Dynamic correlations and domestic-global diversification

Abstract

This study designs a generalized dynamic conditional correlation (DCC) model to examine three dynamic elements involved in the correlations between domestic and world stock markets. Further, we apply this generalized DCC model to the domestic-global portfolio establishment and a comparative analysis with the conventional CCC (constant conditional correlation) is conducted. Our empirical results are consistent with the following notions. First, the domestic and global markets are strongly correlated when both simultaneously find themselves in the same state of volatility. Conversely, this correlation is weaker when a different volatility state characterizes each market. Second, persistence and correlation clustering are presented in domestic-global markets, whereas this self-dependent process involved in cross-market correlations is insignificant in certain markets. Third, market integration leads to an increase in co-movement between domestic and global markets. Finally, although the domestic-global portfolio effectiveness tests demonstrated the superior performance of our generalized DCC model in comparison with the conventional CCC and DCC models, we further indicate that the benefits mainly rely on reductions in risk, rather than increases in mean returns.