Content area
Abstract
The Rational Extraction Model (REM) and the Standard Commodity Model (SCM) are two theoretical approaches to resource market models. The REM focuses on the long-run aspects of resource markets. The SCM concentrates on the short-run features of the markets. This study develops an econometric model by integrating the REM and the SCM to present the full picture of resource markets: the depletion of resources, the adjustment in capacity, as well as the accumulation of inventories.
Two models are estimated in the context of world copper market. These are the "fixed coefficient" model and the "varying coefficient" model. The fixed coefficient model is the traditional econometric model, which assumes that independent variables have a constant effect on the dependent variable. The varying coefficient model allows the effects of independent variables on the dependent variable to vary over time. Both models have the same specification and are based on dynamic disequilibrium theory. The fixed coefficient model is estimated by the OLS method; the varying coefficient model is estimated by Kalman filter technique. The dynamic simulation is performed using the Gauss-Seidel framework.
The models are validated using parametric and nonparametric methods. The models are then used for sensitivity analysis and forecasting analysis. The sensitivity analysis is conducted to examine the impacts of stock adjustments on mineral price responses. The forecasting analysis is conducted to compare the forecasting capabilities of the fixed and varying coefficient models. The results confirm the existence of stock effects on mineral price variation. The varying coefficient model proves to perform better than the fixed coefficient model during the sample and ex post forecast periods, which indicates that it has more potential to be used for forecasting purposes.
