UNIVERSITY OF . ILLINOIS ypRARY AC URBANA- CAMPAIGN £NGLN£ERLtiQ N NOTICE: Return or renew all Library Materjalsl The Minimum Fee for each Lost Book is $50.00. .lllS A O The person charging this maTe¥iaTisVe*|aB©ible for its return to the library from which it was withdrawn on or before the Latest Date stamped below. TheflfHuAaMfc&ftnltfmjfg ft^&f-f feasons for discipli- e University. nary To re; UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN F£ 28 KFB L161— O-1096 UIU UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN URBANA, ILLINOIS 61801 ittlmSBls Tto» JJbnwy of the University t Illinois •I urtena-Charr CAC DOCUMENT 172 ENERGY GROWTH IN THE U.S. ECONOMY By Clark W. Bullard III Bruce M. Harmon Energy Research Group Center for Advanced Computation University of Illinois at Urbana-Champaign Urbana, Illinois 6l801 April 1976 This work was sponsored by the National Science Foundation Digitized by the Internet Archive in 2012 with funding from University of Illinois Urbana-Champaign http://archive.org/details/energybenefitsco172putn ABSTRACT The relation "between energy and gross national product and the potential for substituting capital and labor for energy are discussed, Detailed data for the 1960's are analyzed to determine the relative impacts of changes in technology, lifestyles, and income levels. TABLE OF CONTENTS Page Introduction 1 Energy and GNP 2 Substitution of Energy for Other Factors of Production. . . k Energy Growth from 1963 to 1967 7 Outlook for the Future 13 Summary and Conclusions 1^ References 17 LIST OF FIGURES Figure Page 1. Energy Productivity since 1920 3 2. U.S. Industrial Labor, Electricity and Capital Cost Ratios, 1926 to 1975, Current Dollars (Manufacturing Worker's Hourly Wage, the Industrial Price of a Kilowatt- Hour of Electricity and the Yields on AAA Corporate Bonds) 5 3. Primary Energy and Labor Intensities for the U.S. Economy 1963 and 1967 11 k. Electric Energy and Labor Intensity for the U.S. Economy, 1963 and 1967 12 5. Extension of the Ford Foundation's projections 15 LIST OF TABLES ___________ _____ Table Page 1. Elements of Energy Growth 1963-1967 9 ENERGY GROWTH IN THE U.S. ECONOMY* INTRODUCTION The emerging national policy of energy conservation is potentially in conflict with another national goal of long standing: continuous material economic growth.** In the past, growth in energy consumption and Gross National Product (GNP) have been closely tied, except during periods of rapid social and technological change. If both national goals are to be realized simultaneously the mechanism for decoupling energy and economic growth must be understood. Our purpose here is to examine the historical relationship between energy use and material economic growth as measured by growth in the Gross National Product. We shall investigate the substitution of other primary factors of production, (e.g. , capital and labor) for energy in order to separate energy use from economic growth. Finally, we shall use the wealth of detailed data available for two years, 1963 and I967, to explore the causes underlying the 2U% increase in U.S. energy consumption during that period. *The authors wish to acknowledge the work Mr. Craig Foster of the CAC Energy Research Group who was responsible for the data management and com- puting for this analysis. **The merits of each of these goals are debatable and beyond the scope of this paper. On energy conservation see Freeman (197*0 ; on growth see Nordhaus and Tobin (1970 ) and Daly (197*0. Schurr, et al (i960) discuss this in more detail, noting aberrations during periods of war, depression and the effects of automation. ++ In the long run, however, the two cannot be decoupled, and it is likely that future energy availability and social attitudes will result in changes to one of these goals. ENERGY MP GNP The historical relationship between energy and national product is shown in Figure 1. It is presented as a ratio called energy productivity, indicating the value of goods and services produced per unit of energy consumed. The growth in energy productivity has slowed to nearly zero in recent years , after a steady rise from 1920 to 1950. The behavior of this ratio in recent years has attracted the attention of many researchers, most notably Netschert (1971).