Parametric Estimation of Quadratic Term Structure Models of Interest Rates

Nonlinear filtering techniques and the quasi maximum likelihood estimator (QMLE) are applied to the problem of estimating the parameters of quadratic models for the term structure of interest rates. It is assumed that zero coupon bond yields data have been contaminated by noise, which allows the application of nonlinear filtering techniques. Without the need of real time computation, we can instead apply the smoothing techniques. The asymptotic properties of the QMLE are also analyzed in two ways: the asymptotical optimality under Kullback-Leibler criterion and its consistent conditions in general. Finally Monte Carlo simulation results are presented to confirm the performance of this strategy.