Variance clustering improved dynamic conditional correlation MGARCH estimators

It is well-known that the estimated GARCH dynamics exhibit common patterns. Starting from this fact the Dynamic Conditional Correlation (DCC) model is extended by allowing for a clustering structure of the univariate GARCH parameters. The model can be estimated in two steps, the first devoted to the clustering structure, and the second focusing on dynamic parameters. Differently from the traditional two-step DCC estimation, large system feasibility of the joint estimation of the whole set of model’s dynamic parameters is achieved. A new approach to the clustering of GARCH processes is also introduced. Such an approach embeds the asymptotic properties of the univariate quasi-maximum-likelihood GARCH estimators into a Gaussian mixture clustering algorithm. Unlike other GARCH clustering techniques, the proposed method provides a natural estimator of the number of clusters.