ENERGY FOR ASUSTAINABLE WORLD

(iiild

wi K.\. IM

Kulvr'l II. Willum^

W O R L D R t S C) U R C H S I N S T I T U T L

ENERGY FOR ASUSTAINABLE WORLD

Jose GoldembergThomas B. JohanssonAmulya K. N. ReddyRobert H. Williams

A book-length version of this study with the same title is being published by Wiley-Eastern, NewDelhi, India.

n

W O R L D R E S O U R C E SA Center for Policy Research

I N S T I T U T E

September 1987

Kathleen CourrierPublications Director

Hyacinth BillingsProduction Supervisor

Earthscan, Toyota Motor Sales, World Bank,An Exxon Photo, Renew America ProjectCover Photo

Each World Resources Institute Report represents a timely, scientific treatment of a subject of public concern. WRItakes responsibility for choosing the study topics and guaranteeing its authors and researchers freedom of inquiry.It also solicits and responds to the guidance of advisory panels and expert reviewers. Unless otherwise stated,however, all the interpretation and findings set forth in WRI publications are those of the authors.

Copyright © 1987 World Resources Institute. All rights reserved.Library of Congress Catalog Card Number 87-050961ISBN 0-915825-21-XISSN 0880-2582

Contents

The Energy Crisis Revisited 1

I. Some Surprises on the Demand Side . . . . 5

A Quiet Revolution in the Making 5Future Possibilities 6Relevance to Developing Countries 8Adding It Up 11

II. A Closer Look at the Energy Problem 17

A Temporary Reprieve forOil Consumers 17

Energy and Development 19The Other Energy Crisis 20The Hidden Costs of Conventional

Energy 22Global Insecurity and Middle East

Oil 22Global Climatic Change and Fossil

FuelUse 22Nuclear Weapons Proliferation and

Nuclear Power 23Looking Ahead—Reason for Optimism . . . 25

III. Alternative Energy Futures 27

The IIASA/WEC Numbers and OurNumbers: What They Mean 29

Future Energy Demand 29Future Energy Supplies 31

A More Hopeful Outlook 33

IV. Behind the Numbers 35

A New Approach 35Structural Shifts in IndustrializedMarket-Oriented Countries 36

The Growth of the Service Sector.... 36The Growing Importance of

Fabrication and Finishing 36The Structure of Energy Demand in

Developing Countries 41Energy and Basic Human Needs 43Energy and Employment

Generation 45Energy for Agriculture 48Energy Planning as an Instrument

of Development 49Opportunities for Using Commercial

Energy More Efficiently 49Space Heating in Industrialized

Countries 51Appliances in Industrialized and

Developing Countries 55Automobiles Around the World 59Steelmaking and Technological

Leapfrogging 67Case Studies: Sweden and the United

States 70Sweden 71The United States 71Conclusion 72

Energy Use in Developing Countries—A Thought Experiment 74

V. Changing the Political Economy ofEnergy 87

The Role of Markets in End-UseEnergy Strategies 88

Market Shortcomings 88The Importance of the Market in

Implementing End-Use Strategies . . . 90Market Intervention in the Present

Political Climate 91Promoting More Efficient Use of

Commercial Energy 91Eliminating Energy Supply

Subsidies 91Rationalizing Energy Prices 92Improving the Flow of Information . . . 92Targeting Energy Performance 92Stabilizing Consumer Oil Prices 93Promoting Comprehensive Energy

Service Delivery 94Making Capital Available for Energy

Efficiency Investments 95Assisting the Poor 95Promoting Research and

Development on Energy End-Uses .. 96

Creating and Maintaining EnergyEnd-Use Data Bases 97

Keeping Score on ConservationPrograms 97

Energy and Economic Development 98Energy Efficiency in the Modern

Sectors of Developing Countries . . . . 98Energy and Basic Human Needs 99Energy and Employment

Generation 100Energy for Agriculture 100Political Feasibility 101

International Actions 102Helping Developing Countries

Become More Self-Reliant 102Coping with Global Problems 105Conclusion 108

Notes 109

Appendix: A Top-Down Representation of OurBottom-Up Energy DemandScenario 113

Acknowledgments

T his report presents the major globalfindings of the End-Use Global EnergyProject (EUGEP), of which the authors

are co-organizers. Results of this project arepresented in more detail in a book of the sametitle being published by Wiley-Eastern, NewDelhi, India.

The authors acknowledge support for theresearch leading to this publication from theWorld Resources Institute, the RockefellerBrothers Fund, the John D. and Catherine T.MacArthur Foundation, the Alida and MarkDayton Charitable Trust, the Energy ResearchCommission of Sweden, the InternationalLabor Organization, the Max and Anna Levin-son Foundation, the Swedish InternationalDevelopment Authority, and the Macauley andHelen Dow Whiting Foundation.

The authors are indebted to Sam Baldwin,Deborah Bleviss, Gautam Dutt, HaroldFeiveson, Howard Geller, Frank von Hippel,Roberto Hukai, Eric Larson, Irving Mintzer,Jose Roberto Moreira, Lee Schipper, RobertSocolow, and Peter Steen for fruitful discus-sions, assistance, and encouragement in thecourse of research leading to this report, toAlan Miller for many discussions and reviewsof early drafts that helped shape the presenta-tion of our findings, and to David Sheridanand Kathy Courrier for their diligent editorialefforts in bringing this manuscript to its finalform.

J.G.T.B.J.A.K.N.R.R.H.W.

in

Foreword

T he 1970s saw nothing less than a revo-lution in energy use in the industrial-ized countries. While energy demand

had grown in lockstep with economic growthfor decades, the oil price shocks of 1973-74 and1978-9 shattered that historical relationship. Soprofound was the change that between 1973and 1986 there was no growth in U.S. energyuse while the economy grew by 30 percent.Comparable declines in the energy intensity ofthe economy occurred in Western Europe andJapan.

This change confounded much conventionalenergy analysis, and has fostered a new breedof energy analysts who have been studying theenergy problem, not with a primary focus onsupply, but with detailed analysis of energydemand. Whereas conventional energy analystsposed the question: "What mix of energy sup-plies can best meet projected demand?" (withdemand based on projected macroeconomictrends), this new breed asked instead: "Whatactual uses is energy needed for and how canthose needs be most efficiently met?" Foremostamong these "end-use" analysts have been theauthors of this study, an international team ofJose Goldemberg (Brazil), Thomas B. Johansson(Sweden), Amulya K. N. Reddy (India), andRobert H. Williams (U.S.), with whom WRIhas collaborated for several years.

What the EUGEP (End Use Global EnergyProject) analysts have found, and actual experi-ence has demonstrated, is that it is possible to

get far more bang for the energy buck thananyone dreamed possible before 1973, and farmore than conventional energy analysis stillsuggests today. How much more is, inessence, the subject of this study.

Building on detailed studies of the energyeconomies of the United States, Sweden, India,and Brazil, the EUGEP global energy scenariofor 2020 suggests a per capita energy use in theindustrialized countries of about half what itwas in 1980 (3.2 kilowatts per person ratherthan 6.3 kilowatts). Envisioning a widespreadshift in the developing world from traditional,inefficiently used non-commercial fuels tomodern energy technologies used in high effi-ciency applications, the EUGEP scenario seesan average of 1.3 kilowatts per capita (con-trasted to the present average of 1.0 kilowatts)supporting a living standard up to that ofWestern Europe today. Overall, even with pop-ulation rising to 7 billion, the EUGEP scenarioportrays a world in which global energy use isonly 10 percent higher than it is today, a starkcontrast to conventional projections whichshow an increase of more than 100 percent.

The EUGEP analysis is neither a projectionnor a policy prescription. It is rather an illustra-tion (a rather conservative one in that itemploys only commercially available or near-commercial technologies) of what is technicallypossible. Nonetheless, its policy implicationsare profound. Foremost among these is the fargreater flexibility and freedom of maneuver

governments have if they must plan for a 10percent rather than a 100 percent increment inenergy supply. Instead of having to push onnearly every energy front to meet projecteddemand, the EUGEP message to policymakersis that they can choose.

The EUGEP scenario described in detail inthese pages embodies choices that reflect theauthors' particular values: limited growth inthe use of nuclear power so as to lower nuclearproliferation risks; a balanced energy supplymix to reduce oil imports from the Middle Eastand thereby promote national self-reliance andenhanced global security; and, limits in the useof coal in particular and fossil fuels in generalso as to diminish the extent of greenhousewarming and other environmental risks.

To my mind these are good choices. Individ-ual policymakers may differ in their priorities.What remains crucial in this work is thedemonstration that at or near the present levelof global energy use, plausible energy supplymixes can be identified that would not serious-ly aggravate other pressing global environmen-tal and security problems, as conventionalenergy strategies do.

The work carried out by the authors of thisstudy is closely tied to the research conductedat WRI since 1983. WRI has studied the role ofbioenergy in development and industry, stagedan international conference on the same themeand investigated the impacts of energy sub-

sidies in promoting inefficient energy use inboth industrialized and developing countries.WRI's scientific and policy work on the green-house effect, on the depletion of the earth'sozone shield, and on acid rain, troposphericozone, and other air pollutants, also ties indirectly with the recommendations of thisstudy on how to satisfy economic goals with-out running up against severe environmentalconstraints.

WRI is pleased to have worked closely withthe EUGEP authors in supporting their workand in making the results of their researchavailable. Their effort is an excellent example ofcollaborative research at the international levelon an issue of global importance.

Both WRI and the EUGEP group wish to ex-press their gratitude to the following organiza-tions which generously provided the financialsupport that made this study possible: theRockefeller Brothers Fund, the John D. andCatherine T. Mac Arthur Foundation, the Alidaand Mark Dayton Charitable Trust, the EnergyResearch Commission of Sweden, the Interna-tional Labor Organization, the Max and AnnaLevinson Foundation, the Swedish Interna-tional Development Authority, and theMacauley and Helen Dow Whiting Foundation.

James Gustave SpethPresidentWorld Resources Institute

VI

The Energy Crisis Revisited

T he sharp rises in world oil prices in1973 and 1979 shook the world econ-omy to its foundations. (See Figure 1.)

Between 1973 and 1981 the industrialized worldpaid $1.5 trillion (1984 dollars) more for oil im-ports than it would have paid had the worldoil price remained at the 1972 level. Althoughoil exporters recycled some of this money backto the industrialized market economies throughincreased investments and purchases, the pricerises resulted in a tremendous transfer ofwealth and a huge loss in purchasing powerthat has affected all goods and services. In-creased oil cost is certainly one of the maincauses of the stagflation—simultaneous eco-nomic stagnation and price inflation—that besetthose economies in the 1970s.

Poor countries also suffered. By 1981, low-and medium-income developing countries werespending, respectively, 61 and 37 percent oftheir export earnings on oil imports.1 (See Table1.) These countries use their export earnings topurchase technologies for industrial and agri-cultural development and to pay internationaldebts. The cost of imported oil became a finan-cial drain on the developing world, economicgrowth in many poor countries stalled, and adebt repayment crisis occurred.

The sharp increases in world oil prices alsocontributed to real price increases for otherenergy forms, such as natural gas. Rises inelectricity prices, though, have been due largely

to other factors. Between 1979 and 1985, whileoil prices were falling sharply, the averageprice of electricity increased 21 percent inJapan, 20 percent in the United States, and 12percent in Western Europe. Electricity pricesrose because of the high cost of additional gen-erating capacity. In other words, the marginalcost of electricity is high compared to the cur-rent average price. In the United States, theproblem dates back to the early 1970s, whenthe long-term downward trend in electricityprices reversed itself. (See Figure 2.) The cost ofelectricity generated by coal and nuclear powerrose sharply in the 1970s, largely because ofquality control problems and more restrictiveenvironmental and safety regulations.

The capital costs of energy supplies are gen-erally rising. In the United States, from 1972 to1982 capital expenditures for energy suppliesincreased from 26 to 39 percent of all newplant and equipment expenditures (from 2.5 to4 percent of gross national product) whiledomestic production remained constant.2 Indeveloping countries, the share of domestic in-vestments committed to expanding energy sup-plies increased from 1 to 2 percent of grossdomestic product in 1970 to 2 to 3 percent in1980. Foreign exchange requirements forenergy investments in developing countriestotaled about $25 billion in 1982, or approx-imately one third of the foreign exchange re-quired for all investments.3

Figure 1.

>,ts

D_jCQCQ

ion

Mil

lB

BL

^

1985

60

50-

40 _

3 0 .

20 _

1 0 .

0_

•r

—

1968

The

The World Oil Price and Oil Consumption Rates

All Market Economies

i _ » - •—"N.

X OECD Countries / \

s \ _.. ^̂ / V• • ' * ' • . . - • ' ' ' • / \

/ ^ \

World Oil Price / *-V

/ j/

• —

! 1 I 1

1972 1976 1980 1984 1988Year

rate of oil consumption (in million barrels per day) is shown separately for the member countries ofthe Organisation for Economic Co-operation and Development, for developing countries, as well as for allcountries with market economies. The world oil price shown is the refiner acquisition cost of importedcrudePrices

oil in the United States (in 1985 U.S. dollars per barrel). Since 1977 the data are plotted quarterly.are converted into constant dollars using the gross national product deflator.

Table 1. Net Oil Imports andCountries, 1973-1984

KenyaZambiaThailandKoreaPhilippinesBrazilArgentinaJamaicaIndiaBangladeshTanzania

KenyaZambiaThailandKoreaPhilippinesBrazilArgentinaJamaicaIndiaBangladeshTanzania

1973

111

1732761669868371

308-

47

0.12.2

11.18.68.8

15.92.5

18.110.6

—12.8

Source: International Monetar}

Their

1974

2730

510967570

3,230328193

1,17092

153

4.15.1

20.921.720.940.7

8.327.329.726.538.0

Relation to

(million

1977

5753

8061,930

8594,200

338242

1,750172102

Imports as

4.89.5

23.119.227.534.7

6.032.427.536.120.2

r Fund, 1985

Export Earnings For Eight Developing

Net Oil ImportsU.S. dollars, current prices)

1979

11372

1,1503,1001,1206,920

351309

3,067247174

1981

31663

2,1706,3802,080

11,720302490

-509306

1983

208274

1,7405,5801,7408,890

---

286175

Percentage of Export Earnings

10.28.2

21.620.624.445.4

4.537.739.337.434.8

26.97.8

30.930.036.850.4

3.350.3

-64.652.7

21.220.827.322.835.440.6

---

39.447.0

1984

219454

1,4805,7701,4707,470

---

314156

20.321.420.019.727.827.7

---

33.642.3

Figure 2

|

lent

s 1

1985

(Long Term and Recent (Insert) Electricity Price Trends for the United States

4.S

40_

35-

30_

25-

20_

10_

5_

0-

\

\

\

\

RESIDENTIAL \\vAVERAGE \

INDUSTRIAL ^ ^ - ^

1935 1945

8 -

7 -

6 -

3 -

2 .

1 -

RESIDENTIAL\ V ^ ^ ^ * " ^ /

AVERAGE y ^

^ 'INDUSTRIAL

1960 1965 1970 1975 1980 19

Year

85

1 1 11955 1965 1975 1985

Year

Prices were converted to constant 1985 cents per kilowatt hour using the gross national product deflator.

I. Some Surprises on the Demand Side

A Quiet Revolution in the Making

A s painful as the energy crises of the1970s were, they led to a fundamen-tally new approach for managing

energy problems, one that offers the hope ofavoiding recurring energy crises. For the firsttime, energy decision-makers turned from thehistorical preoccupation with expanding energysupplies to examine how energy can be usedmore effectively in providing such services ascooking, lighting, space heating and cooling,refrigeration, and motive power. Decision-makers have found that energy services can beprovided cost-effectively with much less energythan previously thought necessary, and as aresult the historical close correlation betweenthe level of energy use and economic well-being has been broken. This revolutionarydevelopment has not been reported in dramaticstories in newspapers and magazines because itis not the result of any big government or in-dustry energy projects. Rather, this develop-ment is a quiet revolution in which not one butmyriad solutions to the energy problem are be-ing found, each matched to one of manyenergy end uses. Moreover, the "revolution-aries" are not just corporate heads of energysupply companies and government officials,they are a much more diverse community:manufacturers of energy-using appliances andmotor vehicles, builders of residential and com-mercial buildings, factory owners, home-owners, and others.

Consider some recent accomplishments. Inthe United States the average fuel economy ofnew cars and light duty trucks increased 66percent between 1975 and 1985; by 1985,resulting fuel savings were equivalent to 2.4million barrels per day (mbd) of oil, or nearly60 percent of U.S. oil imports. In Sweden, thealready energy-efficient steel industry, whichaccounts for about one fifth of Swedish manu-facturing energy use, reduced its energy re-quirements per tonne of steel produced by onefourth between 1976 and 1983. In Japan, whererefrigerators account for about 30 percent ofresidential electricity use, energy requirementsfor the average new refrigerator were reducedtwo thirds between 1973 and 1982.4

The aggregate results of such improvementshave been impressive for the countries of theOrganisation for Economic Co-operation andDevelopment (OECD). Oil use in OECD coun-tries fell 15 percent, or 6.1 mbd between 1973and 1985. In the same period, total energy useper capita for OECD countries fell 6 percent,while per capita gross domestic product (GDP)increased 21 percent. Some countries havemade even more impressive advances. Percapita energy use in the United States fell 12percent while per capita GDP rose 17 percent.In Japan, a 6-percent reduction in per capitaenergy use was accompanied by a 46-percentincrease in per capita GDP in this period.

Future Possibilities

Does the experience of OECD countries since1973 represent a new trend or just a one-timeadjustment to the energy price increases since1973? Strong evidence suggests that the currenttechnological revolution in energy efficiencymay persist for decades, because the most effi-cient technologies now commercially availableor under development far outperform existingtechnologies.

In 1986, the most energy efficient four-passenger automobile available was the 1986Chevrolet/Suzuki Sprint, which has an on-the-road fuel economy of 57 miles per gallon (mpg)(4.1 liters/100 kilometers [lhk]), making it near-ly three times as efficient as the average car inuse in the world. In both the United Statesand Sweden, new superinsulated houses re-quire just one tenth as much energy forheating as the average house. (See Figure 3.)Energy-efficient electrical devices now availableinclude: heat-pump water heaters that use onethird the energy of ordinary electric resistancewater heaters, air conditioners that use lessthan one half the energy required by typicalunits now in use, variable-speed controls formotors that reduce electricity requirements forfans, pumps, and other motive power uses by20 to 50 percent, and new lighting technologiesthat can cut lighting electricity use in commer-cial buildings by 50 percent or more.

Among other advanced technologies, theElred and Plasmasmelt steelmaking processesunder development in Sweden would reduceenergy requirements one third or more relativeto what the Swedish steel industry has alreadyachieved. In 1986, Toyota unveiled a prototypecar for four or five passengers with a fueleconomy of 98 mpg (2.4 lhk). (See Figure 4.)

Many of these technologies are now eco-nomically attractive on a life-cycle cost basis.Nonetheless, some will not be readily adoptedbecause more efficient technologies tend to re-quire extra first costs, which many consumersresist paying even if longer-term savings are

high. In such cases, public policies may beneeded to promote commercial adoption.

A trend is now developing toward newenergy-efficient technologies that are "smart"and attractive for many reasons. These technol-ogies will be more readily adopted than thoseoffering only an energy-saving benefit. Historyshows that the technologies that succeed com-mercially have broad appeal. Technical changefaces some resistance, but generally suchresistance fades if a new technology has manyclear advantages over the old.

In industry, the successful new processeshave generally been those that simultaneouslyreduced various costs—for labor, materials, andenergy, for instance.5 This tendency has beenso powerful that major improvements inenergy efficiency have often been made evenin periods of constant or declining energyprices. Such was the case in several importantbasic industries in the United States betweenthe end of World War II and the first energycrisis. (See Table 2.)

For consumer products too, those most likelyto succeed are ones that offer consumersseveral appealing attributes. Developments inlighting illustrate this point. Lighting efficacy(measured in lumens per Watt) has improvedalmost a hundredfold since Thomas Edison'sinitial invention. (See Figure 5.) Lighting tech-nology continually improved as electricityprices were plunging. (See Figure 2.) Changeswere motivated primarily by consideration ofsuch features as durability and light quality,but they incidentally led to improved energyefficiency as well. Rising electricity prices arenow a strong incentive to make more improve-ments, and the opportunities for further im-provements are substantial.

The development of compact fluorescentbulbs illustrates the trend toward energy-efficient technologies that are "smart." Fluores-cent lights used mainly for commercial and in-dustrial applications have efficacies severaltimes those of the incandescent bulbs used in

Figure 3. Cross-Section of a "Stress-Skin" Panel (Left) Used in a Northern Energy Home (Right)

FinishedInterior

Stress-Skin Panel Northern Energy Home

Highly insulating "stress-skin panels" are featured in some super-insulated houses ("Stress-Skin Panels,"Progressive Builder [September 1986]: 23-26), such as the Northern Energy Home (NEH) offered in thenortheastern United States. At the construction site, 4 feet x 8 feet (1.2 meters x 2.4 meters) stress-skinpanels are mounted on a post-and-beam frame. The factory-assembled panels, containing thick (8-inch[20-centimeters]) rigid polystyrene insulation, are made to fit together easily for rapid construction. Doorsand windows are foam-sealed into the panels at the factory. Because of the tight construction, natural airinfiltration is low, so that forced ventilation is used with a heat exchanger that extracts heat from the staleexhaust air and warms the fresh incoming air.

The extra costs for insulation and forced ventilation are more than offset by various savings. Prefabrica-tion of the panels leads to time and labor savings in construction. In addition, temperatures can be keptuniform throughout the house with a couple of small space heaters, which cost much less than a centralfurnace with ductwork. For a ranch-style NEH with a floor area of 1,300 square feet (120 square meters) inthe New York City area, the annual fuel cost for natural gas would be less than $50 per year—compared tomore than $400 per year for heating the average house in the Middle Atlantic region of the United States.

most homes. Compact fluorescent bulbs, high-efficacy bulbs that can be screwed into socketsdesigned for incandescent bulbs, were intro-duced in the 1970s. The first of these bulbsemitted a harsh white fluorescent light thatmany consumers dislike in the home. Theywere also bulky and difficult to fit into manylamps, and they did not have the "instant on"feature that consumers have come to expect.

However, these problems have been largelyovercome during the last several years. (SeeFigure 6.) The most recently introduced bulb(available in Europe) has a soft yellow light likethat of an incandescent, is nearly as small asthe incandescent it would replace, illuminatesnearly instantly, and will last six times as longas an incandescent.

Figure 4. The Toyota AXV, a Prototype Four- to Five-Passenger, Super Fuel-Efficient Car

Toyota AXV

Introduced in late 1985, the AXV has a fuel economy of 2.4 lhk (98 mpg) on the combined ur-ban/highway test administered by the U.S. Environmental Protection Agency. For comparison, theaverage U.S. automobile gets about 12.4 lhk (19 mpg), and the average car in the rest of the world about9.8 lhk (24 mpg). High fuel economy is achieved with the systematic application of presently availabletechnologies: low weight (650 kg [1,430 lb]) from extensive use of plastics and aluminum, low aerodynamicdrag, a direct-injection diesel engine (the kind used in trucks), and a continuously variable transmission.

Source: Toyota press release, October 23, 1985.

Relevance to Developing CountriesRich countries can save far more energy thanpoor countries by adopting more efficientenergy-using technologies. Industrialized coun-tries consume 70 percent of the world'senergy. Yet, it does not follow that little energycan be saved in developing countries. The elite

in these countries—industrialists, commercialtraders, landlords, government and military of-ficials, bureaucrats, professionals, and someskilled craftsmen—account for 10 to 15 percentof the population, but one third to one half ofincome and most commercial energy use.These elites have acquired the energy-using

Table 2. Pre-Energy Crisis Trends in Energy Intensities for Selected Basic Materials in theUnited States

Average Rate of Decline in Energy Useper Tonne Produced

(percent per year)

Material

Raw SteelPortland CementChlorinec

Aluminumd

Period

1947-19711947-19711947-19711954-1971

Final Energy3 Primary Energy1"

1.411.170.402.83

1.191.090.421.89

Source: R.H. Williams, E.D. Larson, and M.H. Ross, "Materials, Affluence, and IndustrialEnergy Use," The Annual Review of Energy 12(1987): 99-144

a. With electricity counted as 3.6 megajoules per kilowatt hour of electricity consumed.b. Here electricity is expressed as the fuel required to produce it in a thermal power plant.c. For 1 tonne of chlorine plus 1.13 tonnes of caustic soda in 50 percent solution.d. Electricity use per kilogram of primary aluminum declined 0.4 percent per year during this

period.

habits of consumers in rich countries, andoften they waste even more energy.

For example, two-door refrigerator-freezersare now becoming popular among Brazil'selite. The new Brazilian units are smaller (340to 420 liters) than typical units in the UnitedStates (500 liters). Yet the Brazilian two-doorsconsume between 1,310 and 1,660 kWh peryear while new U.S. units consume only 1,150kilowatt hours (kWh) per year (as of 1983).Moreover, the most efficient U.S. model, intro-duced in March 1985, is a 490-liter frost-freeunit requiring only 750 kWh per year. Ironical-ly, this unit achieves its efficiency in large partby using a compressor imported from Brazil:the manufacturer exports a high-efficiency lineand markets a less efficient product at home.

In developing countries even poor house-holds, which depend largely on non-commer-cial biomass fuels and have few if any modernamenities, tend to use energy inefficiently. The

poor who use wood for cooking consume threeto ten times as much energy per capita as con-sumers in developing or industrialized coun-tries who have access to modern energy car-riers. (See Figure 7.) In fact, they use about asmuch fuel per capita for cooking as WesternEuropeans use for automobiles.

Fortunately, recent successes in applying heattransfer and combustion principles have led tonew highly efficient wood-burning stoves.6

These new designs, along with standardizedtesting and production techniques, have madeit possible to introduce various low-cost woodstoves compatible with a wide range of culturalsettings. (See Figure 8.)

Further efficiency gains would result fromshifting to modern gaseous or liquid fuels:biogas, producer gas, natural gas, liquidpetroleum gas (LPG), and ethanol. Simplestoves based on such fuels can be 50-percentefficient, while even the best available wood

Figure

1000

ens

Per

Wat

t)

8:y

(L

um

S

1 0 .

