Ambiguity is Detrimental for Long-Run Cooperation

Abstract

Why agents cooperate is an old question that has been widely studied in economics, as well as in other disciplines. In economics, the focus has been mainly put in analyzing whether cooperative behaviors may arise from repeated interactions between strategic individuals. This paper seeks to contribute to the literature by studying the effects of ambiguity, shortly defined as uncertainty about the realization of future states and the probabilities assigned to their realization, on long-run cooperation. Concretely speaking, the infinitely repeated Prisoner’s Dilemma, and the consequent determination of conditions for sustaining long-run cooperative behaviors, is revisited in an ambiguity setting. Results suggest that ambiguity is detrimental to cooperation as it decreases (increases) the expected payoff of the cooperative (non-cooperative) equilibrium and, therefore, (1) it may change the game’s structure from the Prisoner’s Dilemma setting to scenarios in which the expected payoff associated with the cooperative equilibrium is no longer Pareto-superior, and (2) it demands more patient players for sustaining cooperative agreements when the game’s structure remains unchanged. An application is made to the Cournot duopoly and the likelihood of sustaining a tacit collusion. In addition to previous results, it is found that the Cournot duopoly may not longer behave as a particular case of the Prisoner’s Dilemma for high levels of ambiguity. A positive relation between ambiguity and competition is stated. Therefore, our results may serve as an additional theoretical input for antitrust policy design.