Comparing the Performances of GARCH-type Models in Capturing the Stock Market Volatility in Malaysia

Abstract

We conduct empirical analyses to model the volatility of stock market in Malaysia. The GARCH type models (symmetric and asymmetric GARCH) are used to model the volatility of stock market in Malaysia. Their performances are compared based on three statistical error measures tools, i.e. mean squared error, root means squared error and mean absolute percentage error for in sample and out sample analyses. Apart from that, we also determine the factors contribute to the stock market movements. The data is ranging from January 1990 to December 2010. The data is divided into three time frames, i.e. pre-crisis 1997, during crisis and post-crisis 1997. Our results reveal that symmetric and asymmetric GARCH models have different performances in different time frames. In general, for the normal period (pre and post-crisis), symmetric GARCH model perform better than the asymmetric GARCH but for fluctuation period (crisis period), asymmetric GARCH model is preferred. Our results also show that exchange rate and crude oil price have significant impacts on the Malaysia stock market volatility in the pre-crisis and post-crisis periods and but the impact is not significant in the crisis period.