1_

5. Changes in the

Monochromatic

Efficacies of Various Light Sources Over Time

Light 550nm (680 L/W)

Maximum for White Light (220 L/W)

6500 K Blackbody

Melting Tungsten

Cooper-Hewitt

Moore \^

mmmmmam

/

(100 L/W) /

(53 L/W) ' ^

^— Gas-Filled Tungsten

A—Ductile Tungsten

y Osmium^- Nernst

""""* ~" Tantalum

mm

T Carbon Filament

1 11900 1930

Year

This plot shows the evolution of lighting performancefrom one technology to another. The efficacies for several

Source: J.M. Anderson and J.J32-40.

5. Saby, "The Electric Lamp: 100 Years

y Low-Pressure Sodium

/ High-Pressure Sodium

^ ^ f- Metal-Halide

^ ^ ^ ^ ^ Fluorescent

. •

High-Pressure Mercury

y Tungsten-Halogen

y Modern 100 W Tungsten

Modern 60 W Tungsten

1 I1960 1990

within each technology as well as the progressstandard sources are indicated for comparison.

of Applied Physics," Physics Today (October 1979):

10

Figure 6. The Rapidly Changing Technology of Compact Fluorescent Lightbulbs

In recent years manufacturers have been steadily improving the compact fluorescent light bulbs that canbe screwed into ordinary incandescent sockets. These bulbs have efficacies in the range of 40 to 60 lumensper Watt, compared to 11 to 16 lumens per Watt for ordinary incandescent bulbs. Although the first-generation "circline" bulb (on the left) is efficient, it is bulky (it weighs 1 pound [454 grams] and is 8 inches[20 centimeters] across) and thus does not fit easily into many lamp holders. The harsh white light of thecircline is also considered inappropriate for many household applications. Finally, there is a lag after theswitch is flicked before the light comes on—an annoying feature for those accustomed to the instant-onfeature of incandescents. Each of the three bulbs to the right of the circline incorporates innovations thatdeal with one or more of these problems. The bulb shown next to the incandescent on the right, manufac-tured by Osram and now being sold in Europe, has none of the drawbacks of the earlier bulbs. It providesas much light as a 60-Watt incandescent drawing only 11 Watts and lasts six times as long. It weighs just 4ounces (114 grams) and fits into most incandescent lamp holders. The phosphor coatings on the inside ofthe bulb give rise to a light quality similar to that for incandescents. For the impatient, the light is instant-on.

stoves have efficiencies of only 30 to 40percent.

Shifting biomass sources to produce modernfluid fuels can so significantly reduce biomassfeedstock requirements for cooking that extrabiomass would be available for other uses.They could include water pumping and ruralelectrification based on producer gas enginesets, which produce gas from biomass feed-stock and convert it to mechanical power orelectricity in modified gasoline or diesel engine

sets. (See Box 1.) The biomass available for suchpurposes would increase, if the gas stoveswere more efficient. One especially efficient gasstove has just been developed. (See Box 2.)

Adding It Up

To indicate how significant technological op-portunities like those described above are, wehave constructed a global energy scenario forthe year 2020 that focusses on energy enduses. We examined in detail how energy is

11

Figure 7. Per Capita Energy Use Rates for Cooking With Wood and High-Quality Energy Carriers

Wood

U

~ (Ju aZ N

Io

H

SC u s

tuD1XJ

CO

OS

O

- 6 0 0

- 5 0 0

.400

- 3 0 0

_200

_100

For both wood stoves and stoves using high-quality energy carriers, the per capita energy use rate is ex-pressed in Watts (an abbreviation for Watt-years per year). For wood it is also given in tonnes of dry woodper year, assuming 1 tonne of wood = 18 gigajoules, so that 1 tonne per year = 570 Watts.

Source: J. Goldemberg, et ah, "Basic Needs and Much More with 1 kW per Capita," Ambio vol. 14, no. 4-5 (1985): 190-200.

now used, identified ongoing structuraleconomic changes that will shape energy de-mand, and explored technically and economi-cally feasible ways to use energy moreefficiently.

Our analysis indicates that the global popula-tion could roughly double, that living stan-dards could be improved far beyond satisfyingbasic needs in developing countries, and thateconomic growth in industrialized countriescould continue, without increasing the level of

global energy use in 2020 much above the pres-ent level. The level of global energy use iden-tified in our scenario is far below the doublingor tripling of global energy use projected inconventional energy analyses. (See Figure 9.)

Our scenario is not a prediction but, rather, astatement of what is technically and econom-ically feasible. It could come about, facilitatedby appropriate public policies.

Yet, why should attention be given to alter-

12

Figure 8. A High-Efficiency Wood Stove Designed by Applying Modern Scientific and EngineeringPrinciples to the Cooking Task

Air Inlet7.0 x 1.5

Sheet Metal.8mm

FiberglassInsulation

Air Inlet7.0 x 1.5

Fuel Inlet3.0 x 3.0

Cross SectionViewed From the Side

(Dimensions in Centimeters)

Cross Section of the BottomViewed From Above

(Dimensions in Centimeters)

Recent applications of the principles of heat transfer and combustion to fuelwood cooking stove designhave resulted in simple, highly efficient stoves. Typically such improved stoves can pay for themselves infuel savings within a few months.

The stove shown here, made of sheet metal and insulation, is under development at the Indian Instituteof Science at Bangalore. The efficiency of this stove can be more than 40 percent compared to less than 20percent for protected open fires. Hazardous smoke emissions would also be greatly reduced with this typeof stove.

Good performance is due to the use of insulation to reduce heat loss through the walls; a unique combus-tion chamber design that promotes mixing of air with and recirculation of unburned volatiles until com-bustion is complete; and a narrowing combustion chamber and a narrow pot-to-stove channel width to in-crease the gas velocities over the bottom and sides of the pot, thereby improving the heat transfer to thepot.

Source: H.S. Mukunda and U. Shrinivasa, "Single Pan Wood Stoves of High Efficiency, Part I," July 1985; H.S. Mukunda, et.al., "Single Pan Wood Stoves of High Efficiency, Part II," December 1985 (Bangalore, India: Indian Institute of Science),Centre for the Application of Science and Technology to Rural Areas.

13

Box 1. Producer Gas Technology

If biomass is to be an important energysource for development, it must be usedmuch more efficiently than in the past,and it must be converted increasingly tomodern energy carriers such as gas andelectricity. "Producer gas" is one ofseveral biomass-derived modern energycarriers. Through the partial oxidization ofthe biomass feedstock in air, a gas can beproduced that consists largely of carbonmonoxide, hydrogen, and nitrogen. Pro-ducer gas can be made in low-cost con-verters at an overall energy efficiency of60 to 70 percent or more. The energylosses in conversion are usually more thanoffset by the energy savings of the moreenergy-efficient end-use devices that canbe used with gas.

At present, the most common applica-tion of producer gas is to displace oil ornatural gas in industrial boilers. It canalso be used for cooking or to produceelectricity in small engine-generator setsfor agricultural pumping, village electrifi-cation, or rural industry. Producer gaswas widely used in World War II as agasoline substitute in automobiles fittedwith wood- or charcoal-fired producer gasengines. Today, when electricity is gener-

ated using producer gas as fuel, the con-verter employed is often a modified auto-mobile engine.

Interest in producer gas was revived inthe 1970s, and many gasifiers are nowoperating in developing countries. Char-coal is currently favored over wood orother raw biomass feedstocks for engineapplications because the produced gas istar free and less likely to cause enginetrouble than wood-derived gas, and itdoes not require costly gas cleanup tech-niques. However, the large difference be-tween charcoal and raw biomass priceslimits the economic attractiveness of char-coal gasification. In addition, gasifyingcharcoal rather than raw biomass is muchless resource-efficient. For these reasons,gasifier research and development effortsare beginning to be focussed on produc-ing tar-free gas from wood and other rawbiomass sources.*

* E.D. Larson, "Producer Gas, EconomicDevelopment, and the Role of Research,"Rep. No. PU/CEES 187, Center for Energyand Environmental Studies, PrincetonUniversity, Princeton, N.J., April 1985.

Box 2. Infrared Impingement Burner

The Thermoelectron Corporation isdeveloping a high-efficiency infrared im-pingement burner for cooking stoves forthe Gas Research Institute. The unit has ameasured efficiency of 65 to 70 percent,much higher than the 40- to 50-percent ef-ficiency of conventional gas stoves. In ad-dition, the burner's carbon monoxide andnitrogen oxides emissions are far less thanthose from conventional units.* This tech-nology exemplifies the trend toward high-

efficiency products that have attractivefeatures beyond good energy perform-ance, making the products more desirableto consumers.

* K.C. Shukla and J.R. Hurley, "Develop-ment of an Efficient, Low NOx DomesticGas Range Cook Top," Rep. No. GRI-81/0201, Gas Research Institute, Chicago,1983.

14

Figure 9. Alternative Projections of Global Primary Energy Use Disaggregated into the SharesAccounted for by the Industrialized and the Developing Countries

30

Ter

25 _

20 _

is _

10 _

1 I Industrialized Countries

I_J Developing Countries

( ) Per Capita Energy Use (kW)

[ ] Population (Billion)

28.4

24.6

19.1

11.2

10.3

(6.3)[l.H]

(1.0)[3.32]

(3.2)[1.24]

(8.1)[1.53]

(1.3)[5.71]

(8.0)[1.52]

(1.1)[6.19]

(1.2)[5.88]

(9.7)[1.53]

(11.3)[1.52]

(1.60)[6.19]

(1.9)[5.88]

1980Actual

2020This

Study

2020WECLow

2020IIASALow

2020WECHigh

2020IIASAHigh

Primary energy use in TW (an abbreviation for terawatt-years per year) is shown both for alternativeprojections to the year 2020 and for 1980. The projections shown are the one constructed in the presentstudy and both the high and low scenarios advanced in the IIASA (W. Haefele, et al., Energy in a FiniteWorld—A Global Systems Analysis [Cambridge, Massachusetts: Ballinger, 1981]) and WEC (WorldEnergy Conference, Energy 2000-2020: World Prospects and Regional Stresses, J.R. Frisch, ed. [London:Graham & Trotman, 1983]) studies.

"Primary energy" is the energy content of the various sources in naturally occurring forms, before con-version into various useful energy carriers and delivery to consumers as "final energy."

15

native energy futures and new energy policies? pie in either developing or industrialized coun-Haven't market forces already solved the tries? How different would the world be givenenergy problem, once more bringing low oil the energy future we have described ratherprices and a worldwide oil glut? Is such a low than the energy futures usually projected?energy future relevant to the needs of the peo-

16

II. A Closer Look at the Energy Problem

A Temporary Reprieve for OilConsumers

T he collapse of world oil prices in late1985 and early 1986 came as goodnews to the many countries that im-

port oil. (See Figure 1.) To many, it seemed tosignal the end of the energy crisis and theOrganization of Petroleum Exporting Coun-tries' (OPEC) loss of control over the world oilmarket.

Unfortunately, the decline in the price of oilindicates neither of these outcomes. Oil con-sumers have been granted only a temporaryreprieve.

To understand why, consider the relationshipbetween world oil price and the demand forOPEC oil. (See Figure 10.) Between 1975 and1985, the oil price was rising when demand forOPEC oil was above 80 percent of OPEC pro-duction capacity, and they were falling atlower demand levels. Just before the second oilprice shock in 1979, as before the first, a surgein oil demand in the industrialized marketeconomies created tight market conditions,enabling OPEC to exercise its monopoly powerand raise the price sharply. (See Figure 1.) Oil-saving efforts by consuming countries, togetherwith increases in oil production outside OPEC(in the North Sea, Mexico, Alaska, and else-where), then transformed the oil market from asellers' to a buyers' market.

Will oil buyers around the world increasetheir demand in response to today's lowerprices? There is probably no going back to oil-guzzling days. For example, consumers willnot pull insulation out of their attics, and manyof the more energy-efficient oil-using technol-ogies were adopted for reasons that go beyondfuel savings. It won't take much, though, toput OPEC back in the driver's seat.

The U.S. Department of Energy (DOE) pro-jected in 1985 that net oil imports by OECDcountries would increase some 4 million barrelsper day between 1985 and 1995, owing aboutequally to a modest increase in demand and toan expected decline in U.S. oil production.7

The projected decline in U.S. productionreflects not the decline in drilling broughtabout by the recent sharp drop in the world oilprice but, rather, the limits of remainingresources; in fact, the DOE scenario is for aworld oil price that returns to $30 per barrel (in1985 dollars) by 1995. In addition, DOE pro-jects an increase in oil demand by developingcountries with market economies of some 2.5mbd in this period, an increase consistent withrecent trends. (See Figure 1.) The DOE also pro-jects a drop of about 1 mbd in exports fromcentrally planned economies, reflecting the ex-pected decline in Soviet oil production. If, asprojected, other oil producers expand produc-tion only modestly, the increase in demand forOPEC oil would amount to more than 7 mbd.With such an increase, OPEC would again be

17

Figure 10. OPEC Pricing Behavior, 1975 to 1985

3.0

'5

I8

Ic

u

100.

50 _

- 5 0 .

1980

1983 1982

50—r~60

1 170 80

Percentage of Capacity Utilization

90 100

This figure shows the percentage change in the real world oil price from the previous year versus the de-mand for Organization of Petroleum Exporting Countries (OPEC) oil, expressed as a percentage of OPECproductive capacity used. (The percentage of OPEC capacity used is equal to crude oil production dividedby maximum sustainable production.)

At the present level of OPEC production capacity, oil prices would rise again if the demand for OPECoil increased 5 million barrels per day, equal to about 10 percent of the oil consumption rate for the world'smarket economies.

Source: U.S. Energy Information Administration, "International Energy Outlook 1985, With Projections to 1995," U.S. Depart-ment of Energy, (Washington, D.C., March 1986).

operating at more than 80 percent of capacity.If history is a reliable guide, the world oil pricewould then be rising. (See Figure 10.)

The DOE analysis suggests that severaldevelopments, each of which alone would onlymodestly affect the world oil market, couldtogether tighten that market relatively quickly.Although the real world might not evolve ex-

actly the way envisaged by DOE, that scenariomust be close to what would happen underbusiness-as-usual conditions. The ups anddowns in oil prices have not changed theessential problem: oil is a relatively scarce com-modity of considerable value, two thirds of theworld's recoverable oil resources are still con-trolled by a handful of nations in the MiddleEast and by the Soviet Union, and no ready oil

18

substitutes exist for fueling automobiles, trucks,airplanes, tractors, and other vehicles. Whateconomists call "the preconditions of carteliza-tion" continue to be very much a reality in theoil market. These conditions differentiate oilfrom such widely traded commodities as cop-per, tin, uranium, and cotton.

Although the world oil market will eventual-ly tighten, the day when the oil prices mustrise again can be forestalled for years, perhapsdecades, by improvements in energy efficiency.For example, if all cars were as fuel-efficient asthe Toyota AXV, oil use in the world's marketeconomies would be lower by an amount equalto DOE's projected increase in demand forOPEC oil by 1995. (See Figure 4.)

It should come as no surprise that oil-importing countries can keep world oil priceslow by consuming less. The 6-mbd reductionin oil demand achieved by the OECD countriesbetween 1973 and 1985 is largely responsiblefor the present oil glut; if demand were nowjust this much higher, oil prices would be ris-ing. (See Figure 10.)

Energy and DevelopmentWhy worry about oil prices' rising again in adecade or so? Arguably, we shall then bericher and better able to afford higher prices.Such thoughts, though, are cold comfort to thethree fourths of humanity who live in develop-ing countries. It is their grave misfortune tohave to industrialize after the end of the era ofcheap energy, upon which the industrial basesof the already industrialized countries werebuilt.

Per capita income in developing countriesaverages one tenth that in the rich industrial-ized countries, per capita energy use less thanone sixth. Average life expectancy is about 50years, compared to more than 70 in the indus-trialized countries. One of every five personsin developing countries suffers from hunger ormalnutrition. One of every two has littlechance of becoming literate.

If the majority of the world's population thatlives in wretched poverty is to achieve a decentstandard of living, it will need affordableenergy for increasing agricultural productivityand food distribution, delivering basic educa-tional and medical services, establishing ade-quate water-supply and sanitation facilities,and building and powering new job-creatingindustries. In addressing the world's energyfuture, therefore, one question stands above allothers: Will people in developing countries getthe energy they need to sustain a higher stan-dard of living, equal, say, to that enjoyed byWestern Europeans today?

Looked at in conventional terms, the chal-lenge seems formidable. If the average annualgrowth rate in per capita commercial energyuse for developing countries in the 1970s (3.6percent) were to persist until the year 2020, percapital commercial energy use would then be2.3 kilowatts (kW), still far less than theaverage of 4.0 kW for Western Europe in 1975.If, as is expected, the population in developingcountries nearly doubles, the resulting ag-gregate demand for commercial energy wouldreach 15 billion Watts, or terawatt-years peryear, (TW) by 2020, up from 2 TW in 1980. Theincrement in energy use by developing countriesin this period would be 1.3 times the world'stotal energy use in 1980, 3 times its oil produc-tion, 5 times its coal production, 7.5 times itsnatural gas production, nearly 9 times its bio-energy production, and nearly 60 times itsnuclear energy production. It would be ex-ceedingly difficult to meet such energy re-quirements at reasonable costs and withoutmajor environmental or security problems.

Outside the Middle East and North Africa,most developing countries have little or nopetroleum. Except for China, most have littlecoal. Their hydropower and biomass resources,though important, appear paltry when com-pared with the needs implied by a continuationof historical trends.

The cost of expanding energy supplies hasbeen growing steeper, and the outlook is even

19

more grim. The World Bank in 1983 estimatedthat to achieve a scaled-down 2.5-percent an-nual growth in per capita commercial energyuse from 1980 to 1995, developing countrieswould require investments in new energy sup-plies of about $130 billion per year (in 1982dollars) between 1982 and 1992, or about 4 per-cent of GDP.8 Half this investment would haveto come out of developing countries' foreignexchange earnings, requiring an average annualincrease of 15 percent in real foreign exchangeallocations to energy supply expansion. Despitethese investments, oil imports by oil-importingdeveloping countries would increase by nearlyone third, to almost 8 mbd by 1995. Such acommitment of economic resources is complete-ly unrealistic, particularly given the financialvise in which many developing countries findthemselves today, squeezed between high debtcosts and low export-commodity prices.

The staggering cost of providing such in-creases in energy supply has led some analyststo conclude—if rarely to state publicly—that liv-ing standards in developing countries simplycannot be substantially improved in the fore-seeable future.

This judgment rests on the assumption thatto raise their living standards to the WesternEuropean level, developing countries will haveto increase their energy use to the WesternEuropean level. However, large improvementsin living standards can be made with little in-crease in energy use if planners take advantageof cost-effective opportunities to use energymore efficiently. For a wide range of energy-using technologies, saving energy will requireless capital investment than supplying anequivalent amount by conventional means,freeing up scarce capital for other purposes.

The Other Energy Crisis

The "energy crisis" of 1973 to 1981 brings tomind the message Mark Twain wired fromVienna on his first visit to that city: "Viennaisn't what it used to be and never was."

The energy crisis has been seen almost exclu-sively as a problem of oil and the othermodern energy forms that shape life in indus-trialized countries and in modern sectors ofdeveloping countries. But more than half thepeople in the world live in villages and smalltowns in Asia, Africa, and Latin America,where modern energy forms are little used. Inparticular, many villagers cannot afford oilproducts—diesel fuel, gasoline, liquid petro-leum gas (LPG), or kerosene—or the thingsthose products run. For these people, depen-dent largely on fuelwood and other biomassforms, the energy crisis did not fade with theonset of the recent world oil glut; it worsened.

People are burning wood more rapidly thanit is grown, and increasing amounts of humanenergy and time are spent gathering and haul-ing fuelwood. The United Nations Food andAgriculture Organization estimates that some100 million human beings now suffer an acutescarcity of fuelwood and about 1 billion have afuelwood deficit. At present consumption rates,the annual fuelwood deficit will more thandouble between 1980 and 2000, from 407 to 925million cubic meters (m3). Current tree-plantingrates are only about 1 percent of what is re-quired to reverse this trend, and many of theplanted trees don't survive.

The increasing pressure on wood resourcesby a rapidly growing population of poor peo-ple who must rely on fuelwood to heat andlight their homes and cook their food has hadserious environmental repercussions. In severalregions, including the Himalayas and theAndes, the cutting of trees and brush on steepslopes lays bare the land, so the soil erodes atcatastrophic rates and rainwater runs rapidlyoff the land, increasing flooding downstream.In dry regions, fuelwood harvesting has be-come a major cause of desertification, alongwith overgrazing by livestock and drylandfarming. The environmental effects of fuelwoodharvesting have been especially severe in sub-Saharan Africa. The hardest hit countries in-clude Ethiopia, Sudan, Mali, Chad, Niger,Mauritania, Burkina Faso> Kenya, Somalia, and

20

Tanzania. Each year human overuse leads tothe worldwide loss of some 20 million hectares(50 million acres) to desertification—an arearoughly the size of Senegal.

An important development goal should be toreverse the process of deforestation and deser-tification, preserving land and soil resources foreconomic activities, wilderness, and wildlifehabitats.

One important use of these resources wouldbe the production of biomass for energy pur-poses. Biomass is the most widely availableenergy resource in developing countries, and

in fact it accounts for more than 40 percent oftotal energy use there. (See Table 3.)

If biomass is to be an environmentally accep-table energy resource, however, it must be pro-duced on a renewable basis. And if biomass isto make a significant energy contribution todevelopment, it must be made more abundantand used much more efficiently.

More biomass can be made available if it isproduced in highly productive woodlots,energy farms, or energy plantations instead ofmerely being harvested from natural forests,which have relatively low productivity.

Table 3. Summary of Selected National

Country

BangladeshNigerGambiaMoroccoIndiaEthiopiaNepalSomaliaBoliviaSudanThailandTanzaniaChinaBrazilMexicoLibya

DevelopingCountries(average)

Energy Consumption Surveys

Commercial Non-CommercialEnergy Energy

(kilowatts per capita)

0.0380.0350.0980.2670.1650.0190.0090.0920.3400.1590.3050.0600.7780.7371.291.76

0.550

Source: Adapted from D.O. Hall, G.W.ing Countries (Oxford: Pergamon

0.0950.2540.2220.0730.1900.3710.4290.4760.2630.6350.5240.8100.3170.3710.1270.095

0.416

3arnard, and P.A. MossPress, 1982), Table 2-3.

TotalEnergy

0.1330.2890.3200.3400.3550.3900.4380.5680.6030.7940.8290.8701.101.111.431.86

0.966

Biomass

Percentage ofEnergy from

Non-CommercialSources

7188692154959884448063932934

95

43

for Energy in Develop-

21

Making efficient fuelwood cooking stoveswidely available is important in the near termto making scarce fuelwood supplies go farther.(See Figure 8.) For the longer term, it is impor-tant to convert biomass into high-quality gas-eous fuels (biogas or producer gas), liquid fuels(methanol or ethanol), or electricity so thatmuch more useful energy can be extractedfrom a given biomass feedstock than is possiblewith traditional combustion technologies. Ashift to such high-quality biomass energyforms, which could substitute for imported oil,would make the biomass more valuable andthus more profitable as a crop.

The Hidden Costs of ConventionalEnergy

Since the late 1960s, the environmental andsecurity risks associated with the productionand use of conventional commercial energyforms have been important considerations inenergy planning.

In some instances, it is possible to limit theserisks with a variety of control technologies. Forexample, various devices have been employedto reduce harmful air pollutant emissions fromfossil-fuel burning power plants andautomobiles.

But for some serious problems, simple tech-nical fixes such as emissions-control devices donot exist, and risk reduction can be achievedonly by limiting dependence on the trouble-some technology. Three such problems standout as particularly worrisome: global insecurityarising from industrialized countries' overde-pendence on Middle East oil, the potential forglobal climate change owing to build-up in theatmosphere of carbon dioxide (CO2) from burn-ing fossil fuels, and the threat of nuclearweapons proliferation that accompanies thespread of nuclear power around the world.

Global Insecurity and Middle East Oil. If theworld once more becomes hungry for OPECoil, the resulting higher world oil price would

be only one cost. In addition, efforts by the in-dustrialized market countries to assure con-tinued access to Middle East oil supplies intimes of crisis could make the world a muchmore dangerous place. The importance ofassured access to these oil supplies was ar-ticulated by U.S. Defense Secretary CasparWeinberger:

The umbilical cord of the industrialized freeworld runs through the Strait of Hormuz andinto the Persian Gulf and the nations whichsurround it. ...

That Middle East conflicts can engage thesuperpowers and even threaten nuclear war isindicated by the experience of October 1973,when the Soviet Union threatened to intervenein the Arab-Israeli War and the United States,in response, put its nuclear forces on alert.U.S. anxiety about the possible course ofevents around the Persian Gulf led it toorganize a Rapid Deployment Force that couldoccupy strategic areas in the region or confrontany expeditionary force the Soviet Union mightintroduce in case of revolution or war there.

The global security risk inherent in this situa-tion can be reduced if the industrialized marketeconomies avoid becoming too dependentagain on Middle East oil. Adopting oil-efficienttechnologies is the most effective way ofachieving this goal. Very efficient cars shouldbe regarded not just as attractive consumerproducts but also as indirect deterrents of war,even nuclear war.

Global Climatic Change and Fossil Fuel Use.Within decades, we could see a general warm-ing of the earth's surface and other majorchanges in the global climate because of CO2

buildup in the atmosphere and the resulting"greenhouse effect."9 The amount of CO2 inthe earth's atmosphere is now more than 15percent higher than in pre-industrial times.With continued emphasis on expanding fossilfuel use in energy planning, the CO2 contentof the atmosphere would double in 50 to 100years.

