Modeling Price Volatility in Petroleum Products in Ghana*

Abstract

The main purpose of this paper is to analyse and model the price volatility in petroleum products in Ghana. A historical data onmonthly petroleum products (kerosene, gas oil, LPG and premium gasoline) spanning from August,2007 to February, 2016 were used in this study. The methods used were the trend analysis and Generalized Auto-Regression Conditional Heteroscedasticity (GARCH. The results of the analysis revealed that, there was an upward trend existing in all the petroleum products. This implies that the prices of petroleum products increase with time. Seven competing GARCH models were fitted to the returns of the petroleum products. Two models, GARCH (1, 1) and GARCH (1, 2), were considered as the best models by the information criteria. Based on predictive performance, GARCH (1, 2) model was considered as the best model for all the petroleum products. Also, the extent of volatility in the petroleum products, as measured by the coefficients of GARCH(1, 2), indicates volatility persistence (explosive process) and hence the model is adequate. The impacts of news on the petroleum products were tested and it was observed that there were no news impacts in the markets of the petroleum products.