Meiselman on the Structure of Interest Rates: A British Test

Abstract

There has been a surge of interest recently in theories of the term structure of interest rates.' A subject which has such obvious implications for monetary and debt policies is not likely ever to fall into obscurity, but the last few years have been particularly full of debate. One of the stimulating influences in the discussion has been the work of David Meiselman, recently published,2 putting forward a hypothesis with respect to the structure of rates in the market for U.S. corporate bonds. Meiselman's hypothesis, which is based on an expectations theory of bond market behaviour, was convincingly supported by the data he used. Although it is not one which pleases the intuition, Meiselman's conclusion that it efficiently describes the market he studied has not yet to my knowledge been questioned.3 This article has the two-fold purpose of pointing out certain characteristics of his data which reduce the value of his test procedure, and of reporting on a test of the same hypothesis against data taken from the British Government securities market. The results may be briefly summarized: the usefulness of Meiselman's hypothesis when applied to either market, that for U.S. corporate bonds or for British Government securities, is seriously called in question. The proliferation of individual securities in financial markets makes it necessary to adopt methods of classification if they are to be analyzed. Successful explanation of the price relationships between securities must take into account call features, credit-rating of borrower, size of issue, coupon rate, and many other characteristics in addition to the term to maturity. Hypotheses sufficiently complex to explain the entire record of fluctuations in security prices are not yet within the grasp of economic theorists. Notwithstanding the fact that the most powerful explanatory variables in any one case may be unrelated to the term of the security, the latter is still a highly useful classifying characteristic, and it is not surprising that much attention should have been paid to the term structure of interest rates , a purely theoretical notion implying an observable functional relationship between yield and maturity. Of course it is unlikely that we should find a term-structure