22

Higher CO2 concentrations in the atmospherewill almost certainly cause a global warming.Carbon dioxide in the atmosphere, thoughtransparent to incoming solar radiation, pre-vents the escape of heat (infrared) radiationfrom the earth. The precise extent and effectsof this warming are less certain. Global climatemodels, however, indicate that a doubling ofatmospheric CO2 would raise the average tem-perature at the earth's surface by 3° + 1.5°C—enough to invite potentially devastating conse-quences. Moreover, the heating effects of theCO2 build-up are being amplified by a roughlycomparable amount of heating owing to the at-mospheric build-up of other trace gases(methane, chlorinated fluorocarbons, etc.),which increases the urgency of the problem.

With continued expansion of fossil fueluse, the CO2 content of the atmospherewould double in 50 to 100 years.

Any warming would not be uniform butwould instead be much greater at the poles.This differential warming would slow downthe "atmospheric heat engine" that is drivenby the equatorial-polar temperature differenceand would surely change weather patterns (in-cluding rainfall distribution).

No technical solution to the CO2 problem isknown. The only way to mitigate the problemis to reduce combustion of fossil fuels, whichaccount for about four fifths of global energyuse. Because remaining coal resources are fargreater than remaining oil and gas resourcescombined, finding ways to reduce overdepen-dence on coal will become increasinglyimportant.

Several decades from now it may be possibleto replace fossil fuels with energy supplies that

do not contribute to the atmospheric burden ofCO2. But for now the most promising way tolimit the CO2 build-up is to use energy moreefficiently.

Nuclear Weapons Proliferation and NuclearPower. Nuclear weapons and nuclear powerare indissolubly linked. Nuclear power reactorsproduce substantial quantities of plutonium, amaterial usable in nuclear weapons. This plu-tonium becomes much more accessible towould-be weaponmakers when it is recoveredfrom spent reactor fuel and recycled in freshfuel. If a nation without nuclear weapons ac-quires plutonium recycling technology, itthereby acquires nearly all the technology andmaterials needed to make nuclear weaponsquickly. There is no way to eliminate the linkbetween nuclear weapons and nuclear power,but the proliferation risk would be much less ifplutonium-recycling technologies were notused.

Interest in plutonium recycling arises becausewithout plutonium recycling only about 1 per-cent of the nuclear energy stored in naturaluranium can be used in reactors of presentdesign. With plutonium recycling, the usefulenergy recoverable from uranium can be in-creased somewhat with existing reactors. Butexperience with plutonium recycling in thesereactors would facilitate the introduction ofplutonium breeder reactors, which could makeuse of up to 50 percent of the energy stored inuranium. Thus, as uranium supplies becomescarcer, interest in plutonium recycling andbreeder reactors will increase.

Even though ambitious plutonium recyclingprograms are under way in France and a fewother countries, there is no economic justifica-tion for plutonium recycling because uraniumprices are low and fuel reprocessing costs arehigh.10 If growth in nuclear power were suffi-ciently modest, plutonium recycling would notbe economical for many decades at least.

Avoiding plutonium recycling greatlyweakens but does not destroy the link between

23

nuclear weapons and nuclear power. Theplutonium in spent fuel produced in nuclearpower plants can be a tempting source ofnuclear weapons material for countries deter-mined to acquire nuclear weapons. This riskcan be reduced only by limiting the overalllevel of nuclear power development—for exam-

ple, by regarding nuclear energy as the energysource of last resort.

The most effective way to avoid plutoniumrecycling technologies and overdependence onnuclear energy over the next several decades isto pursue more efficient energy use. Doing so

Figure 11. Alternative Projections of Global Primary Energy Use Disaggregated into the SharesAccounted for by Various Energy Sources

25.

20 _

15 _

10 _

5 _

• OTHERIHl HYDROE l NUCLEARE3 BIOMASSIB COAL• NATURAL GASM OIL

I.I JUJ.IJ i.i.i j i j M i j i .mu'. M i IIJJIJ.

1980Actual

2020This

Study

2020WEC

2020IIASA

Primary energy use (in billion Watts, or terawatt-years per year [TW]) is shown both for alternative pro-jections to the year 2020 and for 1980. The projections shown are the ones constructed in the present studyand both the high and low scenarios advanced in the IIASA (W. Haefele, et ah, Energy in a FiniteWorld—A Global Systems Analysis [Cambridge, Massachusetts: Ballinger, 1981]) and WEC (WorldEnergy Conference, Energy 2000-2020: World Prospects and Regional Stresses, J.R. Frisch, ed. [London:Graham & Trotman, 1983]) studies. For the IIASA and WEC cases the projections shown are averages ofthe high and low scenarios described in those analyses.

The convention adopted here for measuring the nonfossil-fuel sources of electricity is that hydroelec-tricity is counted as the energy available in the produced electricity, and nuclear energy is counted as theheat released in nuclear fission, the process from which the produced electricity is derived.

24

would also buy time for the development ofless dangerous technologies for electricitygeneration in the future.

-Reason forLooking Ahead-OptimismThe low-energy demand scenario made possi-ble by the "quiet revolution" in more efficientenergy-using technology is attractive undernarrow economic criteria: energy services areprovided at equal or lower costs than pro-viding the same services with conventional,less efficient end-use technologies and moreenergy supplies. With demand thus lowered,energy prices would also be lower becausethere would be less need for Middle East oiland for costly domestic energy sources.

In addition, when energy demand is low,considerable flexibility is gained in energysupply planning, making it possible to avoidthe environmental and security problems posedby overdependence on Middle East oil, fossilfuels generally, and nuclear power. We haveprepared a supply mix matched to our globallow energy demand scenario for the year 2020to illustrate this flexibility. (See Figure 11.) Likeour low-energy demand scenario, this supplyscenario is not a forecast. It is, however, aplausible energy supply mix not obviously con-strained by supply availability or costs. Itshows what is economically, technically, andpolitically possible—in short, what our choicesare.

25

III. Alternative Energy Futures

T he energy price shocks of the 1970sprompted a number of institutions,both national and international, and

scientists to project future energy demand andsupply. In most of these projections, theenergy future is portrayed as an extension ofthe past: energy demand is seen to rise steep-ly, so that virtually all energy supplies must beexpanded at formidable rates to meet projecteddemand. Most have underestimated the poten-tial of measures to increase energy efficiency,probably because the world had never beforeexperienced anything remotely like the changesrelating to energy that have occurred since thefirst oil shock.

What is striking about these efforts is that astime passed, the demand projections camedown. For example, the International EnergyAgency's demand projections dropped marked-ly between 1977 and 1980. (See Figure 11.)Energy demand projections for the UnitedStates also moved steadily downward. (SeeFigure 13.) The first effort exploring the pros-pects for energy efficiency improvement wasthe Ford Foundation's Energy Policy Project(EPP). In its final report, published in 1974,EPP suggested that "zero energy growth" forthe United States was a possibility near theturn of the century, so that by the year 2000demand would be "only" about 100 quads(quadrillion British thermal units [Btu]) peryear.11 This suggestion was roundly criticizedat the time as "fuzzyheaded," "irresponsible,"and "totally impractical" by a number of ex-

perts, including several on the project's boardof advisors. In fact, it was then virtually im-possible to find any energy analyst outside theEnergy Policy Project who thought U.S. energydemand could be as low as 100 quads per yearin the year 2000. Government agencies and thetrade associations for the oil and gas, electricutility, and nuclear power industries were pro-jecting energy demands of 150 to 170 quads by2000. It was not long, however, before theFord Project's middle-of-the-road "technicalfix" scenario of 125 quads was adopted byenergy analysts as conventional wisdom. TheUnited States has already experienced zeroenergy growth for the period 1973 to 1985, andtoday most forecasters project a U.S. energydemand in 2000 of only 85 to 90 quads peryear.

Why did analysts err on the high side? Themost serious problem with these projections isthat most were based on the assumption thatin the future there would be a close couplingbetween a nation's energy demand and itslevel of economic activity, as measured by itsGDP or gross national product (GNP). Thiscorrelation was indeed strong historically, butno more in the market-oriented industrializedcountries. As we have already noted, a remark-able decoupling of energy and economicgrowth has already been seen for the OECDcountries since 1973. (See Figure 14.)

Certainly the most ambitious analysis of theglobal energy problem in this period, published

27

Figure 12.ji

vale

nt

lion

Ton

nes

of O

il E

q

%

8500-

7500-

6500.

5500-

4500.

3500_

2500-

1500

19

The(World

OECD Primary Energy Demand and Supply, History and

World Energy /Outlook, 1977 —*" /

A

// 'JOECD Energy / v ' j*£r

« « OECD Energyj ^ « Supply |1 i I I I | l60 1970 1980 1990

Year

alternative projections were made in 1977, 1978, andEnergy Outlook [Paris: 1982]).

/

1

i

\i i

2000

Projections

-8900

-6910

Coal Study, 1978

5810 High

5500 Reference Case 1980

5090 Low

4350 Reference Case 1980

4000 High3870 Low

1980 by the International Energy Agency

28

Figure 13. Forecasts of United States Primary Energy Requirements for the Year 2000 Versus theYears the Forecasts Were Made

200-

175-

150-

125 _

on 100.co

75-

50-

25-

D O I ,i FPC

INPC

>AEC

»NSF

• SIERRA CLUB

i LOVINS

• DOI

»EE1

>ERDA EE1 <

Houthhakker0 and Kennedy

• EXXONEIA > • BNL DJA

• FORD ERDA •

(Technical Fix)

• FORD(ZeroEnergyGrowth)

ORAUECE

> LOVINS

_NEPP4

(CONOCO!

• LOVINS I Williams

LOVINS •

(Least-Cost) AUDUBON

LOVINS

I I I I i I I

1972 1973 1974 1975 1976 1977 1978

Year of Publication

1979 1980 1981 1982 1983

Source: U.S. Department of Energy, Office of Policy, Planning, and Analysis, "Energy Projections to the Year 2010: A TechnicalReport in Support of the National Energy Policy Plan," (Washington, D.C., October 1983).

in 1981, was carried out between 1973 and 1979at the International Institute for Applied Sys-tems Analysis (IIASA) by 140 scientists from 20countries.12 Another important study was pub-lished in 1983 by the World Energy Conference(WEC), the result of a cooperative effort of onecentral team and ten regional working teamsinvolving 50 participants with diverse experi-ence.13 Less academic than the IIASA study,the WEC study gives a view of the energyfuture shared by most planners at energy com-panies and agencies and associated institutions.Both studies represent what might be called

the conventional wisdom about energy at thetime they were carried out. Their projections ofglobal energy demand and supply for the year2020 differ markedly from our own. (See Figures9 and 11.)

The IIASA/WEC Numbers and OurNumbers: What They Mean

Future Energy Demand. The IIASA and WECanalysts generated high- and low-energy de-mand scenarios for the year 2020. In their low

29

Figure 14. Primary Energy Consumption, Net Oil Imports, and Gross Domestic Product forOrganisation for Economic Co-operation and Development Countries, Relative to theValues of These Quantities in 1973

1.4

1.3.

1.2 _

1,1-qrHIIII

ro 1o.

. 9 .

.8 _

. 7_

. 6 .

.5

.• ..j/x

GDP . • *

•

^ ^«» Primary Energy** •«• ^ Consumption

\

k̂ Oil Imports

\

^ ^ .

I I 1

1973 1975 1977 1979

Year

1981 1983 1985

Source: For Energy Data: BP Statistical Review of World Energy (London: June 1986). For Gross Domestic Product Data:Organisation for Economic Co-operation and Development, National Accounts of OECD Countries, vol. I (Paris: 1985).

demand scenarios, global energy use wouldnearly double between 1980 and 2020; in theirhigh-demand scenarios, it would increase 2.5to nearly 3 times. For comparison, globalenergy use would be only 11.2 TW in 2020 ac-cording to our scenario, up only slightly from10.3 TW in 1980. (See Figure 9.)

The IIASA and WEC scenarios involve noreduction in the great regional disparities inpresent energy use. In 2020, per capita energyuse in industrialized countries would, accord-ing to both, be at least six times that in devel-oping countries, just as it is today. Puttingaside for the moment any consideration of the

30

analytical basis for these projections, suchstrong growth in energy demand by the indus-trialized countries implies significantly higherworld energy prices. The energy appetites ofthe industrialized countries would thus exacer-bate a problem that is already frustratingdevelopment in developing countries andwould certainly lead to intolerable North-Southtensions.

Economic growth can, however, continue inthe North without putting such pressures onworld energy resources. Analysis based ondetailed studies for Sweden and the UnitedStates indicates that per capita energy usecould be reduced roughly by half in industrial-ized countries between 1980 and 2020 whileper capita GDP grows by perhaps 50 to 100percent. This decoupling of energy and eco-nomic growth reflects both an ongoing struc-tural shift away from energy-intensive eco-nomic activities and the exploitation of oppor-tunities for more efficient energy use. Thus,energy use by the industrialized countriescould decline by 3 TW, from 1980 to 2020.

For developing countries, our scenario showsin the year 2020 a per capita primary energyuse rate of 1.3 kW. Because of the expectednear doubling of population, this implies thatthe developing countries' share of worldenergy use would increase from one third (itspresent value) to two thirds in 2020. Our de-mand projection is between the values given inthe IIASA and WEC low and high scenarios,but developing countries could achieve a rela-tively better living standard with our scenariothan this comparison indicates. In fact, any liv-ing standard up to that of Western Europe inthe mid-1970s could be obtained with about thesame per capita energy use as that prevailingtoday in developing countries. This result isachieved by shifting from traditional, ineffi-ciently used non-commercial (biomass) fuels,which currently account for more than 40 per-cent of developing countries' energy use, tosuch modern energy forms as electricity, liquidand gaseous fuels, and processed solid fuelsand by emphasizing efficiency improvements in

energy-using equipment as economic develop-ment proceeds. With our scenario, the costs ofproviding energy services would be less thanwith more conventional supply-oriented devel-opment strategies, thus freeing up economicresources for other development.

Future Energy Supplies. Meeting the globaldemand levels projected in the IIASA andWEC scenarios would require monumental ef-forts to expand energy supplies. Consider theIIASA scenarios, which specify supplies of 19to 28 TW of energy in 2020. (See Figure 11.)

To provide for both net growth and thereplacement of retired facilities, the IIASAscenarios would require, for example, theopening of one new large nuclear power plant(1 gigawatt electric, GW(e)) every four to sixdays from 1980 to 2020. It would also requirebringing on line new fossil-fuel productioncapacity equivalent to that of the Alaska pipe-line—the equivalent of 2 mbd of oil every oneto two months throughout this period. Such anenormous expansion in energy supplies wouldbe exceedingly costly in terms of global securityand the environment, as well as in terms ofdirect economic costs. The impacts of thisdevelopment are made more vivid by exam-ining each of the major supplies in turn.

Consider oil. To meet IIASA's demandlevels, the Middle East and North Africa wouldhave to produce at or near maximum capa-city—34 mbd, compared to the average of lessthan 15 mbd produced from 1983 to 1985. Thisextent of dependence on Middle East pro-ducers would certainly put the OPEC cartelback in control of world oil prices, and oil im-porters could expect prices probably evenhigher than those demanded by the cartel inthe 1970s. In addition, this dependence wouldcreate a volatile political situation. Military in-tervention in the area by the United States orits allies, sparked perhaps by a supply inter-ruption aimed at diminishing U.S. support ofIsrael, would become much more likely. Asuperpower confrontation or even a major war

31

might ensue. In short, the Middle East couldbecome a nuclear flashpoint.

Under our projection, dependence on MiddleEast and North African oil could probably besustained at the 1983-1985 "world oil glut"level of 15 mbd. Global oil demand wouldprobably be so low that oil supplies availableoutside the Middle East and North Africa atproduction costs of less than $30 a barrelwould be adequate to make up the differencebetween total demand and 15 mbd. The resultwould be both a lower world oil price than inthe IIASA scenarios and greatly enhancedglobal security.

Next consider nuclear power. To meetIIASA's projected demand levels, installednuclear generating capacity would have to in-crease to between 2,200 and 3,700 GW(e) by2020, up from 120 GW(e) in 1980. With thisamount of nuclear capacity, each year 2.1 to3.5 million kilograms of plutonium would berecovered from spent reactor fuel and circu-lated in nuclear commerce around the world.However well meaning the international effortto prevent nuclear weapons proliferation, it isdifficult to imagine how international institu-tions could adequately safeguard virtually allthis material against occasional diversion byeither terrorists or governments. It takes only 5to 10 kilograms to make a nuclear weapon.

In our scenario, nuclear power is consideredan energy source of last resort. Installednuclear capacity would increase to only 460GW(e) by 2020. No net increase in nuclearpower beyond that already planned for theyear 2000 would be necessary. The onlynuclear plants built after the turn of the cen-tury would be those replacing retired plants.Under these circumstances, the economics ofspent fuel recovery, nuclear fuel reprocessing,and plutonium recycling would remain unfav-orable for the entire period and far beyond.

Now consider fossil fuels generally. Overallfossil fuel use would increase substantially withthe IIASA scenarios, and coal use in particular

would increase by 2 to 3.5 times. (See Figure11.) The carbon dioxide (CO2) content of the at-mosphere would double by the second half ofthe next century, and major changes in theglobal climate would be likely well before then.

In our scenario, fossil fuel use would not in-crease, coal use would be reduced about 20percent, and atmospheric CO2 would be 1.3times the pre-industrial level by 2020. If all re-maining recoverable oil and natural gas wereeventually used up, and if coal use were tocontinue declining at the rate of 0.6 percent peryear (the rate of decline for the period 1980through 2020 in our scenario), the level of car-bon dioxide in the atmosphere would eventual-ly reach 1.7 times the pre-industrial level.

The reduced use of oil and coal in our sce-nario would be offset by an increase in the useof natural gas, to the point at which oil andgas uses are equal. (See Figure 11.) A shift tonatural gas would help reduce the CO2 prob-lem because the combustion of natural gasreleases only about half as much CO2 per unitof energy as does coal combustion. Because oilprices, and thus natural gas prices, would notrise much in our low-demand scenario, naturalgas would be preferred to coal, which, whenused in environmentally acceptable ways, en-tails large capital investments.

The shift to natural gas also makes sensefrom the perspective of resource availability.There is probably about as much recoverablegas left in the ground as there is oil. Yet,natural gas is used at only about half the rateof oil.

Our low-demand scenario would certainlynot solve the CO2 problem. An eventual at-mospheric CO2 level 1.7 times the pre-indus-trial level would still induce significant climaticchange. Our scenario would buy time how-ever. If acceptable long-term alternatives tofossil fuels are not developed, there would bemuch more time to adjust to a changed globalclimate than if the IIASA or WEC projectionswere realized. It is more likely that acceptable

32

alternative fuel sources (such as hydrogen pro-duced from amorphous silicon cells14) willbecome available in the next few decades, mak-ing it feasible to phase out fossil fuels morequickly.

A More Hopeful Outlook

Fortunately, energy futures such as thosedescribed by the IIASA and WEC analyses arenot inevitable. It turns out that the worldneeds less energy, much less, than mostenergy analysts thought in the 1970s and early1980s. To meet the energy needs of the world's7 billion citizens in 2020 would not requiredoubling or tripling present energy require-ments by 2020; roughly as much energy as wenow use would be enough if energy plannersworried less about supply expansion and muchmore about efficiency improvements at thepoint of energy use.

We have not shown that the global energybalances we have arrived at are consistent withwhat can be achieved in a particular region or

The world needs less energy, much less,than most energy analysts thought in the1970s and early 1980s.

country; that important exercise remains to becarried out. But our analysis does suggest thatthere are no obvious significant global con-straints to an energy future consistent with thesolutions to other important global problems.Our analysis also stops short of identifying atruly sustainable long-run energy future; wehave not looked beyond the year 2020. Beforethen, however, new technology will makelong-range problems even more manageable.Meanwhile, energy planners should not striveto solve the energy problem for all time, but,rather, they should pursue an evolutionaryenergy strategy consistent with the achieve-ment of a sustainable world.

33

IV. Behind the Numbers

E nergy is only one important globalproblem that must be managed in thedecades ahead if a sustainable world

society is to be achieved. Other pressing prob-lems include the global economic crisis, North-South tensions, widespread poverty in devel-oping countries, population growth, foodscarcity, the risk of nuclear war, nuclearweapons proliferation, environmental degrada-tion, the human role in global climate change,and deforestation and desertification. As thepreceding discussion shows, all these problemsare strongly related to energy use.

The global energy strategy pursued shouldbe consistent with efforts to solve these otherglobal problems. In other words, it should con-tribute to achieving such social goals as eco-nomic efficiency, equity, environmental sound-ness, human welfare, and peace. But moststudies of the global energy problem have notfocussed on these links to other importantglobal problems. Rather the implicit assump-tion in most analyses is that the energy futureis largely determined or is restricted to a rela-tively narrow range of outcomes, with littleroom for changes people may induce. Usingmodels that tightly couple economic growthwith growth in energy use, most global energyanalyses thus project large increases in futureenergy requirements and then focus on the ef-forts needed to expand energy supplies to sat-isfy projected energy needs. In these analyses,the consequent global risks of overdependenceon Middle East oil, fossil fuels generally, and

nuclear power are either ignored or regardedas the necessary price of progress.

In addition, the energy problems of develop-ing countries are usually slighted in mostglobal energy analyses. The burdens of having

Any global energy strategy pursuedshould contribute to achieving such socialgoals as economic efficiency, equity,environmental soundness, human welfare,and peace.

to compete for scarcer, more costly energy sup-plies because of industrialized countries' grow-ing energy appetite are usually ignored. More-over, though nearly half the energy consumedby the three fourths of humanity living indeveloping countries is non-commercialenergy,15 non-commercial energy has not beenan important consideration in most globalenergy analyses. To the extent that developingcountries are considered, their energy futuresare typically decribed as following the pathalready taken by industrialized countries.

A New ApproachOur analysis focusses on end uses rather thanaggregate energy consumption. Because the

35

use of energy is not an end in itself, we havenot studied the use of energy in the traditionalway, using "macro descriptors" of aggregateenergy demand—e.g., the income and priceelasticities economists use to explain energy-consumption behavior. Obviously, energy isuseful only insofar as it provides such servicesas cooking, lighting, heating, refrigeration,mechanical work, and personal and freighttransport in ways that improve the quality oflife. Accordingly, we have tried to understandthe role of energy in society better by scrutiniz-ing the patterns of energy end uses—askinghow and by whom different forms of energyare used today and how the energy end-usesystem might look in the future. Using thisend-use approach, we explore the feasibility ofmodifying the evolution of the energy systemin ways that would facilitate or be compatiblewith the achievement of a sustainable society.

We have identified possible energy featuresfar outside the range normally considered inlong-term energy projections. (See Figures 9 and11.) The end-use approach helps identify prob-lems (e.g., whether progress is being made ineradicating poverty), trends (e.g., structuralshifts in the economy), and new possibilities(e.g., more energy-efficient end-use technol-ogies) that are obscured in energy analysesbased on highly aggregated descriptors.

If total energy demand is not too great, plan-ners have flexibility in charting the energycourse to meet overall energy needs at accept-able costs, to reallocate resources to fulfillunmet social needs, and to select a mix ofenergy supplies that avoids or minimizesdependence on the most troublesome sources.Significant flexibility in energy planning hasalready been achieved by OECD countries,where there has been no increase in energyuse since 1973, despite a large increase ineconomic output. (See Figure 14.) We focus onthe potential for further decoupling energy useand economic growth, in both industrializedand developing countries, and on the resultingflexibility to pursue energy strategies consistentwith achieving a sustainable world society.

Structural Shifts in IndustrializedMarket-Oriented CountriesThe economies of market-oriented industrial-ized countries are undergoing structuralchanges that are reshaping their energydemands. These economies have entered thepost-industrial phase of economic growth, inwhich the service sector grows rapidly relativeto the goods-producing sector while the manu-facture of goods shifts to products character-ized by a high ratio of value added to materialcontent. Both these trends are leading to a lessenergy-intensive mix of economic activity.

Growth of the Service Sector. The role ofservices—in finance, insurance, informationmanagement, marketing, medical care, andeducation, for example—relative to the produc-tion of goods has been increasing for decadesin industrialized market economies. Thischange is reflected in the long-term employ-ment trends of Sweden and the United States.(See Figure 15.)

In the early years of industrialization, theshares of employment accounted for by manu-facturing and services both grew while employ-ment in agriculture declined. Mechanization ofagriculture made it possible for the productionof food and fiber to increase at the same time.Over the last few decades, the service sectorhas been growing relative to manufacturing aswell. This trend is significant because pro-viding services generally requires much lessenergy per dollar of output than manufacturinggoods does.

The shift to services is also reflected in theeconomic output of the goods-producing sec-tor. In the United States, goods production(measured by the value-added index of "grossproduct originating") grew just 0.83 times asfast as GNP between 1960 and 1980. InSweden, it grew 0.6 times as fast.

The Growing Importance of Fabrication andFinishing. A more recent development relatingto goods production is that the demand for a

36

Figure 15. Sectoral Distribution of Employment in

c

8£C0*

a

ploy

!0

c

tio

3

stri

5

inn

9 0 -

80-

70_

60_

50_

40 _

30-

20-

10-

0

\r

• * • •

1930

For theSweden it

Services y. ,S^ •

f \ i \ «.«̂ "

Manufacturing, Miningand Construction

A A"~^-v^

Agriculture

*••• . . .

1 1 1 1 11940 1950 1960 1970 1980

United States

V

a'—'s6

ploy

woc

tio3

stri

b

United States, the employment measure isis the number of employees working inon

the United States (Left) and Sweden (Right)

80

70_

6 0 -

50_

4 0 -

30_

20_

1 0 -

0

•

• • # Agriculture

* # Services —*• •*

• *

/ y \**" """"̂ ^̂ *•

/Manufacturing, Mining » , #

and Construction

1 1 I 1 11870 1890 1910 1930 1950 1970

Sweden

the number of full-time equivalent employees. For: than half time.

wide range of materials, both traditional andmodern, is no longer increasing in physicalterms (measured in kilograms per capita peryear). (See Figure 16, bottom, and Figure 17.)

For traditional materials, this recent develop-ment can be explained in part by the substitu-tion of more modern materials. In addition,both traditional and modern materials are be-ing used much more efficiently—for example,through the development of higher strength ormore durable products. The high-strength steelgirders used recently to repair the Eiffel Towerweighed just one third as much as thosereplaced. Although inter-material competitionand the increased cost of materials have accel-

erated improvements in the efficiency ofmaterials use, this trend is not a new pheno-menon—witness the continual drop in theweight-to-power ratio for locomotives from thebeginning of the 19th Century to the present.(See Figure 18.)

