16 - Lessons about commodity markets and modeling strategies  pp. 455-464


By Jeffrey C. Williams and Brian D. Wright

Image View Previous Chapter Previous Chapter



In this book, we have embodied what we consider to be the crucial stylized facts about technology in an internally consistent marketwide model of a commodity. We have used this model and its extensions to derive new insights regarding market behavior and the effects of policy interventions. Our approach differs from previous work in that it highlights dynamics. Proper consideration of storage impels a distinction between dynamics and comparative statics; all welfare effects, for example, are properly dynamic phenomena. Also the time-series behavior resulting from storage, especially the tendency for positive serial correlation, is inherently interesting.

More specifically, we have set forth and probed a multiperiod model with random weather in which storage connects the periods. The storage itself is relatively simple, the aggregate behavior of the many risk-neutral storers in a competitive industry with constant marginal costs over the whole range of positive storage. In each period, the market is in equilibrium, given the current weather. Such a formulation is firmly in the tradition of microeconomics, with an emphasis on the “whole behaving differently from the parts.” We add two features to this standard industry of microeconomic theory, however. One is that the storers are forward looking - indeed, that their expectations about future prices are rational in the sense of being internally consistent. The other is that aggregate storage cannot be negative. The interaction at the level of the whole industry of this simple and obvious technical feature with the random weather, not to mention the anticipation of such interactions in the future, is sufficient to create a complex dynamic system, even without any seasonality or long-run trends.