Linkages between European National Stock Markets During Trading and Non-Trading Hours

The paper documents differences in the structures of dependence between stock markets during trading and non-trading hours. We study this issue on the basis of analysis of the dynamic dependence between five selected European stock markets during the period from 1997 to the beginning of 2013. The markets are represented by their main stock indices. The analysis is performed by means of Markov-switching copula models with three regimes. The adopted approach enables us to avoid limiting ourselves to elliptical distributions for the bivariate returns, and allows to assess dependence in tails of the bivariate distributions. Moreover, due to the use of copulas with Markov-switching there is no necessity to make any a priori assumptions concerning the time or cause of a change in the dependence structure