Gauging the Fuel Price Volatility using GARCH Models

Abstract

This study examines the fuel price volatility in Pakistan. The daily time series data for two fuel prices, petrol and diesel, were obtained from January 3, 2010, to October 30, 2023. Auto-Regressive Conditional Heteroskedasticity – a univariate time series econometric model is applied where it was found that ARCH and GARCH effects are significant for both fuel price returns at a 1 per cent significance level. The mean reversion coefficients for petrol and diesel are 0.9076 and 0.8895, respectively, showing that petrol prices observe a slow mean reversion as it is closer to 1 (α+β→1). This is further evident with the half-life analysis, which shows the speed of mean reversion and concludes that petrol prices are more volatile as it has eight days to revert to the mean position compared with six days for diesel.