Comparing the Performance of Alternative Generalised Autoregressive Conditional Heteroskedasticity Models in Modelling Nigeria Crude Oil Production Volatility Series
Keywords:GARCH (p, q), GARCH (o, q), volatility measure.
Communication in Physical Sciences 2019, 4(2): 87-94
Authors: A. E. Usoro, C. E. Awakessien and C. O. Omekara
There is no gainsaying the fact that crude oil production remains a major factor to the Nigeria economic growth given its significant contribution to the nation’s gross domestic product. Preponderance of the researches in the oil sector dwell more on oil prices, with less focus on the volatility of crude oil production. What cannot be overemphasized in oil sector is the production volatility effect which is mostly caused by unstable production quantity due to certain nation’s economic, social, political factors. In this paper, volatility of crude oil production was considered, and different Generalised Autoregressive Conditional Heteroskedasticity (GARCH) models were fitted to Nigeria crude oil production volatility series. Data for the work were monthly crude oil quantity data from January 2010 to August 2019 (NNPC ASB) from which the crude oil production volatility was measured. The suggested GARCH models included GARCH (0,1), GARCH (0,2), GARCH (1,1), GARCH (1,2), GARCH (2,1) and GARCH (2,2). Using Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and Schwarz’s Information Criterion (SIC), GARCH (1,2) and GARCH(2,1) competed favourably. The MSE of forecast revealed GARCH (2,1) to perform better for the forecast of crude oil production volatility. Further findings will reveal other alternative models as the crude oil production pattern changes
in the future.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.