Although materials substitution and increasedefficiency of materials use are clearly contribu-ting to the shift away from basic materials, to-day these factors are probably not as importantin reducing demand as the saturation of mar-kets for bulk materials and heavy consumergoods and a shifting of consumer preferencesto products characterized by a higher ratio ofvalue added to material content. Today the af-

37

Figure 16. Trends in Consumption Per Capita (Bottom) and Consumption Per Dollar of Gross NationalProduct (Top) in the United States for Both Traditional (Steel, Cement, Paper) and Modern(Aluminum, Ethylene, Chlorine, and Ammonia) Basic Materials

G.N.P. Per Capita (In 1983 Dollars)

ft.

U

& J3

3

2

11000I

14000

_ 6

ft.

.5 "

_4

.3

_i

2

700-, L

600_

500_

400-

300-

200.

100.

.0

.70

- 6 0

|_50 i ,

. 4 0 Cu

I- 3 0 |

- 2 0

. 10

I I I I I 1 I i r ^1890 1900 1910 1920 1930 1940 1950 1960 1970 1980

Year

The data are for apparent consumption (production plus net imports, adjusted for stock changes), andthe plotted points are for 5-year running averages.

Source: R.H. Williams, E.D. Larson, and M.H. Ross, "Materials, Affluence, and Industrial Energy Use," The Annual Review ofEnergy 12 (1987): 99-144.

38

Figure 17. Trends in the Apparent Consumption Per Capita for Selected Basic Materials in WesternEurope

500.

400.

- 5 0

_ 40

- 3 0

-20

3'S.U

S9

1950 1955 1960 1965 1970 1975 1980 1985

Year

The data shown are aggregates of apparent consumption for the Federal Republic of Germany, France,and the United Kingdom. The ammonia data are for the Federal Republic of Germany and France only.The plotted points are 3-year running averages.

Source: M. Ross, E.D. Larson, and R.H. Williams, "Energy Demand and Materials Flow in the Economy," Energy, the Inter-national Journal, in press.

fluent tend to spend additional income not onextra refrigerators or cars but on such items asvideo cassette recorders and personal computersand software. Moreover, even in replacement

markets, the trend is toward reduced materialintensity. As more affluent consumers replacetheir old appliances and automobiles with moreexpensive new ones, they are buying products

39

Figure 18. Trends in the Weight-to-Power Ratio of Locomotives Between 1810 and 1980

o&eo

v

I1

1,000

800 .

600 .

400 i-

200

100 .

80 -

60 .

40

20

10 .

\

\

\

\

SteamLocomotives

^

1

ElectricLocomotives

1810 1850 1900 1950 1980

The weight-to-power ratio for locomotives decreased nearly 70-fold between 1810 and 1980 as a result ofimprovements in design and materials. In the mid-19th Century iron was replaced by steel in boilers—achange that made possible lighter equipment and higher internal pressures. Between 1810 and 1900 theratio declined 10-fold, and it declined another 4-fold by 1950, when the electric locomotive was introduc-ed. (The gap between 1910 and 1920 results from the disruption of data collection during World War I.)

Similar (albeit less dramatic) improvements have been made in many industrial products. Substitutionand design changes that lead to more efficient use of materials are two of the factors responsible for theleveling off of demand for basic materials in the industrialized market economies.

Source: Economic Commission for Europe, Evolution of the Specific Consumption of Steel (New York: United Nations,1984).

40

containing less material per dollar of price. Thedownward trend in material intensity withprice is illustrated for a wide range of tradi-tional consumer products in Sweden. (SeeFigure 19.)

These changing consumer preferences arereflected in a production shift away from pro-cessing basic materials to fabricating andfinishing increasingly complex goods that arecharacterized by low ratios of material contentper dollar of value added (e.g., high-strengthand corrosion-resistant steels and specialtychemicals for the fast-growing pharmaceutical,electronics, and biotechnology markets).

The high-strength steel girders usedrecently to repair the Eiffel Tower weighedjust one third as much as those replaced.

The concept of the life cycle of a material inthe economy can be helpful in understandingrecent trends in materials use in industrializedcountries. At the beginning of this cycle, whena material is first introduced, consumptionrates are low and there are vast potential mar-kets. In this phase, consumption growsrapidly—usually much more rapidly than theoverall economy. This growth encourages ad-vances in processing technology that generallyincrease productivity, leading to lower pricesand improved product quality, thus stimulatingfurther growth in demand. In the next phaseof the cycle, the ratio of value added to thekilogram content of material increases as moresophisticated products are emphasized. In thisphase, the demand for the material measuredin kilograms of material per dollar of GNPpeaks and begins to decline. Such was the situ-ation in the steel and cement industries in theUnited States in the 1920s and 1930s and inSweden in the 1950s and 1960s. (See Figure 16,

top, and Figure 20.) This trend also occurred inthe late 1970s in the United States for impor-tant modern materials—ammonia, ethylene,chlorine, and aluminum. (See Figure 16, top.)

Materials use per capita often continues togrow after materials use per dollar of GNPbegins to decline. (Compare the top and bot-tom curves in Figure 16.) Eventually, however,materials-intensive markets approach satura-tion, and new markets are largely for specialtyproducts that have little effect on total con-sumption. In this stage, per capita consump-tion levels off and may even begin to decline,as is now happening in the United States andWestern Europe. (See Figure 16, bottom andFigure 17.)

The energy demand implications of this shiftaway from basic materials are profound. Theindustries that process basic materials (petro-leum refining; primary metals; paper and pulp;chemicals; stone, clay, and glass; and food pro-cessing) account for most energy use in manu-facturing, while the fabrication and finishingindustries (the rest of manufacturing) accountfor most value added. These latter industriesrequire only a tiny fraction of the energy re-quired in processing basic materials per dollarof value added. (See Figures 21 and 22.) Thisshift therefore implies a substantial decouplingof industrial energy use from industrial output.In the United States, the shift away from basicmaterials accounted for an average annualreduction of 1.6 percent in the ratio of in-dustrial energy use to GNP between 1973 and1985.16 If this structural shift had not occurred,industrial energy demand in 1985 would havebeen higher by the equivalent of 3.2 millionbarrels of oil per day, or roughly three fourthsof all oil imports that year.

The Structure of Energy Demand inDeveloping Countries

In contrast to industrialized countries, formany years most developing countries willneed to expand materials-and energy-intensive

41

Figure 19. Material Intensity for Household Appliances (Top) and Transport Vehicles (Bottom)Versus Sales Price in Sweden

3 0 .

2 0 .

1 0 .

A

X\ \\ i

(_} Average A A

if RangesCH Washing MachinesA Refrigerators© Refrigerator/Cooler• Refrigerator/FreezersA Freezers

i

NN

• N

A "̂ t" ^ Uu

— —

A

D

\ D"* D

-A.

2000 4000 6000

30-Price (Swedish Kroner)

20_

10 _

O Average

A Autos (1961)A Autos (1981)if Motorcycles• Bicycles

A

> *

50,000 100,000

Price (Swedish Kroner)

150,000 200,000

The downward trend of the materials intensity with increasing sales price shows that as their incomesincrease and they purchase more costly products, consumers tend to buy less materials-intensive (morevalue added-intensive) products.

Source: T.B. Johansson and P. Steen, Perspectives on Energy (Stockholm: Liber Vorlag, 1985). In Swedish.

42

Figure 20. The Material Intensity of Steel and Cement in the SwedishLiving Over Time in Sweden

100 _

•u 75JS0

D

&

2 5 0 -O

late

rial

Int

ensi

ty

1

0

Source: P

Steel —*~

1 9 5 6 / 6 0 ^ ^

1941/45 /

/ N . / / 1951/55

/ | ~*— Cement

1926/30 / 11896/1900 K / \ l f

1 * 1941/45

1886/90 ^ ** •• 1931/35

-\ 1916/20

1

- ' 1886/90

1 i | I | l

1000 3000 5000

Economy Versus

1965

y/ \ 1970

\y~\

Standard of

V \ /• \1975 \

1968

\\

• _1970

17000

Gross Domestic Product Per Capita (1975 U.S. Dollars)

Steen, et a\., Energy—For What and How Much? (Stockholm: Liber Vorlag,1981). In Swedish.

1979

• \\

\\ 197511

1979

t

-

-

—

9000

economic activities—building and operatingmore factories, schools, hospitals, officebuildings, houses, and transport systems andmanufacturing consumer products for domesticmarkets that are far from saturated.

The energy required for this developmentcan vary significantly, however, according tothe development strategy selected. The devel-opment strategy underlying our analysis em-

phasizes the satisfaction of basic human needs,industrial activities that promote new employ-ment, and agricultural production.

Energy and Basic Human Needs. In the 1950s,it was widely believed that maximizing eco-nomic growth was the best way to eradicatepoverty. However, the benefits of rapid eco-nomic growth have not trickled down to thepoor. While rapid growth is necessary for suc-

43

cessful development, it is not sufficient. Amore effective way of fighting poverty is bydirectly allocating resources, including energy,to meet basic human needs for nutrition,shelter, clothing, health, and education.17 Thereis no empirical evidence showing that targetingbasic human needs slows economic growth.18

In fact, theory suggests that satisfying suchneeds would speed growth because farmer andworker productivity would increase.19

If the emphasis in development is on satisfy-ing basic human needs, the societal patterns ofenergy use, the kinds of energy-using equip-ment deployed, and the kinds and amounts ofsupplies produced are all affected. In a needs-based energy strategy, for example, the wideuse of fuel-efficient cooking stoves would be

The industries that process basicmaterials account for most energy use inmanufacturing, while the fabrication andfinishing industries account for mostvalue added.

promoted, both for those who can purchasethem and for those living outside the marketeconomy, who cannot afford them. Taking thisapproach would reduce the increasingdrudgery of gathering and hauling wood,drudgery borne primarily by women and chil-dren. It would also ease pressure on woodresources and perhaps help end their unsus-tainable exploitation. In addition, it might freesome fuelwood resources for rural industrializa-tion and other uses.

Emphasizing efficient stoves would requiremajor new efforts to manufacture, market, anddistribute the stoves to the general population.Such a program would be ambitious, but it

could provide many benefits while using onlya fraction of the financial resources required byconventional attempts to expand energy sup-ply. An annual investment of $1 billion wouldprovide efficient fuelwood stoves to all the 400million rural households of developing coun-tries (assuming that each stove costs $5 andlasts two years on average). Much fuelwoodwould be saved—enough, for example, to pro-duce electricity in biomass-fired power plantsequal to the output of about 80 large nuclearpower plants costing $160 billion.

Emphasizing basic needs would also meantrying to electrify every household. Today,even in countries with rural electrification pro-grams, typically only 10 to 15 percent of ruralhouseholds are electrified. Reaching all house-holds often requires greater emphasis ondecentralized electricity production (for exam-ple, producer-gas engine generator sets forvillages) because centralized electricity produc-tion is simply uneconomical or impractical formany rural needs. A society that emphasizesbasic needs would also seek to bring cleanwater and good sanitation to every household,requiring a pattern of direct and indirectenergy use quite different from the one thatemerges when production is geared to satisfy-ing elite consumers.

More generally, a society determined to meetbasic needs would develop the capacity to pro-duce a wide range of goods and services forthe domestic population. In turn, emphasizingmass domestic markets requires industrialdevelopment with a mix of materials-intensivebasic industries and fabrication and finishingindustries to convert the processed raw materi-als into industrial and consumer products.Such an economy would be, on average, lessenergy-intensive than one that emphasizes theexport of processed basic materials such assteel, aluminum ingot, and basic chemicalsbecause fabrication and finishing activities indeveloping countries are far less energy-intensive than the basic materials processingindustries, just as they are in industrializedcountries. (See Figures 21, 22, and 23.)

44

Figure 21. Final Energy Intensity Versus Manufacturing Value Added for United States Manufactur-ing Industries in 1978

450_

a

re

1Rcoy

>t3

oD

400 Z

350 Z

300 Z

250 Z

200 Z-

150 Z

too Z

50Z

-

o I

PetroleumRefining

"*~~ (17.3)

PrimaryMetals

/ (21.3)

i 1 I | . i i

Chemicalsy , (26.4)

Paper(10.9)

Stone, Clay,-*— Glass

(6.3)

Food

i i i i | • i i | I i i | i i I I | i i i i

/

1 i i i

Other(13.1)

I I i I i i | i i > i

10 20 30 40 50 60 70 80

Contribution to Manufacturing Value Added in Percent

90 100

The number displayed for each bar is that sector's contribution (in percent) to total final energy use inUnited States manufacturing.

Energy and Employment Generation. Theunemployment now rife in developing coun-tries must be a foremost concern in any devel-opment strategy. Unemployment is so seriouslargely because many current production tech-nologies are ill-suited to developing countries'needs. While developing countries are capital-

poor and labor-rich, many available technol-ogies are capital-intensive and labor-saving,designed primarily for industrialized countries.Of course, the labor-intensive technologiesused in Europe and North America duringtheir industrialization in the 19th Centurywould not be competitive today, partly because

45

Figure 22. Final Energy Intensity Versus Manufacturing Value-Added for Swedish ManufacturingIndustries in 1978.

500.

2re

1

450 _

400 _

350 _

cU 300 .

-a

u 250 _

j« 200 _

a150_ |

100 _

50 _

Petroleum- Refining

(7.2)

Pr imaryMetals Paper(18.1) and Pulp

X

Stone, Clay,

/Glass(6.2)

ForestProducts

/ (7 2) Chemicals' ' l7-°) .Food

Other/ (8.0)

1 I I I | I | I I | I I I I | I I I I | I I I I | I I I I | I I I I | I I I I | I I I I | I I I I0 10 20 30 40 50 60 70 80 90 100

Contribution to Manufacturing Value Added in Percent

The number displayed for each bar is that sector's contribution to total final energy use in Swedishmanufacturing, in percent.

they are too energy-intensive for today'senergy prices. But the employment implicationsof alternative technologies and strategies for in-dustrialization must be taken into account.

Gainful employment provides the purchasingpower that enables people to satisfy some oftheir basic needs. In rural areas, the lack ofsuch employment not only keeps per capita in-

46

Figure 23. Final Energy Intensity Versus Manufacturing Value Added for Brazilian ManufacturingIndustries in 1980

2u

3

u

3

cr

5

40 1

3 5 :

•S 30 _coU

25 _

2 0 ~

15 _

10 _

Non-FerrousX a n d Other Metals

(7.8)

Ceramics(5.0)

Cement(6.9)

Ironand Steel

/ (20.3)

Non-Metallic

/Minerals

(4.6)

Paper' (5.9)

Iron Alloysy (2.D

i <

10

Food

. Chemicals(11.5) Other

/ (14.2)

| i I i i | I I I i | I I I I | I I I 1 | I I I I | I I I i | I I i I |

20 30 40 50 60 70 80

Contribution to Manufacturing Value Added in Percent

90 100

The number displayed for each bar is that sector's contribution (in percent) to total final energy use inBrazilian manufacturing.

come extremely low but also induces people tomigrate to already overburdened cities.

Of course, some industries generate far morejobs than others. Most basic materials process-Ing industries have both low labor intensitiesmd high energy intensities, the fabrication and

finishing industries have high labor intensitiesand low energy intensities. In the industrialsector of the state of Karnataka in India, for ex-ample, 18 electro-metallurgical firms consumedtwo thirds of all industrial electricity anddirectly employed 4,000 people. In contrast,1,200 other firms used the remaining one third

47

of industrial electricity but provided employ-ment for 250,000.20

Naturally, developing countries cannot basedevelopment exclusively on fabrication andfinishing industries. Yet, planners should guardagainst over-investment in industries with lowemployment potential. For this reason, invest-ments in basic industries aimed at export mar-kets should be approached cautiously. If, in-stead, basic industries are expanded largely toserve domestic markets and are complementedby new fabrication and finishing industries,many more jobs would be created and muchless energy would be required. Moreover,when manufactured exports are required toearn foreign exchange, much more valuewould be added and employment created iffinished materials instead of processed basicmaterials (for example, aluminum-intensiveautomobiles instead of aluminum ingot) wereexported.

Energy supply choices can also significantlyaffect unemployment. In particular, it takes farmore labor to plant, grow, harvest, and pro-cess biomass for energy production than toproduce fossil fuels. The Brazilian alcohol pro-gram, which produced about 11 billion liters ofethanol from sugar cane in 1985, illustrates thepoint. The program requires an investment ofonly $6,000 to $28,000 per job, compared to theaverage of $42,000 for Brazilian industry and$200,000 per job for the oil-refining/petrochemi-cal complex at Camarcari. In 1985, the ethanolprogram directly generated an estimated 475,000full-time jobs in agriculture and industry, alongwith another 100,000 jobs indirectly in com-merce, services, and government.21

Brazil's ethanol program is also cost-effective.The cost of producing this ethanol has beenestimated at $50 to $56 (in 1983 U.S. dollars)per barrel of gasoline replaced when the sub-sidies are removed and a realistic exchange ratebased on the parity value of exported goods isused.22 This cost is competitive with gasolineproduced in Brazil from imported crude oil atthe 1981 world oil price, a price that will likely

be reached again in the next decade. Theethanol program has also dramatically reducedoil imports and associated foreign exchange re-quirements, displacing 55 percent of the gaso-line that would otherwise have been demandedin 1985.

Among biomass's other appeals as an energysource are its wide availability in rural areasand its suitability for conversion to usefulenergy in relatively small-scale systems. Bothfeatures facilitate rural industrialization, whichin turn creates jobs where they are mostneeded.

Energy for Agriculture. A major challenge fordeveloping countries in the decades ahead is tofeed rapidly growing populations. To meet thischallenge, the United Nations Food and Agri-culture Organization (FAO) has called for mod-ernizing traditional agriculture so as to doubleagricultural production by the year 2000.23

An important concern in modernizing agri-culture is the implications of such change foremployment. Traditional low-yield agricultureis labor-intensive, and modern, mechanized,energy-intensive, high-yield agriculture isnot—or so goes the conventional wisdom.Some analysts thus worry that in manydeveloping countries modernization will exacer-bate unemployment or underemployment andwill further increase inequities in income. Afterall, countries that have already passed fromagrarian to industrial economies have watchedtheir agricultural work forces shrink. On theother hand, if agricultural output is increasedusing traditional labor- and land-intensivemethods, agricultural production will demandmore forest and marginal lands, aggravatingthe already serious deforestation anddesertification.

Comparing modern and traditional agricul-tural techniques in this way may be simplistic,however. Agricultural modernization is notsimply a "black box" through which capital,energy, and other inputs can be substituted forlabor to increase yield per hectare.24 Rather,

48

many different technological possibilities existfor modernizing various aspects of agriculturalproduction. Increased energy and capital inputsdo not necessarily reduce labor inputs or in-crease crop output. The outcome depends onwhat combination of technologies is used, ashas been shown for rain-fed rice culture, whichaccounts for half the Asian rice production andyielded 411 million tonnes in 1982. (See Figure24.) The pumping of water for irrigation in dryperiods is a good example of how increaseduse of energy can promote employment gener-ation. Multiple cropping is then possible, andmultiple cropping increases labor requirements.

Agricultural modernization is not simplya "black box" through which capital,energy, and other inputs can besubstituted for labor to increase yield perhectare.

Although modernizing agriculture need notentail job losses, it does require extra energyfor mechanical tillers, tractors, irrigationpumps, fertilizer production, and other meansof crop production. According to FAO, meetingthis goal would require increasing the commer-cial energy used for agriculture at an averageannual rate of 8 percent, from the equivalent of0.74 million barrels of oil per day in 1980 to 3.5mbd in 2000.25 Despite the high growth rate,however, the amount of extra energy requiredis not large in absolute terms. Indeed, the in-crement required between 1980 and 2000, about2.8 mbd, is no more than the amount of oil theUnited States saved between 1978 and 1981.Providing this much extra energy for agricul-ture need not be particularly burdensome—forexample, in terms of foreign exchange require-ments for oil-importing countries. In fact, agri-culture's increased requirements could besatisfied largely through energy-efficiency im-

provements in major oil-using sectors such astransportation. (See Table 4.)

There are also alternatives to petroleum formeeting agriculture's energy requirements.Agriculture's energy demands could providedeveloping countries an incentive to developsynthetic fuels from biomass, such as methanolderived by thermochemical processes fromorganic residues or wood. Methanol in anamount equivalent to 2.8 mbd of oil could beproduced with 50-percent overall conversionefficiency, using 40 percent of the organicwastes—crop residues, animal manure, andfood-processing wastes—generated in develop-ing countries today. Alternatively, methanolcould be produced from wood from treesgrown on energy plantations. At an annualyield of 10 tonnes per hectare, roughly 66million hectares—about 3 percent of theforested area in developing countries—wouldbe required to meet agriculture's increasedenergy requirements. Clearly, the availability ofenergy need not constrain the production in-creases FAO projects.

Energy Planning as an Instrument of Devel-opment. Instead of simply retracing the devel-opment path once traveled by the industrializedcountries, developing countries must chart newcourses reflecting their unique resource capabil-ities and constraints. Although developmentgoals cannot be achieved by energy planningalone, our analysis has shown that energyplanning can be a powerful instrument ofdevelopment, in light of the central role ofenergy—in meeting basic human needs, in gen-erating employment, and in meeting agricul-tural goals.

Opportunities for Using Commer-cial Energy More Efficiently

The economies of both industrialized anddeveloping countries present abundant oppor-tunities for making commercial energy usemore efficient. It usually costs less to save aunit of energy through more efficient use than

49

Figure 24. Paddy Output, Human Labor, Animal Labor, Direct and Indirect Inanimate Energy, andFixed Capital Costs of Selected Technologies for Rice Production

4000

13 3000.

£ 2000 -

o 1000.

Paddy Output Human Labor Animal Labor

-

=

111£38

• • A %<

T VI V2 V3

1200

600 . -

B.

11IVI V2 V3

3600

2700

1800

900

0

Direct Inanimate Energy

D.

•

•

Indirect Inanimate Energy

500

400

300 .

100 .

200 .

V2 V3

Fixed Capital

VI V2 V3

-

F.

i—|

BBSSHSSSS

VI V3 VI V2 V3

T = Traditional Technology: No hybrid seeds, chemical fertilizer, pesticides, or herbicides; no use of oilfor transport vehicles or land preparation. Instead traditional seed varieties and organic fertilizers areused, with draught animals used for land preparation, manual transplanting, and manual harvesting,threshing, and winnowing.

VI = Variant 1: Differs from Traditional Technology in that it uses modern biological-chemical inputs(hybrid seeds, fertilizer, insecticides, and herbicides).

V2 = Variant 2: Differs from Variant 1 by using tractors for plowing and mechanical driven vehicles fortransport while retaining draught animals for harrowing.

V3 = Variant 3: Uses power threshers and replaces the tractors of Variant 2 with power tillers ("walkingtractors") for plowing and harrowing.

Source: A.K.N. Reddy, "The Energy and Economic Implications of Agricultural Technologies: An Approach Based on theTechnical Options for the Operations of Crop Production," PU/CEES 182 (Center for Energy and EnvironmentalStudies, [Princeton University, Princeton, New Jersey, 1985]): and ILO World Employment Programme Research Work-ing Paper 2-22/WP 149 (Geneva: International Labor Organization, June 1985).

50

Table 4. Distribution

AgricultureIndustryTransportHouseholdCommercialMiningRoad ConstructionElectricity ProductionOther

Sources: For Ethiopia:in a Forest olReddy, "An

of Petroleum Consumption by End Use (in Percent)

Ethiopia, 1979

5.512.364.70.73.00.55.35.62.4

100

India, 1978

13.86.5

55.320.3

4.1

100

R. Hosier, et al., "Energy Planning in Developing Countries: Blunt AxeProblems?" Ambio, vol. 11, no. 4(1982): 180-187. For India: A.K.N.

End-Use Methodology for Development-Oriented Energy Planning inDeveloping Countries, with India as a Case Study," PU/CEES 181,and Environmental Studies, Princeton University, Princeton, New

Center for EnergyJersey, 1985.

to produce an additional unit by expanding theenergy supply. Four of the many areas thatshow promise for efficiency improvements arespace heating, appliances, automobiles, andsteelmaking.

Space Heating in Industrialized Countries.Space heating accounts for 60 to 80 percent offinal energy use in residential buildings in in-dustrialized countries. Prospects are good forcost-effectively reducing energy requirementsfor space heating to a minor fraction of house-hold energy use by reducing heat losses andimproving heating system efficiencies.

SHELL MODIFICATIONS IN NEW HOMESHeat losses can be reduced by increasing in-sulation, by adding extra glass panes to win-dows, and by reducing natural air infiltration.Indoor air quality can be maintained in a tighthouse by replacing natural ventilation withmechanical ventilation and by using a heat ex-changer to transfer heat from the stale exhaust

air to incoming fresh air or to water fordomestic heating.

A good measure of a house's energy perfor-mance is the heating system's energy output,adjusted for floor area and climate variations.In the United States and Sweden, improvedenergy performance has been demonstrated invarious groups of new houses incorporatingenergy-saving features. (See Table 5.) Enormousimprovements compared to both the existinghousing stock and to typical new constructionare possible in both countries. New superin-sulated houses, such as the Northern EnergyHome (NEH) marketed in New England, haveextremely low energy requirements. (See Figure3.) A NEH in the New York City area with 120square meters of floor space could be heatedwith electric resistance heaters for just 1,400kWh per year, roughly the same amount ofelectricity that a typical new refrigerator-freezerconsumes.

