Finding Costs in the U.S. Petroleum Industry: Assessing the Opposing Effects of Technological Change and Depletion with Error Correction Modeling

A common claim in the nonrenewable resource literature is that improvements in technology may largely offset the effects of increasing scarcity over time. Along with Cuddington and Moss (CM) (1999), this paper provides perhaps the first empirical evidence on this issue by analyzing the determinants of the average finding cost for additional U.S. petroleum reserves. Both analyses cover the1967-90 period and use a new technology index constructed by Moss (1993). This paper proposes a new approach for estimating average finding costs in a model with depletion and/or technology variables. Earlier estimations of finding cost functions are possible spurious regressions because they use regression techniques that are inappropriate in the presence of nonstationary variables. This paper uses error correction modeling (ECM) to avoid the possible spurious regression problem. With the new technology variable, it is possible to isolate the separate effects of depletion and technological improvement on finding costs for additional proven reserves. The ECM facilitates counterfactual simulations for average finding costs in a scenario with no technological improvement while recognizing that annual reserve additions in this scenario might be higher or lower than they would have been with ongoing technological improvements. The simulations suggest that technological advance has been very significant in offsetting what would otherwise have been sharply rising costs for additions to U.S. natural gas reserves. Technology's impact on crude oil finding costs have been more modest, presumably because that segment of the industry is more mature.