Bayesian methods for Markov switching time series models

Economical as well as financial time time series tend to exhibit structural changes which occur at time points which are unknown a priori. We model these strucutral breaks by including a hidden Markov process indicating the presence or absence of a structural break. Estimation is carried out by MCMC methods. We investigate in detail an important application of these model to Markov-Switching ARCH-models.