| Popis: |
ZISIMOS KOUSTAS [*] This article tests the long-run neutrality of money proposition using quarterly U.S. data over the period from 1960:1 to 1996:2 and the methodology suggested by King and Watson (1997), paying particular attention to the integration and cointegration properties of the variables. Comparisons are made among simple sum, Divisia, and currency equivalent (CE) monetary aggregates using the Anderson et al. (1997a, 1997b) series of Divisia and CE monetary aggregates. (JEL E40, E50, C32) I. INTRODUCTION The aim of this article is to investigate a long-run question about the role of money in monetary policy--whether changes in the conduct of monetary policy influence the growth rate of output. As Lucas (1996, p. 661) puts it, "So much thought has been devoted to this question and so much evidence is available that one might reasonably assume that it had been solved long ago. But this is not the case." In fact, recently Fisher and Seater (1993) and King and Watson (1997) contribute to the literature on testing quantity-theoretic propositions (by developing tests using recent advances in the theory of nonstationary regressors) and show that meaningful long-run neutrality and superneutrality tests can only be constructed if both nominal and real variables satisfy certain nonstationarity conditions and that much of the older literature violates these requirements. In this paper, in the spirit of Serletis and Koustas (1998) and Koustas and Serletis (1999), we test the long-run neutrality and superneutrality of money, using the King and Watson (1997) methodology, paying particular attention to the gains that can be achieved by rigorous use of microeconomic and aggregation-theoretic foundations in the construction of money measures. This is accomplished by making comparisons among simple sum, Divisia, and currency equivalent (CE) monetary aggregates using the Anderson et al. (1997a, 1997b) data--see Anderson et al. (1997a, 1997b) for more details. We also pay attention to the integration and cointegration properties of the variables, because meaningful neutrality and superneutrality tests critically depend on such properties. The organization of the article is as follows. The next section briefly discusses the problem of the definition (aggregation) of money. In section III, we investigate the univariate time-series properties of the variables, as these properties are relevant for some problems of potential importance in the practical conduct of monetary policy as well as for estimation and hypothesis testing. In section IV, we test the long-run neutrality and superneutrality of money propositions, using the King and Watson (1997) structural bivariate autoregressive methodology, paying particular attention to the integration and cointegration properties of the data. The final section concludes. II. THE MANY KINDS OF MONEY The monetary aggregates currently in use by the Federal Reserve are simple-sum indices in which all monetary components are assigned a constant and equal (unitary) weight. This index is M in (1) M = [[[sigma].sup.n].sub.i=1] [x.sub.i] where [x.sub.i] is one of the n monetary components of the monetary aggregate M. This summation index implies that all monetary components contribute equally to the money total, and it views all components as dollar-for-dollar perfect substitutes. There is no question that such an index represents an index of the stock of nominal monetary wealth but cannot, in general, represent a valid structural economic variable for the services of the quantity of money. Over the years, there has been a steady stream of attempts at properly weighting monetary components within a simple-sum aggregate. With no theory, however, any weighting scheme is questionable. Barnett (1980) derived the theoretical linkage between monetary theory and aggregation and index number theory. He applied economic aggregation and index number theory and constructed monetary aggregates based on Diewert's (1976) class of "superlative" quantity index numbers. … |
Plný text