mous strain of financing requirements for new plants. Despite many cancellations and deferrals of powerplants, new plant financing increased from about $10 billion in 1970 to over $28 billion in 1982, of which more than two-thirds had to be financed externally (45).

The biggest incentive for utilities to cancel and postpone more nuclear than coal plants, was the more than fivefold increase in the constant dollar cost of nuclear plants from plants completed in 1971 to plants scheduled for completion in the 1980's compared to the approximately threefold increase in the constant dollar cost of coal plants (55,56). Another incentive to favor coal plants was their shorter leadtime, an average of 40 to 50 percent fewer months than for a nuclear plant (1).

Looking ahead, the prospects for substantial numbers of new central station powerplants appear fairly uncertain. The prospects for more nuclear plants appear even more uncertain. The reasons why this is so are laid out in the rest of the chapter. The next section describes the uncertainty about the future growth rates in electricity demand. With some assumptions about the future it is reasonable to expect that the fairly slow growth rates (1 or 2 percent per year) of the past few years will continue. With equally plausible assumptions, however, electricity load growth could resume at rates of 3 to 4 percent per year. The sources of uncertainty are described in the next section.

The third section explains why utilities can afford to wait awhile before ordering powerplants in large numbers, since reserve margins are now so high. Sooner or later, however, as this section points out, some number of new powerplants will need to be built to replace aging powerplants and meet even modest increases in electric load.

The fourth section of the chapter presents an argument that there may be systematic biases in

rate regulation that discourage those types of generating capacity that are of high capital cost and high risk relative to other types. Over the long run such rate regulation would discourage further construction of large coal and nuclear plants even when increased load and replacement of existing plants would make it sensible to construct central station plants, for which capital costs are high relative to fuel cost.

The fifth section of the chapter lays out the uncertainties involved in constructing and operating a nuclear plant which discourage utilities from ordering more nuclear plants even when they decide to order more central station powerplants. Construction costs have risen much faster than general price increases and vary severalfold from plant to plant even when built the same year. In addition, there is a financial risk of at least several billion dollars from an accident that disables a powerplant and more from one that causes damage to public health and property. To date insurance is available to cover only a fraction of this risk.

Given the uncertainties of demand and nuclear construction cost, utility decisions to cancel nuclear powerplants and some coal plants have been sensible, and in the short-term interests of the ratepayers. Over the long run, however, if ratemaking discourages electric-generating technologies of greater capital cost and greater risk, further investment in nuclear powerplants could be discouraged even if it were in the longer term interests of ratepayers.

The final section of the chapter describes the choices utilities have and the choices they seem to be making. Under a few specific assumptions about changes in outside circumstances and rate regulation incentives, utilities could order nuclear plants again. It appears now, however, that they will avoid central station construction as long as possible and then build coal plants.

From 1973 to 1982, annual increases in electricity demand averaged 2.6 percent. If these

growth rates were to continue for the next 20 years, they would provide no more than a weak

stimulus to further building of central station powerplants, including nuclear powerplants. (See the detailed discussion of capacity requirements in the next section.) It would be possible for most utilities, with some effort, to avoid building central station powerplants altogether until the late 1990's by encouraging conservation, load management, cogeneration, and small sources of power from hydro and wind; or by purchasing from U.S. utilities with excess capacity or from Canadian utilities; or by keeping existing plants online past normal retirement age. These strategies are discussed later in the chapter.

What are the chances that the average electricity demand growth rate will be significantly higher or lower than 2.6 percent per year? A significantly lower growth rate would make it difficult to justify any major construction of central station powerplants. A significantly higher growth rate would make a strategy of little or no powerplant construction difficult to sustain.

Published projections of electricity demand reflect considerable uncertainty about future growth rates. As is clear from figure 4 above, the utilities' own estimates of future peak demand have dropped each year since 1974 and now average an annual increase of 2.9 percent from 1983 to 1992 (70). Some studies (e.g., the Energy Information Agency and Starr and Searl of EPRI) project higher rates of electricity growth than the electric utilities do, although none project more than 4 percent annual growth through 1990 (27, 51,83). Only one (Edison Electric Institute) projects more than 5 percent from 1990 to 2000. Several studies (e.g., Audubon and the Solar Energy Research Institute) on the other hand, project very low rates of annual electricity growth of O to 1.5 percent (77,84).

One of the reasons for this range is different assumptions about the future growth rates in GNP. The projections of faster electricity growth assume a range of 2.5 to 3.0 percent annual GNP growth per year (51). The projections of slower electricity growth assume somewhat slower growth rate in real GNP, a range of 2.0 to 2.8 percent per year (77). In general, however, all these projections assume that the United States has a "mature" economy and increases in real

GNP faster than an average of 3.0 percent per year are not likely.

