Theoretical and Empirical properties of Dynamic Conditional Correlation Multivariate GARCH

摘要: In this paper, we develop the theoretical and empirical properties of a new class multivariate GARCH models capable estimating large time-varying covariance matrices, Dynamic Conditional Correlation Multivariate GARCH. We show that problem conditional variance estimation can be simplified by univariate for each asset, then, using transformed residuals resulting from first stage, correlation estimator. The standard errors stage parameters remain consistent, only need modified. use model to estimate up 100 assets S&P 500 Sector Indices Dow Jones Industrial Average stocks, conduct specification tests estimator an industry benchmark volatility models. This demonstrates very strong performance especially considering ease implementation