INTEREST-RATE CONTROL AND NONCONVERGENCE TO RATIONAL-EXPECTATIONS

This paper investigates the feasibility of a monetary policy aimed at pegging the nominal rate of interest. It shows that under general conditions such a policy would produce the well-known cumulative process, despite the fact that there exists a well-behaved rational expectations equilibrium with no tendency for inflation to accelerate or decelerate. The cumulative process shows up as the failure of learning to converge to rational expectations. Specifically, the paper shows, first in a conventional IS-LM model with an expectations-augmented Phillips curve and then in a micro-based finance constraint model, that if people follow any learning rule based on experience that satisfies a weak condition, then the sequence of temporary equilibria under a policy of interest pegging cannot converge. The nonconvergent path that will be observed accords with the familiar cumulative process, in that inflation accelerates if the market rate of interest has been pegged below the natural rate.