1 Structural Models of Commodity Prices Craig Pirrong 1.1 INTRODUCTION The transparency of fundamentals in commodity markets (in contrast to equity or currency markets, for instance) holds out the promise of devising structural models of commodity price behavior that can illuminate the underlying factors that drive these prices, and which perhaps can be used to value contingent claims on commodities. There has been much progress on these models in recent years, but the empirical data show that real-world com- modity price behavior is far richer than that predicted by the current generation of models, and that except for non-storable commodities, structural models currently cannot be used to price derivatives. The models and empirical evidence do, however, point out the deficien- cies in reduced form commodity derivative pricing models, and suggest how reduced form models must be modified to represent commodity price dynamics more realistically. They also suggest additional factors that may be added to the models (at substantial computational cost) to improve their realism. This chapter sketches out the current state of fundamental models of commodity markets. It starts with a taxonomy of commodities, and then proceeds to discuss models for storable and non-storable commodities, and structural models for each. 1.2 A COMMODITY TAXONOMY Although the catchall term “commodity” is widely applied to anything that is not a true asset, it conceals tremendous diversity, diversity that has material impacts on price behavior and modeling. The most basic divide among commodities is between those that are storable, and those that are not. The most important non-storable commodity is electricity (although hydro generation does add an element of storability in some electricity markets). Weather is obvi- ously not storable – and it is increasingly becoming an important underlying in commodity derivatives trading. Most other commodities are storable (at some cost), but there is considerable heterogeneity among goods in this category. Some are continuously produced and consumed, and are not subject to significant seasonality in demand; industrial metals such as copper or aluminum fall into this category. Some are continuously produced and consumed, but exhibit substantial Craig Pirrong, University of Houston, Houston, TX 77204. Tel: 713-743-4466. Email: cpirrong@uh.edu Risk Management in Commodity Markets: From Shipping to Agriculturals and Energy Edited by H´ elyette Geman c  2008 John Wiley & Sons, Ltd COPYRIGHTED MATERIAL