Abstract

In this paper, we establish an empirical model for oil production in the lower 48 states that represents its economic, physical, and institutional determinants. We estimate a vector error correction model for oil production in the lower 48 states that specifies real oil prices, average production costs, and prorationing by the Texas Railroad Commission. These modifications enable us to generate a model that accounts for most of the variation in oil production in the lower 48 states between 1938 and 1991. The inconistencies between the historical record and the assumptions that underlie economic and Hubbert models imply that these models cannot be used to forecast future supplies. Furthermore, these inconsistencies indicate that the ability of a competitive market to make a smooth transition from oil to alternative energy sources may be less effective than currently believed.