<p>We investigate the existence and structure of Markov-perfect equilibria of discrete-time dynamic games in which players are risk averse and have time preferences consistent with discounting. We establish the existence of a Markov-perfect equilibrium when each player strives to maximize the expected exponential utility of the present value of the time stream of rewards. Also, we give sufficient conditions for a Markov-perfect equilibrium to be myopic, namely to be a sequence of Nash equilibria of static games. The myopia results are applied to a dynamic oligopoly model in which firms choose prices and production quantities, encounter stochastic demand and hold inventories.</p>