Multivariate volatility analysis of VW stock prices

Multivariate GARCH{type models are proposed which allow for asymmetry and size eeects in the conditional variance{covariance matrices of a system of time series variables. These models are applied for analyzing the properties of VW common and preference stock prices. The relationship between the conditional variances of these variables is investigated using suitable impulse responses or conditional moment prooles. It is found that there is a clear asymmetry in the volatility of the series which are seen to react quite diierently to positive and negative shocks in the market. Also some diierences in the reactions of preference and common stocks are uncovered. No signiicant evidence is found for size eeects, that is, the way the variables respond to unexpected shocks in the market depends more on the sign of the shocks than on their size.