51

Table 5. Space Heat Requirements in Single Family Dwellings(kj per square meter per degree-day)a

United StatesAverage, housing stock 160New (1980) construction 100Mean measured value for 97 houses in Minnesota's Energy Efficient Housing

Demonstration Program 51Mean measured value for 9 houses built in Eugene, Oregon 48Calculated value for a Northern Energy Home in New York City areab 15

SwedenAverage, housing stock0 135Homes built to conform to the 1975 Swedish Building Coded 65Mean measured value for 39 houses built in Skanee 36Measured value, house of Mats Wolgast* 18Calculated value for alternative versions of the prefabricated house sold by Faluhus8

Version No. 1 83Version No. 2 17

Sources: For U.S. housing stock average: R.H. Williams, G.S. Dutt, and H.S. Geller, "FutureEnergy Savings in US Housing," Annual Review of Energy 8(1983): 269-332. For newU.S. construction and nine houses in Oregon: J.C. Ribot et al., "Monitored Low-Energy Houses in North America and Europe," in What Works: Documenting EnergyConservation in Buildings, J. Harris and C. Blumstein, eds., proceedings of the SecondSummer Study on Energy Efficient Buildings (Washington, D.C.: American Council foran Energy Efficient Economy, 1983), 242-256.

a. The required output of the space heating system per unit floor area per heating degree day(base 18° C).

b. The Northern Energy Home (NEH) is a superinsulated home design based on modular con-struction techniques with factory-built wall and ceiling sections mounted on a post andbeam frame. The energy performance was estimated using the CIRA computer program(personal communication from D. Macmillan of the American Council for an Energy Effi-cient Economy, Washington, D.C.). The house has 120 square meters of floor area, triple-glazed windows with night shutters, 20 centimeters (23 cm) of polystyrene insulation in thewalls (ceiling), 0.15 air changes per hour (ACH) natural ventilation plus 0.35 ACH forcedventilation via 70 percent efficient air-to-air heat exchanger, and an internal heat load of 0.65kilowatts. The indoor temperature is assumed to be 21° C in the daytime, set back to 18° Cat night.

c. In 1980 the average values for fuel consumption, floor area, and heating degree days were98.5 gigajoules, 120 square meters, and 4,474 degree days, respectively, for oil heated singlefamily dwellings. For conversion of fuel use to net heating requirements, a furnace efficiencyof 66 percent is assumed (L. Schipper, "Residential Energy Use and Conservation inSweden," Energy and Buildings, vol. 6, no. 1(1984): 15-38.).

d. For a single story house with 130 square meters floor area, no basement, electric resistanceheat, an indoor temperature of 21° C, and 4,010 degree-days.

52

Table 5. Continued

e. The average for 39 identical 4 bedroom, semi-detached houses (112 square meters of floorarea, 3,300 degree-days).

f. The Wolgast house has 130 square meters of heated floor space, 27 centimeters (45 cm) ofmineral wool insulation in the walls (ceiling), quadruple glazing, low natural ventilation plusforced ventilation via air preheated in ground channels. Heat from the exhaust air isrecovered via a heat exchanger. For 3,800 degree-days.

g. The Faluhus has a floor area of 112 square meters. The more energy-efficient Version No. 2(with extra insulation and heat recuperation) costs 3,970 SEK (US $516) per square metercompared to 3,750 SEK (U.S. $488) per square meter for Version No. 1. The annual electrici-ty savings for the more efficient house would be 8960 kWh per year. The cost of savedenergy (assuming a 6 percent discount rate and a 30-year life for the extra investment)would be 0.20 SEK per kilowatt hour (U.S. $0,026 per kilowatt hour). For comparison elec-tricity rates for residential consumers in Sweden consist of a fixed cost independent of con-sumption level of about 1,200 SEK (U.S. $156) per year plus a variable cost of 0.25 SEK(U.S. $0,032) per kilowatt hour.

Although the costs of improved energy per-formance in new construction vary significantlyfrom builder to builder, the cost of savedenergy is generally less than the cost of pur-chased heating fuel or electricity. (See Box 3.)Surprisingly, growing evidence indicates thatthe net extra cost of even superinsulatedhouses may not be very large compared to thecost of conventional houses. Consider two ver-sions of a Faluhus prefabricated house inSweden that are identical except for theirenergy use characteristics and first costs. (SeeTable 5.) For the more efficient version, one ofthe most energy-efficient houses built, the costof saved energy is still less than the cost ofelectricity in Sweden at current rates, eventhough these rates (based largely on hydro-power) are very low—far below marginal costs.

The economics of improving energy efficiencymay actually be even more favorable than suchcalculations suggest, because the additional in-vestment brings several benefits—such as in-creased comfort or reduced maintenance re-quirements. Superinsulated houses generallyare not drafty. And the efficient version of theFaluhus uses a forced ventilation system thatfilters incoming air.

HEATING SYSTEMS FOR NEW HOMES Thenoteworthy achievements in reducing heat lossthrough superinsulated house designs havebeen matched in recent years by improvementsin the design of furnaces and heat pumps.Manufacturers in Europe and North Americaare marketing high-efficiency natural gas andoil furnaces that use one third less fuel thanconventional furnaces do. Even though thesefurnaces typically cost $600 to $700 more thanother new furnaces, they are cost-effective inthe United States wherever the heating re-quirements exceed half the average for the ex-isting housing stock.26

In superinsulated houses, though, thesehigh-efficiency furnaces would be cost-effectiveat present prices only in very cold climates—forexample, in Canada, Sweden, or in the NewEngland, Upper Midwest, and Northern RockyMountain regions of the United States. Forsuperinsulated houses in moderate climates,where space heating is needed only on thecoldest days, a different approach is possible: itis feasible to heat with a few small wall-mounted space heaters. These heaters wouldbe far less costly than a central furnace andheat-distribution system, thus offsetting much

53

of the extra cost of the superinsulated buildingshell. In such houses, the internal temperaturevariation is less than in ordinary houses withconventional heating systems.

With new high-performance heat pumps,substantial energy savings are possible in elec-trically heated houses. In 1982, the most effi-cient unit available in the United States pro-vided 2.6 units of heat per unit of electricityconsumed, that is, its coefficient of perfor-mance (COP) is 2.6.27 This unit requires onethird less electricity than the typical heatpumps in use, which have a COP of about 1.7.

Most heat pumps sold in the United Statesextract heat from outside air. In very coldweather such heat pumps perform no betterthan resistance heaters. Good performance ispossible in cold weather, though, with heatpumps that extract heat from warmer sources.An ingenious system employed in new Swed-ish housing extracts heat from the warm ex-haust air in air-tight, mechanically ventilatedhouses and uses it mainly to preheat domestichot water. A very high COP of 3 is achievedby such systems, which are cost-effective evenin Sweden's well-insulated houses. A COP of 3is also achieved with large district-heating heatpumps in Sweden that draw heat fromsewage, lake, or sea water.

SHELL MODIFICATIONS FOR EXISTINGHOUSES Improving the thermal performanceof existing houses is crucial because they willdominate the industrialized countries' housingstock for many decades.

Thermal design improvements are generallymore costly for existing houses than for newones, and some housing features cannot bemodified easily. Yet numerous cost-effectiveways of improving energy performance arewidely available today for existing houses.They include insulation, storm windows, clockthermostats, and various retrofits for theheating system. Even so, these conventionalmeasures do not exhaust the possibilities.

Box 3. Economic Figures of Merit

In this study, three figures of merit areused to assess the cost-effectiveness ofenergy-saving investments: the cost ofsaved energy, the life-cycle cost, and theinternal rate of return.

The cost of saved energy is an indexhaving energy price units that permits aready comparison between investments inenergy efficiency and energy supply alter-natives. Simply stated, the cost of savedenergy is the annual repayment on ahypothetical loan taken out to pay for aninvestment to save energy, divided by theexpected annual energy savings. Bothprincipal and interest charges (correctedfor inflation) are included, and the termof the loan is equal to the expected life ofthe investment. An investment in energyefficiency is cost-justified if the cost ofsaved energy is less than the cost of theenergy supply alternative beingconsidered.

The life-cycle cost is the present valueof all costs associated with providing aparticular energy service, with futurecosts discounted to present worth usingan inflation-corrected market rate of in-terest. A comparison of the life-cycle costsfor different energy systems provides asimple means of identifying the leastcostly energy strategy.

The internal rate of return is the real(inflation-corrected) rate of return realizedfrom the dollar value of the energy savingsresulting from an energy-saving invest-ment. The decision rules with this indexare that: (1) to be considered cost-justified,the internal rate of return for an invest-ment must be greater than some thres-hold "hurdle rate" that represents theopportunity cost of capital and (2) theinvestment offering the highest rate ofreturn would be favored.

54

Detailed measurements in the late 1970srevealed that obscure defects in houses' ther-mal envelopes allow far greater heat lossesthan were predicted by traditional heat-lossmodels. Fortunately, auditing procedures aidedby instruments such as house pressurizationdevices and infrared viewers now permit thesedefects to be identified quickly.28 Such auditsare much more costly than walk-through, non-instrumented audits that many U.S. gas andelectric utilities offer. However, many of theobscure defects discovered with instrumentscan be corrected on-the-spot with low-costmaterials. When such improvements are madeat the time of the audit, the costly diagnosticprocedure becomes a very cost-effective way tosave significant amounts of energy, even inhouses with attic and wall insulation and stormwindows.

This "house doctor" concept was tested inthe Modular Retrofit Experiment by gas utilitiesin New Jersey and New York in collaborationwith the Buildings Research Group at Prince-ton University. In this experiment, a one-day,two-person house doctor visit saved, on aver-age, 19 percent of the gas use associated withspace heating.29 Later conventional shell-modification retrofits brought the total fuelsavings to an average of 30 percent, for anaverage total investment of about $1,300. Thereal rate of return on this investment, in fuelcosts savings, was nearly 20 percent.30

The achievements demonstrated commerciallyin the Modular Retrofit Experiment do notrepresent all that can be done through shellimprovements in existing dwellings. Onereason is that energy-saving shell improve-ments become more cost-effective when theyare accompanied by home improvements madefor reasons other than energy savings. For ex-ample, if old windows are replaced to reducedrafts, facilitate cleaning, or make a room morepleasant, energy-saving windows are economi-cally attractive compared to conventional re-placement windows, even when replacementwindows cannot be justified on the basis of theexpected energy savings alone. Over the years,

there will be many opportunities to incorporatenew energy-saving features this way in typicalhouses.

A second reason is that new technologies forenergy savings will continually be commer-cialized. One experiment by Princeton Univer-sity researchers suggests the possibilities. Theseresearchers exploited unconventional retrofitopportunities that brought about energy sav-ings of two thirds in a house regarded as"thermally tight" by U.S. standards before itwas modified.31

HEATING SYSTEMS FOR EXISTING HOUSESThe economics of efficient furnaces or heatpumps tend to be much better in the replace-ment market than in new housing because theheat loads in the former are relatively large.Even after major shell improvements, heatingrequirements greatly exceed those in newsuperinsulated houses. For example, if newenergy-efficient condensing gas furnaces wereintroduced after the shell improvements weremade in the houses of the Modular Retrofit Ex-periment, the fuel requirements for spaceheating could be reduced to 44 percent ratherthan only 70 percent of the pre-retrofit level.Installing a new condensing furnace instead ofa new conventional furnace that costs $1,000less would result in a 15-percent rate of returnon the extra investment.32

Appliances in Industrialized and DevelopingCountries. Water heating accounts for between10 and 35 percent of residential energy use inindustrialized countries. In the United Statesand Sweden, refrigeration accounts for onefourth of residential electricity use. Lightingtypically accounts for about half as much elec-tricity use as refrigeration.

WATER HEATING Recently, gas-fired waterheaters that use only half as much fuel as con-ventional units have become available, and themost efficient new heat-pump water heatersuse only one third as much electricity as con-ventional electric resistance units do.33

55

REFRIGERATION Recent innovations in refrig-eration technology have led to new refriger-ators, refrigerator-freezers, and freezers thatuse far less electricity than units now in wideuse. The most energy-efficient refrigerator-freezers on the market in Europe, Japan, andthe United States, for example, require only 1.3to 1.6 kWh per liter of cooled volume annually,compared to 3.5 kWh per liter for the averageunits (450 to 500 liters) now common in theUnited States. Still more efficient units areunder development. (See Table 6.)

The cost of saved energy for the more effi-cient units now on the market is far less thanthe cost of electricity,34 and the initial costs ofmore efficient units may fall as the new tech-nology becomes commonplace. As with newhouses, synergistic energy-saving strategies canhelp reduce first costs. For example, invest-ments to reduce heat losses may be offset bythe reduced costs of lower-capacity motors andcompressors.

The wide use of energy-efficient refrigerationtechnology could significantly affect electricitysupply requirements. For instance, if the ex-isting U.S. stock of refrigerator-freezers werereplaced by the most efficient models availablecommercially in 1982, the energy saved would beequal to the output of about 18 large nuclear or coalpower plants, or the output of 18 GW(e) of base-load electrical generating capacity.

LIGHTING The incandescent bulb has domi-nated residential lighting since it was first in-troduced in 1879. The more efficient fluorescentlighting has not caught on for home usebecause most people prefer the soft yellowlight of the incandescent bulb to the harsherwhite light of fluorescents. As noted earlier,though, new high-efficiency compact fluores-cent lightbulbs are three to five times as effi-cient as incandescents, they last several timesas long, and the newest do not have the draw-backs of the earliest models. (See Figure 6.)Although they cost far more than incandes-cents, these more efficient bulbs are cost-effective in many circumstances.

An indication of the overall possibilities forenergy savings in the residental sector is pro-vided by a comparison of the actual averageenergy use levels for the residential sectors ofSweden and the United States with the levelsthat would be realized in hypothetical superin-sulated, all-electric houses having a full set ofthe most energy-efficient appliances now avail-able commercially. Enjoying more amenitiesthan the average household today, these hypo-thetical households would nevertheless use lessthan one fourth as much energy per capita asthe average household. (See Table 7.)

APPLICATIONS IN DEVELOPING COUN-TRIES Investments in energy efficiency mayactually be more important for developingcountries than for industrialized countries. Thisfinding is contrary to the conventional wisdomthat poor countries cannot afford to make thecapital investments needed to improve the effi-ciency of energy use. Although more energy-efficient end-use devices do tend to be morecostly than conventional units, the extra invest-ment is typically much less than what wouldbe required for an equivalent amount of energysupply expansion. Thus, investments in energyefficiency can lead to a net overall capital sav-ings by society.

For example, compare the cost in Brazil ofsaving 1 kW by investing in efficient compactfluorescent light bulbs with the cost of theextra kW of hydroelectric supply that would berequired if incandescent bulbs were used in-stead. (See Figure 25.) The cost in each case, asa function of the discount rate, is the dis-counted present value of all required invest-ments over a 50-year life cycle.

At a 10-percent discount rate, the cost of ex-panding supplies per kW is about three timesthe cost of saving one kW by investing in moreefficent bulbs. Moreover, the benefit of invest-ing in energy-efficient bulbs actually increasesas the discount rate rises. This savings occursin part because the purchases of efficient lightbulbs are spread over the life of the energysupply facility, so that the present value of

56

Table 6. Energy Performance of Refrigerators and Refrigerator-Freezers

Brand (Origin)

Average, 45 U.S. ModelsHitachi (Japan)Gram (Europe)Kenmore (United States)National (Japan)Laden (Europe)Bosch (Europe)Gram (Europe)Gram (Europe)

Average, 488 U.S. ModelsBosch (Europe)Amana (United States)National (Japan)Amana (United States)Whirlpool (United States)Electrolux (Europe)Amana (United States)Kelvinator (United States)Toshiba (Japan)Amana/KelvinatorPedersenSchlussler

Model

(1983)617AK215

564.86111211

40.830KS2680SR

K395prototype13

(1983)KS3180ZLTSC18E

291(HV/T)ESR14E

TR1120Cprototypeprototype

GR411conceptual0

prototype11

prototype6

Type

1 door1 door1 door1 door1 door1 door1 door1 door1 door

2 door2 door2 door2 door2 door2 door2 door2 door2 door2 door2 door2 door2 door

Capacity(liters)

363169215311207305255395200

518310510290402487315510510411510510368

SpecificElectricity Usea Electricity Use3

(kilowatt hour per (kilowatt hourliter per year)

1.941.361.261.190.990.950.860.800.52

2.461.771.711.651.581.541.511.431.391.311.140.940.68

a. Electricity use values may not be directly comparable, because they are based ontests that may vary from country to country.

b. Refrigerator with automatic defrost but no freezer. Thisat the Physics Laboratoryrefrigerator manufacturer

oer year)

703230270370205290220315104

1275550870480635750475730710540580480252

standardized

prototype was designed and analyzedIll, Technical University of Denmark, in cooperation with theBdr. Gram A/S, Vojens, Denmark, and the compressor

systems manufacturer Danfoss A/S, Norborg,Hoick, "Development ofNo. 4 for the Period SepLyngby, Denmark, 1983)

Denmark (J.S. Norgard, J. Heeboll,and controland J.

Energy-Efficient Electrical Household Appliances: Progress Rep.t. 15 1982 to March 15, 1983," Technical University of Denmark,

c. This requirement is the estimated consumption that would result from combiningcient compressor utilizedof the Amana prototypeAmerican Council for an

in the Kelvinator prototype with the other energy savin[H.S. Geller Energy Efficient Appliances (Washington, D.CEnergy Efficient Economy, 1983)].

the effi-5 features

57

Table 6. Continued

d. Frost-free unit designed for two California utilities (PG&E and SCE) at the Physics Labora-tory III, Technical University of Denmark. The energy-saving features include: integratingthe condenser into the inner side of the outer cabinet, to reduce the condenser temperaturewhile simultaneously eliminating the need for anti-sweat heaters; using separate compres-sors for the refrigeration and freezing compartments; and increasing slightly the amount ofinsulation (P.H. Pedersen, "Reducing Electricity Consumption in American Type CombinedRefrigerator/Freezer," in Procedings of the 37th Annual International Appliance Technical Con-ference (Lafayette, Indiana: Purdue University, Indiana, in press)).

e. Horizontal unit designed by Larry Schlussler at the University of California at Santa Barbara(L. Schlussler, "The Design and Construction of an Energy-Efficient Refrigerator," TheQuantum Institute, University of California at Santa Barbara, June 1978).

these purchases declines rapidly with the dis-count rate. In addition, the total investment forenergy supply expansion must be made beforethe new power plant is completed, and theaccumulation of interest charges during con-struction causes the present value of the sup-ply investment to rise with the discount rate.The difference between these two costs riseswith the increasing discount rate, so that thesocietal benefit of the energy-saving bulbslooks better and better as the discount raterises—a result that holds for a wide range ofenergy-saving investments.35 Surprisingly,then, investments in energy efficiency improve-ment will often make even more sense incapital-short developing countries than in richindustrialized countries.

Investments in energy efficiency may bemore important for developing countriesthan for industrialized countries.

Significant gains are possible from shifting toan energy strategy focussed on end uses.Analysis of opportunities for more efficient

energy use throughout the electrical sector inBrazil has shown, for example, that invest-ments totaling less than $10 billion over theperiod 1985-2000 in more efficient refrigerators,street lighting, lighting in commercial build-ings, motors, and variable speed drives for in-dustrial motors would eliminate the need toconstruct 22 GW(e) of electrical capacity costingsome $44 billion.36

Investments in energy efficiency offer variousbenefits in addition to such direct economicbenefits. For example, in southeastern Brazil,untapped hydroelectric resources are quitelimited. When they are exhausted, much morecostly thermal electric power plants will beneeded. Improving the efficiency of electricity-using devices will delay this costly shift.

These points also apply to bioenergy, forwhich a major concern is the constraint on thebioenergy potential imposed by the low effi-ciency of photosynthesis. At some point, thelimits of land will prevent expanding the use ofbioenergy resources, bioenergy productioncosts will rise, and a shift will be required tomore costly, less secure, or more environmen-tally troublesome resources. However, invest-ments to promote the efficient conversion ofbiomass to gas, electricity, and other high-quality energy forms and the more efficient useof these energy forms (through, say, efficient

58

Table 7. Final Energy Use in the Residential Sector (Watts per capita)

End-Use

Average Household at Present

United States, 1980b Sw

All Electric, Four-PersonHouseholds with the

Most EfficientTechnology Available in

1982-1983*

Space HeatAir ConditioningHot WaterRefrigeratorFreezerStoveLightingOther

89046

2807923624180

n 1972-1982°

900_

1801726263063

United States

60d

65f

43g

25212118'75

Swed«

65e

-110h

817169

41

Total 1,501 1,242 328 266

a. With 100 percent saturation for the indicated appliances, plus dishwasher, clothes washer,and clothes dryer.

b. Total consists of 360 watts of electricity and 1,140 watts of fuel.c. Total consists of 350 watts of electricity and 890 watts of fuel. Fifty percent of the electricity

is for appliances and 50 percent is for heating purposes.d. For an average-sized, detached, single-family house (150 square meters of floorspace);

average U.S. climate (2,600 degree-days); a net heating requirement of 50 kilajoules persquare meter per degree-day.6 (Table 5): a heat pump with a seasonal average coefficient ofperformance (COP) of 2.6 (the highest efficiency for new air-to-air units).

e. For a Faluhus (Table 5) in a Stockholm climate (3,810 degree-days).This house uses a heatexchanger to transfer heat from the exhaust airstream to the incoming fresh air.

f. For the average cooling load in air-conditioned U.S. houses (27 gigajoules per year) and aCOP of 3.3 (the COP on the cooling cycle for the most efficient heat pump available in1982).

g. For 59 liters per capita per day of hot water (at 49° C) (or 910 kilowatt hours per year percapita) and the most efficient (COP of 2.2) heat pump water heater available, 1982.

h. For 1,000 kilowatt hours per year per capita hot water energy use via resistive heat. Am-bient air-to-water heat pumps are not competitive at the low Swedish electricity prices,

i. Savings achieved by replacing incandescent bulbs with compact fluorescent bulbs.

gas stoves and electrical devices) would makeit possible for bioenergy to play a much largerrole in development.

Automobiles around the World. In 1982 auto-

mobiles used about 10 mbd,or one sixth of alloil used worldwide, equal to about threefourths of all oil produced in the Persian Gulfregion that year. Automobiles are therefore

59

Figure 25. Life-Cycle Costs for Both Providing Peaking Hydroelectric Power and Saving ElectricityThrough Use of Compact Fluorescent Bulbs as Alternatives to Incandescents

4000.

3000_

e1

2000_

1000_

GenerationPlus Transmission

Compact FluorescentUsed 5 Hours Per Day

10

Discount Rate (Percentage Per Year)

I

15 20

The life-cycle cost shown is the discounted present value of the peaking electricity produced in ahydroelectric power system in Brazil or saved through the installation of compact fluorescent bulbs, indollars per kilowatt measured at the household. All capital investments over the 50-year estimated lifetimeof the hydroelectric power plant are included.

The cost of hydroelectric power supplies increases with the discount rate because of interest charges ac-cumulated during construction. The cost for the more efficient light bulbs declines with the discount ratebecause replacement bulbs do not have to be purchased until they are needed.

Source: J. Goldemberg and R.H. Williams, "The Economics of Energy Conservation in Developing Countries," in EnergySources: Conservation and Renewables, D. Hafmeister, H. Kelly, and B. Levi, eds., AIP Conference Proceedings,number 135 (New York: American Institute of Physics, 1985): 33-51.

prime candidates for energy-efficiencyimprovements.

Improved automotive fuel economy couldhave a significant impact on oil use in indus-trialized market-oriented countries, which ac-count for four fifths of the world's cars and

more than five sixths of the oil consumed bycars. (See Table 8.) About one fourth of all oilused in these countries fuels automobiles.

The automobile is rapidly becoming a majoroil user in developing countries as well. Thenumber of automobiles in developing countries

60

Table 8. Oil Use by Automobiles in 1982

Region Number of Autos

Industrialized CountriesUnited StatesCanadaOther OECDEastern Europe, USSR

Subtotal

Developing CountriesAfricaLatin AmericaAsia(Asia except China, India)

Subtotal

World

a. In the United States the averageliters per 100 kilometers) in 1982same amount of annual driving i

(millions)

104.510.5

143.420.1

278.5

6.9823.359.928.94

40.25

318.8

Persons per Car

2.22.33.6

18.9

4.1

71.315.7

246.781.2

82.3

14.0

auto had a fuel economy of 16.25 miles

Oil Usea

(million barrelsper day)

4.00.43.70.52

8.6

0.180.610.260.23

1.04

9.6

per gallon (14.5and was driven 9,533 miles (15,340 kilometers). Here thes assumed for autos

fuel economy assumed for Canada is the same as foreconomy for other regions is assumed to be 24 mileskilometers).

in other countriesthe United States.

as well. The averageThe average fuel

per gallon (9.8 liters per 100

grew at an average rate of 7.3 percent per yearbetween 1975 and 1984, compared to 2.6 per-cent in the industrialized market-oriented coun-tries. Moreover, the number of automobiles indeveloping countries is projected to increase150 to 200 percent by the year 2000. (See Table9.) It this increase occurs, the developing coun-tries' share of the world's automobiles wouldincrease from one eighth to about one fifth. Ifthe average fuel economy of these new carswere the same as that outside the UnitedStates today—24 miles per gallon, or 9.8 litersper hundred kilometers—the increasing numberof autos by itself would lead to a 1 percent an-nual increase in total oil use in the developing

market-oriented countries between 1982 andthe year 2000.

Aggressive efforts in both industrialized anddeveloping countries to improve automotivefuel efficiency could forestall a rise in the worldprice of oil. Fortunately, technical opportunitiesfor doing so abound.

In the late 1970s, most experts thought thatonly relatively modest improvements in fueleconomy were feasible. For example, a majorstudy in 1979 by the U.S. National Academy ofSciences (NAS) concluded that the likely tech-nological limit on fuel economy achievable over

61

Table 9. The Automobile Population: Historical Data

198419831982198119801979197819771975

Growth Rates1977-1984(percentper year)

MIT Study3

OECD Study

Averages

Growth Rates1984-2000(percentper year)

MIT StudyOECD StudyAverage

Developing CountriesLatin

Africa America Asia Subtotal

and Projections

Industrialized Countries

OECDHistorical Data (millions)

7.34 25.60 11.57 44.597.45 24.78 10.11 42.346.98 23.35 9.92 40.256.64 21.17 9.18 36.996.33 18.70 8.25 33.285.85 17.21 8.01 31.075.26 16.29 6.65 28.204.92 15.50 6.86 27.284.34 13.45 5.52 23.31

5.9 7.4 7.8 7.3

Projections to the Year '.

