Cross sectional efficient estimation of stochastic volatility short rate models

We consider the problem of estimation of term structure of interest rates. Filtering theory approach is very natural here with the underlying setup being non-linear and non-Gaussian. Earlier works make use of Extended Kalman Filter (EKF). However, as indicated by de Jong (2000), the EKF in this situation leads to inconsistent estimation of parameters, though without high bias. One way to avoid this is to use methods like Efficient Method of Moments or Indirect Inference Method. These methods, however, are numerically very demanding. We use Kitagawa type scheme for nonlinear filtering problem, which solves the inconsistency problem without being numerically so demanding.