2
income statistics had been available, it is doubtful
whether they would have improved the overall accuracy of the
forecast because of the difficulty of providing reliable forecasts of the explanatory variables. (e.g. price, personal income, industrial output,)-
It was assumed that electricity demand was determined by demographic factors, income per capita and the price of electricity deflated by the consumer price index; and two approaches to the problem of finding an equation linking demand and these factors were used, one quarterly and the
other annual.
(b) Arriving at an Equation
Despite several different specifications, including attempts to allow for a lag in the response of demand for
electricity to growth in incomes, it was not possible to obtain a useful quarterly equation (using as variables, price, population and the quarterly quantum index of exports as a proxy for income).
Moderate success, however, was achieved using annual data. In this case two methods were employed -one using
per capita statistics, that is per capita sales of electricity and GDP, and the other using statistics relating to the whole population. The relationship between per capita statistics and demand more accurately represents the demand for electricity because demand for electricity associated only with demographic change is proportional and therefore income growth reflecting solely population growth would have a smaller effect
on estimated demand than would income growth involving
per capita gains: