A Little More Evidence from the Time Series

IN AN earlier article in this Journal and in several other places, evidence has been presented supporting a theory of the demand for money that is a part of the "wealth adjustment process." The posited demand function has successfully passed a large number of tests in competition with more than a dozen alternatives, representing the bulk of substantive work on the demand for money in the past thirty years. Though no series of tests is "definitive," the evidence from tests against alternatives is of crucial importance in establishing the economic relevance of the particular demand function. I regard such tests as preliminary to —and far more important than—"Chow tests," "Theil-Nagar tests," "Durbin-Watson tests," and other sophisticated statistical procedures for establishing the relevance of particular hypotheses.4 However, the accumulating evidence suggests that the use of refined statistical procedures may now be desirable. I welcome the opportunity presented by the comments of Courchene and Shapiro to present some of the available evidence on the points that they raised.