Dynamic Asset Correlations Based on Vines

We develop a new method for generating dynamics of conditional correlation matrices between asset returns. These correlation matrices will be parameterized by a subset of their partial correlations,whose structure will be described by an undirected graph called 'vine.' Since such partial correlation processes can be specified separately, our approach provides very flexible and potentially parsimonious multivariate processes. We introduce the so-called 'vine-GARCH' class of processes and describe a quasi-maximum likelihood (QML) estimation procedure. Compared to other usual techniques, particularly for the Dynamic Conditional Correlation family, inference is simpler and can be led equation per equation. We compare our models with some DCC-type specifications through some simulated experiments and we evaluate their empirical performances by exploiting a database of daily stock returns