Resources for the Future
Krautkraemer and Toman
6.3 Models of Depletable Energy Supply Behavior
Behavioral models of depletable energy or other resource supply seek to develop empirical counterparts to theoretical efficiency conditions like (4) through (6) in order to describe how supply decisions (reserve development and extraction) respond to economic incentives (prices and cost drivers). Much of this work has focused on oil and gas supply, though some applications to mineral resources (e.g., nickel) also can be found in the literature. Relatively little of this kind of modeling work has been done for coal, since the depletion effects central to the development of the dynamic efficiency conditions discussed above seem to be of very limited relevance to coal supply (with large reserves and continued technical progress in resource exploitation, see Darmstadter 1999).
The problem of developing empirical behavioral models based on the theory of efficient depletable energy supply is quite difficult for several reasons. One problem, noted above, is the poor data on reserves and costs. But even with better data, it is inherently complicated to develop reliable empirical models for complex dynamic process, and resource depletion is inherently dynamic. Moreover, because behavior depends so much on expectations of unknown future influences in this dynamic framework, the empirical analyst must formulate a hypothesis about the nature of the expectations driving behavior.
Finally, the theory we have discussed previously applies most directly to an individual decisionmaking unit and perhaps even to decisions within a specific resource-bearing location. Once the analysis is extended to cover multiple extraction and investment decisions from heterogeneous deposits, some difficult aggregation problems arise. Care is needed not to omit or mix together key influences. For example, the yield from exploratory drilling likely depends on both how much the best resource prospects have been depleted by past activity and what advances in drilling technique can do to raise discoveries per unit of effort.
Basically, two types of modeling approaches are found in the literature: process optimization simulation models and econometrically estimated models. Process optimization simulation models specify and numerically solve sets of equations based on first-order efficiency conditions