The use of the rational expectations hypotheses is now so standard in both the microeconomic and the macroeconomic literature that authors no longer feel the need to justify and explain the hypothesis. Nevertheless, this chapter discusses the arguments for using the rational expectations hypothesis in the light of recent work by myself and others on the question of how agents learn to form rational expectations (Bray and Kreps, 1981; Bray, 1982). See also Blume, Bray, and Easley (1982) for a survey of the literature on the subject.

Sargent and Wallace (1976) provide a variety of reasons for using the rational expectations hypothesis. They argue that it provides identifying restrictions for econometric work and a partial explanation for the finding that macroeconometric models fail tests for structural stability. Most fundamentally, they write that the rational expectations hypothesis “accords with the economist's usual practice of assuming that people behave in their own best interests.” This could mean that people act in their own best interests given their perception of the nature of the world; indeed, this is the nature of the Bayesian theory of decision making under uncertainty that underlies so many economic models. However, something more is entailed in the rational expectations hypothesis. Lucas (1976) writes: