Bayes inference in common Markov switching trends models

This paper proposes a Bayesian approach to estimate a common trends model in which the growth rate of the common stochastic trend is subject to regime shifts. By applying the Gibbs sampler we can draw from the conditional posterior distributions and evaluation of the nontrivial likelihood function is avoided. Examples with simulated and real data demonstrate the practicality and e ciency of the method. key words: common trends, Markov switching, Markov Chain Monte Carlo Method JEL Classi cation: C11, C15, C32, E32 University of Vienna, Department of Economics, Hohenstaufengasse 9, 1010 Vienna, Phone ++431 4277 374-30, email [email protected] I thank Sylvia Fr uhwirth-Schnatter for many helpful comments and discussions and Klaus Neusser for helpful comments. All remaining errors or omissions are mine.