Non-renewable resource price forecasting: a comparison of methods

Abstract

Dramatic changes in U.S. coal prices during the 1970's and 1980's have brought into question the ability of currently available forecasting models to predict non-renewable resource prices. This thesis compares two types of forecasting models used to predict non-renewable resource prices. Each model is assessed based upon theoretical and practical considerations. The models evaluated are a conventional time series model and an optimization model.

The thesis finds each model to have inherent advantages and disadvantages within the areas of comparison. The conventional time series model is assessed as relatively simple to develop and easy to use for rough approximations of future prices, but lacking in its reflection of the current body of economic theory on non-renewable resources. The optimization model is evaluated as more thoroughly embracing some of the advances made in the theory, but requiring an extensive commitment of resources for its development.