Stock Index Prediction Using the Variance Modeling Approach

Abstract

The importance of stock index forecasting can never be underestimated in the world of investments. However, for a rational investor who is reasonably risk-averse, the ability to forecast volatility in the index is at least as important, if not more, as the capability to predict the return. The current study measured the volatility in the KSE 100 Index using the ARCH-GARCH modeling techniques. Daily figures of the Index were taken from January 2010 to August 2020 leading to 2,660 observations. It was found that KSE 100 Index was characterized by ARCH effects meaning that there were certain periods that were more volatile followed by relatively tranquil periods. In operational terms, the ARCH(4) configuration delivered better results than ARCH(1) or any other ARCH model. The best GARCH model, on the other hand, for the KSE 100 Index was found to be GARCH(1,1). Finally, analysis of the TGARCH and the EGARCH portrayed that asymmetries were there in the positive and negative news for the KSE 100 Index. Hence, bad news was found to have a significantly larger impact on the volatility of the Index than good news.