A Data-Driven Approach for Modeling Stochasticity in Oil Market

TL;DR

This paper introduces a System Dynamics-based model for oil prices that incorporates expectations about future demand and supply changes, addressing limitations of traditional models that focus only on historical data.

Contribution

It presents a novel modeling approach that includes non-traditional factors like expectations, improving the accuracy of oil market predictions.

Findings

Model aligns well with real data trends

Incorporates expectations into stochastic modeling

Enhances understanding of oil price fluctuations

Abstract

Global oil price is an important factor in determining many economic variables in the world's economy. It is generally modeled as a stochastic process and have been studied through different techniques by comparing the historic time series of demand, supply and the price itself. However, there are many historic events where the demand or supply changes are not sufficient in explaining the price changes. In such cases, it is the expectations on the future changes of demand or supply that causes heavy and quick influences on the price. There are many parameters and variables that shape these expectations, and are usually neglected in traditional models. In this paper, we have proposed a model based on System Dynamics approach that takes into account these non-traditional factors. The validity of the proposed model is then evaluated using real and potential scenarios in which the proposed model follows the trend of the real data.