Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 Historical Perspective _ r ~ l ~ Regulation _ ,,,,,,,, :: ; ' """' ""'':':""" "" ' - Technology . ~ If ~ ~ ~Sihihily4~1l~FD6Is _ ~,'::: '2:2.2'. ', :::.'.2:'' .2.2.'' ..... ... , ~ , ~ _ . ~, f1'^~dent~ai 1 ~ , , , ,.,., } ~ . .~ ~ . . . l.~i.~`ct~ta:t ' . 1' SU PP LY | Conservation |~ l ~ , Electri' :ity Income _- Using Devices DEMAND This chapter provides ~ historical perspective on the relationship of electricity to economic growth. It deals with the shaded portions of the above reproduction of Figure 1.1. Early patterns are brief ly noted. A more extended discussion of the period at ter World War II ~ ~ ~ or ~ ._~, goes into the correlation of electricity use with gross national product (GNPy, patterns of composition of economic sectoral electricity use, changes in output, and the effects of price changes. This material forms the basis for comment on the likely continuity of prior relationships and possible changes in them. The chapter, in 15
16 conj unction with further discussion in Chapter 5, helps to support two of the principal conclusions that the study draws: o Electricity use and Gross national product have been, and probably will continue to be, strongly correlated. o Changes in the composition of national output toward less . electricity-intensive Hoods and services have been offset by growth in the intensity of electricity use within all the major use sectors so that the combined effect on electricity demand _ growth has not yet been great. However, if the trend toward a - leveling off in sectoral electricity intensity growth that began in the late 1970s continues, future shifts toward less electricitY-intensive goods and services are likely to dampen _ . . ~ electricity demand growth relative to national output. HISTORICAL PATTERS: 1902 TO 1983 The Growth of Electric ity and Other Energy Consumption Electricity is such a versatile energy form that its use and the number of its applicat ions have g rown rapidly throughout the tweet ieth century. Electricity consumption continues to grow more rapidly than that of other energy forms. As a consequence, the proport ion of the nation's primary energy supply used to produce electricity has expanded substantially--from near zero at the turn of the century to 36 percent in 1983. The growing importance of electricity as a component of total energy supply can be seen in Figure 2-1, which shows the growth in primary energy input, in British thermal units (Btu), for the production of electricity. Another way of measuring the comparative growth of electrical and nonelectrical energy is by directly comparing the energy delivered by electricity (instead of by the primary energy consumed in its generation) and the energy delivered by other forms (mostly coal and coke, oil products, and natural gas). The average annual growth rates shown in Table 2-1 are based on such data. Clearly electricity consumption has grown at a higher rate than has the consumption of other energy forms throughout the twentieth century, including the most recent decade. In fact, over the last decade electricity use continued to g row, while that of all other energy declined (an unprecedented occurrence in the long historical record) . Nevertheless, the rate of growth of electricity consumption itself fell sharply during the recent past compared with its growth rates in all earlier periods. That the rate of growth of electricity use has tended to decline, particularly compared with the early periods of its introduction, is not surprising. As the base f rom which growth is measured becomes larger, even ever-growing absolute inc remeets can translate into smaller percentage growth rates. By the same token, the early rates of
17 80 70 - 4J m o 60 . _ ._ ~5 5 50 of o - :~ 40 an o C' 30 G UJ of FIGURE 2-l 1983. Total Energy Primary Energy Input hi! I i/ to Electricity \ ~_~~ 10 O Nonelectric Enerov ,~ 1910 1920 1930 . 1940 1950 1960 1970 1980 YEAR Historical trends in U.S. energy consumption, 1902 through SOURCES: U.S. Bureau of Mines, as presented in Towards Pro ject Independence: Energy in the Coming Decade, prepared for the Joi nt Committee on Atomic Energy, U. S. Congress, 94th Congress, 1st Session (December 19751; Edison Electric Institute, Historical Statistics of the Electric Utility Industry through 1970, and Statistical Yearbooks; U.S. Department of Energy, Annual Energy Review, various issues.
18 TABLE 2-1 Average Annual Growth Rates in Total Energy, Electricity, and Nonelectric Energy Consumption for Selected Periods, 1902 through 1983 (Percent per Year) Nonelectric Period Total Energy Electricity Energy 1902-1912 6.1 15.5 5.6 1912-1920a 2.9 10.8 3. 3 1920-1930a 1.2 7.3 1. 3 1930-1940 0.7 4.6 0. 5 1940-1950 3.5 7.9 3 0 1950-1960 2.8 8.1 2.2 1960-1973 4.1 6.7 3. 4 1973-1983 -0. 5 2.0 -1. 7 Average annual g rowth rates for total energy and for nonelectric energy were computed from British thermal units (Btu) of consumption. The growth rate for electricity was computed from kilowatt hour figures. Because of rapidly improving efficiency of electric power generation in the early years, electricity kilowatt hours grew much faster than Btu input for electricity generation. This aspect of the computation is the reason that, for these two periods, the growth rate of total energy appears to be lower than the growth rate of both of its components. SOURCES: U.S. Bureau of Mines, as presented in Towards Project Independence: Energy in the Coming Decade, prepared for the Joint Committee on Atomic Energy, U.S. Congress, 94th Congress, 1st Session (December 1975~; Edison Electric Institute, Historical Statistics of the Electric Utility Industry through 1970, and Statistical Yearbooks; U. S . Department of Energy, Annual Energy Review, various issues.
19 growth of a newly emerging product or industry, such as electricity in the first part of this century, will loom large compared to those of already established quantities, such as population, GNP, or the use of other energy f arms . In considering electricity growth in historical perspective, it is instructive to review time trends, as such, and to compare the growth races of electricity with those of other energy forms. However, the focus of this study is electricity in economic growth. Accordingly, our goals are to try to discern trends in the relationship between electricity and economic growth and to consider some of the factors underlying these trends. Electricity Growth in Relation to the Growth of Gross National Product How should the relationship between electricity use and economic growth be expressed? Our approach is to look first at the relationship in aggregate terms, that is, in terms of total electricity made available in the United States, regardless of source, and of GNP, expressed in constant dollars. Later this relationship will also be examined, although only for the years following World War II, in terms of the major sectors of electricity use: residential, commercial, and industrial. The discussion addresses the nature of the aggregative relationship, changes in this relationship over time, and sectoral relationships compared to the aggregative, or national, one. It is well known that regional disaggregations will exhibit diversity, but this chapter does not address that effect. When standard statistical techniques are used to measure the relationship between annual levels of electricity use and GNP (in constant dollars) , certain regular features of the historical record appear. Perhaps the most significant characteristic of the relationship is its stability over appreciable segments of time. This stability is indicated in Figure 2-2 (with additional detail in Figure 2-3), which displays lines of regression for four periods covering most of the twentieth century to date. * For any point on these lines one may calculate an average electricity intensity, that is, electricity consumption per unit of GNP. Changes along these lines relate increments in electricity use to increments in GNP; each period is marked by a stable linear relationship, which, though showing some annual fluctuations, indicates a strong tendency toward a constant incremental intensity of electricity use within each period. Equally signif icant is the fact that there are only a few changes in the slope (and the level) of the regression line over the long histor ical record. Clearly d iscernible changes in slope occur red following World Wars ~ and II. Even the Great Depression did not result in a change in slope; the level of the regression line did shif t upward, however, reflecting the fact that during this period GNP . *A regression is the best functional relationship between two (or more) cor related var. tables as j udged by a part icular stat i stical c r ite rion, such as the criterion of ordinary least squares.
