicated that the output adjustment is significantly correlated with the previous period’s two-sides profit changes comparisons. In the group matchings (ingroup matchings and outgroup matchings), group identity further strengthens the role of enhancement for collusion.

Group identity can influence significantly the player’s quantity choices. In this study we reassess the representation of group identity by applying group contingent other-regarding preferences. First, the influence of group identity varies unsympathetically across different devices of repeated Cournot interactions, so it cannot be explained through a well-behaved preference function. Second, this study suggests that group identity plays a key role in the preference over strategies of norms. Simulation results generated from a norm model estimated at the subject level provided insight into the repeated interactions and the group identity that motivate the collusion.