On the optimality of equilibrium when the market structure is incomplete

Abstract

In the Arrow-Debreu theory of general equilibrium, it is assumed that markets exist for current goods and also for goods to be delivered at future dates and in uncertain events. Under this assumption, and if economic agents face no transaction costs, all economic decisions will be made at one time and markets will open only once. . In the real world, however, we observe that few markets for contingent futures goods exist at any one time, and that markets reopen many times. In recent years, several authors have constructed general equilibrium models that reflect this incomplete and sequential aspect of real world trading. Two main approaches have been taken: the temporary equilibrium approach and the rational expectations approach. The temporary equilibrium approach assumes that economic agents have given expectations of future prices and investigates whether there exist prices which clear current markets (for a discussion of this approach, see Arrow and Hahn [2] and Grandmont [5]). The rational expectations approach, on the other hand, regards expectations as variables and investigates whether there exists a set of current prices and expected prices such that all markets, both current and future, are cleared (for a discussion of this approach, see Radner [ 11 I). Most of the work that has been done on incomplete and sequential markets has been concerned with establishing the existence of equilibrium. The subject of the optimality of equilibrium has received considerably less attention. Now it is clear that a temporary equilibrium will not generally enjoy any optimality properties since the actions of economic agents are