20 450 400 o c 300 o . _ ._ - z O 250 u' 0 200 - c: ~ 150 c' J 100 50 o , _..... l / An/ /1930~1 946 / . ~ '' World War I / 1902 1912 100 1 947-1 983 / Post-World War I I I S;h if t / Depression Shift t non-'n7q 1 1 , 1 _ 1 1 _ 1 200 300 400 500 600 700 GNP (billions of 1972 dollars) FIGURE 2-2 Electricity consumption vs GNP in the United States, with lines of regression by periods, 1902 through 1983. NOTE: GNP is expressed in constant (1972) dollars. Data for 1902 through 1928 have been converted from constant (1958) dollars in U.S. Department of Commerce, Bureau of the Census, Historical Statist ics of the United States, Colonial Times to 1970, Bicentennial Edition, Part 2, Series F1-5, p. 224; for 1929 through 1980 from Council of Economic Advisers, Economic Report of the President, February 1984, p . 222; for 1981 through 198 3 f ram Counc il of Economic Advi sers f or the Joint Economic Committee, Economic Indicators, March 1985, p. 2. Electricity consumption is expressed as "electricity made eve; ]=h7^ in the United States." ~onceptua ~ by this quantity includes utility generation and nonutility generation ~ industrial self- and co-generation), and net imports. Electricity data sources are Edison Electric Inst itute ( 1973, 1984a) . SOURCE: Compilation and f igure by Energy Study Center, Electric Power Research Institute, Palo Alto, California.
21 400 - o -° 300 . _ Q of o z 200 o > G C: 100 LL] J UJ o Years 1902-1 912 1 91 2-1 920 1920-1 929 1929-1932 1930-1 946 19441946 1946-1 947 1 947-1 983 Trend Slope 0.29 0.s8 o.ss 2.12 torso. / 38 037 35 / 36 ' 29 32 '` 2 1~/ i\ World War I-t 17 a`9ttO5 02~1 \~ Post-World War I I / Transition Slope 2.19 0.20 0.14 3.30 49 / 487/ 47 · I 46Lis 42 a/ i J _ , I ~ ' 0 200 300 400 500 600 GNP (bil lions of 1 972 dollars) FIGURE 2-3 Electricity use and GNP--the transitions. NOTE: GNP is expressed in constant ( 1972) dollars . Data for 1902 through 1928 have been converted f rom constant (1958) dollars in U. S . Department of Conunerce, Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1970, Bicentennial Edition, Part 2, Series F1-5, p. 224; for 1929 through 1949 from Council of Economic Advisers, Economic Report of the President, February 1984, p. 222. Electricity consumption is expressed as "electricity made available in the United States. " This quantity includes utility generation and nonutility generation ~ industrial self- and co-generation), and net imports. Electricity data sources are Edison Electric Institute (1973, 1984a) . SOURCE: Compilation and f igure by Energy Study Center, Electric Power Research Institute, Palo Alto, California.
22 decreased relatively more than did electricity use. The record of the past decade poses the question of whether we are once again witnessing a change in slope, an issue that is addressed later in this chapter. The specif ic f indings embodied in Figures 2-2 and 2-3 may be summarized as follows: 1. From 1902 to 1912, (a period during which data on electricity use are available only for every f if th year), the national economy tended to use an additional 0.29 kilowatt hours (kWh) per additional dollar of GNP, measured in constant (1972) dollars. * 2. A transition to a new slope occurred between 1912 and 1920; the 1917 observation shown on Figure 2-3 appears to be transitional. 3. Between 1920 and 1929 the incremental use of electricity per unit of GNP averaged 0.58 kWh per dollar, twice the value that prevailed between 1902 and 1912. 4. Following 1929, GNP dropped by almost one-third, while electricity use declined only slightly. Consequently, average electricity intensity increased. However, the slope of the line for the years 1930 through 1946 did not change signif icantly f rom that for the years 1920 through 1929. Thus, incremental intensity of electricity use remained the same, even though average electricity i ntens ity rose . 5 e Another transition occurred following World War II, and the new trend line has persisted ever since (with a critical question remaining about the most recent past). The new slope shows on the average an increment of 2.12 kWh per additional constant (1972) dollar of GNP, about three and one-half times that characterizing the relationship observed between 1920 and 1946. POST-WORLD WAR I I TRENDS: 1 9 4 7 TO 1 9 8 3 The Growth of Electricity Use and of Gross National Product The relationship between increases in electricity use and increases in GNP is shown for the full post-World War II period in Figure 2-4. The relationship appears to have persisted through the entire period, with the possible exception of a break since the mid-1970s. Although observations for the most recent years fall below the trend line, this fact is still consistent with a characteristic feature of the relationship, that is, a tendency for individual years to exhibit a cyclical pattern around the long-term trend line, as the f igure shows. However, to conclude that the data points af ter the mid-1970s are nothing more than a manifestation of a persistent cyclic pattern is only one way of interpreting the record for recent years. There has been a strong decreasing trend in the ratio of the annual percentage *A discussion of measures of electricity use and load demand r in terms of energy (kilowatt hours) and power ~ kilowatts), appears in Appendix B.
23 2800 _ 2600 2400 2200 ~ 2000 - o O 1800 - - - 0 1 600 z 1400 8 ~ 1 200 J t 000 LU 800 400 _ 600 _ 200 o / 1 9114 1980 /983 1982 of-\ / .ge1\ ran 19,3 1947 ~ 1 1 1 1 1 1 1 1 ' 1 0 200 400 600 800 1000 1200 1400 1600 1800 GNP (billions of 1972 dollars) FIGURE 2-4 Electricity consumption vs GNP in the United States, 1947 through 1984. NOTE: GNP is expressed in constant (1972) dollars. Data from Council of Economic Advisers, Economic Report of the President, February 1984, p. 222; Council of Economic Advisers for the Joint Economic Committee, Economic Indicators, March 1985, p. 2. Electricity consumption is expressed "= ran; -~1 <:- =~= " This quantity ~ ~ "electricity made available in the United States. " This quantity includes utility generation and non- utilitv generation ~ industrial self- and co-generation), and net imports. Electricity data sources are Edison Electric Institute (1973, 1984a). SOURCE: Compilat ion and f igure by Energy Study Center, Electric Power Research Institute, Palo Alto, California.
