Rational Expectations and Short-Run Neutrality: A Reexamination of the Role of Anticipated Money Growth

Abstract

In the rarified air surrounding timeless, general-equilibrium models, perfectly flexible wages and prices, rational economic agents, and the absence of information and adjustment costs ensure there is no need to distinguish between anticipated and unanticipated fluctuations in exogenous policy variables. Economic agents adjust completely and immediately to all changes anticipated or not; money is always and everywhere neutral. In the heavier air closer to the ground, dynamic models which allow for imperfect and costly information, as well as adjustment costs, have (appropriately) placed considerable emphasis on the expectations of economic agents. Such models have forced researchers to confront directly the analytical and empirical timedimensioned relationship between the imperfect information and incomplete adjustment characterizing the short run and the full information and complete adjustment characterizing the long run. Such a focus naturally gives rise to the need to distinguish in the short run between anticipated (expected) and unanticipated (unexpected) fluctuations in exogenous policy variables. The full analytical implications of this line of inquiry have been worked out by Lucas, Sargent, and others.' The basic proposition regarding monetary policy emanating from what has come to be called the rational expectations, equilibrium business cycle hypothesis, is that only unanticipated movements in monetary policy will have real (albeit temporary) effects on output and employment; anticipated movements in monetary policy will only have price effects. The major empirical work supporting these a priori theoretical predictions has been developed in a series of papers by Robert Barro (1977, 1981; Barro and Rush, 1980). Given the implications of his findings and the controversy which has surrounded this entire body of work, it is not surprising that analytical and empirical critiques abound.2 Our focus in this paper is on the empirical work conducted by Barro, and most especially Barro and Rush (1980, hereafter B-R). Particular emphasis is placed on the problems associated with the partitioning of monetary growth into anticipated and unanticipated components. In the next section we briefly review the B-R work and various critiques which have been directed at it. The third section develops an alternative approach to the partitioning problem, presents our empirical work and compares it to the results yielded by the B-R approach. The final section summarizes our major findings and develops the resulting implications.