123.616 57 27 100

111.8

6.65.2

5.9

277.83270.82258.44257.59253.01247.55240.43232.71216.68

2.6

EasternEurope

22.0021.2020.0818.9617.0915.3114.2712.5810.10

8.3

>000 (millions)

362.8383

372.9

1.72.0

1.9

49.646

47.8

5.24.7

5.0

World

344.50334.55318.77313.54303.38293.93282.90272.57250.09

3.4

536529

532.5

2.82.7

2.8

Fraction inDevelopingCountries

0.1290.1270.1260.1180.1100.1060.1000.1000.093

0.2300.189

0.210

Source: For MIT study: Massachusetts Institute of Technology International Automobile Pro-gram, The Future of the Automobile (CambridgeOECDOutlook

Massachusetts: MIT Press,study: Organisation for Economic Co-operation and Development,for the World Automobile Industry (Paris,1983).

a. For non-Organisation for Economic Co-operation and Development regions thein this study are actually for developing countriesned economies (including China).

excluding

1984). ForLong Term

projectionsChina) and for centrally plan-

62

the next several decades was 37 mpg (6.4 lhk)because of its estimate that the total cost ofowning and operating a car would rise sharplyat higher fuel economy levels.37

Soon after the study was published, how-ever, the more efficient Volkswagen Rabbitdiesel and the Honda City Car were intro-duced. In the last couple years, the number offuel-efficient models available has grown con-siderably. (See Table 10.) The most energy-efficient model available in the United States atthis writing is the four-passenger, gasoline-fueled Sprint, which has an estimated on-the-road fuel economy of 57 mpg (4.1 lhk), some50 percent higher than the NAS estimate of the"technological limit."38

Even these impressive new cars exploit onlya fraction of the presently available technologyfor improving fuel economy. Most of the im-provements to date are from reducing weightthrough use of lightweight materials, reducingrolling resistance through radial tires, reducingaerodynamic drag, and using pre-chamberdiesel engine and lean-burn gasoline engines.

Options for making further gains with pres-ent technology include shifting to direct-injec-tion diesel engines (the kind used in trucks) orspark-ignited, direct-injection diesel engines(which have multifuel capability), introducingthe continuously variable transmission (CVT),introducing the feature of engine-off duringidle and coast, using lightweight materials moreextensively, and further reducing aerodynamicdrag. Several prototype cars indicate what canbe achieved using some of these technologies.(See Table 10.) The most recently introducedprototype, the Toyota AXV, a four- to five-pas-senger lightweight car with a direct-injectiondiesel engine, CVT, and a drag coefficient of0.26, gets 98 mpg (2.4 lhk). (See Figure 4.)

But doesn't good fuel economy imply a slug-gish vehicle that will make highway entry orpassing difficult or dangerous? Not necessarily.The Volvo Light Component Project 2000 vehi-cle gets 65 mpg (3.6 lhk) but requires only 11

seconds to accelerate from 0 to 60 mph (96 km/hour), compared to 17 seconds for the popularautomatic Chevrolet Cavalier, which has a fueleconomy of only 28 mpg (8.4 lhk).39 (See Table10.)

Doesn't high fuel economy mean a tiny,cramped vehicle? Again, not necessarily. Mostof the new highly efficient cars and prototypeslisted carry four or five passengers comfortably.

Aren't fuel-efficient cars unsafe? Here too,good engineering design can improve thestructural strength and safety of even verysmall cars. For example, the lightweight (707kg) Volvo LCP 2000 can withstand 35-mph(56-km/hour) front and side impacts and30-mph (48-km/hour) rear impacts—meetingstricter safety standards than do cars currentlysold in the United States.40 Moreover, ad-vanced technology still under development willmake it possible to build "heavy" fuel-efficientcars, such as the Cummins/NASA Lewis cardesign. (See Table 10.) This 80-mpg (2.9-lhk)car, equipped with a CVT and an advancedmultifuel-capable, direct-injection adiabaticdiesel engine and turbocompounding, wouldweigh 1,360 kg, approximately the averageweight of new cars sold in the United Statestoday.

One problem posed by diesel engines is thatthey emit on average about 100 times theweight of particulates produced by gasolineengines of comparable performance. However,Mercedes Benz and Volkswagen have devel-oped emission-control devices that permitdiesel cars to meet the strict 1986 Californiaparticulate limit of 0.2 grams per mile.41 More-over, the use of diesel fuel is not necessary torealize high fuel economy. Several highly effi-cient cars have gasoline engines, and spark-assisted diesel engines can have the fueleconomy of diesels operating on gasoline oralternative fuels (e.g., ethanol or methanol).(See Table 10.)

Often the technologies added to cars to im-prove fuel economy cost more than the tech-

63

Table 10. Fuel Economy for Passenger Automobiles

Car

Commercial1985 VW Golf, Jetta1986 Honda CRX1985 Nissan Sentra1985 Ford Escort1986 Chevrolet

Suzuki Sprint

Prototype1"VW Auto 2000°Volvo LCP 2000d

Renault EVE + e

Toyota LtwghtCompact'

DesignCummins/NASA

Lewis Car8

dieselgasolinedieseldiesel

gasoline

dieselmultifueldiesel

diesel

multifuel

5.04.34.24.3

4.1

3.63.43.4

2.4

2.9

Fuel Fuel Economy3 Maximumliters per 100 miles per Power

kilometers gallon (kilowatts)

47 3954 4555 4155 39

57 36

666970

3939/66

37

98

81

42

52

Curb Weight(kilograms)

1,029779850945

676

780707855

650

1,364

PassengerCapacity(persons)

5255

4-52-44-5

4-5

5-6

a. For diesel autos, fuel economy as measured in U.S. Environmental Protection Administra-tion (EPA) test procedures; the combined fuel economy consists of a weighted average of 55percent urban driving and 45 percent highway driving. For gasoline autos the fuel economyis a more recent EPA rating in which the EPA test values are modified to conform better toactual road performance: the fuel requirements (in liters per hundred kilometers) determinedin the EPA test are divided by 0.9 (0.78) for urban (highway) driving. For diesels the EPAtest is a good indicator of actual performance.

b. Fuel economies for the European and Japanese prototypes were converted to equivalentEPA test values, using conversion factors recommended by the International Energy Agen-cy. Both the European Urban Cycle Test and the Japanese 10-Mode Test values for fuel re-quirements were converted to the EPA Urban Cycle Test values by dividing by 1.12. TheEuropean 90 kilometers per hour test value is equal to the U.S. highway test value. Fuel re-quirements measured in the Japanese 60 kilometers per hour test were multiplied by 1.15 toobtain the equivalent U.S. highway test value.

c. Three-cylinder, direct-injection (DI), turbocharged (TC) diesel engine; more interior spacethan the Rabbit; engine off during idle and coast; use of plastics and aluminum; drag coeffi-cient of 0.26 (H.W. Grove and C. Voy, "Volkswagen Lightweight Component Project Vehi-cle Auto 2000," SAE Technical Paper No. 850104, presented at the SAE International Con-gress and Exposition, February 25-March 1, 1985).

64

Table 10. Continued

d. The 39-kilowatt car has a 3-cylinder, DI, TC, diesel engine. The 66-kilowatt car has a3-cylinder, heat-insulated, TC, intercooled diesel engine with multifuel capability. Extensiveuse of aluminum, magnesium, and plastics; drag coefficient between 0.25 and 0.28 (R.Mellde, "Volvo LCP 2000 Light Component Project," SAE Technical Paper No. 850570,presented at the SAE International Congress and Exposition, February 25-March 1, 1985).

e. Supercharged, DI diesel engine; engine off during idle and coast; drag coefficient of 0.225(Renault USA press release).

f. DI diesel engine, continuously variable transmission (CVT), wide use of plastics andaluminum, drag coefficient of 0.26 (Toyota press release, October 23, 1985).

g. Four-cylinder, DI, spark-assisted, multifuel capable, adiabatic diesel with turbocompounding;CVT; 1984 model Ford Tempo body (R.R. Sekar, R. Kamo, and J.C. Wood, "AdvancedAdiabatic Diesel Engine for Passenger Cars," SAE Technical Paper No. 840434, presented atthe SAE International Congress and Exposition, February 27-March 2, 1984).

nologies they replace. The costs of addingvarious fuel-saving technologies to a gasoline-powered 30-mpg (7.9-lhk) Volkswagen Rabbitcosting $7,000 have been calculated by vonHippel and Levi.42 (See Figure 26.) Theestimated extra cost for adding all measures,which would raise the fuel economy to 90 mpg(2.6 Ihk), is $1,725. This analysis indicates thatthe total cost of owning and operating a carwould be roughly constant over the entirerange from 30 mpg (7.9 Ihk) to 90 mpg (2.6Ihk), a finding in sharp contrast to the 1979NAS study, which suggested that the costwould rise sharply after a fuel economy of 37mpg (6.4 Ihk) was reached.43

The von Hippel-Levi analysis may actuallyhave overestimated the costs of fuel economyimprovements. One reason is that most of thefuel-saving measures being explored by manu-facturers offer consumer benefits other thanjust fuel savings, so that charging the extracosts exclusively to fuel economy is inappro-priate. Consider the multiple benefits of theongoing trend toward the use of more plasticsin cars. Analysis by John Tumazos of Oppen-heimer & Company indicates that greater useof plastics will reduce the cost of owning andoperating a car in the United States by $150 to$250 per year because of longer product life

from cheaper repairs, corrosion resistance, andlower insurance rates. The CVT would elimi-nate the noticeable jerkiness in shifting withautomatic transmissions.44

The cost estimates of von Hippel and Levimay be too high for another reason too: thecosts of improvements are not simply additive.Some extra costs may be offset by savingsthrough related technological innovations. Forexample, the extensive use of plastics in autobodies can cut fabrication and assembly costs.Such savings opportunies have led developersof the Volvo LCP 2000 to conclude that in massproduction this car would cost the same as to-day's average subcompact.45

How much first costs will increase with fueleconomy is still uncertain, though the argu-ments cited here suggest that the net increasedfirst costs may be modest. High fuel economycan almost certainly be achieved without in-creasing life-cycle costs, as indicated by vonHippel and Levi.

If the goal for the average fuel economy ofnew cars in the mid-1990s were 60 mpg to 65mpg (3.9 to 3.6 Ihk), the average fuel economyfor all cars on the road would be about 48 mpg(4.9 Ihk) in the year 2000. If there were 112

65

Figure 26. The Cost of Driving Versus Automotive Fuel Economy

u

3

ki SB00 32 ^

3

§.2C C 41 "-C W

O (0

^

8J.S

"3! o3 • ; ;

2 u -as j »

£ Q Dei o U Q U > H i«

I I I IMiles Per Gallon (United States)

50I I I

70I

Base Vehicle Purchase Price ($7,000)

Vehicle Fees and Taxes _.

Incremental Price •for Improvements

Insurance

Garaging, Parking, and Tolls

Repairs, Parts, and Maintenance

I

20 10I

7.5I

3.5

O

t r90

-30

40

- 2 0

-10

I2.7

Liters Per 100 Kilometers

c<u

U

The indicated energy performance is based on computer simulations of an automobile having variousfuel economy improvements added in the sequence shown at the top of the graph. The base car is a 1981Volkswagen Rabbit (gasoline version).

The figure shows that the reduced operating costs associated with various fuel economy improvementsare roughly offset by the increased capital costs of these improvements over a wide range of fuel economy.

Source: F. von Hippel and B.G. Levi, "Automotive Fuel Efficiency: The Opportunity and Weakness of Existing Market Incen-tives," Resources and Conservation (1983): 103-124.

66

million autos in developing countries in 2000with this average fuel economy, the corre-sponding fuel use would be 1.4 mbd, only 40percent higher than in 1982, despite a 180 per-cent increase in the number of cars. (See Table9.) With 533 million autos worldwide in 2000getting 48 mpg (4.9 lhk), global fuel use bycars would be about 6.7 mbd, 2.9 mbd lessthan in 1982. This reduction approximatelyequals the 2.8 mbd that FAO estimates isneeded to double agricultural production indeveloping countries by the year 2000.46

It is doubtful, though, that market forcesalone would quickly lead manufacturers to pro-duce and consumers to buy these highly effi-cient cars. For one thing, consumers wouldprobably not enjoy significant direct economicbenefits by buying such a car. If the cost ofowning and operating a car would remainessentially constant over the entire range from30 mpg (7.9 lhk) to 90 mpg (2.6 lhk), thenmarket forces certainly won't promote highfuel economy, because few consumers will paymore initially in return for savings in operatingcosts. Hence, market forces might push con-sumers to buy cars with fuel economies up toabout 30 mpg (7.9 lhk), but not much beyond.Even if the von Hippel-Levi estimates of thecost of fuel economy improvements provehigh, market forces may still provide only aweak incentive to seek high fuel economy,because for highly efficient cars, fuel costsrepresent a tiny fraction of the total cost ofowning and operating a car. (See Figure 26.)

The value of fuel-efficient automobiles tosociety at large is much more clear-cut: theywould reduce oil imports, help keep the worldoil prices from rising, and promote globalsecurity.

In addition, if developing countries requiretheir own car manufacturers to produce highlyefficient cars, they would become more com-petitive in global car markets. The HyundaiExcel, a high-quality, low-cost subcompact carrecently introduced in North American marketsfrom Korea, is proving to be a strong competi-

tor to U.S., Japanese, and Western Europeanmanufacturers. Brazil and Taiwan are alsoevolving into competitive, world-class carmanufacturers. If these countries become notonly automobile exporters but also exporters ofhighly efficient cars, they could promote theuse of such cars in industrialized countries andenjoy the benefits of the lower world oil pricesresulting from oil savings there.

Fuel savings front more efficient carscould be used to double agriculturalproduction in developing countries by theyear 2000.

Steelmaking and Technological Leapfrog-ging. About five sixths of all steel is producedin industrialized countries, where it accountsfor a significant fraction of manufacturingenergy use, for example, one sixth in Swedenand one seventh in the United States.

THE TECHNOLOGICAL POSSIBILITIESEnergy efficiency improvements have been pur-sued throughout the history of steelmaking, asis indicated by changing coking coal require-ments for the blast furnace, the single largestenergy user in the industry. In a blast furnace,hot air is blown through a stack of pieces ofcoke and chunks of iron ore and oxygen istransferred from the iron ore to the carbon,producing a combustible blast furnace and li-quid pig iron, which is removed from the bot-tom. In 1804, about 5.5 tonnes of coking coalwere required to produce a tonne of pig iron inEngland. The requirements in the UnitedStates had been reduced to 1.6 tonnes by 1913and 0.9 tonnes by 1972.

Besides the energy needed for blast furnaceoperation, additional energy is required forother steps in conventional steelmaking: orepreparation and coke manufacture, operation of

67

the steelmaking furnace, casting, rolling, andfinal fabrication. At each stage, there are op-portunities to improve energy efficiency. Thetheoretical minimum amount of energy re-quired to produce a tonne of steel is 7 giga-joules (GJ) from iron ore and 0.7 GJ fromscrap.

At present, steelmaking in Sweden and theUnited States is based on a 50-50 mix of ironore and scrap, so that the theoretical minimumis about 3.9 GJ per tonne of raw steel. The ac-tual energy used to produce raw steel was 27GJ per tonne in the United States in 1980 and16 GJ per tonne in Sweden in 1983. (See Table11.)

The Swedish steel industry has a betterenergy performance record, no doubt partlybecause it is relatively small and must be in-novative to secure its place as a producer ofspecialty steels for export.

Potential energy-efficiency improvements insteel production can be illustrated by compar-ing four alternative technological structures forthe Swedish steel industry. (See Table 11.)"Modern technology" refers to changes basedon plans for the 1980s, which have largelybeen fulfilled. "Maximum heat recovery"represents the full exploitation of presentlycommercial technology for heat recovery, most-ly by use of combustible gases in cogeneration

Table 11. Unit Energy Requirements for Raw Steel Production3

Alternative New Technologies

Electricity*Oil and GasCoal

Total

SwedishAverage,

1976

2.97.6

11.9

SwedishAverage,

1983

3.22.8

10.3

Modern MaximumTech- Energy

nologyb Recovery0

(gigajoules per tonne)

22.3 16.3

1.84.39.0

15.1

1.82.29.0

13.0

Elredd

1.31.39.4

Plasma-Smelt6

4.21.33.3

U.S.Average,

1980

2.07.5

17.5

11.9 8.7 27.0

a. The mix of iron ore and scrap feedstocks is assumed to be 50/50, which is approximately thepresent average for both Sweden and the United States. The theoretical minimum energyrequired to produce a tonne of steel with this mix is 3.9 gigajoules.

b. This energy structure should result from changes planned in the late 1970s by the Swedishsteel industry. Most of these improvements have already been achieved.

c. Same as "Modern Technology," except that the potential for energy recovery with presentlycommercial technology would be more fully exploited.

d. Same as "Modern Technology," except that the blast furnaces are replaced by a processcalled Elred, which is under development by Stora Kopparberg AB [S. Eketorp, et al., TheFuture Steel Plant (Stockholm: National Board for Technical Development, 1980)]

e. Same as "Modern Technology," except that the blast furnaces are replaced by a new pro-cess called Plasmasmelt, which is under development by SKF Steel AB. (S. Eketorp, et al.,The Future Steel Plant).

f. Here electricity is evaluated at 3.6 megajoules per kilowatt hour (i.e., losses in generation,transmission, and distribution are not included).

68

(the combined production of heat and electri-city). The Elred and Plasmasmelt processes areironmaking processes now under developmentin Sweden.

However dramatic the energy-efficiency im-provements offered by Elred and Plasmasmelt,interest in these technologies stems not somuch from these features per se but, rather,from the prospect of overall cost reduction andenvironmental benefits. Specifically, with theseprocesses:

• Powdered ores can be used directly, with-out having to agglomerate the ore intosinter or pellets. (Because lower- and lower-quality ores are being exploited, the re-quired preliminary processing now leavesthe ore concentrated in powdered form.)

• Ordinary steam coal can be used, greatlyreducing the need to process more costlymetallurgical coal into coke.

• Various individual operations can beintegrated.

These new ironmaking processes near com-mercialization are by no means the ultimate inimproving energy use and total productivity inthe steel industry. Direct casting, directsteelmaking, and dry steelmaking all attempt tointegrate separate operations to save capital,labor, and energy.47 If the dry steelmakingtechnique is mastered, it will likely become theindustry norm because with powder metallurgythe finished product is exceptionally uniform inquality.

APPLICATIONS TO DEVELOPING COUN-TRIES Steel demand can be expected to growrapidly in developing countries as they indus-trialize. In expanding their steel industries,should developing countries adopt the maturetechnologies of the industrialized countries? Orshould they instead develop more advancedtechnologies? Conventional wisdom holds that

developing countries cannot afford the risksassociated with new process development andshould instead stick with the tried and true.

Nevertheless, the reasons for pursuing ad-vanced technologies are powerful. In the steelindustry energy-saving innovations are neededto offset new higher energy prices. Yet, the in-dustrialized countries are not providing ad-vanced technologies because their own declin-ing steel demand creates a poor economicclimate for innovation. (See Figures 16 and 17.)

Additionally, the comparative advantages ofusing human, financial, and natural resourcesare often quite different in countries of theSouth than in the North. Many of the indus-trial technologies commercialized in the Northare capital-intensive and labor-saving—charac-teristics not well-suited to the South, wherelabor is cheap and abundant and capital costlyand scarce. Many developing countries are alsoblessed with largely undeveloped and relativelylow-cost hydroelectric resources, whereas mostindustrialized countries must rely on morecostly thermal sources for increased electricalcapacity. Similarly, biomass is a promisingsource of chemical fuels for many developingcountries, requiring decentralized developmentstrategies quite unlike the centralized strategiesthat have been pursued by the countries rich infossil fuels.

For such reasons, developing countriesshould examine the range of advanced tech-nological possibilities and pursue those com-patible with their development goals andresources.48 For some industrializing countrieswith little coal but rich water resources, theelectricity-intensive Plasmasmelt process mightbe an attractive technology. For countries withneither of these resources but with significantnatural gas, direct-reduction ironmaking (inwhich iron ore is converted into sponge iron attemperatures below the melting point, using awide variety of reductants, including naturalgas) may be more appropriate. Or entirely newtechnologies tailored to conditions in the Southmay be preferable.

69

Brazil's experience with charcoal-based steel-making demonstrates the feasibility of "techno-logical leapfrogging," the introduction of acompetitive advanced industrial technology inthe South before it is introduced in the North.In industrialized countries, coke began toreplace charcoal as a reducing agent for iron-making in the mid-18th Century. The shift tocoke led to much larger blast furnaces thanwere possible with charcoal because coke canbetter resist crushing under the load of the fur-nace charge.

Although most of the world's steel industryis now based on coke, 37 percent of Braziliansteel production, 4.9 million tonnes, was basedon charcoal in 1983. This anomaly reflects thescarcity of high-quality coking coal in Brazil.Although charcoal-based steel production iswidely viewed as anachronistic, it producesbetter quality steel because charcoal has fewerimpurities than coke. Brazilian charcoal-basedsteel competes well in world markets becausethe industry is far more advanced than thecharcoal-based industry abandoned long ago byindustrialized countries.

Brazil's demonstration that charcoal-basedsteel produced by blast furnaces processinghundreds of tonnes per day can compete withcoke-based furnaces processing thousands oftonnes per day is an important lesson fordeveloping countries generally. The technologyis labor-intensive, it is well-matched to theresource bases of countries rich in biomass butpoor in fossil fuels, and the scale of its installa-tions often permits increments in productivecapacity more appropriate to the size of localmarkets than giant coke-based facilities.

The major cost item in charcoal-based iron-making is charcoal, which accounts for 65 per-cent of total pig iron production costs in Brazil.Charcoal for steelmaking there is producedmainly with wood from plantations of pine andeucalyptus, for which the planted area ex-ceeded 5 million hectares in 1983.

Technical developments relating to charcoal-

based steel production are advancing rapidly.(See Table 12.) Plantation yield has roughlydoubled over the last decade, and a further50-percent increase is expected. Over the lastdecade, charcoal yields from wood have alsoimproved about 20 percent, and another10-percent increase is expected shortly. At thesame time, charcoal requirements for iron-making have been lowered. Thanks to theseimprovements, the land area required for agiven level of steel production is soon expectedto be just one fifth that required in the 1970s.

Clearly, technological leapfrogging can be ac-complished in developing countries. But newtechnologies must be compatible with localresources and support broad developmentgoals as well as reduce direct costs.

Case Studies: Sweden and theUnited States

To show the potential impact on future energydemand in industrialized countries of (1) struc-tural economic shifts toward less energy-intensive activities and (2) the possibilities formore efficient energy use, we have constructeddetailed demand scenarios for Sweden and theUnited States in the year 2020, details of whichare presented elsewhere.49

These scenarios were constructed "from thebottom up." The starting point was the devel-opment of a detailed picture of present energyconsumption disaggregated by end-uses foreach energy-consuming sector: residences,commercial buildings, transportation, and in-dustry. Then, future activity levels (such aspassenger-kilometers of air travel or liters ofhot water per household) were estimated usingthe most likely population projections andalternative economic growth assumptions,based on historical trends modified accordingto identifiable structural shifts. Finally, the pro-jected activity levels were matched to energyintensities (e.g., kilojoules of kerosene perpassenger-kilometer of air travel or kilojoules ofnatural gas per liter of hot water) characteristic

70

Table 12. Parameters Relating to the AnnualCharcoal in Brazil

Wood Yield on Plantations(tonnes per hectare per year)a

Wood-to-Charcoal Conversion Rate(cubic meters per tonne)

Specific Charcoal Consumption(cubic meters per tonne of pig iron)

Required Area for Plantations(thousand hectares)

Investment Required to Establish Forest(million U.S. dollars)

a. Air dry tonnes (25 percent moisture).

Production of 1 Million

70s Decade

12.5

0.67

3.5

336

201.6

Tonnes of

80s Decade

25

0.80

3.2

128

76.8

Steel Based on

Near Future

37.5

0.87

2.9

71

42.6

of selected energy-efficient technologies con-sidered economical on a life-cycle cost basis.(The new technologies were assumed to be in-troduced at the rates of capital stock turnoverand growth.) Finally, future aggregate energydemand estimates were obtained by multiply-ing the activity levels by their correspondingenergy intensities and summing up activities inall sectors.

Sweden. Although per capita gross domesticproduct in Sweden is comparable to that in theUnited States, final energy use per capita isonly about three fifths as large—averaging 5.3kW per capita in both 1975 and 1980. AlthoughSweden is generally seen as a model energy-conserving society, there are major oppor-tunities for energy savings.

Our analysis shows that with use of energy-efficient end-use technology now commerciallyavailable, Sweden's per capita final energy usecould be reduced to about 3.5 kW with a

50-percent increase in the per capita consump-tion of goods and services and to 4.2 kW witha 100-percent increase in goods and services. Inlooking to the year 2020, it may be more ap-propriate to consider instead advanced end-usetechnologies still under development that areestimated to be cost-effective. In this case, percapita final energy use could be reduced toabout 2.7 kW or 3.3 kW, depending onwhether the per capita consumption of goodsand services increases 50 or 100 percent. (SeeFigure 27 and Tables 13, 14, and 15.)

The United States. U.S. per capita final energyuse averaged 9 kW in 1980, two and one halftimes that of Western Europe and Japan andten times that of developing countries. With 5percent of the world's population, the UnitedStates accounts for one fourth of global energyuse. Thus, U.S. energy consumption has a ma-jor impact on such global energy-related prob-lems as cartel control of the world oil marketand the security difficulties that go with it, the

71

Figure 27.

2500_

2000-

V

^ 1500 _

'eta

jouh

1000-

5 0 0 -

Final Energy Demand in Sweden for Four Levels of Energy Intensity at 100, 150,Percent of the 1975 Level of Consumption of Goods and Services

. . . . . . . 150Actual Level, 1975

. 1 0 0

T1 [HI

Average Technology 1990 Forecast Best Advancedof 1975 by National Technology Technology

Industrial Board

and 200

. 10

_ 8

6 -

53

Per

Cap

ita ]

L. 0

carbon dioxide build-up in the atmosphere,and the great disparity in commercial energydistribution between rich and poor countries.

