On the volatility estimation of daily price returns of Nigerian Stock Market

Abstract

The stock market is exposed to risk and this risk which is the risk of losses in position brought about the movements on market variables like prices and volatility. Modelling the volatility in daily stock prices entails studying the particular error distribution that is most appropriate for the model. Considering a particular Nigeria stock market, this study estimates both the symmetric and asymmetric volatility models. The ARMA-GARCH, ARMA-EGARCH and ARMA-TGARCH. These models are employed with the error distributions such as normal distribution, student tdistribution and skewed student t- distribution. The ARMA (2,1)-EGARCH (1,1) with student tdistribution was seen to be the most appropriate model. A volatility forecasting accuracy was determined by using the mean absolute scaled error (MASE) to predict the values of the stock market prices for the next 20 years and the result showed that the model was appropriate for predicting volatility. Hence volatility prediction would help in achieving a sound policy decision. R-Code is used to fit the ARMA-GARCH, ARMA-EGARCH and ARMA-TGARCH models (as in the appendix).