24 growth of electricity use to that of GNP, and this fact is f requently cited as evidence that the relationship between the two has changed. Figure 2-5 shows that the ratio of percentage electricity growth to percentage GNP growth has f alien f rom an average of about 2 before 1973 to about 1 today. This tendency toward convergence is consistent with the postwar linear relationship between the two variables. The electricity use-GNP line of regression for 1947 to 1983 shows an increment of 2.12 kWh of electricity for every constant (1972) dollar increment of GNP. In the early postwar period, when average electricity intensity was comparatively low (about 0.6 to 0.8 kWh per dollar) , the high incremental electricity intensity (2.12 kWh per dollar) led to much higher electricity growth rates than GNP growth rates. As average electricity intensity has increased--to 1.57 kWh per constant (1972) dollar in 1983--the effect of the incremental electricity intensity (2.12 kWh per dollar) has relatively decreased, leading toward a convergence in growth rates.* A critical question before us, then, is whether the long-standing post-World War II trend has been broken by another of the historically infrequent transitions, but for the first time toward a decline in the incremental intensity of electricity use. To shed more light on this question, we next examine some of the underlying forces that determine electricity use in relation to national output. Such influences include the trends of electricity use in the major consuming sectors, the effects of changes in the composition of national output, and the effects of changes in energy prices. Electricity Use in the Major Consuming Sectors Electricity use is ordinarily classified by three major consuming categories: 0 Residential, that is, private households o Commercial, that is, nonindustrial business establishments o Industrial, that is, agriculture, mining, construction, and manufacturing. ** Table 2-2 shows the changing importance of each sector as reflected by its percentage of total electricity consumption over the postwar period. The residential sector sharply increased its share of *In other words, because of the nonzero intercept in the relationship (which appears as the offset of the regression line from the origin in Figure 2-4), the percentage growths of electricity and GNP along the regression line are more nearly equal where both quantities are large, as in the later years, than where both quantities are small, as in . earlier years. **These def initions differ somewhat from the Edison Electric Institute (EEI) sector definitions. However, for our purposes, the differences are not great enough to warrant concern, and so we have used the EEI statistics, with no change, to fit our categories.
25 LL cD Z 12 I o LL] llJ CD 6 En UJ Cat 6 Z ~ 6 in O - 10 8 6 - ~n a, ~4 ._ to 2 Q Z . 1' ~- Electricity a I' · · ~_ . ·. · ~:. . · ·.- GNP ---~\ 1 1 1 1950 1960 1970 1980 U' 6 4 G ~I ~ > ~N ~ \` O G 1 1 _ O ~ G ~O l l l l l 1950 1960 1970 1 980 YEAR FIGURE 2-5 (a) Growth rates of U.S. electricity use and GNP, (b) ratio of the growth rates. SOURCES: Based on data f rom Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, various issues; U.S. Department of Commerce, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues.
26 TABLE 2-2 U.S. Electricity Sales by Sector (Percent of Total) a Year Residential Commerc iamb Industrials . 1950 20.6 20.1 ~ 59.3 1955 22.3 17.3 60.4 1960 25.4 18.4 56.2 1965 26.6 22.6 , - 50.7 . . 1970 29.9 24.7 45.4 1975 32.2 26.7 41.1 1980 33.5 27.2 39.3 1983 33.9 28.3 37.8 aIncludes industrial self-generation. bSmall light and power, street and highway lighting, other public authorities, railroads and railways, and interdepartmental transfers. PLarge light and power and industrial self-generation. SOURCES: Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, various issues. l
27 electricity use from one-fifth to over one-third of the total. The commercial sector share increased sharply during the 1960s, and in 1983 it stood at 28 percent of the total. The industrial share, starting at 59 percent of the total, dropped dramatically after 1955, to 38 percent of the total by 1983. These postwar trends in sectoral shares of electricity use parallel the underlying trends in the economic measures for each sector. That is, growth in disposable personal income (DPI), the residential sector surrogate for gross product originating (GPO) in the other two sectors, and commercial sector growth outpaced that of industrial output over the entire period.* In fact, if we examine the relationships between electricity use in these sectors and their respective economic measures, as in Figure 2-6, we find the same stable linear relationship (with cyclical variation) as is seen in the aggregate economy. Also, as in analyzing the aggregate case, one gains a different perspective when comparing the ratio of the percentage growth rates of electricity use and economic output measures. Figure 2-7 shows the same trend toward convergence between the sectoral percentage growth rates as was observed in the total economy. For further insight into these trends, we examine in more detail the postwar patterns of electric ity use within each of the three sectors. Trends in Residential Use of Electricity The trend in residential electricity consumption since World War II falls into three distinct periods, corresponding roughly to the decades of the 1950s, 1960s, and 1970s, as shown in Figure 2-7. During the 1950s, the growth rate of electricity consumption was very high but steadily decreasing f rom about 14 to 8 percent per year. During the 1960s, this growth rate slowly accelerated to about 10 percent per year toward the end of the decade. It then dropped in the post-embargo period to an average of about 5 percent per year, until the late 1970s and early 1980s when it dropped further. Figure 2-8 illustrates the changing pattern of residential electricity use in the first two decades after the war. In 1950 electricity went primarily to four end uses: lighting (29 percent), water heating (24 percent), refrigeration (17 percent), and cooking (11 percent). Between 1950 and 1960 electricity use per customer more than doubled. Over 50 percent of this increase is attributable to the three largest incremental end uses: refrigeration, water heating, and televisions. The introduction of frost-free refrigerators and the trend toward larger refrigeration units more than doubled the electricity consumption for this appliance by 1960. The penetration of electricity into the water-heating market doubled between 1950 and *GPO, the statistic used for both commercial and industrial outputs, a measure of value added derived from the national income and product accounts; it emphasizes the sectoral origin of GNP.
28 800 600 400 200 o 600 400 - o~ z 200 8 - - ~: J o 800 600 400 200 REStDE.~TIAL - ,~r _' 1 1 1 400 600 800 1,000 DPI (billions of 1972 dollars) _ CC\IMERCIAL '! 1 1 1 9.: ·_ - · ~ A/ i/ 400 600 800 1,000 GPO (billions of 1972 dollars) I ~ DUSTS I A L ·, - ·' .~ 200 300 400 500 GPO (billions of 1972 dollars) FIGURE 2-6 Electricity use-economic measure relationships, by economic sector, 1947 through 1983. NOTE: Different scales are used for the three sectors to highlight the linearity of the electricity use-economic output relationship within sectors. Based on data f rom Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, Department of Commerce, Bureau of Economic Analysis, The . National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues. various issues; U. S. SOURCE: Compilation and f igure by Energy Study Center, Electric Power Research Institute, Palo Alto, California.
29 15.0 12.5 10.0 7.5 5.0 \ Electricity _ \\ a RESIDENTIAL o 12 10 8 6 2 - 10.0 7.5 5.0 2.5 o 1950 1 960 _ trinity 970 1980 b COMMERCIAL 1950 1960 1970 1980 ma\\ Electricity ·.... C INDUSTRIAL ~ · - By, . . GPO ·.. ~ \-. 1950 1960 1970 1980 YEAR FIGURE 2-7 Growth rates of electricity sales and sectoral output indicators, 19J,7-1983: (a) residential, (b) commercial, (c) industrial. SOURCES: Based on data from Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, various issues; U.S. Department of Co~Tunerce, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues .
30 1,845 kWhlCustomer 29°,~o 24.,` ,7% 60 ,3% 7,066 kWh/Customer 1 10% /15% Three Largest / / Incremental Percent of Total / / Period End Uses Incremental Use / / . - I I 195~1960 Refrigeration 22 / / / Water heat 20 / / / Televisions 13 / / / 1960-1970 Space heat 28 / / 18% Air conditioning 19 / Refrigeration 15 / / / // / / / / 3,854 kWh/Customer / 796 15~1~6 I% I Water Heat 1/ // L / / Lighting Water Heat Refrigeration Cooking . _ Space Heat _ Other 1950 'it/ 20% 1o% / / /7% / /8% ~7% Ret rigera~ion Cooking 496 Space Heat - Televisions Air Conditionino F reezers Other 1 960 At// Lighting Water Heat Ret rigeration Cook ing Space Heat Televisions Room Air Conditioning Central Air Conditioning Freezers Drying D ishwashers Other 1970 3 FIGURE 2-8 Residential electricity use patterns, 1950, 1960, and 1970. SOURCE: Tansil and Moyers (1974~.
