EMPIRICAL DISTRIBUTION AND MODELLING OF SPOT PRICES OF NIGERIA CRUDE OIL
J. F. Ojo and R.O. Olanrewaju
Department of Statistics,
University of Ibadan, Ibadan, Nigeria.
E-mail: jfunminiyiojo@yahoo.co.uk
Abstract: This paper attempts to provide a suitable time series model for daily West Texas Intermediate (WTI) Spot prices for Nigeria crude oil having determined the appropriate probability distribution for the series. The identified probability distribution for the series formed an integral part of the model. Two thousand eight hundred and eighty (2880) sample points of the daily returns of the WTI Spot prices from June 14,2005 to October 17,2016 were considered. Sixty-one (61) probability distributions were fitted to the series. Generalized Error Distribution (GED) emerged as the ideal distribution that explained the abstraction of reality for the return series due to its smallest test statistic(0.04337, 9.1145, and 158.62) for Kolmogorov-Smirnov, Anderson-Darling and Chi-Squared, respectively, from all the considered distributions goodness-of-fittests. The Generalized Auto Regressive Conditional Heteroscedasticity (GARCH) volatility model with GED as the error term (GED-GARCH)was used to capture the leverage effects. GED-GARCH(1,1)with minimum Akaike Information Criteria(AIC),Bayesian Information Criteria(BIC),and Hannan and Quinn Information Criteria(HQIC) calculated tobe (7.573, 7.583, 7.577), respectively, was found better to explain the fluctuation characterized by the WTI Spot prices. It was discovered that 19.2% of the present variance shock(either positively or negatively) was realized in succeeding period, and that the volatility clustering was the major leverage effects expect for years between 2007 and 2008; and 2014 and 2015 that experienced sharp deviation of leverage effects. The study concludes that GED is the appropriate distribution. We recommend GED-GARCH for modeling Spot Prices of Nigeria crude oil because of its ability to capture fluctuation inherent in Spot prices of Nigeria crude oil.
Keywords: Crude Oil Spot Prices, Leverage Effect, Volatility Clustering, Generalized Error Distribution, West Texas Intermediate.