* The period considered by Netschert (19^-7-1970) was characterized by relatively small fluctuations, many of which might be attributable to climatic conditions or to uncertainty in the energy and GNP statistics. To effect a substantial decoupling of energy and GNP growth, substantial changes in energy productivity are needed. For example, if energy productivity could grow at 3% per year, the same rate of economic growth could be sustained with zero energy growth. Alternatively, a slower rate of economic growth could be sustained while actually decreasing U.S. energy consumption. As a reference point, notice that the "zero energy growth by the year 2000" scenario of the Ford Foundation Energy Policy Project requires an average annual increase in energy pro- ductivity of 2.27 % over a 25 year period. The mechanism for changing energy productivity is apparent from *See also Linden (1976) **A 10 percent variation in space heating requirements is not unrealistic, N0AA (1973). Since on the average 20% of U.S. energy use is for space heating, a 12 percent fluctuation in energy demand is possible. In addition, there is reason to believe that errors in the measurement of energy consumption and GNP may account for variations of similar magnitude (See Mor gens tern, 1950). Note that most of the fluctuations in Figure 1 during the last 30 years would fit within a i5 percent envelope). 1920 1930 1940 1950 I960 1970 1980 Fig. 1 Energy productivity since 1920, Source: Bullard (191 h) \ the relatively simple equation linking energy demand to four key variables: E = £-y_ g P (1) Here E in the energy use during a given period, P is U.S. population, g is per capita GNP during the given period, y_ is a vector (whose elements sum to l) of normalized dollar values of goods and services making up the GNP , and e is a vector denoting the energy required, directly and indirectly, to pro- duce a dollar's worth of each good and service. To decouple energy from GNP, changes are necessary in y_, the market basket of goods and services (e.g. , lifestyles), or e_, the technology of producing goods and services, so as to require less energy. Efficiency of production, as well as efficiency of consumption (of goods and services) are equally important. Examples of technological changes include not only the reduction of direct energy inputs to a manufacturing process, but also the substitution of other inputs: material, capital, labor, for energy. These are discussed in more detail in the next section. SUBSTITUTION OF ENERGY FOR OTHER FACTORS OF PRODUCTION Buring the last twenty years, the price of energy has declined relative to the price of capital and labor (see figure 2). Under such conditions, the latter factors would be replaced by increased energy use on the part of producers and consumers alike. Technological change^ however, appears to have kept this effect from significantly influencing the energy-GNP ratio. Substitution of labor for energy has characterized an even longer per- iod. Automation in industry and commerce is perhaps the most visible example, but there are many other less obvious instances, such as the replacement of retail clerks by highly illuminated spacious shopping areas with energy intensive self-explanatory packaging. The same substitutions have taken k < Q. < O / q cvi in o 10 CD 0) o CO m -m o cc m < m 0) o 0> m ro 0> O ro 0) in CM LO "=C r»» CT> U_ r- O O LU I— o 1— 1 ko cc C\J Q. cr> r— _l«— ^ •— I CO co oc o OhZ •— i CO O 1— ID CO >=C Q CC ^ LU hh H- CO LU CC o 3: o O h— D_ q: _i «o cC LU <_> 1— CD >— - —i z: o o: o z: =5 «=C O CO or Q 5- _J 1— CO LU H- 1 - l-H o ct: >- i— i LU d: y uj I— o; hi o O 1— LU 3 _J Q lu o z: 2= - ODh CQ h— t— i «=C O C_> _l I— — -LU i — CO 3D "LL- n CS CO O c ■z. cc o i— t • LU 3 u CMQiO s- CC _J =3 (UDM o S- O ^C CO 3 C7) o b X3QNI 0I1VU 1S00 place in the home, in the form of labor saving appliances and synthetic fabrics. The substitution of energy for capital is more recent, but quite signif- icant. Perhaps the best example is the inadequate insulation of buildings, resulting from incentives to minimize the down payment on housing. Similarly, many appliances with a low first cost have very inefficient electric motors.* These trends might reverse as energy prices rise, with the ultimate sub- stitution of capital for energy being the use of capital intensive solar energy systems to replace consumption of conventional forms of energy. The standard economic production equations relate output to various ex- pressions involving the inputs of labor and capital, the primary factors of production. Great attention is paid to the productivity of each factor: that is, to how much of it is required to produce a unit of output; and to mar- ginal productivity, the change in output with respect to a unit change in the use of a factor. The ratio of the marginal productivities for capital and labor equals the ratio/of their unit prices. Thus, a time series for the average