FORECASTING PERFORMANCE OF SOME GARCH MODELS ON HOLIDAY-INDUCED VOLATILITY IN NIGERIA STOCK EXCHANGE PRICE RETURNS UNDER DIFFERENT ERROR DISTRIBUTIONS

Abstract

Time series data occasionally depend on factors among which are holidays (such as Mother’s days, Children’s days, Democracy days, Independent Days, Valentine’s days to mention but few) which number of researchers did not put into consideration. This paper aimed at evaluating the forecast performance of some asymmetry GARCH models (EGARCH, GJR-GARCH, and APARCH) on holiday-induced volatility in Nigeria stock exchange price returns under three different error distributions of innovation: Normal, Skewed student’s t, and Generalized Error Distribution (GED). Based on minimum value of Root Mean Square Error (RMSE), EGARCH (1,1) model under Skewed student’s t is found to be the best model. In addition, there exists consequences of all the holiday’s that falls on Thursday’s (with effect 0.002803; indicating that for any unit of holiday on Thursday(s), the volatility of NSE price series returns will significantly increase by 0.002803). Volatility clustering and persistence are found in the models. More so, leverage effect is found in EGARCH model under the three error distributions of innovation.