The Role of Autocorrelation to Volatility of Stock Markets during the Asian Crisis

In this paper we explore the persistent pattern how to impact volatility in stock market. We use TAR-ANST GARCH model to obtain a better approximation of the volatility pattern. An extension to the ANST GARCH model with the threshold variable is proposed and estimated. In contrast to the general regime-switching model focusing on the shock mean equation or assigned the threshold variable as residual term which is set zero as threshold value, we consider asymmetric response in volatility to the stock return autocorrelation and apply threshold space to refine volatility equation. It is found that different levels of autocorrelation are related to stock return volatility. The volatility is getting the less information from innovation term and previous volatility under the case stock return presents persistent property. We observe stock return in 10 countries before, on, and after 1997 Asian financial crisis. The result reveals that a not significant difference exists on the crisis