ABSTRACT
Energy elasticities seem to be the dominating tool for forecasting energy demand. This chapter explores the microeconomic foundations of energy elasticities and their special problems for econometric estimation. A look into the technological structure of the energy sector, however, suggests that energy elasticities may be a dangerous shortcut for drawing conclusions about long-term developments for energy demand.
An analysis of the forecasting properties of microeconomic energy demand models with econometrically estimated parameters reveals theoretical deficiencies, instability of the estimated parameters and poor forecasting quality over longer forecasting periods. This class of forecasting models is compared with technology-oriented structural models of the energy sector. The main emphasis in these models is given to the identification of technological options for improving the efficiency of the transformation of primary energy to end-use energy and useful energy and the substitution of useful energy and capital for obtaining specific energy services. Finally a statistical methodology is proposed which combines the two approaches taking into account the relative merits of the two types of models depending on the forecast horizon.
