A MULTIVARIATE GARCH MODEL FOR EXCHANGE RATES IN THE US, GERMANY AND JAPAN

After the so-called Asia crisis in the summer of 1997 the stock markets were shaken by an increased volatility transmission phenomenon around the world. In this paper we will therefore analyse the daily stock returns in New York, Germany and Japan for a period of 2 years (June 21st, 1996 to June 22nd, 1998). We estimate a VAR-GARCH in mean model investigate the multivariate volatility effects between the time series. We are also interested in the question of whether or not the volatility of the 3 stock returns will feed back on the returns of the stock returns themselves. Using the marginal likelihood criterion for model selection we find a VAR(1)-GARCH(2,2)-M (1) model. The model is estimated using MCMC methods and the coefficients show a quite rich transmission pattern between the stock markets. Comparing the models before and after the Asia crisis we see that the dynamic structure of the VAR model has changed. Keywords: GARCH and VAR-GARCH-M models, MCMC models, posterior, pseudo marginal likeliho ods, model selection.