production worker's wage divided by the cost of capital shows which factor is being sub- stituted for the other. Such a series is presented in Figure 2. It indicates that labor was marginally more efficient with respect to capital from the period of the Great Depression until 1955- This implies that capital was being substituted for labor. Since that time, the ratio has been rather neutral. Energy should also be considered as a primary factor of production be- cause of its obvious physical importance in production processes. A time *See Allen (197*0 series of the ratio of the cost of capital to the cost of a kilowatt hour of electricity* is also shown in Figure 2. The data indicate neutrality with regard to substitution until about 1955. Since then, the cost of capital has risen much faster than that of electricity. This trend helps explain the promotion of off-peak-period uses of electricity, such as electric heat and exterior night lighting by electric utilities. It also helps explain such things as the general tendency in the U.S. not to install adequate build- ing insulation, the trend toward less-energy-efficient electric motors of fractional horsepower, and the movement toward centralization by many in- dustries. ** The most consistent pattern in Figure 2 is the wage-price of electricity ratio. It grew by a factor of twelve in the forty years since 1935, in- dicating a continous pressure to substitute electricity for labor. The capital-electricity ratio shown in Figure 2 indicates that the rising relative value of capital would be a deterrent to the substitution of machines and electricity for labor. However, the share of the average production dollar allocated to capital and electricity, to all energy for that matter, is a small fraction of that going to labor. ENERGY GROWTH FROM 1963 to 1967 U. S. energy demand rose 25% during this U-year period. In this section we will attempt to attribute this growth to changes in each of the four Electricity was chosen for comparison because it is almost totally available for mechanical work, and represents the most logical substitute for labor. The same conclusion results when the average price for all energy is used. **Centralization allows full-time use of capital equipment which is capital- efficient, but requires large transportation networks which are energy- inefficient . variables of eq. (l). We focus on the years 1963 and 1967 because they are the only ones for which a detailed energy input-output model for the U.S. economy is available. The model, discussed in detail by Bullard and Herendeen (1975) describes at a 360-sector level of detail the technology of production and the mix of goods and services produced. At this level of re- solution, the energy intensity (Btu required directly and indirectly per unit of goods and services produced) is determined as a function of over 135,000 parameters describing the technology of production. Similarly, the options for consumers to substitute competing goods and services having dif- ferent energy intensities can be modeled at a meaningful level of detail. As seen in Fig. 1, the energy/GNP ratios for the 2 years are not sub- stantially different. In order to find out if there were offsetting changes in technology and consumption patterns, the effect of each change was eval- uated. Expanding equation 1 we obtain o E + AE = eygP + AeygP + eAygP + eyAgP + eygAP + 0(A ) (2) where the first term on the right hand side is the 1963 energy consumption, and the next four terms represent the first order contributions of changes in each of the four variables and the remaining terms indicate the higher order interactions. Performing the calculation, it was found that the higher order terms were negligible: the first order terms accounted virtually all of the total 2k% increase in energy use. The results summarized in Table 1 show that the technological and lifestyle changes were small and essentially offsetting and the contribution of population growth was relatively small at 5%. Most of the growth in U.S. energy demand could be attributed to growth in per capita GNP (l8$). The most significant lifestyle (consumption pattern) changes accounting 8 First-Order Effect % Increase in Energy Use Technology - 1.2 Lifestyle 2.5 GNP/ capita IT. 7 Population 5-0 Table 1. Elements of Energy Growth 1963-1967, for the increased energy demand were purchases of more airline tickets, chem- ical products, radio and TV equipment, and the increased use of electricity in homes. The data show most products became less energy intensive (Btu/ $ worth) but it is not possible to separate this effect (perhaps due to energy economies of scale in centralized production) from the possible increased re- quirements for transportation of goods to final consumers.