The Volatility Structure of the Fixed Income Markets under the HJM Framework

Here I consider the dynamics of interest rate processes in the multi-factor model specified in Heath, Jarrow and Morton (1992). Despite its flexibility and theoretical advances, the number of empirical studies using the HJM model remains inadequate, principally because of the difficulty estimating models in this class, which are high-dimensional, nonlinear, and involve latent state variables. Here I treat the estimation of a broad class of HJM models as a nonlinear filtering problem. I adopt the local linearization filter of Jimenez and Ozaki (2003), known to have some desirable statistical and numerical features, and estimate the model using maximum likelihood. The estimator is applied to the U.S., U.K. and Australian markets. Different two- and there-factor models are found to be best for each market. The contributions of the factors towards overall variability in interest rates and the financial rewards claimed by each factor are found to differ considerably over markets