MODELLING AND FORECASTING DAILY RETURNS VOLATILITY OF NIGERIAN BANKS STOCKS

Abstract

This study models and forecast daily return volatility of Nigerian bank stocks. Data on daily closing prices for fifteen Nigerian banks were collected between 4th January, 2005 and 31st August, 2012. Daily returns series were then computed for each bank from price, stationarity of the resulting series and normality were tested. Different autoregressive models were fitted for the mean equation. From the mean equation, ARCH effect was tested using Lagragian Multiplier test. To capture the volatility pattern, three symmetric models which are ARCH(1), ARCH(2) and GARCH(1,1) and two asymmetric models EGARCH(1,1) and TARCH(1,1) were considered.. Post estimation and performance evaluation metric was done using the RMSE, MAE and MAPE. The results showed that the return series were stationary but not normally distributed with presence of ARCH effect. Furthermore, results of post estimation revealed that these models were competitive. However, EGARCH (1, 1) predicted daily return volatility of majority of Nigerian bank stocks compare to other volatility models considered. This is an indication of the suitability of asymmetric volatility models compared to symmetric models.