Macroeconomics Uncertainty and Performance of GARCH Models in Forecasting Japan Stock Market Volatility
Wei-Chong Choo, See-Nie Lee, Sze-Nie Ung

Abstract
Since the introduction of ARCH/GARCH methods have been widely examined. However, the role played by macroeconomics environment in volatility forecasting has been ignored. This paper investigates the behavior of Japanese stock market volatility with respect to a few macroeconomic variables including gold price, crude oil price and currency exchange rates (Yen/US$). A comparison study has also been carried out on the performance of GARCH models and Ad Hoc methods. This empirical study employs the daily data over 12 years. The result reviews that macroeconomic variables used in this study have no impact on the volatility of Japanese stock markets and the simplest GARCH (1,1) model yields the best result. Further comparison on the best performing model suggest that GJRGARCH (1,1) model is superior to GARCH (1,1) model in one-step ahead forecast.

Full Text: PDF