The term structure of commodity prices.

Summary This dissertation presents the work carried out since the doctoral thesis by placing it in a framework that provides a general view of the literature on the term structure of commodity prices. It begins with a first part devoted to the theoretical analysis of the term structure. In order to understand the relationship between spot and futures prices, I first turn to the traditional theories of commodity prices - the normal discount theory and the storage theory. Originally developed in a context where the notion of term structure was less important, because the maturity of futures contracts was not very distant, these theories appear however to be limited when the horizon of analysis increases and one wishes to take into account the entire term structure. A number of my works therefore aim at extending the traditional theories to the long term. They also aim at improving the understanding of the dynamic behavior of the price curve in a commodity market. The theoretical analysis of the term structure of commodity prices having been carried out, it was interesting to consider how to exploit its teachings for modeling purposes. This is the purpose of the second part of the paper, which presents a new term structure model - called an asymmetric model - justifies the underlying assumptions, and positions it in relation to other models. The third part aims at assessing the quality of the modeling carried out: it focuses on the results of empirical tests carried out on the basis of term structure models. Simulations allow me first to highlight the influence of the assumptions concerning the stochastic process chosen for the state variables and the number of state variables, in three different models. I then present methodological work developed in response to the specific estimation difficulties presented by term structure models of commodity prices. Finally, the performance of the models, i.e. their ability to reproduce the observed price curve, is discussed, allowing us to present the empirical results obtained with the asymmetric model. Two applications have been considered in the literature for term structure models of commodity prices: dynamic hedging and valuation. With respect to hedging, the reflection on the use of term structure models is motivated by the desire to hedge long-term commitments on the physical market with the help of short-term futures contracts. The work then focuses on determining the hedging strategy to be implemented and on analyzing its effectiveness. In the field of valuation, term structure models have been used based on real options theory. Most of the research undertaken has focused on the investment decision in the context of mineral commodities, as the irreversible nature of the investment is particularly pronounced for the latter. My research work has focused on each of these two applications. They have been the subject of several publications. They are also the subject of ongoing research and are at the origin of several projects.