Informational requirements of collusion detection

A sensible economic interpretation of the European Communities' Directorate General (DG) IV's competition policy is as follows. Suppose a structural oligopoly is modelled as playing a repeated game. If the path of prices and quantities can be justified as the non-cooperative equilibrium of some stage game in every period, then the industry's behaviour is ‘permitted:’. On the other hand, if the path of prices and quantities can be justified only as an equilibrium of the repeated game in history-dependent strategies (e.g., ‘trigger’ or ‘carrot-and-stick’ strategies), the industry's behaviour is ‘forbidden’, and subject to antitrust redress. Stated differently, if the level of profits being attained can only be reached via threats to punish deviations from a tacitly collusive path (see chapter 6), then the behaviour is ‘forbidden’. These concerns are also relevant to antitrust policy in the United States and other nations.

Only this game-theoretic interpretation per se, rather than any issues of how closely this interpretation fits DG IV policy, is explored here. A welfare analysis of the merits of a policy which fits this interpretation is also outside our scope. Our concern is with the prior question of whether a policy so interpreted can be implemented given sensible models of the information available to antitrust authorities.