7Expectations at the Short End of the Yield Curve: An Application

The recent interest in the expectations hypothesis and the term struc-ture of interest rates has been marked by an abandonment of the search for accurate forecasting ' that characterized early empirical work on the term structure. Economists have by and large followed David Meiselman's [15] lead in accepting the proposition that even if expec-tations prove to be inaccurate, they may still determine the yield struc-ture. Thus, the literature's emphasis has shifted to attempting to explain the process by which expectations are formed.2 This paper presents an inspection of the accuracy of the expectations implicit in the yield curve using essentially the same approach followed by Frederick Macaulay [13] in one of the earliest empirical studies of the term structure. Our return to this approach is motivated by several factors. First, existence of accurate forecasting provides a particularly convincing type of evidence confirming the expectations hypothesis. Second, as Kessel [12, p. 7] has noted, the reappearance of a seasonal in money market rates in the 1950's provides a new body of data with which to conduct tests along the lines of Macaulay's. Third, cross-spectral analysis provides a set of tools well suited to performing the required tests. NOTE: The author thanks Melvin Hinich, Phillip Cagan, Gregory Chow, Stanley Diller, and Jack Guttentag for helpful comments on an earlier draft of this paper. 1 Kessel [12] is the most notable exception