Per capita final energy consumption could bereduced to 4.3 kW with a 50-percent increasein per capita GNP or to 4.6 kW with a 100-per-cent increase in per capita GNP if cost-effec-tive, energy-efficient technologies are used. (SeeTables 16, 17, 18, and 19.)

Because U.S. energy use is currently so high,the projected reduction in aggregate U.S.energy use between 1980 and 2020 is very sig-nificant globally—about 0.71 to 0.87 TW-yearsper year, which is equal to 25 to 30 percent of

all energy used in developing countries in1980.

Conclusion

Given these results for Sweden and theUnited States and the broad applicability of thetechnologies considered in these analyses, it isreasonable to expect that a 50-percent reductionin per capita final energy use, from 4.9 to 2.5kW, is achievable on average for all industrial-ized countries between 1980 and 2020, at thesame time that standards of living improve sig-nificantly. This projection forms the basis forour scenario for industrialized countries. (SeeFigure 9.) Our analyses for Sweden and the

Table 13. Final Energy Use Scenarios for Swedena

1975

Consumption of Goods and ServicesUp 50 Percent

Present Best AdvancedTechnology Technology

Fuel Electricity Total Fuel Electricity Total Fuel Electricity Total(Petajoules per yearb)

Residential0

Commercial"1

TransportationDomesticInternational

BunkersIndustry

ManufacturingAgriculture,

Forestry, andConstruction

Total

295104

223

47

410

65

1141

6136

11

137

14

259

356140

6111

234 183

47 32

547 293

79 40

1400 622

6539

11

153

21

289

12650

363

194 137

32 25

446 230

61 37

911 467

5437

7

141

13

253

9040

144

25

371

50

720

a. The population is assumed to be 8.3 million in all cases. The per capita gross domestic prod-uct in 1975 was $8,320.

b. One petajoule equals 1015 joules. One petajoule per year is equivalent to 448 barrels of oilper day.

c. Heated residential floor space is assumed to increase from 36 square meters per capita in1975 to 55 (73) square meters per capita with a 50 percent (100 percent) increase in percapita consumption of goods and services.

d. Commercial buildings' floor space is assumed to increase from 12.7 to 16.4 (20.2) squaremeters per capita for a 50 percent (100 percent) increase in per capita consumption of goodsand services.

United States suggest that this scenario wouldbe technically and economically feasible. Andalthough it is not a prediction, it is consistentwith plausible values of future energy prices,income elasticities, and price elasticities asso-ciated with a 50- to 100-percent increase in percapita GDP between 1980 and 2020. (SeeAppendix.)

As noted, many efficiency improvements willbe made automatically as the capital stock

grows and turns over. Accordingly, the recentdownward trend in per capita energy use inmarket-oriented industrialized countries is like-ly to continue. Nevertheless, a 50-percentreduction in per capita energy use will prob-ably not occur unless governments clear awaysome of the obstacles—market imperfectionsand other institutional impediments—to ex-ploiting energy efficiency opportunities.

Probably the greatest uncertainty legarding

73

Table 13. Continued

Residential0

Commercial11

TransportationDomesticInternational

BunkersIndustry

ManufacturingAgriculture,

Forestry, andConstruction

Total

Fuel

7811

210

40

353

54

751

Consumption of

Present BestTechnology

Electricity

7341

13

_

191

29

347

Goods andUp 100 Percent

Total(Petajoules

15154

223

40

544

83

1098

Fuelper yearb)

505

160

29

275

49

569

Services

AdvancedTechnology

Electricity

5838

9

-

175

19

299

Total

10843

169

29

450

68

868

our scenario is how far the centrally plannedindustrialized countries of Eastern Europe andthe Soviet Union will pursue energy efficiency.To date at least, they have adopted few energy-efficiency improvements. However, pressuresto use energy efficiently are mounting. Oil pro-duction in the Soviet Union has peaked andwill probably decline slowly in the future. Theeasy-to-exploit coal resources in the westernpart of the Soviet Union are being exhausted,so that coal production is shifting to remoteSiberian sources. And even before the Cher-nobyl accident, nuclear power was proving tobe more costly than expected and its expansionwas well behind schedule. Moreover, becauseaverage per capita energy use in these coun-tries is about the same as in market-orientedindustrialized countries, and the level ofamenities made possible by energy are prob-ably higher in the West than in the East, itmay be true that what can be achieved in a

few countries such as Sweden and the UnitedStates is feasible in any industrialized country.

Energy Use in DevelopingCountries—A Thought ExperimentThe energy demand situation is completely dif-ferent for the three quarters of the world'spopulation who live in developing countriesand account for one third of world energy use.At present, per capita final energy use indeveloping countries averages about 0.9 kW, ofwhich about 0.4 kW is non-commercial energy,most of it used by the two thirds of the popu-lation who live in rural areas.

Current patterns of energy end use in vari-ous developing countries are not nearly aswell-understood as those in industrializedcountries. The available data are not nearly as

74

Table 14. Final Per Capita Energy Use Scenarios for Swedena

With Consumption ofGoods and Services Up

50 percent

Residential13

Commercial0

TransportationDomesticInternational

Bunkers

IndustryManufacturingAgriculture,

Forestry, andConstruction

Total

1975

1.360.53

0.89

0.18

2.09

0.30

5.34

0.480.19

0.74

0.12

1.70

0.23

3.48

1.42

0.19

2.75

With Consumption ofGoods and Services Up

100 percent

vanced[inologyCapita)

0.340.15

0.55

0.10

Present BestTechnology

0.580.21

0.85

0.15

AdvancedTechnology

0.410.16

0.65

0.11

2.08

0.31

4.19

1.72

0.26

3.31

a. The population is assumed to be 8.3 million in all cases. The per capita gross domestic prod-uct in 1975 was $8,320.

b. Heated residential floor space is assumed to increase from 36 square meters per capita in1975 to 55 (73) square meters per capita with a 50 percent (100 percent) increase in percapita consumption of goods and services.

c. Commercial buildings' floor space is assumed to increase from 12.7 to 16.4 (20.2) squaremeters per capita for a 50 percent (100 percent) increase in per capita consumption of goodsand services.

comprehensive, and it is inherently more dif-ficult to project long-range energy demand forrapidly industrializing countries than formature industrialized countries where mostenergy-intensive activities are growing onlyslowly or not at all.

In the face of such problems, a different ap-proach for estimating long-term energy require-ments in developing countries makes sense.Imagine for argument's sake a developing

country having a standard of living roughlyequal to that in Western Europe, Japan,Australia, and New Zealand in the late 1970s.49

(See Table 20.) In other words, the averagefamily lives in a reasonably well-constructedhouse with running water, plumbing forsewage, an easy-to-use cooking fuel (such asgas), electric lights, and all basic appliances—refrigerator-freezer, a water heater, a clotheswasher, and a television set. There is oneautomobile for every 1.2 households on aver-

75

Table 15. Disaggregated Energy Use Scenarios for Swedish Transportation and Manufacturing

Total

Consumption of Goods and ServicesUp 50 Percent

Consumption of Goods andServices Up 100 Percent

DomesticTransportation

Automobiles13

Trucksc

Rail FreightOther

1979a

14850

450

Present BestTechnology

86434

61

AdvancedTechnology

(Petajoules per

54294

58

Present BestTechnology

year)

10147

765

AdvancedTechnology

6532

765

252 194 145 223 169

1975

ManufacturingPulp and Paper6

Iron and Steelf

Fabrication andFinishing

CementChemicalOther

21211954

403290

Total 547

1487661

363690

446

1226154

323665

371

1488690

4750

122

544

1227276

404794

450

a. Transportation data disaggregated by end-use are not available for 1975.b. The number of person-kilometers of car travel is assumed to increase from the 1979 level of

41 billion by 25 percent (50 percent), for the case of a 50 percent (100 percent) increase inthe per capita consumption of goods and services.

c. The volume of truck freight is assumed to increase from the 1979 level of 14 billion tonne-kilometers by 12.5 percent (25 percent), for the case of a 50 percent (100 percent) increase inthe per capita consumption of goods and services.

d. The volume of rail freight is assumed to increase by 60 percent (120 percent), for the case ofa 50 percent (100 percent) increase in the per capita consumption of goods and services.

e. Production in the paper and pulp sector was limited to a 50 percent increase over the 1975level in both cases because of limited raw material supply.

f. Because of the considerable competition faced by the export-oriented Swedish steel industry,the output of the steel industry is assumed to be only 23 percent (44 percent) higher than in1975, if per capita consumption of goods and services increases 50 percent (100 percent).

76

Table 16. Alternative Scenarios for Total and Per Capita Final Energy Use in the United States

Actual Energy Use in 1980

Residential13

CommercialTransportation0

Industry"1

Total

Residential13

CommercialTransportation0

Industryd

Total

FuelTotal

Electricity(Exajoules per

8.04.3

20.823.7

56.8

2.62.0

-3.0

7.6

Totalyear3)

10.76.3

20.826.7

64.4

Per Capita(kilowatts)

1.50.92.93.7

9.0

Projected Energy Use in 2020

Gross National Product per CapitaUp 50 Percent

Gross National Product per CapitaUp 100 Percent

Total TotalFuel Electricity Total Per Capita Fuel Electricity Total Per Capita

(Exajoules per year3) (kilowatts) (Exajoules per year3) (kilowatts)

3.31.4

12.414.1

31.2

2.01.80.24.7

8.7

5.33.2

12.618.8

39.9

0.60.31.32.0

4.3

3.31.4

14.315.1

34.1

2.01.80.25.1

9.1

5.33.2

14.520.2

43.2

0.60.31.52.2

4.6

a. One exajoule equals 1018 joules. One exajoule per year is equivalent to 448,000 barrels of oilper day.

b. See Table 17.c. See Table 18.d. See Table 19.

age, and air travel per person averages 350kilometers per year. Moreover, an industrialand service infrastructure is in place to sustainthis standard of living.

Now further suppose that the energy-usingtechnologies furnishing energy services have

efficiencies comparable to those of the mostenergy-efficient technologies commerciallyavailable or of advanced technologies thatcould be available in the next decade. (See Table21.) Most of the technologies indicated here—the water heaters, light bulbs, cement plants,paper mills, nitrogen fertilizer plants, etc.—are

77

Table 17. Final Energy Use Scenario for the U.S. Residential and Commercial Sectors8

Residential Sector*

End Use

Space HeatAir Conditioning

CentralRoom (#HH w)

Hot WaterRefrig/FreezerFreezerRangeDryerLightsMiscellaneous

Millions of1980

Electricity

14.3

22.224.826.193.033.744.438.381.6

Fuel

66.7

--

55.3--

37.111.8

-

Units in Use2020

Electricity

43.4

62.419.043.4

119.2119.243.443.4

119.2

Fuel

75.8

--

75.8--

75.875.8

-

Annual Energy1980

Electricity

0.29

0.240.090.330.560.160.120.160.290.34

Fuel

6.02

--

1.64--

0.280.06

--

Use in Exajoules2020

Electricity

0.19

0.370.040.140.250.240.120.140.130.34

Fuel

1.93

--

0.67--

0.400.33

--

Total 2.59 8.00 1.96 3.33

Total

Commercial Sector0

Annual Energy Use in ExajoulesCommercial Floor Area Electricity Fuel Electricity Fuel1980 2020 1980 2020(billion square meters)

4.23 6.37 2.03 4.34 1.80 1.40

a. In accord with U.S. conventions, final energy use is defined as primary energy use less lossesin the generation, transmission, and distribution of electricity. Losses associated with petro-leum refining and transport are counted as final energy use by the industrial and transportsectors, respectively.

b. Because of a 30 percent increase in the population and a decline in the average householdsize from 2.8 persons in 1980 to 2.4 persons in 2020 the number of households increases from82 to 119 million from 1980 to 2020. It is assumed that the amount of heated floor space percapita increases from 50 square meters in 1980 to 58 square meters in 2020. One hundred per-cent saturation is assumed for the ownership of all major appliances except air-conditioning.The assumption that the ownership level is thus independent of gross national product sim-plifies the analysis with little loss of generality, because of the already high level of use ofmajor energy-using appliances. Household amenities that are income-sensitive are of littleconsequence in terms of energy use. For air conditioning the saturation is assumed to be twothirds, because in most parts of the country where it is not yet common, air conditioning isnot needed.

c. The projection for commercial floor space is based on a regression against service sectoremployment for the period 1970-1979, which indicates that for a 65 percent increase in servicesector employment from 1980 to 2020, there would be a 50 percent increase in commercialfloor space.

78

Table 18. Alternative Final Energy Use Scenarios for the U.S. Transportation Sector

Transport Mode

Automobiles and Light TrucksCommercial Air Passenger TransportIntercity Truck FreightRail FreightOther

Total

Fuel

12.31.6b

1.6d

0.6f

4.7

1980Electricity(Exajoules

___--

Fuelper year)

3.5a

3.5C

1.9e

0.7s

4.7

2020Electricity

_--

0.2s

-

20.8 14.3 0.2

a. It is assumed that the number of light vehicles per person aged 16 and over is 0.80 (com-pared to 0.78 in 1980), that the average light vehicle is driven 17,000 kilometers (10,600miles) per year, the same amount as in 1980, and that the average fuel economy of lightvehicles is increased to 3.1 liters per 100 kilometers (75 miles per gallon) by 2020.

b. For 420.6 billion person-kilometers at 3.82 megajoules per person-kilometer.c. It is assumed that revenues grow 1.64 times as fast as GNP, continuing the trend of the

1970s; revenue per people-kilometer remains constant at the 1980 level; and the averageenergy intensity of air travel is reduced in half between 1980 and 2020.

d. For 825 billion tonne-kilometer of intercity truck freight at 2.0 megajoules per tonne-kilometer.e. It is assumed that the volume of truck plus rail freight grows 0.86 times as fast as gross

national product, continuing the trend of the 1970s; the truck-rail mix of freight doesn'tchange; and the energy intensity of truck freight is reduced in half between 1980 and 2020.

f. For 1,345 billion tonne-kilometer at 470 kilajoules per tonne-kilometer.g. It is assumed that, by 2020, one half of rail freight is electrified and that the final energy in-

tensity of electric rail freight is one third of that for diesel rail freight.

Probably the greatest uncertainty is howfar the centrally planned industrializedcountries of Eastern Europe and theSoviet Union will pursue energyefficiency.

on the market today. None requires technolog-ical breakthroughs, and all will probably becost-effective at present energy prices.

To complete the "thought experiment," mul-tiply each activity level in Table 20 by the cor-responding specific energy requirement inTable 21, and then sum them up. (See Table 22and Figure 28.) The results are remarkable. Inthis hypothetical developing country, peopleenjoy a Western European standard of livingwith a total final energy demand of about 1kW per capita, only slightly more than atpresent.

How is this possible? Great improvements inliving standards can be achieved without in-creasing energy use much, in part by adoptingthe more energy-efficient technologies now

79

Table 19. Alternative Final Energy Use Scenarios for the U.S. Industrial Sector

20201972 1980 Gross National Gross National

Product per Product perCapita Up Capita Up50 Percent 100 Percent

Industrial Output (billion 1972 dollars)3

Annual Fuel Use (exajoules)b

Annual Electricity Use (exajoules)Fuel Intensity (megajoules per 1972 dollar)Electricity Intensity (megajoules per 1972

dollar)Final Energy Intensity (megajoules per 1972

dollar)

406.125.2

2.362.1

5.7

67.8

464.723.73.0

51.0

6.4

57.5

83714.1C

4.7c-d

16.8

5.6

22.5

107415. lc

5.1c'd

14.1

4.7

18.8

a. Industrial output is gross product originating in industry (manufacturing, construction, min-ing, and agriculture, forestry, and fisheries). Following the trend established in the 1970s,industrial output is assumed to grow 0.83 times as fast as gross national product.

b. Includes wood.c. Final energy use for the year 2020 is determined by assuming that (1) the outputs of the

basic materials processing (BMP) and mining, agriculture, and construction (MAC) subsec-tors of industry grow only as fast as the population between 1980 and 2020, because of de-mand saturation, and (2) the average energy intensity of each industrial subsector [BMP,MAC, and other manufacturing (OMFG)] is reduced in half between 1980 and 2020, viaenergy efficiency improvements. Assumption (1) implies that from 1980 to 2020 the BMPand MAC subsectors grow 30 percent and the OMFG subsector grows 130 percent and 230percent, associated with a 50 percent and 100 percent increase in per capita gross nationalproduct, respectively.

d. The electrical fraction of final energy use in industry in 2020 is assumed to be 0.25 (up from0.11 in 1980). This increase is based on the assumed persistence of the relationship estab-lished in the 1970s between the electrical fraction of final industrial energy use and time.

becoming available in industrialized countries.In addition, enormous increases in energy effi-ciency arise simply by shifting from traditionalinefficiently used non-commercial fuels (such ascattle dung, crop residue, and wood) tomodern energy carriers (such as electricity, gas,and processed solid fuels). The importance ofshifting to modern carriers is evident from thefact that in Western Europe (where non-com-mercial fuel use is very low) per capita final

energy use (for all purposes except spaceheating) in 1975 averaged 2.3 kW, only twoand a half times what it was in developingcountries, even though the per capita grossdomestic product in Western Europe was tentimes as large as in developing countries.

In the 1-kW scenario, the residential sectorcontributes much less to total energy use thanit does now. The residential sector today ac-

80

Table 20. Activity Levels for a Hypothetical Developing Country in a Warm Climate, withAmenities (Except for Space Heating) Comparable to Those in the WE/JANZa Regionin the 1970s

Activity

Residential13

CookingHot WaterRefrigerationLightsTVClothes Washer

CommercialTransportation

Automobiles

Intercity BusPassenger TrainUrban Mass TransitAir TravelTruck FreightRail FreightWater Freight

ManufacturingRaw SteelCementPrimary AluminumPaper and PaperboardNitrogenous Fertilizer

AgricultureMining, Construction

Activity Level

4 persons per householdTypical cooking level with liquid petroleum gas stovesc

50 liters of hot water per capita per dayd

One 315-liter refrigerator-freezer per householdNew Jersey (USA) level of lighting1 color TV per household, 4 hours per day1/HH, 1 cycle/day5.4 square meters of floor space per capita (WE/JANZ average, 1975)

0.19 autos per capita, 15,000 kilometers per auto per year (WE/JANZaverage, 1975)

1850 person-kilometer per capita per year (WE/JANZ average, 1975)3175 person-kilometer per capita per year (WE/JANZ average, 1975)e

520 person-kilometer per capita per year (WE/JANZ average, 1975)'345 person-kilometer per capita per year (WE/JANZ average, 1975)1495 tonne-kilometer per capita per year (WE/JANZ average, 1975)814 tonne-kilometer per capita per year (WE/JANZ average, 1975)one half OECD Europe average, 1978s

320 kilogram per capita per year (OECD Europe average, 1978)479 kilogram per capita per year (OECD Europe average, 1980)9.7 kilogram per capita per year (OECD Europe average, 1980)106 kilogram per capita per year (OECD Europe average, 1979)26 kilogram of nitrogen per capita per year (OECD Europe average,

1979-1980)WE/JANZ average, 1975WE/JANZ average, 1975

a. WE/JANZ stands for Western Europe, Japan, Australia, New Zealand, and South Africa.b. Activity levels for residences are estimates, owing to poor data for the WE/JANZ region.c. Equivalent in terms of heat delivered to the cooking vessels, to using one 13 kilogram can-

nister of liquid petroleum gas per month for a family of 5.d. For water heated from 20 to 50°C.e. In 1975 the diesel-electric mix was in the ratio 70/30.f. In 1975 the diesel-electric mix was in the ratio 60/40.g. The tonne-km per capita of water freight in 1978 in OECD Europe is assumed to be reduced

in half because of reduced oil use (58% of Western European import tonnage and 29% ofthat of exports were oil in 1977) and emphasis on self reliance.

81

Table 21. Assumed End-Use Technologies and Their Energy Performance Levels for theHypothetical Developing Country Described in the Text

Activity

ResidentialCookingHot WaterRefrigerationLightsTVClothes Washer

CommercialTransportation

AutomobilesIntercity BusPassenger TrainUrban Mass TransitAir TravelTruck FreightRail FreightWater Freight

ManufacturingRaw SteelCementPrimary AluminumPaper and PaperboardNitrogenous Fertilizer

AgricultureMining, Construction

Technology, Performance

70 percent efficient gas stove3

heat pump water heater, coefficient of performance of 2.5b

Electrolux refrigerator-freezer, 475 kilowatt-hours per yearCompact fluorescent bulbs75-watt unit0.2 kilowatt-hours per cyclec

performance of Harnosand Building (all uses except space heating)d

Cummins/NASA Lewis Car at 3.0 liters per 100 kilometers6

three-fourths energy intensity in 1975£

three-fourths energy intensity in 1975s

three-fourths energy intensity in 1975h

one-half U.S. energy intensity in 1980'0.67 megajoules per tonne-kilometer'electric rail at 0.18 megajoules per tonne-kilometerk

60 percent of OECD energy intensity'

average, Plasmasmelt and Elred ProcessesSwedish average, 1983m

Alcoa process"average of 1977 Swedish designs0

ammonia derived from methanep

three-fourths WE/JANZ energy intensity9

three-fourths WE/JANZ energy intensity9

a. Compared to an assumed 50 percent efficiency for existing gas stoves; 70 percent efficientstoves having low NOX emissions, have been developed for the Gas Research Institute inthe United States (K.C. Shukla and J.R. Hurley, "Development of an Efficient, Low NOX

Domestic Gas Range Cook Top," GRI/81/0201, Gas Research Institute, Chicago, 1983).b. The assumed heat pump performance is comparable to that of the most efficient heat pump

water heaters available in the United States in 1982.c. Typical value for U.S. washing machines.d. The Harnosand Building was the most energy-efficient commercial building in Sweden in

1981, at the time it was built. It used 0.13 gigajoules of electricity per square meter of floorarea for all purposes other than space heating.

e. See Table 10.f. A 25 percent reduction in energy intensity is assumed relative to the 1975 average of 0.60

megajoules per person-kilometer for intercity buses, owing to the introduction of adiabaticdiesels with turbocompounding.

82

Table 21. Continued

g. A 25 percent reduction in energy intensity is assumed relative to the 1975 average of0.60 (0.20) megajoules per person-kilometer for diesel (electric) passenger trains, owingto the introduction of adiabatic diesels with turbocompounding (electric motor controltechnology).

h. A 25 percent reduction in energy intensity is assumed relative to the 1975 average of1.13 (0.41) megajoules per person-kilometer for diesel buses (electric mass transit), ow-ing to the introduction of adiabatic diesels with turbocompounding (electric motor con-trol technology).

i. A 50 percent reduction in energy intensity is assumed relative to the 1980 U.S.average value of 3.8 megajoules per person-kilometer for air passenger travel, owingto various improvements.

j . The assumed energy intensity is one-third less than the simple average today inSweden for single unit trucks (1.26 megajoules per tonne-kilometer) and combinationtrucks (0.76 megajoules per tonne-kilometer), to take into account improvements ow-ing to use of adiabatic diesels with turbocompounding.

k. The average energy intensity for electric rail in Sweden, with an average load of 300tonnes and an average load factor of about 40 percent.

1. A 40 percent reduction in fuel intensity is assumed, reflecting such innovations as theadiabatic diesel and turbocompounding.

m. Assuming an energy intensity of 3.56 gigajoules of fuel and 0.40 gigajoules of electrici-ty per tonne, the average for Sweden in 1983.Assuming an energy intensity of 84 gigajoules per tonne of fuel (the U.S. average in1978) and 36 gigajoules of electricity, the requirements for the Alcoa process now be-ing developed.Assuming an energy intensity of 7.3 gigajoules of fuel and 3.2 gigajoules of electricityper tonne, the average for 1977 Swedish designs.Assuming an energy intensity of 44 gigajoules of fuel per tonne of nitrogen in am-monia, the value with steam reforming of natural gas in a new fertilizer plant.Assuming a 25 percent reduction in energy intensity, owing to innovations such asthe use of advanced diesel engines. WE/JANZ stands for Western Europe, Japan,Australia, New Zealand, and South Africa.

n

o.

counts for 22 percent of total energy use inBrazil, 58 percent in India, and 87 percent inTanzania, but it accounts for only 8 percent oftotal energy use in the 1-kW scenario, and inabsolute terms residential energy use is far lessthan at present for each of these countries. (SeeFigure 28.) In contrast, although the industrialsector at present accounts for only 37, 28, and4 percent of total energy use in Brazil, India,and Tanzania, respectively, it accounts for 56

percent of total energy in the 1-kW scenarioand involves more energy use per capita thanin these countries at present. (See Figure 28.) Ina sense, modernizing residential energy usefrees up an enormous amount of energy thatcan be used to support industrial growth.

Our thought experiment is not intended toestablish amenity-level targets for developingcountries or dates to meet them. Rather, our

83

Table 22. Final Energy Use Scenario for a Hypothetical Developing Country in a Warm Climate,with Amenities (Except for Space Heating) Comparable to Those in the WE/JANZ3

Region in the 1970s, but with Currently Best Available or Advanced EnergyUtilization Technologies

Activity

ResidentialCookingHot WaterRefrigerationLightsTVClothes WasherSubtotal

CommercialTransportation

AutomobilesIntercity BusPassenger TrainUrban Mass TransitAir TravelTruck FreightRail FreightWater Freight (Including Bunkers)Subtotal

ManufacturingRaw SteelCementPrimary AluminumPaper and PaperboardNitrogenous FertilizerOther6

Subtotal1

AgricultureMining, ConstructionTotal

Average Rate of Energy UseElectricity Fuel

(Watts per capita)

29.013.53.83.12.1

5122

4.52.0

34

34

10726328

Total

8522

12

286

1111

65121

4

210"

2132

50276

7754262436

21242941

_59839

288

5504559

1049

a. WE/JANZ stands for Western Europe, Japan, Australia, New Zealand, and South Africa.The activity levels are those indicated in Table 20 and the energy intensities are those givenin Table 21.

b. This is the residual difference between the manufacturing total and the sum for the manu-facturing subsectors identified explicitly.

c. It has been estimated that at Sweden's 1975 level of gross domestic product, final energy de-mand in manufacturing would have been 1.0 kilowatt (half the actual value) had advancedtechnology been used [P. Steen, et al., Energy—For What and How Much? (Stockholm: LiberForlag, 1981). In Swedish.]. The value assumed here is 45 percent less, because the averageper capita gross domestic product was 45 percent less for Western Europe than for Sweden in1975. In addition, 22 percent of final manufacturing energy use is assumed to be electricity,the Swedish value for 1975.