31 1960, while the household saturation level for televisions increased from 13 percent in 1950 to 90 percent in 1960. Electricity use per customer nearly doubled again between 1960 and 1970. However, as distinguished from the earlier decade, almost one-half of this increase is attributable to the increased penetration of electric space heating and air conditioning (from 2 to 8 percent and from 13 to 38 percent of all households, respectively). Both of these applications consume large amounts of electricity, so that even modest market penetrations can have a large influence on consumption levels. Refrigeration (accounting for 18 percent of electricity use in 1970) again accounted for a relatively large part of the absolute increase in electricity use as the trend toward using frost-free refrigerators continued. Moreover, in general, household appliances were becoming more electricity-intensive. From 1970 to 1980 average consumption per household increased only 28 percent, as may be seen in Figure 2-9. One half of this increase occurred by 1973. Again the main contributors to incremental consumption were air conditioning, refrigeration, and space heating. Within the residential sector electricity now accounts for about 58 percent of gross energy (28 percent of net energy) consumed.* More than half of this electricity is used in applications for which there is essentially no competition from other fuels, applications such as air conditioning, lighting, and most appliance operation. Since 1980, electricity use per customer has decreased slightly. However, electricity continues to make signif icant inroads in space heating and air conditioning. While about 17 percent of the total occupied housing stock today uses electricity as its primary heating source, 50 percent of new single-family housing units (and a greater percentage of multifamily housing units) incorporate electric heating systems, up from 28 percent in 1970. Of all occupied housing units, 57 percent now have air-conditioning systems of some type, but only 27 percent are equipped with central air-conditioning systems. However, about two-thirds of new single-family homes are built with central air- conditioning systems, which indicates that such electricity penetration will continue (U.S. Bureau of the Census, 1984; U.S. Energy Information Administration, 1984~. Part of these trends can be attributed to the fact that much of the housing stock expansion is occurring in warmer climates, where electric heating and air conditioning are more prevalent. Almost three-fourths of all recent housing starts are in the southern and western regions of the country, although these regions accounted for only one-half of existing housing stock in 1973. However, both electric space heating and air conditioning are increasingly penetrating all regions of the United States (U.S. Bureau of the Census, 1984; U.S. Energy Information Administration, 1984~. *Gross energy, which refers to primary energy input, includes the energy lost in generating, transmitting, and distributing electricity, whereas net energy does not.
32 Period Three Largest Incremental End Uses 1970-1980 Air conditioning Ret rigeration Space heat 7,066 kWh/Customer 10% 15% 18% 7% 16% 12% 5% 4% 12% Lighting Water Heat Refrigeration r , Cooking Space Heat Air Conditioning Freezers . Drying Other 1970 Percent of Total Incremental Use 25 19 18 9,025 kWh/Customer '.~ Lighting - - - Water Heat Refrigeration Cooking - - - Space Heat - 15% Air Conditioning 5% 5% Freezers Drying at% Other t980 FIGURE 2-9 Residential electricity use patterns, 1970 and 1980. SOURCES: Tansil and Moyers (1974); Geller (1983); {J.S. Energy Information Administration (1982).
33 Given the increasing penetration of electric space heating and air conditioning, the slower growth in residential electricity use suggests that significant conservation measures are being used within this sector. Certainly, new housing is built to be more energy-efficient and many older houses are being retrofitted with energy-saving features. In addition, there is a trend toward manufacturing more energy-efficient appliances (Meyers and Schipper, 1984~. Trends in Commercial Use of Electricity Electricity consumption in the commercial sector has grown faster than that in the other sectors since 1960 (Figure 2-7~. The large increase in commercial electricity use between 1960 and 1973 (at a 9.5-percent average annual growth rate) has been attributed to increases in the use of mechanical air-conditioning systems and to new standards for building illumination, which resulted in increased lighting requirements (Solar Energy Research Institute, 1981~. From 1973 to 1983, commercial electricity use continued to increase but at a much slower average rate of 3.1 percent per year. It appears that much of this increase was due to an increased penetration of electricity use in space heating. While it is estimated that about one-fourth of all commercial buildings are now heated with electricity, almost one-half of all commercial buildings constructed since 1973 use electric heat (Hirst et al., 1983~. Electricity represents 65 percent of gross energy (35 percent of net energy) consumed in the commercial sector, making this sector by far the most "electrified." More than one-third of commercial sector electricity is used for lighting, about one-third for air conditioning, one-fifth for space heating, and the remaining one-tenth or so for water heating and miscellaneous appliance operation (for example, cooking and refrigeration). The declining growth of commercial sector electricity use during a period when output growth remained relatively strong points again to improvements in efficiency of use. New commercial buildings are generally constructed to be more energy-eff icient than were older buildings (U.S. Energy Information Administration, 1984~. Improved lighting systems and the introduction of computerized energy management systems will also increase the efficiency of energy use. On the other hand, the trend toward greater use of electric heating will probably continue; and the increased automation of of f ice services will add to electr ical loads, not only for equipment operation but also for waste heat removal. Trends in Industrial Use of Electricity Electricity use in the industrial sector grew at an average rate of 8 percent per year between 1950 and 1960. From 1960 to 1973, the growth
34 rate of industrial electricity use averaged 4.7 percent per year before falling to an average of 0.6 percent per year from 1973 to 1983.* Electricity represents 34.8 percent of the gross energy (13.5 percent of net energy) consumed in the industrial sector. Manufacturing accounts for 85 percent of total electricity use in this sector, with agriculture, mining, and construction activities accounting for the remainder. In 1981, electricity represented about 46 percent of gross purchased energy (20 percent of net purchased energy) in manufacturing compared to 28 percent gross (10 percent net) in 1958. Manufacturing can be divided into two large groups : the process industries, in which the physical or chemical properties of materials are altered, and the nonprocess, or fabrication, industries. The latter group, which accounts for about 75 percent of manuf acturing value added, is largely electrif fed already, with 57 percent of its combined gross energy requirements being satisf fed by electricity. These industries require energy mainly for mechanical drive and low-temperature heating. The process industries, on the other hand, use much of their energy for high-temperature heating or steam raising, needs that historically have been satisf fed by the use of fossil fuels. Electricity generally accounts for 25 to 30 percent of gross energy use in these industries. However, electrotechnologies are increasingly penetrating high-temperature materials processing, and they may make further inroads (Burwell, 1983~. We will examine the manufacturing sector in greater detail below in considering structural economic changes. Electricity use has continued to increase in agriculture and mining because of increased energy requirements for irrigation and mineral extraction. In agriculture, energy requirements for irrigation are increasing as groundwater supplies are being depleted and a greater percentage of water requirements must be satisfied by pumping from greater distances. In mining, depletion effects and environmental regulations have increased electricity requirements, mainly in oil and gas extraction and coal mining. Electricity use in construction is negligible (Werbos, 1984~. *The electricity consumption of the federal government's gaseous diffusion plants for uranium enrichment represented about 20 percent of industrial sector demand growth in the 1950s, but has declined since then, except for a brief increase in the early and mid-1970s. Uranium enrichment operations are not expected to have an important effect on industrial electricity consul tion in the near future. Excluding electricity consumption for uranium enrichment, the average annual growth rates of industrial sector electricity demand are 6.8 percent per year from 1950 to 1960, i. 5 percent per year from 1960 to 1973, and 0.7 percent per year from 1973 to 1983.
