A Test of Fischer's Theory of Monetary Misperceptions and the Business Cycle in the Presence of Long-Term Contracts

This article uses multi-period ex ante anticipations of money supply growth to estimate the parameters of a model, suggested by Stanley Fischer, in which money affects real variables only through multi-period errors in anticipations. This model is tested against an alternative, first evaluated empirically by Robert Barro, in which money affects real variables only through single period errors in anticipations. The two models are compared using the "P" test procedure for non-nested models suggested by Davidson and MacKinnon. The small sample properties of the test are unknown. Random experiments are performed to approximate these properties. On the basis of estimated small scale distributions, the Fischer model rejects the Barro model at conventional levels, but is not rejected by it.