Modeling Asymmetry and Persistence Under the Impact of Sudden Changes in the Volatility of the Indian Stock Market

In this paper, we compare the performance of Inclan and Tiao's (IT) (1994) and Sanso, Arago and Carrion's (AIT) (2004) iterated cumulative sums of squares (ICSS) algorithms by means of Monte Carlo simulation experiments for various data-generating processes with conditional and unconditional variance. In addition, we investigate the impact of regime shifts on the asymmetry and persistence of volatility from the vantage point of modeling volatility in general and, in particular, in assessing the forecasting ability of the GARCH class of models in the context of the Indian stock market. We apply the Iterated Cumulative Sums of Squares (ICSS) algorithm to identify the points of sudden changes in the volatility of the Indian stock market. We find that that when endogenously determined regime shifts in the variance are incorporated in the Generalized Autoregressive Conditional Heteroscedasticity (GARCH) model and GJR-GARCH model, the estimated persistence and asymmetry in the volatility of returns come down drastically. This suggests that ignoring regime shifts in the model may results in an overestimation of the persistence of volatility. In addition, we find that sudden changes in the variance are largely associated with domestic and global macroeconomic and political events. The out-of-sample forecast evaluation analysis confirms that volatility models that incorporate regime shifts provide more accurate one-step-ahead volatility forecasts than their counterparts without regime shifts. These findings have important policy implications for financial market participants, investors and policy makers.