Evidence from Global Markets Based on A Bayesian Analysis

This paper examines the hypothesis that both stock returns and volatility are asymmetric functions of past information derived from domestic and US stock-market news. The evidence finds the presence of negative autocorrelation, which is consistent with the dominance of positive-feedback trading behavior. By employing a double-threshold autoregressive GARCH model to investigate four major index-return series, we find significant evidence to sustain the asymmetric hypothesis of stock returns. Specifically, we find that negative news will cause a decline in national stock returns that is larger than the gain caused by good news of an equivalent magnitude. This also holds true for the conditional variance. The return appears to be more volatile and persistent when bad news hits the market than when good news does. This may be also attributable to the wealth effect argued by Black (1988)