The Relevance of an Aggregated Benchmark for Testing Stein's Theory: Comment

Abstract

Jerome Stein utilized three-month interest rates as well as spot and three-month forward exchange rates to classify periods as speculative, basic-balance normal or interest-differential normal. The various benchmarks employed in testing Stein's theory have been constructed with data that aggregates thq forward positions of commercial banks over all maturities. In his comment,1 Mark R. Eaker claims that, since Stein only considered the three-month forward rate, these aggregated benchmarks are inappropriate. He argues that only the three-month forward positions are relevant for testing the theory. However, this argument is more an indictment of Stein's theory than of the tests. The ideal indication of the speculative activity of professional risk-bearers in a given period is the sum of the present value of the changes in the risk-bearers' foreign exchange positions at every maturity, with these changes in foreign exchange positions discounted by the interest rate relevant for each maturity. Restricting Stein's theory to merely an indication of the risk bearers' three-month forward speculative position renders his classification scheme sterile. For example, if the spot rate and three-month forward premium were both rising. Stein's theory would denote a bullishly speculative period. In cases where the aggregated ideal benchmark is according to Eaker, the sum of the discounted present value of all changes in foreign exchange positions is in the opposite direction of changes in the three-month forward position. In this situation, for Stein's methodology to accurately indicate an increase in the three-month forward position, it must classify as bullishly speculative a period when professional risk-bearers have undertaken net bear speculation. Thus to the extent that the ideal aggregated benchmark is for testing Stein's theory, his classification scheme is inappropriate for describing the general speculative nature of a period. For his classification scheme to be a meaningful delineation of speculative and normal periods, Stein must have implicitly assumed that changes in the three-month speculative position were in the same direction as changes in the total net speculative position.