James B. Bullard presents a synopsis of the recent work on the microfoundations of rational expectations, with special emphasis on the implications for macroeconomic policy. The theory of rational expectations does not specify a method of expectations formation, sometimes called a learning mechanism. Bullard demonstrates that, as a result, the application of the theory can be ambiguous. This is an important fact about modern macroeconomic literature because the treatment of expectations plays a crucial role in deriving policy implications in standard macromodels.