Measuring Risk in Fixed Income Portfolios using Yield Curve Models

We propose a novel approach to measure risk in fixed income portfolios in terms of value-at-risk (VaR). We obtain closed-form expressions for the vector of expected bond returns and for its covariance matrix based on a general class of dynamic factor models, including the dynamic versions of the Nelson-Siegel and Svensson models, to compute the parametric VaR of a portfolio composed of fixed income securities. The proposed approach provides additional modeling flexibility as it can accommodate alternative specifications of the yield curve as well as alternative specifications of the conditional heteroskedasticity in bond returns. An empirical application involving a data set with 15 fixed income securities with different maturities indicate that the proposed approach delivers accurate VaR estimates.