Modelling and Forecasting of Crude Oil Price Return Volatility from 2006-2023: An Application of the Garch Models

Abstract

The volatility dynamics of Nigerian crude oil prices from 2006 to 2023 were investigated in this work using the generalized autoregressive conditional heteroskedasticity (GARCH) modeling technique. Strong evidence of time-varying conditional heteroskedasticity in the dataset was found using diagnostic tests. The model information criteria showed that the dependencies were well represented by the parsimonious GARCH (1,1) formulation. Based on maximum log-likelihood and information criterion values, exponential GARCH (EGARCH) showed the best in-sample fit. The projection predicts prices beginning at a very high level in 2024 and then rapidly falling, with a steep downward slope in the expected price trajectory. The model predicts that by the end of 2025, crude oil prices will be significantly lower than they were at the start of the projection period. Overall, our empirical findings give good support for using EGARCH techniques to forecast volatility in Nigerian crude oil returns. As a result, practitioners now have access to efficient prediction tools that have been verified using rigorous statistical approaches to help them estimate future price risk.