* Figure 3 shows the energy and labor intensities of goods and services, and how they changed from 1963 to 1967.** The chart has been corrected for inflation, and shows that production of most goods and services required slightly less energy and much less labor in 1967 than in 1963. These are the results of changes in the producing sectors of the economy, consistent with the long-term increase in the wage/energy price ratio shown in Figure 2. Figure k is similar to Figure 3 except that electricity intensity is plot- ted against labor intensity, showing that electrical energy was more generally substituted for labor than was total primary energy. This substitution effect increased the use of electricity by about 2.5 percent during the period 1963-1967. This is not surprising because electricity is nearly totally available for mechanical work. Since energy prices were relatively stable during the period, it is likely that the reduction in energy requirements resulted either from anticipated increases in energy prices, or from technological change. The rapid energy price increases experienced since the 1973 embargo represent a significant departure from historical trends in the relative price of energy. Based on these new prices, the Ford Foundation Energy *This would show up as increased purchases of transportation services by final consumers (a "lifestyle" change). **Calculations based on methods described by Bullard and Herendeen (1975) 10 ex en t— t- d CL a a >- co r- UJ LO UJ ai en ^ i i x >- CVJ -■ 3* CM I— 0"09in h- 0'89E h- O'9/.S 1— CTTi9l oc_) - L 2 a: r- co oo 2uj a - m o r- a: LU r» V «— • r- UJ in O) •— i n: rr CJ C\J (T) u; -P _3 >- 10 CO CO ?■ »— < CO ItJ (O CO r-» s o CT) (11 ec 3" cr en I— H- - en r- UJ lo to en en ±c i i x >- en CT. fj 1- >- C_) t- HI -^ CO o a (X iLl i- t- < ) h- iki f) j -J UJ a. o CC 1- O 2 < s '^*2 ee-e wrs o/."i ge-i $ uiAiiGb ^96t/snig ooo'oi— j.irsN3iNr -d H § o CJ CO 0) -p u o >> -p •H CO c 0) -p 5h o ■3 'd a} !>> u a o •H ?H -P o id •H 89-0 Air3IcJ133"13 TiE'O 00 - 12 Policy Project identified a set of energy conserving technological changes which are economically feasible under present conditions. To ascertain the relative magnitude of these changes, the model was used to determine what energy demand would have been in 1967, had these technological changes been in effect then. The resulting energy demand of U3.8 quads represented a 28$ savings below the actual energy use level. The technological changes predicted by the Ford Foundation are not expected to be fully implemented before 1985 however. This is due to the general comple- mentarity of capital and energy and to the constraints imposed by the age struc- ture of existing capital stocks. Since capital stocks, expecially in the form of machinery, can be viewed as devices for converting energy into heat or useful work, the complementarity of capital and energy will dominate short run responses to energy availability. In the longer run, the two factors are substitutable, as discussed earlier, and significant energy savings might be accomplished. OUTLOOK FOR THE FUTURE It is important to recognize that changes in technology and lifestyles are limited: there are thermodynamic limits to energy requirements for pro- duction processes, and the elements of the lifestyle vector are constrained to sum to unity. So long as per capita GNP and population increase expo- nentially, these variables will dominate eq.(l)in the long run. Technological and lifestyle changes can have significant impacts only during transition periods. In this respect, the Ford Foundation study was somewhat misleading. Their so-called "zero energy growth" scenario was characterized primarily by technological and lifestyle changes occurring between 1975 and 2000. These changes did sufficiently decouple energy and GNP growth to bring energy 13 growth to zero in the year 2000 while the 3.5% annual GNP growth was virtually unahated. Figure 5 illustrates what the Ford Foundation scenarios would imply for the long run if the prescribed changes saturated the potential for energy conservation. The exponential growth of GNP would dominate, and the net effect of switching to smaller cars, changing lifestyles etc., would only he to shift the curve to the right. The dilemma is not unlike a dieter wishing to limit calories and switching to low calorie food while increasing the total quantity of food eaten. The calorie intake can be slowed only for a