35 The Effects of Structural Change on Electricity Intensity Measuring Structural Change Structural change may be indicated by different economic indicators, such as the composition of GPO, employment, and final demand. Table 2-3 shows that the share of GPO accounted for by industry fell f ram 38 percent in 1950 to about 31 percent in 1983 . However, this change is accounted for almost wholly by declines in the relative importance of agriculture, mining, and construction. The manufacturing share of total domestic output has remained at about one-fourth during this entire period. GPO in the commercial sector, a very broad classification that encompasses all output originating outside the industrial sector, grew from 62 to 69 percent. These figures encompass transportation; communications; electric, gas, and sanitary services; wholesale and retail trade; finance, insurance, and real estate; personal services; and government operations. However, GPO composition is not the only way of looking at structural change. Another perspective is gained by looking at employment trends, shown in Table 2-4. As a share of total employment, industrial sector employment fell f rom about 45 percent in 1950 to about 29 percent in 1983. Commercial sector employment rose f ram 55 percent of total employment in 1950 to about 71 percent in 1983. This information tells us that the commercial sector has been absorbing more of the growing labor force. This trend is especially true of part-time workers. On the other hand, the rapidly increasing share of commercial sector employment relative to output growth in that sector implies that growth of labor productivity (that is, output per unit of labor input) has been slower in this sector than in the industrial sector. Still another way to view structural change is through the mix of final demand for goods, services, and structures, shown in Table 2-5. Since 1960, the share of goods in final demand has remained at about 45 percent, after dropping from 49 percent in 1950. Services have increased from about 39 to about 47 percent of final demand, while structures in 1983 accounted for 8 percent of final demand, down from 12 percent in 1950. These data imply that the relative consumption of final goods has not changed much over the last 20 years. What has changed is the value composition of these goods, through various intermediate markups from such service activities as distribution and sales that have increased over time. Thus, compared to demand composition in 1950, a greater part of the final demand of goods now consists of value added in the services sector and a decreasing share is attributable to value added in industry.
36 TABLE 2-3 Gross Product Originating (GPO) in the U.S. Economy for Selected Years, 19 50 to 198 3 (Percent of Total) Sector 1950 1960 1970 1983 Industrial 38. 0 36. 0 34.230.7 Agriculture 5.5 4.4 3.22.6 Mining 2.1 1.8 1.81.4 Construction 5.5 6.3 5.03. 3 Manuf acturing 24. 8 23. 5 24.323.4 Commercial 62.0 64.0 65.869.3 SOURCES: U . S . Department of Conunerc e, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues. TABLE 2-4 Employment in the U. S . Economy for Selected Years, 1950 to 1983 (Percent of Total) Sector 1950 1960 19701983 Industrial 44.9 38.9 34.728.7 Agriculture 10.9 7. 0 4.03. 3 Mining 1.6 1.1 0.81.0 Construction 5.8 5.3 5.35.1 Manufacturing 26.5 25.4 24.619.3 Commercial 55.1 61.1 65.371.3 SOURCES: U. S . Department of Commerce, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues.
37 TABLE 2-5 Gross National Product (GNP) by Major Type of Product for Selected Years, 1950 to 1983 (Percent of Total) Product Type 1950 1960 1970 1983 Goods ~48.9 - 45.5 44.9 44. 9 Services 38.8 42.4 44.4 47.1 Structures 12.3 12.1 10.7 8. 0 SOURCES: U. S . Department of Commerce, Bureau of Economic Analyst s, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues.
38 Of the several ways of looking at structural change, the most significant for analyzing electricity use is the GPO (value-added) measure, since it provides a measure of total productive activity--embracing both labor and capital inputs--within any particular sector. GPO analysis shows manufacturing to have generally maintained its share of output over the postwar period. The shif t in output has been from agriculture, mining, and construction toward selected commercial sector activities. The following two sections adopt GPO as the best measure of structural change. Electricity Intensity of the Sectors Electricity intensity is def iced as total kilowatt hours (kWh) of electricity consumed divided by the aggregate economic output measure of a sector, that is, a measure of average electricity use per unit of output. * Again, constant-dollar GPO is used to measure real output in the industrial and commercial sectors and constant-dollar DPI is used to measure real income for residential users. Industrial output, at 1.80 kWh per constant (1972) dollar, is now (1983) three times as electricity-intensive as commercial output, at 0.60 kWh per constant (1972) dollar. The residential sector consumed 0.69 kWh per constant ( 1972) dollar of DPI in 1983. Throughout the postwar period, the intensity of electricity use in all three sectors has increased, as is shown in Figure 2-10. Between 1950 and 1983, industrial electricity intensity increased 80 percent, commercial sector intensity increased more than 180 percent, and residential intensity increased about 260 percent. Thus, electricity use was growing faster than output or income in every sector. Although there were large increases in average electricity intensity over this total period, almost all of the growth occurred prior to 1973. Industrial and commercial sector electricity intensities increased slightly in some years after 1973, but in 1983 had fallen back to their 1973 levels. In the residential sector, electricity intensity was about 8 percent greater in 1983 than in 1973, but has not shown an appreciable increase since 1977. Elect r ic ity Intensity Within Manuf acturing Shifts in the output mix within a sector can have an important effect on the intensity of electricity use. This effect is particularly notable in manufacturing, where the electricity intensity of industry g roups can vary widely . *Recall the discussion above of average and incremental electricity intensities.
39 2.0 1.8 - ~ 1.6 o cat on 1 4 ~. - ~ 1.2 id us Z 1.0 cr ~ 0.8 LL 0.6 0.4 0.2 FIGURE 2-l 0 1984. lo, ~ i _ ;1 ~ . Industrial ,/ r~ - ~ I \_ / l I / a_ Aggregate Economy Residentia - Commercial oL I I I I I I 1970 i 1 950 1 960 1980 1984 YEAR Electricity intensities in the U.S. economy, 1947 through SOURCES: Based on data from Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, various issues; U.S. Department of Commerce, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues.
