MODELLING AND FORECASTING OF NIGERIAN CRUDE OIL PRICES

USING BOX–JENKINS TECHNIQUE

Yakubu .M. Yeldu¹, Shehu .L.² and ³Mukhtar .G.

^{1, 2,3}Department of Statistics, College of Science and Technology

Waziri Umaru Federal Polytechnic, Birnin Kebbi, Kebbi State, Nigeria.

E-mail: yeldouyakubu@gmail.com, yeldouykb@yahoo.com

Abstract

Box – Jenkins modelling approach has been applied for the time series analysis of Weekly (Forcados, Nigeria) Spot Price FOB (Dollars per Barrel) from August 07, 2000 to September 02, 2013. Basic statistical properties of these series were investigated. After taking the first order difference the time series seems to be stationarity. Autocorrelation and partial autocorrelation plots were used to make tentative identification of the form and order of Box – Jenkins Autoregressive Integrated Moving Average (ARIMA) models. Initially several non – seasonal ARIMA models were postulated for further analysis. These models were then estimated for their adequacy based on the significance of the parameter estimates, mean square and Box – Pierce (Ljung – Box) statistics. Based on these criterion ARIMA (2, 1, 2) model fitted adequately and was also used for short term forecasting. The estimated model is: .