On Rational Expectations in a Model of Inflation M.A.Virasoro Dipartimento di Fisica, Universit`a di Roma, La Sapienza, I-00185 Roma, Italy May 29,1994 Abstract In economic models, expectations about future parameters influ- ence the behaviour of the agents. To have a well-defined model one has to write down a closed system of equations that relates expected and real values. Infinite rationality with perfect foresight provides a simple and appealing recipe. Other possibilities include adaptive ex- pectations and/or evolving, adaptive agents. In this paper we examine critically some of these ideas in a simple model for inflation. 1 Introduction In this note we want to make some observations on a model originally pro- posed by Sargent and Wallace [1] as a possible explanation of hyperinflation and later challenged and modified by Marcet and Sargent [2]. In both versions the dynamical equations have two fixed points at which inflation remains sta- tionary. In the Sargent-Wallace version the agent uses rational expectations to form an estimate of the future rate of inflation whereas in the Marcet- Sargent version he forms forecasts using a least squares regression of price on the once lagged price. 1