Forecasting the volatility of the Saudi stock market index

Abstract

This study presents an empirical study of the forecasting performances of a set of ARCH- type models on the Tadawul All Share Index (TASI), some of which are symmetric and asymmetric, to consider the conditional variance and leverage effect. By adopting an econometric point of view, the study aims to (i) model and forecast the volatility of the Saudi stock market, (ii) compar the forecasting performance of the symmetric and asymmetric models, and (iii) test the models' forecasting accuracy prior to, during, and after crises to find an appropriate model for each case. We use the daily closing price data for the TASI from January 2010 to December 2019. in order to take into account the oil crisis that affected the Saudi stock market in 2014, the data were divided into three window periods to evaluate the in-sample and out-of-sample data for volatility fitting and forecasting, the first covered the time before the crisis from January 2010 to June 2014, the second covered the time during the crisis from January 2014 to December 2016, and the third covered the time after the crisis from January 2017 to December 2019. The out-of-sample periods are six months from each period to evaluate the volatility forecasts. The results show that symmetric models are the most suitable for modelling and forecasting TASI volatility for the periods before and during crisis despite the leverage effect, but for the period after a crisis, asymmetric models are the best. We also conclude that there is no preferred model for forecasting TASI volatility since the order of the forecast accuracy of the models was mixed. The results of this study will help narrow the gap in the scarcity of studies on the volatility of the Saudi stock market. They will also be helpful to local and global investors when analysing and making economic and financial decisions related to the Saudi stock market, risk management, asset pricing, and derivatives.