40 Figure 2-11 compares the electricity intensities of the six most electricity-intensive manufacturing industries with the average intensities for the rest of manufacturing and with those of the entire manufacturing sector. It can be seen that those six industries combined are about six times as electricity-intensive as the remaining industries. This f igure also compares the 1973 and 1981 levels of electricity intensity for the industries; as a whole, the six industries slightly increased their electricity intensity over this period. There was also a slight reduction in electricity intensity for the rest of manufacturing and a slightly larger decrease for the entire manufacturing sector. The decrease for the entire manufacturing sector is consistent with the fact that the combined output share of the six most electricity-intensive industries has been decreasing as a percentage of total manufacturing output. Figure 2-12 illustrates this trend, and the figure also shows that the trend continued in the post-embargo period. Thus, even with no change in the electricity intensity of the two groups of industries, electricity intensity of manufacturing tends to fall because of shifts in output mix away from the electricity- intensive group. Counteracting this tendency before 1970 was the increasing electricity intensity of both groups of industries. Since 1970, however, the historical increase in electricity intensity for the six industry groups as a whole has leveled off, while that for the rest ~ 15 percent, as Figure 2-13 shows. The primary reason for the recent drop among the latter group of industries is that these industries generally are already highly electrified, and thus efficiency improvements have outweighed any incremental penetration of electricity use (Burwell, 1983~. For manufacturing as a whole, average electricity intensity has fallen over 10 percent since 1970 f rom a combination of both effects. of manufacturing has fallen a - ~t Pr ice Movements Changes in Energy Pr ices The trend in energy prices for the 40-year period before 1973, as illustrated in Figure 2-14, was one of generally stable or decreasing prices for most fuels. Electricity prices, in particular, declined- throughout the entire period. The rapid price decline for electricity has been attributed to the increasing economies of scale in electricity generation and distribution over this period and to improvements in the efficiency (heat rate) of generation. Electricity prices were also favorably affected by the stability of primary energy input costs over the period. Since 1973 a number of forces have combined to reverse the historical trend of declining electricity prices. First, there was the great increase in oil prices that accompanied the Arab oil embargo of 1973. This event drove petroleum product prices up immediately and also adversely affected the price of electricity in those regions of the country that depended on oil for a significant portion of their
41 1 - 1, 1 1 1 1 1 1 1 1 CO CD ~ ~ O Slop ZL6t ul 'Ode ~° Jellop Ad AMP) ~lISN31N1 A1101813313 ·- al - ~ ~ oh l:5 _ , ~ (q us C: ~ oh Id cad - N X can 3 A: to I O Cry ~ _ CO Cal Cal ~ at - O ~ . - - So ~ 3 · - - - U] U] ~C) O · - U] 3 U] ~ Q. H Jo A: ·_. ·_l V ·_' U1 O a 1 ~ U] ~; 3 C' ~ O 2 o ~1J U] U] 3 ~5 ~: ·e O U] 3 O U] CQ Q U] U] · S ~ ~ · - £ O ,c o 3 aJ c ~U: C5 · · - m ~ ~ Q ' 00 3 3 O ~ ~ ' V] O `- - C~ ~ V1 s~ >, S ~: c: ~ · - V · - ~ ~ O PJ ~ O ~ ~i C: · U] ~ O . U] :5 m ·e ~ a o o a) .~. o £ ~ ~ as ~ m
42 32 ~ 30 o o LL ~ 28 in G 26 . ~ i: \J ~ 24i 1 ~I 1 1 1 1 \/ 1~50 1960 1970 1980 1983 YEAR FIGURE 2-12 Six-SIC share of constant dollar manufacturing GPO, 1947 through 1983. . SOURCE: Derived f rom unpublished data provided by the U. S . Department of Commerce, Bureau of Economic Analysis.
43 120 a: ~ O J us J_ _ O ° _ /~ 1' 100 - u' no J J o ~ 80 rat O 60 FIGURE 2-1 3 100) . SIC 22, 26, 28, 29, 32, 33 (combined) /_ r~?N / _.\ ~ A/ rho l ! / '\~ / Other Manufacturing - 1°~5 1 960 1 965 1 970 1975 1980 YEAR Electricity intensities in manufacturing (Index: 1971 SOURCES: Electric ity data: U . S . Department of Commerce, Bureau of the Census, Annual Survey of Manuf actures: Fuels and Electric Energy Consumed. Economic data: U. S . Department of Commerce, Bureau of Economic Analyst s, unpublished data provided by the Bureau .
44 30C O 200 Go 11 to x z LL 1 . Gasoline 100 _ Fuel Oil ·. E lectricity Natural Gas \ \~ .. ~ 1 ·- ~ 1 1 1 1 1 1 of_ 1920 1 930 1940 1 950 1 960 1970 1980 1984 YEAR FIGURE 2-14 Trends in real energy prices to U. S. personal consumers, 1935 through 1984. NOTE: The price indicator used is the Consumer Price Index. SOURCES: U. S. Department of Commerce, Bureau of the Census, Historical Stat ist ics of the United States; U. S. . Department of Labor, Bureau of Labor Statistics,. Handbook of Labor Statistics 1978, Bulletin 2000; and Monthly Labor Review, various issues.
45 generation requirements. Figure 2-15 compares energy price trends over this period for residential and industrial consumers. The 1973 change was followed during the rest of the 1970s by a sustained rise in the price of natural gas and the second oil price shock of 1979 to 1980. The electric utility industry undertook programs during this period (with federal prodding) to cut back on both oil and natural gas as generation sources. A second fundamental change occurred during this period: the apparent exhaustion of the economies of scale and improvements in heat rate that had led to lower per-unit costs of generation over the longer period. Power plant construction projects were also being increasingly affected by inflation and delays so that the average cost of electricity generation in many utility systems has risen as many of these new plants have come on line. An additional factor was the sharp increase in environmental regulations for power plants during the 1970s and early 1980s. Throughout this troubled period, however, electricity price increases, on average, were only moderate. In real terms, the consumer price index (CPI) for electricity rose only 18 percent between 1973 and 1983, while the indexes for both fuel oil and natural gas more than doubled. The same trends are evident for producers. The producer price index (PPI) for electricity rose 44 percent in real terms from 1973 to 1983, whi le the index for petroleum rose 136 percent and the index for natural gas more than tr ipled. Price increases for electricity undoubtedly have led to increases in the eff iciency of electricity ut ilization (Edison Electric Institute, 1984b) . However, there has also been a trend toward greater electricity use resulting f rom the substitution of electricity for oil and gas, which have been increasing in price snuch faster than electric ity. Figure 2-16 shows that the ratios of electricity price to the prices of oil and natural gas continued their historical decreasing trend in the post-embargo period. Studies have shown that electric resistance heating becomes more cost-effective in residences than fuel heating at existing furnace efficiencies when the ratio of electricity price to competing fuel prices reaches three or below (Burwell et ale, 1982~. In the aggregate, the ratio of electricity price to heating oil price has been below three since 1978 while the ratio of electricity price to natural gas price is approaching three. Of course, using electric heat pumps, electricity can be cost-effective at electricity price ratios above three. The same price trend is apparent in the industrial sector. Although energy-using technologies in this sector are quite diverse, studies have shown that the same cost-effective price ratio thresholds for electricity and oil and electricity and natural gas have already been achieved in several industries as well (Burwell, 1983~. Price Elasticities The basically favorable relative price ratios of electricity to other fuels must be weighed against the fact that electricity price
46 300 to o 11 ret Cal 200 - X UJ Z UJ ~ 100 z 90 8 80 70 8 400 11 ~ 300 a' - X 200 z a: Q G o .tI: Personal Consumers Irk 1 970 1 975 100 ~ 90 _ 80 1 _ JO 1 _ / / / I me Industrial Producers fuel oil __ Natural gas E lectrici~y ~ 1980 1984 Natural gas ~ ., ~ / Petroleum `_ - ~ ·. . . ! -A ~ Electricitv 1970 197 5 1980 1984 YEAR . FIGURE 2-15 Trends in real energy prices, 1967 through 1984. SOURCES: U. S. Department of Labor, Bureau of Labor Statist ics, Handbook of Labor Statistics, Bulletins 2000 and 217S; and Monthly Labor Review, various issues.
