Bayesian Estimation of Multivariate Stochastic Volatility Modeled by Wishart Processes

This paper picks up a model developed by Philipov and Glickman 2006 for modeling multivariate stochastic volatility via Wishart processes. The implementation and estimation of the parameters using Bayesian techniques is analyzed profoundly with a simulation study. Special focus is put on the estimates of the time-dependent covariance matrix. Employing Gibbs sampler in combination with Metropolis Hastings algorithm the estimation of the latent variables is feasible and appropriate with respect to statistical properties. Still further research is necessary, however the preliminary findings are very promising.