High and Low Frequency Correlations in Global Equity Markets

This study models high and low frequency variation in global equity correlations using a comprehensive sample of 43 countries that includes developed and emerging markets, during the period 1995-2008. These two types of variations are modeled following the semi-parametric Factor-Spline-GARCH approach of Rangel and Engle (2008). This framework is extended and modified to incorporate the effect of multiple factors and to address the issue of non-synchronicity in international markets. Our empirical analysis suggests that the slow-moving dynamics of global correlations can be described by the Factor-Spline-GARCH specifications using either weekly or daily data. The analysis shows that the low frequency component of global correlations increased in the current financial turmoil; however, this increase was not equally distributed across countries. The countries that experienced the largest increase in correlations were mainly emerging markets.