Abstract
A fundamental problem occurring in the analysis of energy demand concerns the aggregation of the fuel input. Since fuel is consumed to produce heat, the appropriate units of measurement are heat units. Energy consumption statistics, however, are not measured in terms of the usable heat released from fuels but in terms of heat content. The ratio of usable heat yielded and heat content of a given amount of fuel is known as the efficiency of fuel use. Physical and chemical characteristics determine the efficiency with which fuels can be used and these, in general, will be different for every fuel. If efficiency values are known, the heat content of fuels can be converted into the usable heat yielded by fuels. Some estimates of relative efficiency can be obtained from engineering sources but these are essentially approximate and arbitrary in nature. Another approach is to use index numbers but since, in practice, price ratios do not necessarily reflect efficiency ratios this may not be a satisfactory procedure. An alternative methodology is presented in this thesis in which fuel input, measured in terms of heat content, is expressed as the dependent variable in a regression model. If fuels are used with different efficiencies the amount of fuel input necessary to produce a given level of output, other factors remaining constant, will vary according to the composition of fuel input. A model is therefore developed in which independent variables are defined that account explicitly for the composition of fuel input. In addition to providing estimates of price, temperature and output elasticities, the model statistically estimates relative efficiencies of fuel use. The model was applied to the U.K. market for thermal energy for the period 1960 to 1974. It performed favourably against an alternative model utilizing representative engineering estimates of efficiency.