Year: 2025
Type: article
Abstract: The global crude oil market is influenced by geopolitical, supply-demand, and financial factors with important macroeconomic considerations. This research used a multivariate linear regression and an ARIMA time series model to explore past price behavior and forecast short-term trends. The primary variables involved are the macroeconomic variables, financial indicators, and measure of geopolitical risk. After addressing stationarity, this paper found the selected ARIMA model performed well and predicted generally decreasing oil prices one year ahead, with increasing confidence intervals around those predictions reflecting increasing uncertainty concerning future prices. Residual diagnostics support the adequacy of the model, but the model is constrained by structural breaks (e.g., financial crises, pandemic shocks) in data, and the omission of relevant exogenous variables. The results of the analysis reaffirm that the long-term price behavior of crude oil is driven by supply-demand fundamentals but highlight the pressure for hybrid models with machine learning algorithms that account for nonlinear relationships and structural breaks. The research provides actionable information for both policymakers and investors as they navigate volatile markets with important geopolitical risk factors.
Source: Theoretical and Natural Science
Author: Ruen Ge
Institution: University of Oxford
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W4410128779