Learning to believe in sunspots

Abstract

A NUMBER OF AUTHORS have shown that competitive economies may possess equilibria, that is, expectations equilibria in which purely extrinsic uncertainty affects equilibrium prices and allocations.2 Such results demonstrate that it does not require a lack of faith in the rationality of market participants to believe that competitive markets may be subject to purely speculative fluctuations, driven solely by expectations.3 The mere existence of equilibria as solutions to a system of marketclearing conditions, however, might not be judged sufficient to indicate that competitive markets with participants could ever be subject to speculative instability. The equilibria represent states of affairs in which agents act differently in the case of different realizations of the sunspot variable, and it is for each agent to do so. But it might be thought unlikely that the beliefs of all the participants in the market could ever come to be coordinated so as to bring about an equilibrium of that kind. It is to believe that sunspots convey information about future states of affairs once the economy is in a equilibrium, but why would agents ever begin to believe in such a thing, so as to create the conditions under which the belief is rational? In order to address such a question, one must go beyond the mere statement of the conditions for equilibrium and discussion of what states of affairs satisfy them; one must specify an explicit dynamic process according to which the beliefs of agents adjust when out of equilibrium. Any exercise of this kind is necessarily unsatisfactory, as there is no univocal meaning for the postulate of rational behavior outside of an equilibrium. It may well be the case that