Multivariate Modelling of Cross-Commodity Price Relations Along the Petrochemical Value Chain

Abstract

We aim to shed light on the relationship between the prices of crude oil and oil-based products along the petrochemical value chain. The analyzed commodities are tied in an integrated production process. This characteristic motivates the existence of long-run equilibrium price relationships. An understanding of the complex price relations between input and output products is important for petrochemical companies, which are exposed to price risk on both sides of their business. Their profitability is linked to the spread between input and output prices. Therefore, information about price relations along the value chain is valuable for risk management decisions. Using vector error correction models (VECM), we explore cross-commodity price relationships. We find that all prices downstream the value chain are cointegrated with the crude oil price, which is the driving price in the system. Furthermore, we assess whether the information about long-run cross-commodity relations, which is incorporated in the VECMs, can be utilized for forecasting prices of oil-based products. Rolling out-of-sample forecasts are computed and the forecasting performance of the VECMs is compared to the performance of naive forecasting models. Our study offers new insights into how economic relations between commodities linked in a production process can be used for price forecasts and offers implications for risk management in the petrochemical industry.