A Monte Carlo comparison of Bayesian testing for cointegration rank

This article considers a Bayesian testing for cointegration rank, using an approach developed by Strachan and van Dijk (2007), that is based on Koop, Leon-Gonzalez, and Strachan (2006). The Bayes factors are calculated for selecting cointegrating rank. We calculate the Bayes factors using two methods - the Schwarz BIC approximation and Chib's (1995) algorithm for calculating the marginal likelihood. We run Monte Carlo simulations to compare the two methods.