| Sumario: | This paper presents empirical evidence on the temporal stability of residential electricity demand in the United States for the period 1962 through 1975. Residential demand was chosen because of the relative ease in specifying the model and the availability of required data. The period also includes both pre- and post-energy crisis years, which permits a limited comparison of pre- and post-energy crisis electricity consumption patterns. Pooled cross-state data for the period is used to avoid data limitations which would result from the use of time-series data only. This approach both increases degrees of freedom and reduces problems of multicollinearity. <BR> In order to investigate temporal variations in residential electricity demand in the United States, a variety of regressions--both individual yearly cross-state regressions and pooled regressions--were carried out. Results of individual yearly cross-state regressions for each of the 14 years from 1962 through 1975 showed that regression coefficients varied substantially. Analysis of covariance based on these regressions indicates that the source of instability in yearly cross-state demand functions seems to be attributable to temporal differences in the intercepts of these demand functions. The 1962-75 period was divided into two basic subperiods: the pre-energy crisis period of 1962-72, and the post-energy crisis period 1973-75. Furthermore, three-year periods within the 1962-72 period were examined. The regressions from these pooled regressions indicate that the electricity demand function has not been stable between the two subperiods 1962-72 and 1973-75. The source of instability in the demand functions between the two subperiods appears to lie more with the slopes than with the intercepts of the demand functions. Finally, decline in the absolute magnitude of the electricity price coefficients in the 1973-75 period as compared to those for the 1962-72 period has important implications for the...
|
|---|