Modeling covariance breakdowns in multivariate GARCH

This paper proposes a flexible way of modeling dynamic heterogeneous covariance breakdowns in multivariate GARCH models through a stochastic component that allows for changes in the conditional variances, covariances and implied correlation coefficients. Different breakdown periods will have different impacts on the conditional covariance matrix and are estimated from the data. We propose an efficient Bayesian posterior sampling procedure and show how to compute the marginal likelihood. Applied to daily stock market and bond market data, we identify a number of different covariance breakdowns which leads to a significant improvement in the marginal likelihood and gains in portfolio choice.