The projections disagree more significantly about the likely future relationship between growth rates in GNP and growth rates in electricity demand. Projections of faster electricity growth assume that electricity will increase faster than GNP. Projections of slower electricity growth asume that electricity demand will increase significantly less fast than GNP.

The ratio between electricity growth rates and GNP growth rates has indeed dropped since the 1960's. As is shown in figure 5, electricity growth rates were about double GNP growth rates in the 1960's and approximately equal to GNP growth rates (except for recession years) in the 1970's. Those expecting fast growth rates in electricity regard the late 1970's as an anomaly and expect a resurgence of a ratio of electricity to GNP growth of more than 1.0. Those expecting slow growth rates in electricity assume that the ratio of electricity growth to GNP growth will fall still further, well below 1.0. They expect that electricity will continue to behave like other forms of energy for which ratios to GNP have dropped steadily since 1973.

The sources of uncertainty in electricity demand forecasts are very evident from a look at

the uncertainty surrounding the factors underlying the forecasts. The uncertainty exists both in conventional macroeconomic approaches to forecasting which relate electricity growth to expected changes in GNP, electricity prices and the prices of competing fuels. (This is sometimes referred to as the top-down approach.) There is comparable uncertainty about the factors underlying engineering or end-use projections of electricity use. This approach (sometimes called bottom-up analysis) identifies possible technical changes in the use of electricity that are economically feasible-such as improvements in appliance and electric motor efficiency, opportunities for fuel switching, and new electricity-using industrial technologies-and then estimates the likely market penetration of these changes.

The Top-Down Perspective on
Electricity Demand-Sources
of Uncertainty

One of the advantages of top-down analysis of electricity demand is that uncertainty is confined to only a few powerful variables-future growth in GNP, changes in electricity prices, and changes in the prices of competing fuels and the responsiveness of electricity demand to each.

The Influence of Economic Growth.-Future growth of GNP is a major source of uncertainty, both because income and industrial production are assumed by economists to have major impacts on electricity demand, and because of some deep uncertainties about the future direction of the economy. Even the fairly narrow range of GNP growth rates of 2 to 3 percent that has been assumed by the major electricity demand projections implies a range of electricity demand growth rates (assuming no price influence) of about 2 to 3 percent over the long run if electricity demand follows the income response patterns identified in the past (79). Many observers concede the range is even wider. Those with private misgivings about the future health of the economy accept the possibility of an annual rate of GNP growth lower than 2 percent. Optimists about economic renewal and increased productivity suggest the potential for a higher rate of growth.

Regional uncertainties about economic growth are more extreme than national ones. Income and industrial output have fallen in some regions as a result of the recent recession and the extent of long-term recovery from the recession in these regions is unclear. Rapid population growth is expected to occur in the South and Southwest.

Electricity Prices.-Future electricity prices and their impacts are a second source of uncertainty about electricity demand growth. This is both because there is disagreement about future change in electricity prices and because there is uncertainty about how electricity demand responds to electricity prices.

From 1970 to 1982, average electricity prices increased in constant dollars at about 4 percent per year, reversing a 20-year trend of décreasing real prices (26). There is considerable disagreement about the future course of electricity prices even though they should be easier to project than oil or gas prices because they are largely determined by regulatory rules that are predictable. The Energy Information Administration in the Department of Energy (DOE) has consistently projected very slow increases in real electricity prices of less than 0.5 percent per year until 1985 and 1.4 percent per year after that (27). The Office of Policy Planning and Analysis, also of DOE, projected somewhat more rapidly increasing electricity prices, at 2.4 percent per year until 1995 with level prices after that (20). Finally, Data Resources Inc. (DRI) has projected sharply increasing electricity prices for both industrial and residential users of 3.7 percent per year until the mid-1980's, slower increases until 1990 and less than 0.5 percent per year from 1990 to 2000 (18).

Forecasts of electricity prices disagree principally about the future cost of coal for electricity, the future construction cost of nuclear and coal powerplants and the future rate regulation policies of Public Utility Commissions. Stabilizing of electricity prices in the 1990's is expected to occur because of a growing share of partially depreciated plants in the rate base and little new construction. (See the discussion of these factors in later sections of this chapter.)

Response of Electricity Demand to Electricity Prices.-There is generally less agreement about the impact of electricity prices on electricity demand than there is about the impact of changes in GNP. Most analysts agree that the short-run response of electricity demand to an increase in electricity prices is very limited-10 to 20 percent of the price increase. Based on comparisons from State to State, however, analysts estimate the long-run response to be much greater-50 to 100 percent of the price increase. (The response to a price increase is always a decrease in demand.) For example, an increase of 2 percent in electricity prices would be expected to result in a short-run decrease in electricity demand of 0.2 to 0.4 percent (from what it would have been otherwise), but a long-run decrease in electricity demand of 1 to 2 percent.

