A dynamic hedging approach for refineries in multiproduct oil markets

A multiproduct portfolio hedge ratio strategy for oil futures is investigated using a multivariate GARCH model based on dynamic conditional correlation and an error correction model (DCC-ECM-MVGARCH). By considering the characteristics of refiner profits from crack spread and the mutual relations among crude oil, gasoline and heating oil spot and future prices, we estimate the time-varying optimal hedge ratios for the oil-cracking margin. In addition, a naïve strategy, a traditional OLS model and dynamic B-GARCH model are selected to compare with our model for hedge effectiveness. Comparison of hedge effectiveness for in-sample and out-of-sample data reveals that the dynamic DCC-ECM-MVGARCH model is more sensitive to market fluctuations, provides a more accurate description of changes in volatility and has more advantages than other models. Therefore, the empirical results prove that application of the DCC-ECM-MVGARCH model for hedging of oil market portfolio can play an important role in avoiding the double risk of crude oil and oil product markets for refineries.

Research highlights
► From the perspective of refineries, multiproduct hedging strategies are investigated by selecting the spread for crude oil, gasoline and heating oil as a representative measure of cracking profits.
► For the first time, multivariate GARCH based on dynamic conditional correlation and an error correction model (DCC-ECM-MVGARCH) is applied to estimate dynamic multiproduct hedge ratios for oil markets.
► By comparing the in-sample and out-of-sample hedge effectiveness with a naïve strategy, a traditional OLS model and a B-GARCH model, it is demonstrated that the DCC-ECM-MVGARCH model is more suitable for refining operations and can thus reduce potential market risk.