A new framework for games in continuous time is proposed. The continuous-time model conforms as closely as possible to the conventional discrete-time framework. Indeed, continuous time is viewed as "discrete time, but with a grid that is infinitely fine." The paper presents several examples illustrating the difficulties that arise in continuous-time game theory. Theorems relate the equilibria of continuous time games to the equilibria of approximating discrete time games. A variety of industrial organization applications are studied, yielding sharp predictions. Applications include continuously repeated games, preemption models, and patent races. Copyright 1989 by The Econometric Society.

This paper examines conditions for the uniqueness of an equilibrium price distribution in stochastic macroeconomic models with rational expectations.A model is developed in which many price distributions, each with a finite variance, satisfy the equilibrium requirements of rationality.Hence, the condition that the variance of the equilibrium price distribution be finite, or equivalently, that the conditionally expected price path be stable, does not guarantee uniqueness.In such cases it is shown that an arbitrary random quantity which is widely publicized can become a leading indicator of prices and, consequently, influence the behavior of actual prices.However, by extending the finite variance (stability) condition to a minimum variance condition, these nonuniqueness problems can be avoided.Such stability or minimum variance conditions suggest a kind of collective rationality which, although not unreasonable, has not yet been fully analyzed in rational expectations models.