Prices of Competing Fuels.-Very few analysts have attempted to estimate the long-run response of electricity demand to changes in the prices of other (and competing) forms of energy of which the principal competitor is natural gas. Of these attempts, the consensus is that electricity demand would be expected to increase (over the long run) about 0.2 percent for every 1 percent increase in natural gas prices.

The Energy Information Administration projects that natural gas prices will increase at more than 10 percent a year (in constant dollars) until 1985 and more slowly, at 5 percent per year after that until 1990 (27). DRI, on the other hand, projects

somewhat slower increases of about 3 percent per year through the 1980's and 1990's (18).

In some areas, especially New England, oil is the chief competitor to electricity and the chief source of uncertainty. Oil prices are now higher than natural gas prices and are projected to increase but more slowly than natural gas prices.

The Impact of Prices on Demand.-The combined effect of uncertainty about future electricity and natural gas (and oil) prices and uncertainty about how electricity demand responds to changes in electricity prices is enough to explain a range of uncertainty in electricity demand from very slow growth to quite rapid growth. This is illustrated in table 3. If GNP is projected to grow at 2.5 percent (the midpoint of the range assumed in current forecasts) and the long-run response to increases in income is assumed to be 100 percent, the effect of price and price response assumptions is to produce a projection of 1.1 percent annual increase in electricity demand, at the low end, and of 4 percent at the high end. It would be possible, for example, for electricity demand to grow at 4 percent per year, if electricity prices increase at no more than 1 percent per year (in constant dollars) while natural gas prices increase at 10 percent per year, and there is a relatively small long-run decrease in demand in response to an electricity price increase (defined as a long-run elasticity of -0.5).

Timing of Response to Prices.-Unfortunately no analyses have been published of the long-run

response of electricity demand to electricity prices in the crucial decade since the 1973 oil embargo. The estimates mentioned above were all based on data up to 1972. The chief reason is because this analysis cannot be done without consideration of the timing of the long-run response. Not only have electricity prices (in constant dollars) changed from decreasing to increasing, but competing fuel prices (which also affect electricity demand) changed from decreasing to increasing even more dramatically.

The length of time it takes for the long-run price response to be felt is crucial to making any estimate of this response. If the "long run" is 3 to 4 years, we have already seen much of the response to the price increases of the 1970's. If the "long run" is 10 years or longer, we are just now beginning to witness the effects of actions taken in response to those price increases.

This lack of understanding of how long it takes for the full long-term response to increases in electricity prices makes it difficult to predict what still remains of consumer and industrial responses to the increasing electricity prices of the last decade. If the response takes 10 years or longer, the effects will last until the early 1990's.

Electricity Rate Structure.-The uncertainty of price impacts is further complicated because forecasts of average electricity price do not fully capture the potential price changes that will influence electricity demand. Decisions of industry and consumers are also influenced by the price of an additional unit of electricity, that is the marginal price of electricity. Utilities have recently begun to shift from "declining block rate" structures (in which each additional block of units of electricity costs less than previous blocks) to increasing block rate structures (in which each additional block costs more than previous blocks). There is no current survey of data on utility rate structures, but a crude estimate can be made that as many as one-fifth of all utilities may have increasing block rates for households. The number of such utilities is likely to continue to grow.

Regional Differences in Demand Response to Price Increases for Individual Utilities.-Another source of uncertainty is that individual utilities will have very different experience in electricity prices

from the national average. A recent regional analysis of projected changes in electricity prices shows a mixture of declining electricity prices in some regions and increasing electricity prices in others (48). Real electricity prices in the Mountain region are forecast to drop by an average of more than 3 percent per year (in constant dollars) until 1987 and then stay nearly stable until 2000. Meanwhile, in the West South Central region electricity prices are projected to increase by an average of 4.6 percent per year until 1991 and then taper off slowly until the year 2000. Price changes as different as these will inevitably induce a wide regional variation in demand growth rates. Declining rates in the Mountain region should eventually stimulate increases in electricity demand while the opposite occurs in the West South Central region. This regional variation in both present and projected electricity prices will complicate and perhaps delay industry's investments to improve efficiency.

The Bottom-Up Perspective on
Electricity Demand-Sources
of Uncertainty

Within an overall framework of economic growth rates and changes in relative energy prices, bottom-up or end-use analysis offers a closer look at how electricity customers might actually change their patterns of electricity use in response to prices and income changes. Industrial customers purchase the most electricity, about 38 percent of the approximately 2.1 billion kWh sold in 1981.* Residential customers are close behind with 34 percent of all sales in 1981. Commercial customers and other customers purchased 24 and 4 percent, respectively.

Given a range of plausible assumptions about how customers are likely to change their patterns of electricity use over the next two decades, growth rates in electricity demand that range from 1 percent per year to as high as 4 percent per year are possible.

From a close look at each sector it is clear that a few variables are far more important than

*1981 data are used because industrial purchases had fallen to 35 percent of the total in 1982 as a result of the economic recession.