Abstract:

Markov switching model is one of the most useful tools to describe the process generating changes in regime. However, a sharp structural break will be appeared sometimes and it is not considered in the Markov switching model. It could be not enough to describe the detailed features of time series only with the single Markov switching model. In order to get a precise description about changes in regime, we introduce the CMR test, which can detect the structural break, to the Markov switching model and call it revised Markov switching model in this paper. We investigate the returns series of Chinese Shenzhen Stock Exchange with the single and revised Markov switching models, and find that the revised Markov switching model is much effective to describe the features of returns series than the single Markov switching model. Furthermore, the empirical analyses indicate that the behavior of returns series changed quite ‘dramatically’ between 2007 and 2008 in Chinese Shenzhen Stock Exchange.