Priors and Bayesian Parameter Estimation of Affine Term Structure Models

This article investigates problems arising with near unit root behavior and different market micro-structure noise assumptions for affine term structure models. We show that with increasing serial correlation the Fisher information matrix approaches a singularity, such that the standard deviation of the level parameter is going to explode. We apply Markov Chain Monte Carlo simulation techniques in connection with regularized priors, as proposed in Schotman and vanDijk (1991) and Pooter et al. (2006) to simulate the joint posterior distribution of the model parameters