From Ad Hoc Bond-Risk Measures to Variance–Covariance Forecasts

Abstract

It is common practice among investors to assess the risks of a fixed-income investment by looking at certain bond characteristics—and at two in particular. The duration provides insight on a bond’s contribution to the interest-rate risk of a portfolio, and the credit spreads provide information about credit risks. Linking these measures together in order to obtain one forecast for the overall price behavior of a bond and bond portfolio is not trivial, however. Although the popular method of taking products, the so-called duration-times-spread rule, is effective and superior to using spread duration, it stops short of delivering complete risk estimates. In this article, the authors propose a method that does provide a complete risk estimate and describe how bond characteristics can be converted into return variance and covariance forecasts. <b>TOPICS:</b>Fixed income and structured finance, risk management, portfolio construction, statistical methods <b>Key Findings</b> ▪ This article proposes a method to convert the durations and credit spreads of bond portfolios into return variance and covariance forecasts. ▪ The conversion method is a direct extension of the duration-times-spread rule. ▪ The method helps bridge the gap between front-office portfolio management tools and back-office risk control systems.