47 RESI DENTIA L o 8: 7 6 Hi: ~ 5 LL A: )4 _ 3 _ 2 _ 1 9 :\~~\~ 8 7 ~ \ \ Electricity/Oil\ o Etectricity/Gas 6 LU 5 4 \' I N 1) USTR I A L -\ : _ 3 _ ElectricitY/Distillate Oil ~\ Electricity/Gas Electricity/ \ ~A Residual Oil \ I l I . 1 l l I I , 1960 1970 1980 1984 1 960 1970 1980 1984 YEAR YEAR FIGURE 2-16 Electricity price ratios in the United States, 1960 through 1984. NOTE: Price ratio calculated as ratio of actual prices in dollars per million Btu. SOURCES: Based on data f rom Edison Electric Institute, Statist ical Yearbook of the Electric Utility Industry; American Gas Association, Gas Facts; U. S. . Department of Commerce, Bureau of the Census, Annual Survey of Manufactures: Fuels and Electric Energy Consumed; rJ.s. Department of Energy, Energy Information Administration, Monthly Energy Review, various issues.
48 increases themselves- also tend to discourage electricity use. It is not simply a matter of dividing a fixed energy market among electricity and other energy forms; the total market for energy may be diminished as a result of rises in all energy prices. Econometricians have tried for years to disentangle the complex of own-price elasticity, cross-price elasticity (with other energy forms), delays in pr ice responses, and nonpr ice f actors such as income and changes in end-use technology. Powerful statist ical tools are employed, but the results leave much to be desired. The available data base does not yet cover enough experience with high energy prices, and, in addition, theory provides very little guidance in this complex task. Bohi (1981) provides a critical review of the methods, data, and results of the leading econometric analyses of the demand for energy forms. He finds that even for electricity, which has been subject to extensive study, 'there is wide disagreement about the responsiveness of demand to changes in prices and incomes, and surprisingly broad gaps in the understanding of the nature of this process" (p. 551. Each sector is considered separately by Bohi. In 25 studies of the residential sector, the spread of long-range, own-price elasticities was found to be -0.45 to -1.89, with a consensus value of -1.0 (that is to say, a 10-percent price increase would produce a 10-percent consumption decrease). However, after considering the methods and data employed in the studies, Bohi concludes that the best estimate for long-range residential price elasticity is -0.70 (p. leg, Table 7-1; long-range effects are usually considered as achieved in up to 10 years). Five studies on the commercial sector are reported. Some used more than one approach, so that there are nine different sets of results. Price elasticities ranged from -1.0 to -1.60. Bohi declines to choose the most likely value, saying simply that "commercial demand appears to be price elastic in the long run...." (p. 79~. Review of a broad range of industrial electricity demand studies, which used a variety of approaches, yielded a range from -0.51 to -1.82 with a consensus estimate "around -l.30.n But Bohi notes that "...one has difficulty in placing much confidence in the consensus estimate" (p. 90~. His own judgment, after examining the various studies, is that the price elasticity of industrial demand is between -0.5 and -1.0 (p. 159, Table 7-1). Sweeney (1984) concludes that "the long-run delivered price elasticity of demand for electricity probably exceeds [that is, is more negative than] unity but may be as low as -0.7 (p. 36). Bohi discusses cross-elasticity estimates but does not present numerical values. He notes that problems in the data tend to make estimates of cross-elasticities unreliable. The most that can be safely concluded, therefore, is that own- and cross-price elasticities exist that are nonnegligible, but hard to establish precisely. As a result there are counteracting price influences on electricity demand--in particular, electricity's own price and electricity's price movements compared with those of other energy forms. In addition, of course, there are the sizable effects on the growth in electricity demand produced by the overall growth in the
49 national output of goods and services. The net aggregative ef feats of all of these forces are assessed in the next section comparing pre- and post-embargo trends in electricity consumption. CONTINUITY AND CHANGE: PRE- AND POST-EMBARGO TRENDS The foregoing discussion shows that growth rates of electricity use have slowed in recent years from the high growth rates of earlier periods. Data for the postwar period are presented in Table 2-6. GNP growth, averaged over recent years, has also slowed from the higher rates achieved over most of the postwar period (Table 2-6~. In light of the strong association that has long been observed between electricity and GNP, viewing electricity growth rates only with respect to time can give misleading impressions. Nevertheless, the ratio of electricity growth rates to GNP growth rates has been gradually declining (Figure 2-S), a point to which many analysts have also drawn attention. Although this trend is consistent in principle with a linear relationship between electricity use and GNP, the question remains whether the degree and rate of convergence are consistent with the trend that has characterized the entire postwar period. Electricity price changes are frequently cited as the reason for a shift in the relationship. The econometric studies summarized above show that when the price of electricity increases it tends to slow the growth of electricity demand. However, the more recent historical period over which these statistical analyses were performed also contained the counteracting influences of rising competing fuel prices, which tend to counterbalance to some degree the effect of the electricity price increases. The extent of the competing fuel price influences on the historical relationship is less well established. Our examination of the data leads us to believe that by far the most important contributor to the slower growth rates in electricity demand over the last decade has been lower economic growth. Others have come to a similar conclusion. The econometric analysis of Hogan (1984) shows that the primary reason for the lower growth rates in electricity demand during the 1970s was slower economic growth. He attributes only about 30 percent of the drop in electricity demand since 1972 to electricity price increases. The Edison Electric Institute (1984b) reached similar conclusions regarding the magnitude of price effects at the aggregate level. However, Hogan notes that his results capture "only part of the eventual adjustment we can expect in the gradual replacement of energy-using equipment" (p. 27~. Thus, it can be expected that the energy price changes already experienced will continue to affect demand growth in the future. The central question is, of course, what the net effect of all factors--price, income, structural change, technological advance, and so forth--has been on electricity demand in recent years and what these influences portend for the future. In our judgment, at the present time there is no clear answer to this question. Figure 2-17b
50 TABLE 2-6 Average Annual Growth Rates of Electricity and Gross National Product (GNP) and Their Ratios over Selected Postwar Periods Rat lo of Electricity Electr ic ity GNP and GNP Per lad Growth Rate Growth Rate Growth Rates 1947-1960 8.07 3.52 2.29 1960-1973 6.70 4.17 1.61 1973-1983 1.99 2.04 0.98 SOURCES: Based on data f rom Edison Electric Institute, Statistical Yearbook of the Electric Utility Industry, various issues; U.S. Department of Commerce, Bureau of Economic Analysis, The National Income and Product Accounts of the United States, 1929-76, Statistical Tables; and Survey of Current Business, various issues.
