Minimum distance estimation and testing for interest rate models

We present some methods for the estimation and testing of usual ergodic interest rate models based on the observation of the short interest rate on the monetary market. First, we develop a test of type Kolmogorov-Smirnov for ergodic diffusion processes. We extend the results to the case where some parameters are estimated. Thereafter, we study a minimum distance estimator, based on the Lμ2 norm of the empirical process, which can be more robust than the usual ones (MLE, Bayes, MME) to some miss-specifications of the model.