We apply univariate GARCH models to construct a computationally simple filter for estimating the conditional correlation matrix of asset returns. The proposed Variance Implied Conditional Correlation (VICC) exploits the polarization result that links the correlation between two standardized variables with the variances of linear combinations thereof. In a Monte Carlo study, we show that the VICC yields accurate correlation estimates for common choices of the correlation dynamics. We also provide an empirical application to cross hedging that confirms the effectiveness of the VICC.

Original languageEnglish
Pages (from-to)200-222
Number of pages23
JournalThe European Journal of Finance
Volume26
Issue number2-3
DOIs
Publication statusPublished - 11 Feb 2020

ID: 47468638