Vega-Redondo (1997) showed that imitation leads to the Walrasian outcome in Cournot Oligopoly. We generalize his result to aggregative quasi-submodular games. Examples are the Cournot Oligopoly, Bertrand games with differentiated complementary products, Common-Pool Resource games, Rent-Seeking games and generalized Nash-Demand games.

Abstract not found

This paper explores the impact of memory in Cournot oligopolies where firms learn through imitation of success (as suggested in Alchian (1950) and modeled in Vega-Redondo (1997)). As long as memory includes at least one period, the long-run outcomes correspond to all the quantities in the interval between the Cournot quantity and the Walras one. There is a conceptual tension between the evolutionary stability associated to the walrasian outcome, which relies on inter-firm comparisons of simultaneous profits, and the stability of the Cournot-Nash equilibrium,

We study a symmetric Cournot duopoly in which firms behave according to a decision program capturing the notion of "copy your competitor but only if it pays to do so." The basic intuition behind the decision program is that a firm initially mimics the competitor but sticks to the imitated behavior only if it pays o#. In the first stage of the decision program a firm imitates the opponent's output as an expression of naive imitative experimentation: Can I achieve a higher profit with the competitor's output? In the second stage of the decision program, the firm evaluates whether the first stage paid o# or not, staying with the competitors former output in case if it was profitable and switching back to previous output otherwise. The dynamics of the decision induces a finite Markov chain. We also assume that firms are bounded rational in the sense that with a small probability they make mistakes when applying their decision program. We study the long run behavior of the Cournot duopoly when the noise goes to zero employing stochastic stability analysis. We find that the collusive outcome is the unique support of the long run distribution.