Volatility Threshold Dynamic Conditional Correlations: An International Analysis

Abstract

This paper proposes a modeling framework for the study of changes in cross-market comovement conditional on volatility regimes. Methodologically, we extend the Dynamic Conditional Correlation multivariate GARCH model to allow the dynamics of correlations to depend on asset variances through a threshold structure. The empirical application of our model to a sample of international stock markets in 1994-2011 indicates that the periods of market turbulence are associated with an increase in cross-market comovement. The modeling framework proposed in the paper represents a useful tool for the study of market contagion.

Keywords: Dynamic Correlations, Volatility Thresholds, Comovement, Contagion

JEL Classification: C50, F37, G11, G15