Alternative Duration Specifications and the Measurement of Basis Risk: Empirical Tests

TLDR

Since the early 1970s, rising interest rates and volatility have spurred renewed interest in fixed‑income risk measurement, leading to many extensions of Macaulay duration and widespread commercial immunization programs that claim to explain returns and protect portfolios. The study empirically tests the explanatory power of seven recently proposed duration measures for basis risk in fixed‑income securities. The analysis shows that the seven duration measures are empirically indistinguishable, fail key risk‑factor tests, and provide no advantage for performance comparisons or immunization strategies.

Abstract

Dramatic increases in interest rate levels and volatility since the early 1970s have renewed interest in fixed-income securities and have motivated much study of the appropriate measure of risk for bonds. Many authors have expanded on Macaulay's (1938) concept of duration as a surrogate for risk measurement and have developed a variety of new measures of duration. Numerous commercial duration-based immunization programs are now available purporting to explain returns, to provide good measures of risk, and to protect fixed-income portfolios from volatile interest rates (e.g., Leibowitz 1979; Leibowitz and Weinberger 1981). This paper tests the explanatory power of a number of recently proposed duration measures. We show that despite the flood of articles and commercial programs claiming superiority for particular measures of duration, the measures studied were virtually indistinguishable empirically. In fact, none of them did much better than This paper empirically investigates seven duration measures and their role in explaining price volatility caused by interest rate movements, that is, basis risk. The data analysis does not support the important testable implications of the various durations as measures of basis risk. All seven durations fail linearity and single risk factor tests from which their advantages are supposed to derive. Fixed-income performance comparisons and immunization strategies based on these durations have not been worthwhile.