Modeling global and local dependence in a pair of commodity forward curves with an application to the US natural gas and heating oil markets

The goal of this paper is to present a model for the joint evolution of correlated commodity forward curves. Each forward curve is directed by two state variables, namely slope and level, and the model is meant to capture both the local and global dependence structures between slopes and levels. Our framework can be interpreted as an extension of the concept of cointegration to forward curves. The model is applied to a US database of heating oil and natural gas futures prices over the period February 2000-February 2009. We find the long-run slope and level relationships between natural gas and heating oil markets, analyze the lead and lag properties between the two energy commodities, the volatilities and correlations between their daily co-movements and evaluate the robustness of these observations to the turmoil experienced by energy markets since 2003.