51 identifies three different interpretations (depicted by lines A, B. and C) of the recent trend in electricity use as a function of GNP. The first interpretation (line A in Figure 2-17b) is that no shift has occurred in the underlying long-term relationship, but that the data for recent years represent a "down phase" of the cycle that has persistently characterized the long-term relationship. In fact, over the postwar period it could have been inferred several times that shifts had occurred in the slope of the relationship given its cyclic movements (Figure 2-17a). Nevertheless, these inferences would have been incorrect, as shown by data for subsequent years. If the cycle continues as before, there will soon be a period when the consumption of electricity will exceed trend values. However, if the growth rate of electricity use does not exceed the growth rate of GNP over the next few years (that is, if the "down phase" is not succeeded by an "up phase"), this interpretation must be considered incorrect. A second interpretation is that a permanent shift has occurred in the relationship, one toward a diminished increase in electricity use per dollar increment in GNP. This interpretation corresponds to a downward shift in the slope of the electricity-GNP trend line, and in fact the relationship can be read in such a way as to support this belief (Figure 2-17b, line B. which is an extension of the uppermost arrow on Figure 2-17a). Still another interpretation is that the increase in the rate of structural change between the industrial and commercial sectors and within manufacturing in recent years will neither be corrected nor proceed at the same rate in the future. If the future shift were to revert to the slower historical postwar rate and the sectoral electricity intensity relationships continue to hold, then the effect on the electricity-GNP trend would be a parallel downward shift in the postwar trend line (an intercept shift), leaving the slope coefficient intact (Figure 2-17b, line C).* ~ *If this result occurs, it would be a mirror image of the transition f ram the relationship of the 1920s to that of the 1930s. Incremental electricity intensity did not change between the two periods, as can be seen f rom the parallel lines of regression for the two periods (Figure 2-31. Because the percentage decreases in GNP in the years immediately following 1929 were larger than the percentage decreases in electricity use, average electricity intensity actually increased. However, as the years passed the trend of the 1930s turned out to be parallel to that of the 1920s. The net result of the change in average intensity was simply a shift in the line parallel to itself (mathematically, a shift of the intercept) . At the t ime that the shi f t occur red, the re was of course no way of knowing that this outcome would result. Likewise, only time will tell whether the line of the future will fall below and parallel to the line representing 1947 through 1973. Such behavior would ref. lect a decrease in average intens ity as a result of pr ice effects, conservation, and the permanent decline of some energy-intensive industries, with any lingering effects of past price rises being offset by the tendencies toward increased electrification already di scussed.
52 2400 2200 2000 Mann - 1400 a1980 // THE RECORDt9824| ~ //1981 I _ \ /19834 2000 - {/- c 1800 ;/ 01600: ~j: 1, /; 1 200 _ ~ooo: I: 800 ~.t -6oor ~ 400 :1947; 200 1 1 1 1 1 1 1 200 400 600 800 1000 1 200 1400 1 600 1 200 1400 1 600 1800 > 2800 2600 2400 \ 2200 \2000 1 800 1600 b SOME POSSIBLE FUTU R ES A / .~ i// 1981 // \ _ · / ha/ / 1 973 / ~ 1982 1 1 ~ GNP (billions of t972 dollars) GNP (billions of 1972 dollars) FIGURE 2-17 Electricity vs GAP: (a) the 1947-1983 record, (b) some possible f uture relet ionships . NOTE: The long-term trend line in the left f igure is the same as in Figure 2-4. The lines with the arrowheads in this f igure indicate how the trend seemed to be changing at various times in the past based on short-term movements of the data. However, these movements turned out to be aberrations, and there was always a reversion to the underlying long-se rm trend. The f igure to the right depicts three possible interpretations of the recent past, none of which can be proved or disproved at the present time. Line A is a continuation of the basic long-term trend of the main f igure; line B is a continuation of the short-term trend starting in 1976; line C is a new trend line. Line C assumes that the decline in average electricity intensity that occurred through 1984 represents a lasting change, but that incremental intensities will revert to the basic trend (see text for further discussion and for a historical parallel) . Cyclic and random variations (e.g., because of weather) around any future trend line will still occur, no matter what the trend line turns out to be.
53 It will be several years before these questions are resolved. The post-embargo years are still too few to provide definitive answers about trend shifts. In the meantime, however, the historical record suggests that the electrification of the economy will continue. Indeed, electricity use has continued to increase in all sectors over the post-embargo period while fuel consumption more generally either has been stable or, as in most cases, has fallen, as is shown in Figure 2-18. Furthermore, our examination of the major consuming sectors indicates that substantial potential remains within these sectors for the continued penetration of electricity in many uses. Generally, the rates of growth in electricity use will depend on the strength and growth of the economy. That much is clear. The exact quantification of this relationship for the current period and its relevance to future trends are important questions that remain to be settled.
54 lot 8: 6: 2 1 ,0 1 1 , 1 RESIDENTIAL ~ l tar _- ~1~~~ ~ ___ ________ I . . .. - · .... Petroleum 1 A t I 1 1 , 1 960 1 965 1 970 1 975 ro 8 c o ._ ._ c o ~4 _ of o C: Rae z LU o l2r 10: 8g 6 4 2 _ 1 980 ~ 984 COMMERCIAL I ~: lo_ - ~| Naturai aas Petroleu m · · . · . 19601965 1970 1975 1980 1 984 INDUSTRIAL I Petroleum atu ral gas _ ' _ ~.-- /1 ~ ~ O I I 1 ! 1960 1965 1970 YEAR Coal 1975 1980 1984 FIGURE 2-18 Gross energy use by economic sector, 1960 through 1984 . ,. ~ SOURCES: U. S. Department of Energy, Energy Information Administration, State Energy Data Report, DOE/EIA-0214 (83~; and Monthly Energy Review, var ious issues .
~5 REFERENCES Bohi, D. R. 1981. Analyzing Demand Behavior: A Study of Energy Elasticities. Baltimore: The Johns Hopkins University Press for Resources for the Future. Burwell, C. C., et al. 1982. Electric Home Heating: Substitution for Oil and Gas. Oak Ridge, Tenn.: Institute for Energy Analysis, Oak Ridge Associated Universities. ORAU/IEA-82-3 (M). March. Burwell, C. C. 1983. Industrial Electrification: Current Trends. Oak Ridge, Tenn.: Institute for Energy Analysis, Oak Ridge Associated Universities. ORAU/IEA-83-4 (M). February. Edison Electric Institute. 1973. Historical Statistics of the Electric Utility Industry Through 1970. EEI 73-34, p. 9. Edison Electric Institute. 1984a. Statistical Yearbook of the Electric Utility Industry, 1983. Edison Electric Institute. 1984b. What happened to the 40 percent? Electric Perspectives. Pp. 28-31. Spring. Geller, H. S. 1983. Energy Efficient Appliances. American Council for an Energy-Efficient Economy and the Energy Conservation Coalition . June . Hirst, E., et al. 1983. Recent changes in U.S. energy consumption: what happened and why. Annual Review of Energy 8:193-245. Hogan, W. W. 1984. Patterns of Energy Use. Cambridge, Mass.: Energy and Environmental Policy Center, Harvard University. E-84-04. May. Meyers, S., and L. Schipper. 1984. Energy in American homes, changes and prospects. Energy 9~6) :49S-504. Solar Energy Research Institute. 1981. A New Prosperity: Build ing a Sustainable Energy Future. Andover, Mass.: Brick House Publishing. Sweeney, J. L. 1984. The response of energy demand to higher prices: what have we learned? AEA Papers and Proceedings 74~2~:31-37. Tansil, J., and J. C. Moyers. 1974. Residential demand for electricity. Pp. 3 7 5- 3 8 5 in Ene rgy: Demand , Conse rvat ion , and Institutional Problems, Michael S . Macrakis, ed. Cambridge, Mass. MIT Press.
56 U.S. Bureau of the Census. 1984. Characteristi General Housing Characteristics. States, Tables 1330, 1339, and 1347. U.S. Energy Information Administration. 1980. DOE/EIA-0314. June. ics of New Housing and Statistical Abstract of the United 1982. Housing Characteristics U.S. Energy Information Administration. 1984. Energy Conservation Indicators, 1983 Annual Report. DOE/EIA-0441~83~. October. Werbos, P. J. 1984. Documentation of the PURHAPS Industrial Demand Model: Volume 1, Model Description, Overview, and Assumptions for the Annual Outlook 1983. Energy Information Administration. DOE/EIA-0429/1. April.