Rational Expectations, Random Parameters and the Non-Neutrality of Money

Abstract

In the Sargent-Wallace (Sargent, 1976, 1979; Sargent and Wallace, 1975, 1976) analyses of output and inflation in a stochastic environment, it has been demonstrated that there is no role for discretionary monetary policy. In effect, variations-for example, in the supply of money-will have no real effects on the economy; any kind of cyclical monetary policy is otiose. Money is neutral. The Sargent-Wallace models are founded on the supposition that economic agents form their expectations (essentially, just of inflation or of the price level) rationally along the lines suggested by Muth (1961) which were amplified and extended formally by Lucas and Rapping (1969) and Lucas (1970, 1972, 1973). The structural equations in the SargentWallace models are log-linear and are based on the supposition that there are no rigidities in the goods and factors markets. Additionally, at any moment, all economic agents have the same information set as that possessed by the authorities. It has now been established in the literature that if any one of these conditions is rescinded, there is a role for active monetary policy. For example, it has been noted by Shiller (1978) that, in the Sargent-Wallace model, if the square of the unanticipated inflation variable enters the supply schedule/Phillips Curve, then the variance of output will depend upon the money supply. It has also been demonstrated by Persson (1979) that, if the log-linear Sargent-Wallace type of economic system is used to derive results appertaining to the effect of money supply changes on the level (as opposed to the log of the level) of output, then, once again, the neutrality of money no longer holds. There is an extensive literature showing that rigidities in the form of wage contracts, for instance, and the fairly obvious case of a divergence of information sets in favour of the authorities, will make the variance of output a function of the stock of money, so reinstating the case for discretion instead of rules in the conduct of monetary policy. So these conclusions are well documented (see for example Shiller, 1978; Fischer, 1977; Phelps and Taylor, 1977; Sargent and Wallace, 1975; and Woglom, 1979). However, the neutrality theorem that emerges from models of the SargentWallace genre has never been challenged successfully on its own ground ruies.1 Indeed, in the latest paper that is available, Minford and Peel (1980), in a study of monetary stabilization policy in the context of different supply schedules, are able merely to repeat the analysis and conclusions of Sargent and Wallace (especially Sargent, 1979) 2 It is the purpose of this paper to accept all of the economic assumptions of the Sargent-Wallace models under rational expectations. The only amendment we introduce into their models is to assume that instability in the economic system is not imparted by random shocks to some or all of the