Markov Puppy Dogs and Other Related Animals

It is shown that steady state Markov perfect equilibria of discrete time, infinite horizon, quadratic, adjustment cost games differ from equilibria of their infinitely repeated counterpart games with zero adjustment costs even though no adjustment costs are paid in the steady state. In contrast to continuous time games, the limit of these equilibria as adjustment costs approach zero is the same as the equilibria of their static counterpart games. A classification scheme is presented and it is shown that the taxonomy is identical to that of analogous two stage games such as those analyzed by Fudenberg and Tirole (1984). This classification is useful in that it implies that steady state equilibria need not be explicitly calculated to analyze qualitatively the effects of adjustment costs in strategic environments. Is is also argued that estimated conjectural variations parameters may capture a well defined property of strategic interaction in a dynamic game.