Grouping Stock Markets with Time-Varying Copula-GARCH Model

The aim of this work is to find the dynamics of interdependencies and similarities between European, American and Asian stock markets. The investigation covers daily returns of 36 market indices. In order to examine the dependencies between these data, the Markov regime switching copula model with two regimes is considered. For the dynamic clustering purposes, the time varying Spearman ratio obtained from the regime switching copula model is taken to construct the dissimilarity measure between any two markets. To demonstrate the dynamics of the changes, three sub-periods are considered: the period before the global financial crisis (from October 2002 to July 2007), the period of the crisis itself (from July 2007 to December 2008) and the post-crisis period (from January 2009 to April 2012). Taking dynamical relationships into account, all stock markets can be divided into four clusters: North and South America, Western Europe, Eastern Europe and Asia. However, in each of these main clusters similarities between financial markets vary with time.