Dynamic complexity in duopoly games (Q1088928)

We discuss the occurrence of periodic and chaotic phenomena in infinite horizon duopoly games where firms maximize their discounted sum of profits and use Markov-perfect equilibrium strategies. In the alternating case their corresponding actions are then constrained orbits of a map similar to a Cournot tâtonnement. We show that any behavior is possible for small discount factors and that one obtains the Cournot tâtonnement as the discount factor goes to zero. In the simultaneous case, we show by mean of an example that many Cournot tâtonnements can be viewed as Markov-perfect equilibria of an infinite horizon game and that one obtains the Nash equilibrium of the static game as the discount parameter goes to zero. We are thus led to study in detail the dynamical properties of a Cournot tâtonnement.