A standard capital asset pricing model is extended to allow for stochastic shifts in the volatility of the news process. This model is then estimated on bivariate stock market data to separate two exogenous news processes -- a world and a domestic. The results indicate that the influence of the world news process on the Swedish stock market has increased significantly over the period 1970-1995. I also find that the foreign influence is much stronger when the volatility of the world news process is high. Furthermore, when the world state shifts to high risk, the Swedish stock market immediately reacts by a large fall, estimated to 7.0%. The bivariate model is also estimated on a set of other national stock markets. * I thank Magnus Dahlquist, Lars E.O. Svensson and Anders Warne for valuable comments. I am grateful to Karin Russel and the Morgan Stanley Capital International for providing data. # Institute for International Economic Studies, Stockholm University, S-106 91 Stockholm, S...