Assessing the Performance of Zero-Truncated Count Distributions in Modelling Brent Crude Oil Price in Nigeria
B. E. Omokaro
Department of Statistics, Delta State Polytechnic, Otefe, Oghara, Delta State, Nigeria.
C. O. Aronu *
Department of Statistics, Chukwuemeka Odumegwu Ojukwu University, Uli, Anambra State, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
This study examines the suitability of Zero-Truncated Count Distributions (Zero-Truncated Poisson (ZTPP) and Geometric-Zero-Truncated Poisson (GZTP)) for modelling the variability in Brent Crude Oil prices. Using secondary data from the Central Bank of Nigeria's Statistical Bulletin, the descriptive statistics of Brent Crude reveal a mean price of $47, with a standard deviation of $32, indicating significant price volatility. The study compares the performance of GZTP and ZTPP distributions in terms of model fit and predictive accuracy, using metrics such as the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Mean Squared Error (MSE). Results show that ZTPP consistently outperforms GZTP in terms of AIC and BIC, indicating a better fit, while both models exhibit similar MSE values. The analysis also highlights the impact of sample size on model performance, with MSE decreasing as data dimension increases, but stabilizing beyond a sample size of 200. The findings suggest that ZTPP offers a more reliable trade-off between model complexity and accuracy, making it a more effective choice for forecasting Brent Crude price variability, particularly in economic and financial modelling.
Keywords: AIC, BIC, MSE, modelling, model fit, predictive accuracy, sample size