FORECASTING WORLD CRUDE OIL PRICE: USING ARIMA MODEL

Abstract

A serious concern has been risen much interest into the investigation of its price swing as well forecasting. The study used the dynamics of monthly Brent oil price from Nov 1994 to Dec 2011.  The data grouped into two parts. The first twelve years used for the model construction and the next twelve months (Jan 2012 –Dec 2012) used for validating forecasting accuracy. They were subject to log transformation as well differencing to make stationary, besides testing of autocorrelation and residual analysis to determine among family of ARIMA models.  ARIMA (2,0,1) was fitting well for forecasting the volatilities price. The study suggested choosing alternative models such as ARCH and GARCH models to have a best accuracy of forecasting oil prices due to prevailing outliers.