Forecasting Nigeria’s Crude Oil Prices Using the Box–Jenkins Arima Framework: A Computational Time Series Analysis

Abstract

Crude oil price forecasting is crucial for macroeconomic planning and fiscal stability in Nigeria, where petroleum revenue constitutes a major source of government income and foreign exchange earnings. However, crude oil prices are highly volatile time-series data, exhibiting sharp fluctuations driven by global demand shifts, supply disruptions, and geopolitical uncertainties, making accurate prediction both challenging and necessary for informed economic decision making. This study applies the Box–Jenkins Autoregressive Integrated Moving Average (ARIMA) framework to model and forecast monthly crude oil prices using Eviews and Python-based computational approach. Exploratory analysis was conducted through time-series plots, while the Augmented Dickey–Fuller (ADF) test confirmed the need for differencing to achieve stationarity. Model identification relied on Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) diagnostics, followed by estimation and comparison of candidate ARIMA models. Results indicate that ARIMA (3,1,2) provides the best fit and reliable short-term forecasting performance, demonstrating the effectiveness of both Eviews and Python-driven ARIMA modeling for volatile economic time-series analysis.