Testing for Structural Breaks in Correlations: Does it Improve Value-at-Risk Forecasting?

Abstract

In this paper, we compare the Constant Conditional Correlation (CCC) model to its dynamic counterpart, the Dynamic Conditional Correlation (DCC) model with respect to its accuracy for forecasting the Value-at-Risk of financial portfolios. Additionally, we modify these benchmark models by combining them with a pairwise test for constant correlations, a test for a constant correlation matrix, and a test for a constant covariance matrix. In an empirical horse race of these models based on five- and ten-dimensional portfolios, our study shows that the plain CCC- and DCC-GARCH models are outperformed in several settings by the approaches modified by tests for structural breaks in asset correlations and covariances.