short period while the composition of the diet is being changed. SUMMARY AND CONCLUSIONS Growth in energy consumption cannot, in the long run, be decoupled from growth in material affluence. In a world with limited energy resources, there- fore, it seems reasonable to expect an eventual transition to a state of zero population growth where well-being is no longer defined in terms of material flow rates. Because of the associated value-shift such a transition may take generations to effect without significant trauma, and it is quite pos- sible that the market will be incapable of providing signals through the price system early enough to assure a smooth transition. It is apparent from analysis of historical data that i) there have been no incentives to substitute labor or capital for energy, ii) technological and lifestyle changes during the 1963-19&7 period had a very small impact on energy use, iii) the substantial rates of growth in population and material affluence have required a matching growth in energy demand. Since energy prices were relatively stable during the 1960's the data can tell us nothing about the magnitude of potential energy savings that might be realized lU 250 - 200 .— * 3 • 1 H- m in o ~ 150 > e> i/ cc u 2 P f/ <5K 100 £/<%> — 50 - vS PROJECTIONS BASED ON 3.5% '^HISTORICAL FORD INCREASE IN DATA PROJECTIONS GNP/YEAR r» i i i i i i 1950 I960 1970 1980 1990 2000 2010 2020 2030 Fig. 5 Extension of the Ford Foundation's projections. 15 through technological and lifestyle changes induced by energy price increases . A prompt reversal of the historical trends in the price of energy re- lative to labor and capital would send early signals through the market that would be consistent with the inevitable value changes associated with the transition to a less energy intensive way of life. In the short run these relative price changes would provide incentives for substituting other factors of production for energy, thereby attenuating the growth in energy use, as we enter a period characterized by great uncertainties shrouding energy supplies: the possibility of public rejection of nuclear fusion technology; the chance that C0„ effects on the ozone layer may necessitate curtailment of fossil fuel use; unforseen events on the international scene; etc. It is clear that a sound national energy policy should include an instrument for control- ling energy prices to internalize many of these costs and risks. We suggest an energy tax as a means for resolving these problems. Work- ing through the price system it would retain freedom of choice. Accompanied by income tax reform, it need not be regressive. Accompanied by a program for disseminating information about its intentions and impacts, it could be quite effective. The key to its success lies in the formula for distributing the revenues from the tax in a way that offsets the political reaction of capital and labor interests. Since the initial increase in the relative price of energy would be perceived as a reduction in the returns to capital and labor, the first reaction of these interests would be to oppose an energy tax. A more detailed description of these problems is given by Hannon (1975) • 16 REFERENCES Allen, J. 19lh. "The Craft of Electric Motors," Environment , Vol. 16, No. 8, pp. 36-39. Bullard, C. , Energy Conservation Through Taxation , CAC Document 95 , Center for Advanced Computation, University of Illinois at Urbana- Champaign, Urhana, Illinois, February 197*+. Bullard, C. and Herendeen, R. , "Energy Impact of Consumption Decisions" Proc. IEEE , March 1975- Daly, H. "Steady-State Economics vs Growthmania: A Critique of the Orthodox Conceptions of Growth, Wants, Scarcity, and Efficiency", Policy Sciences 5, (197*0 pp. 1U9-I67. Freeman, D. S. , A Time To Choose, America's Energy Future , Final Report of The Ford Foundation Energy Policy Project, Ballinger, Cambridge, MA, 197*+. Hannon, B., "Energy, Growth & Altruism", 1975 Mitchell Award - First Prize, Limits to Growth '75 Conference, Woodlands, Texas. Morgenstern, On the Accuracy of Economic Observations , Princeton University Press, Princeton, NJ (1950). National Oceanic & Atmospheric Administration, U.S. Department of Commerce, "Variability of Seasonal Total Heating Fuel Demand in the U.S.", Report to the Energy Poling Office, Executive office of the President, Sept. 18, September 18, 1973. Netschert, B. , "Energy Consumption and Gross National Product" (mimeo) National Economic Research Associates Inc., Washington, D.C. 1971. Schurr, S. H. , et. al. , Energy in the American Economy 1850-1975 , Johns Hopkins Press, Baltimore i960. 17