Modeling and forecasting the crude oil demand in Sri Lanka : an econometric approach

Abstract

This study examines the effect of economic variables, Gross Domestic Product (GDP), Foreign Direct Investment (FDI), Population and Oil Price on oil consumption in Sri Lanka using an Error Correction Model. Yearly data of oil consumption, Gross Domestic Product (GDP), Foreign Direct Investment (FDI), Sri Lankan population and crude oil price during the period 1988 – 2013 were used in the analysis. All the data have been obtained by the online data sources of World Bank and United States energy information administration. This research involves estimating the elasticity of Gross Domestic Product (GDP), Foreign Direct Investment (FDI), Sri Lankan population and crude oil price on crude oil consumption in Sri Lanka. Unit root test confirmed that series are not stationary in its levels but they are stationary in first difference. Therefore the study uses the Engle-Ganger cointegation method to create a dynamic short run model. Also Chow - break point test was used to test the significance of a structural break down in the data set and the dummy variable was significant in allowing for the structural change. The Vector Error Correction (VEC) model finds that Gross Domestic Product (GDP), Foreign Direct Investment (FDI), population and oil price are determinants of the oil demand. It shows that in the long run only FDI increases the overall oil demand while GDP and population increase the oil demand in the short run. By using the selected model, oil demand was forecasted and the Mean Absolute Percentage Error (MAPE) of the fitted model was found less than 5 percent. Therefore the fitted model is recommended as the suitable model to forecast oil demand. As the crude oil storage is a common problem in Sri Lanka, forecasting oil demand can be used to find the solutions for the challenges in the petroleum sector.