Kalman Filtering of Generalized Vasicek Term Structure Models

We present a subclass of Langetieg's (1980) linear Gaussian models of the term structure. The bond price is derived in terms of a finite set of state variables with correlated innovations. The subclass contains a reformulation of the double decay model of Beaglehole and Tenney (1991), enabling us to clarify interpretation of their parameters. We apply Kalman filtering to a state space formulation of the model allowing measurement errors in the data. One, two and three factor models are estimated on US data over 1987-1996 and the results indicate the subclass of models can fit the US term structure