84

Figure 28. Final Energy Use Per Capita by Sector and Energy Carrier, For India in 1978, Tanzania in1981, Brazil in 1982, and the 1-Kilowatt Scenario Summarized in Table 22

1200

96 I I 99 | I 88 I 180(65)1 1(90)1 (37)

Industry Residential Transportation I T B 1-kW

Total

For the three sets of columns on the left, the numbers at the top are the sectoral shares in percent of totalfinal energy use. For the four columns on the right (showing total final energy use per capita for India [I],Tanzania [T], Brazil [B], and the 1-kW scenario), the numbers at the top are the total final energy use percapita, and the numbers on the columns are the carrier shares in percent. The numbers in parenthesis onthese columns give the noncommercial energy share of total final energy use in percent.

Source: For India, A.K.N. Reddy, "An End-Use Methodology for Development-Oriented Energy Planning in Developing Coun-tries, With India as a Case Study," PU/CEES 181 (Center for Energy and Environmental Studies, Princeton University,Princeton, New Jersey, 1985). For Tanzania, M.J. Mwandosya and M.L.P. Luhanga, "Energy Use Patterns in Tanzania,"Ambio, vol. 14, no. 4-5, (1985): 237-241. For Brazil, Republica Federativa do Brasil Ministerio das Minas e Energia,Balanco Energetico National (Brasilia, Brasil: 1983).

purpose is to show that it is possible to achievea standard of living in developing countries atany level along a continuum from the presentone up to the level of Western Europe in the

1970s, without greatly increasing average percapita energy use above the present level, ifmodern energy carriers and energy efficiencyare emphasized as economic development pro-

85

ceeds. On the basis of such considerations, weassumed for our global energy scenario anaverage per capita level of final energy use of 1kW for developing countries in 2020. (SeeFigure 9.)

As in the case of our scenario for the indus-trialized world, this scenario is not a predic-tion. But it is consistent with plausible valuesof future energy prices, income elasticities, andprice elasticities associated with major increasesin living standards in developing countries tothe year 2020. (See Appendix.)

Implementing end-use-oriented energy strat-egies in developing countries will not be easy.Large amounts of capital will be required tobring about a shift to modern energy carriersand efficient end-use technologies. But ouranalysis indicates that for a wide range of ac-tivities it would be less costly to provideenergy services using more efficient end-usetechnologies than to provide the same serviceswith conventional less-efficient end-use tech-nologies and increased energy supplies.

Energy can become an instrument ofdevelopment instead of a brake, as it sooften is today.

Most important, end-use-oriented energystrategies are compatible with the achievementof broad development goals. With emphasis onmodern energy carriers and energy-efficientend-use technologies, it becomes possible tosatisfy basic human needs, to expand the in-dustrial infrastructure radically, and allow ma-jor improvements in living standards beyondthe satisfaction of basic needs. Althoughenergy alone will not produce these results, itcan become an instrument of development insteadof a brake, as it so often is today.

V. Changing the Political Economyof Energy

T he long-term energy outlook would befar more hopeful than conventionalprojections suggest if emphasis in

energy planning shifted from expanding supplyto improving energy use, as described inChapter IV. The costs of providing energy ser-vices would be lower, freeing up economicresources. North-South tensions would beeased by the more equitable use of globalresources that would result. In developingcountries, living standards could be increasedconsiderably beyond subsistence in the nextseveral decades without energy supply con-straints. Finally, end-use energy strategieswould permit considerable flexibility in thechoice of energy supplies, making it possible tomitigate the global problems posed by in-creased use of oil, fossil fuels generally, andnuclear power. Yet such end-use energy strat-egies are probably not feasible unless govern-ment intervenes in the market.

How much market intervention would beneeded to implement end-use energy strat-egies? The difference between more conven-tional projections and our scenario should notbe taken as a measure of the effort required toshift course. (See Figures 9 and 11.) Conven-tional projections have not adequately ac-counted for structural shifts in industrializedmarket countries, so that even with no new in-terventions in the market future energy de-mand would grow more slowly than these pro-jections indicate.51 In addition, many newenergy-efficient technologies, especially those

offering benefits beyond energy savings, willbe adopted without market intervention. Thentoo, many technologies even more efficientthan those considered in our analysis will cometo market in the decades ahead. (Most of thetechnologies on which this analysis is basedemerged only in the last few years.) In fact,our end-use-oriented energy scenario for theyear 2020 is consistent with plausible values ofincome and energy price elasticities and energyprices not much higher than those at present,if complemented by public policies adequate toinduce economy-wide energy efficiency im-provements at average rates of the order ofone-half percent per year. (See Appendix.) Thus,although new public policy initiatives areneeded to bring about an energy future like theone described here, the effort required wouldnot be Herculean.

Although this chapter offers general guide-lines for energy-policy-making and includessome specific policy proposals, it stops short ofproviding a blueprint for implementing end-useenergy strategies. Specific policies must betailored to the cultural and political conditionsof different countries. In addition, experiencewith policies to promote innovation in energyend-use is still too limited to show clearly whatthe best approaches are. Moreover, end-use-oriented energy policies should be continuallychanged over time, not only to make theadjustments suggested by experience but alsoto reflect changing needs. The use of energythroughout the economy is determined by a

87

combination of user preferences, incomes, theresourcefulness and technical know-how ofthose who produce and sell energy-using tech-nologies and related services, and other factors.Further, user preferences as well as the techno-logical and institutional opportunities for satis-fying consumer needs are varied and everchanging. In short, patterns of energy use de-pend on decisions by large numbers of con-sumers, each continually confronted withscores of energy-related decisions. This patternof decision-making is far more complex thanthat for the production of energy, which

Patterns of energy use depend ondecisions by large numbers of consumers,each continually confronted with scores ofenergy-related decisions. In contrast, theproduction of energy generally involvesonly a few energy forms and a relativelysmall number of producers.

generally involves only a few energy forms anda relatively small number of producers. None-theless, policy-makers must try promising newapproaches, assess how well different ap-proaches work in different situations, andmake continual adjustments in light ofexperience.

The Role of Markets in End-UseEnergy StrategiesMarket Shortcomings. Despite the economicattractiveness of end-use energy strategies, themarket cannot be relied on to promote thembecause of (1) existing policies that further theexpansion of energy supply (market biases), (2)the reluctance of many consumers to makecost-justified energy-saving investments (marketfriction), and (3) the inherent inability of mar-kets to meet the challenges of poverty, the ex-

ternal social costs of energy production anduse, and the welfare of future generations(market failings).

New public sector interventions are neededpartly because past interventions have biasedthe market in favor of energy supply expan-sion. Historically, either energy producers havebeen subsidized or prices have been held downto benefit consumers.

Government support for energy producershas come in various, sometimes ingenious,forms: tax breaks—including depletion allow-ances, intangible cost write-offs, accelerateddepreciation, and investment tax credits—aswell as a variety of other subsidies, hidden andovert. For example, almost every industrializedcountry has subsidized the development andcommercialization of nuclear power, certainaspects of nuclear plant operations, and eventhe export of nuclear technology.

One problem with energy supply subsidies isthat they keep energy prices below the truelong-run marginal costs of energy supplies andthus encourage economically inefficient con-sumption. A more serious problem is that suchsubsidies make energy supply investmentsespecially attractive to investors, making capitalfor other investments scarcer. Subsidies toenergy producers may not even increaseenergy production. Indeed, unless they aredirected to research activities having no pros-pect of direct commercial payoff, energy supplysubsidies tend to decrease net energy yieldsbecause increased energy production will bemore than offset by increases in the energyopportunity costs of the non-energy inputsinduced by the subsidy.52

Governments have also set energy pricesbelow market-clearing levels for certain classesof consumers—to promote particular patterns ofeconomic growth or, as in the 1970s, tocushion the impacts of market price increaseson certain consumer groups or on the economygenerally. The prices of kerosene or liquidpetroleum gas in many developing countries

are controlled to protect the poor, who needthese fuels for lighting or cooking. In manycountries, energy prices have been keptespecially low for energy-intensive industries,such as those involving the production ofprimary aluminum or chemicals. In recentyears, developing countries have often usedsuch energy price subsidies to promote a pat-tern of development that emphasizes the ex-port of processed basic materials.

It is necessary to control prices in marketswhere suppliers have monopolies, such as elec-tricity markets. However, electricity prices inmany parts of the world are set below econom-ically efficient long-run marginal costs and in-stead are based on average costs (i.e., the elec-tricity price equals the cost of productiondivided by the total electricity sales, with ratesfor different consumer groups adjusted toreflect differences in cost of service). In anumber of developing countries, revenuesdon't even cover total costs.

Overall, efforts to subsidize energy producersand consumers have promoted economic ineffi-ciency and slowed adjustment to the new eco-nomic realities of energy. Very often these in-efficiencies extend far beyond the targetedsectors. In India, for example, where the priceof kerosene was kept low to protect the poor,the subsidy had to be extended to diesel fuelas well in order to keep diesel fuel consumersfrom switching to kerosene (which can be usedin diesel engines) and thus diverting kerosenesupplies from the poor. Extending the subsidyto diesel fuel created excess demand for diesel-based truck freight, heightening the country'sdependency on imported oil.53

Short-run oil price fluctuations notwithstand-ing, the era of cheap energy is over. It endedin 1973, and government subsidies and pricecontrols only put off the day when energy pro-ducers and users adjust to this unpleasantreality.

Market friction also inhibits cost-effective in-vestments in energy-efficiency improvements.

If consumers were economically efficient, theywould seek to minimize life-cycle costs in theirpurchases of energy-using equipment—that is,they would choose for space heating, refrigera-tion, cooking, travel, or another energy servicethose end-use technologies for which the dollarsavings from reduced energy costs represent areturn on the required additional investmentcomparable to the return from alternative in-vestment opportunities. In practice, though,many energy consumers invest in energy-effici-ency improvements only if they can be assuredof much higher returns. To put it more simply,

Efforts to subsidize energy producers andconsumers have promoted economicinefficiency and slowed adjustment to thenew economic realities of energy.

consumers typically must be assured of pay-back periods of two or three years or less54—farshorter than the 10-to 15-year payback periodstypical of investments in new energy supplies.

Why aren't consumers more willing to investin energy efficiency? Information about the po-tential cost savings from investments in energyefficiency is often inadequate or unreliable, ac-quiring the right information and making theappropriate investment is a "hassle," capital tofinance such investments may be scarce, andfuture energy prices may be uncertain. In addi-tion, if the buyers of energy-using devices arenot also the ones who will use the devices,they may be reluctant to make the extra in-vestments in energy efficiency. For example,landlords or builders who purchase appliances,furnaces, and air conditioners rarely take muchinterest in the operating costs that tenants orhomebuyers will be shouldering.

Like the market biases discussed above, theseforms of market friction boost energy demand

89

levels higher than they would be if life-cyclecosts were minimized, and thus they drawcapital to the expansion of energy supply fromother purposes.

Even a properly operating market cannotredress poverty, which involves considerationsof equity, not economic efficiency. The bestthat can be expected from an efficient market isrelatively rapid economic growth. It has oftenbeen argued that while the rich get richer withmore rapid growth, the poor get richer too. Butthe "trickle down" approach to developmenthas not effectively addressed poverty in devel-oping countries. In industrialized countries,high energy prices have induced middle-andupper-income households to become moreenergy efficient, but they also created severeeconomic hardships for people too poor to in-vest in more efficient cars and appliances or inthermally tighter homes.

Decisions made in free markets also do notreflect social costs that are not accounted for inmarket prices. These so-called external socialcosts include: the loss of self-reliance in marketpower and in foreign policy because of overde-pendence on oil imports; the risk of war, evennuclear war, as a consequence of the danger-ous dependence of the industrialized marketeconomies on Middle East oil; the risk ofnuclear weapons proliferation associated withthe availability of weapons-usable materialsfrom nuclear power fuel cycles; and the risk ofclimatic change associated with the atmosphericbuild-up of carbon dioxide from fossil fuelcombustion.

Finally, the market does not look after the in-terests of future generations—witness theprivate sector's lack of interest in basic and ap-plied research, for which the potential paybackextends far beyond business planning horizons.Privately funded research and development(R&D) is concentrated, instead, largely on im-provements in existing products and processes,which promise near-term benefits. Private firmsalso cannot capture the benefits of R&D thatprovides only generic information. In addition,

private firms have little incentive to pursueresearch and development in areas in whichmuch of the potential payoff involves broadsocial benefits—such as research aimed atunderstanding the problems of the poor andhow these problems might be alleviated orresearch on indoor air pollution and the sideeffects of energy-efficiency improvements.

The Importance of the Market in Implement-ing End-Use Strategies. Although market in-tervention is needed to implement end-useenergy strategies, market mechanisms need not

A properly functioning market has a farmore important role in implementing end-use energy strategies than in executingconventional energy strategies.

be abandoned. To the contrary, a properlyfunctioning market has a far more importantrole in implementing end-use energy strategiesthan in executing conventional energystrategies.

The complexities of energy end-use decision-making generally can be dealt with far more ef-fectively by those who know exactly what theyneed and can afford (that is, the buyers) andby those who know what the energy-usingdevices cost to produce (the sellers) than bybureaucracies. Bureaucracies are notoriously in-effectual in keeping track of the needs andpreferences of a multitude of users and inreplacing buyers' and sellers' free-wheeling in-teractions with their own procedures andrules—witness the emergence of black marketswhenever bureaucracies allocate resources. Ac-cordingly, policy should be aimed at improvingthe market's ability to allocate resourcesrelating to energy end-use and interveningmore actively only where the market mecha-nism is inherently weak or incapable of im-

90

plementing social goals. Far more importantthan any specific detail in these initiatives isthat government create a favorable environ-ment for energy-efficiency improvements.

Market Intervention in the Present PoliticalClimate. On the surface, at least, the presentpolitical climate does not seem favorable to in-terventionist policies. The political pendulumhas swung away from government interven-tion, especially in the United States and GreatBritain. But even in a capitalist country such asJapan, with a long tradition of governmentguidance of the economy, and in such socialistcountries as China and Hungary, there hasbeen movement toward laissez faire policies.Democratic socialists throughout WesternEurope are rethinking their policy agendas andseeking alternatives to nationalization, pricecontrols, the expansion of the welfare state,and other approaches. In many parts of theworld, terms such as "privatization," "deregu-lation," "fiscal belt-tightening," and "freemarket economics" are currently fashionable.

To put these observations in perspective, afew aspects of the current political swing needto be noted. Why has it occurred? Certainly,three of the many factors involved stand out.First, there has been widespread disillusion-ment with the poor performance of publiclyowned enterprises. Second, many people areworried about rising government costs and arewary of the measures governments have takento manage them (either raising taxes or print-ing money). Third, fiscal austerity and laissezfaire policies popular in the market-orientedcountries of the North have been imposed onmany developing countries by the InternationalMonetary Fund and international banks as con-ditions of obtaining more credit.

Still, the political pendulum has a way ofswinging back. All through the 1920s, for ex-ample, it was swinging toward laissez fairepolicies and austerity; in 1929, it headed backswiftly. It is impossible to say exactly what im-petus may change policy direction. Perhaps itwill be another oil price shock sometime in the

1990s. Perhaps it will be a combination of fac-tors. In many industrialized countries, chronicunemployment and slow productivity growthpersist. In addition, the manufactured goodsexported from developing countries are becom-ing increasingly competitive in the markets ofindustrialized countries. These factors maycombine with high oil prices to induce eco-nomic stagnation or worse. Under such cir-cumstances, the public sector would have tointervene to improve economic performance,and policies for stimulating improved energyefficiency would then logically come into play.

And, of course, one important reason for in-tervening in the market is to remove the biasesarising from past interventions, therebystrengthening the market mechanism in energydecision-making, a goal that should be widelyshared even in the present political climate.

Promoting More Efficient Use ofCommercial Energy

Given the need to eliminate market biases,reduce market friction, and compensate for in-herent market failings, what are the elementson an end-use energy policy needed to pro-mote more efficient use of commercial energy?In addressing this question, we focussed onthe market-oriented industrialized countries,although much of our analysis is also relevantto the modern sectors of market-orienteddeveloping countries.

Eliminating Energy Supply Subsidies. Despitethe economic distortions they may cause, exist-ing subsidies to energy supply industries areoften considered too entrenched to remove.Consequently, some "political realists" haveoften proposed compensating for their ill ef-fects by extending similar subsidies for in-vestments in energy-efficiency improvementsand solar energy.

Such alternative subsidies, however, alsopose problems. It is generally difficult toremedy one market distortion with another.

91

The compensating measure will tend to offsetonly partially the effects of the original distor-tion, and it may have unintended, undesirableside effects. Moreover, subsidies are notneeded to make competitive a wide range oftechnologies relevant to end-use energystrategies.

A more appropriate policy would be to chal-lenge the conventional political wisdom andeliminate existing subsidies for energy supplieswhile simultaneously implementing policiesthat promote minimal life-cycle costs in marketdecisions relating to energy. In general, energysubsidies should be used only to solve theproblems the market cannot solve—for exam-ple, to alleviate the energy problems of thepoor and promote research and development.

In principle, current efforts to promote fiscalausterity and free-market economics shouldwork to eliminate energy supply subsidies, buteven in this political climate, doing so will notbe easy. For example, the historic tax legisla-tion passed by the U.S. Congress in late 1986preserves major subsidies long enjoyed by theU.S. oil industry, even though the legislationwas designed to eliminate distorting subsidiesand promote fairness.

Rationalizing Energy Prices. Economic effi-ciency would be enhanced and consumers' in-centives to use energy efficiently would be in-creased if energy prices reflected the high costsof new energy supplies—that is, if controlsdesigned to keep energy prices artificially lowwere eliminated and utility rate structures wereredesigned to sensitize consumers to the costsof new supplies.

However, policies enacted to bring energyprices in line with long-run marginal costsmust be carried out in conjunction withpolicies that address the problems that theoriginal pricing policies were designed to solve.For this reason, and because entrenched in-terests may fiercely oppose price rationaliza-tion, it may be necessary to phase in pricereforms slowly.

Improving the Flow of Information. Lack ofinformation about energy-saving opportunitiesimpedes investments in cost-effective energy-efficiency improvements. Government can helpimprove market performance by enhancing theflow of such information to consumers.

One option is making generic informationavailable through "energy extension services,"akin to the agricultural extension services thathave been so successful in facilitating the trans-fer of productivity-enhancing technologies fromi\\? Moratory to the farmers' fields in the in-dustrialized countries.

Energy utilities might also be required to of-fer advice that reflects customers' uniqueneeds, as determined by energy audits. Mostlarge gas and electric utilities in the UnitedStates are required to offer such audits to resi-dential customers. The challenge is to ensurethat customers receive accurate and useful in-formation. Measurements made in instru-mented audits can be far more reliable thanpaper-and-pencil audits, but such informationis also much more costly to obtain. However,these higher costs can often be justified ifaudits are carried out in conjunction with cor-rective actions, as in the case of the "housedoctor" concept developed in the UnitedStates. (See Chapter IV.) The success of anyaudit program depends on the reward system;a law simply requiring utilities to conductaudits without making it profitable for them toprovide good ones probably won't succeed.

Another way to improve the flow of informa-tion would be to require that certain energy-intensive products be labeled to indicate theirenergy performance at the time of sale. Label-ing is useful for products whose energy per-formance is readily measurable, relativelyunambiguous, and easily understood. Candi-dates for such labels are automobiles, varioushousehold appliances, and even whole houses.

Targeting Energy Performance. Many of thesame energy-intensive products eligible formandatory labeling are also candidates for

92

energy performance targets. Such targets helpprotect—among others—those energy userswho are stuck with energy-using equipmentpurchased by builders or landlords. They alsogive manufacturers signals, clearer and steadierthan those of fluctuating prices, that energy ef-ficiency matters to both consumers and societyas a whole.

In some cases, such targets offer major socialbenefits that market forces alone cannot bringabout. For example, compare the private andthe social benefits of improving automotive fuel

Consumers would not be any better orworse off with cars of higher fueleconomy, but society would be muchbetter off.

economy. With today's technology, the fueleconomy of automobiles could be improvedfrom the current global average of 18 mpg (13liters per 100 kilometers) to 80 mpg (3 lhk) orbetter. (See Chapter IV.) Although the total costper mile or kilometer of owning and operatinga car declines rapidly with improved fueleconomy up to about 30 mpg (8 lhk), it re-mains roughly constant at higher fuel econ-omies. (See Figure 26.) The consumer has no in-centive to seek fuel economies better thanabout 30 mpg (8 lhk) because savings owing toreduced fuel requirements are just about offsetby the higher first costs for fuel economy im-provements and because at high fuel economylevels, fuel accounts for such a small fraction ofthe total cost of owning and operating a car. Inother words, consumers would not be anyworse off with cars of higher fuel economy,but they wouldn't be any better off either.

In contrast, society would be much better offif consumers had more efficient cars. If oil im-

ports were lower, the world oil price wouldprobably be lower too. In addition, the worldwould likely be more secure, with a smallerlikelihood of conflict over access to Middle Eastoil. The prospect of generating such enormoussocial benefits without burdening the consumerprovides a powerful rationale for public sectorintervention to promote high fuel economy incars.

Various policy instruments could be used tobring about market shifts to high-efficiencyproducts. Energy performance might be regu-lated to promote high automotive fuel economyor high-efficiency appliances. Taxes might ac-complish much of the same goal; for example,devices performing worse than the average fornew devices of their kind might be taxed at thetime of purchase, with the penalty increasingin proportion to expected extra life-cycle energyrequirements. Utilities might be required tooffer rebates to customers who buy householdappliances that are more efficient than theaverage for new appliances, with the rebatesincreasing proportionally with the expectedenergy savings. To the extent that such rebatesallow a utility to defer investments in morecostly new energy supplies, the consumer, theutility, and all the utility's rate-payers wouldbenefit. A growing number of U.S. utilitiesoffer such rebates for certain appliances.

Stabilizing Consumer Oil Prices. The marketcannot be relied on for regulating the world oilprice. The world oil market has been character-ized by wildly fluctuating prices with alter-nating periods of shortage and glut. (See Figure1.) The onset of a glut following each priceshock gives investors the misleading impres-sion that the oil crisis is over and encouragesincreased oil consumption, setting the stage forstill another shock. Moreover, the periods ofoversupply, as we have seen in the mid-1980s,diminish the sense of urgency needed to in-duce oil importers to reduce their long-runvulnerability to supply disruptions. In general,uncertainties about future oil prices discourageinvestments in energy efficiency.

93

Efforts by major oil importers to makespecific end-use technologies (such as the auto-mobile) more energy-efficient could be success-ful initially in driving oil imports down andthus forestalling a rise in the world oil prices.However, by exerting downward pressure onthe world oil prices, they could stimulate oilconsumption in other areas. (See, for example,Figure 10.)

What is also needed is more stable consumeroil prices. Stability could be achieved with avariable tariff on oil imports or a tax on oilproducts that shields consumers from the vicis-situdes of the world oil market. Ideally, such atariff or tax should keep consumer oil pricesconstant or slowly rising in time, so it wouldhave to be adjusted continually for realchanges in the world oil price and for generalinflation.

Stable or slowly rising consumer oil priceswould make the economic environment for in-vestments in energy efficiency and other alter-natives to imported oil more predictable andthus more favorable. They would help preventsharp upturns in the world oil price and makethe world more secure.

Setting and administering such a tax or tariffwould not be easy. The relative merits of tariffsand oil product taxes, the appropriate level ofconsumer oil prices, issues of equity, the effecton the overall economy, and other issues wouldhave to be taken into account. So would im-portant international issues, discussed below.

Promoting Comprehensive Energy ServiceDelivery. It is relatively straightforward for aconsumer in an industrialized country to pur-chase natural gas, oil, or electricity. Well-estab-lished systems exist for making such trans-actions. The quantities exchanged are easy tomeasure, and both buyer and seller understandthe values of the commodities.

Making energy-saving investments is not sosimple. Marketing energy efficiency requiresdiagnosing the individual consumer's energy

needs and finding ways to meet them in themost cost-effective manner. The consumermust be educated about the need to makeenergy-saving investments—difficult when theexpected savings are often uncertain. Financingmust often be provided for new equipment orrequired contract work. After-purchase per-formance in the field should be monitored toascertain actual savings and the informationthus acquired used to modify energy-savingstrategies.

The energy price shocks of the 1970s promptedthe formation of energy service companies insome countries to help industry cut energycosts. The delivery of such technical assistanceto individual consumers or small businesses,however, remains haphazard at best.

One way to provide the needed assistance isto convert energy utilities into "energy servicecompanies" that market heating, cooling, light-ing, etc., much as they now market electricityor natural gas. Some U.S. electric and gasutilities already provide advice on investmentsin energy efficiency, arrange for contractors tocarry out the necessary work, finance such in-vestments with low- or zero-interest loans, andoffer rebates to consumers who purchaseenergy-efficient appliances or to sellers whopromote these appliances.

Accustomed to accumulating large quantitiesof capital, utilities are well-positioned to investin energy efficiency. They also have the admin-istrative structures for channeling capital toessentially all households and businesses.Through a utility's billing system, for example,customers can pay "life-cycle cost bills" in-stead of fuel or electric bills if they receiveloans from the utility for energy-efficiency in-vestments. Finally, utilities are in a good posi-tion to undertake such difficult tasks as theretrofitting of existing buildings with energy-efficiency improvements. For example, utilitiescould provide a comprehensive "one-stopretrofit service" that includes audit and post-retrofit inspection, coordination of contractorwork, and long-term financing.

94

Of course, energy utilities won't becomeenergy service companies until utility regu-lators develop publicly acceptable ways offinancially rewarding them for facilitating cost-effective energy-efficiency improvements intheir markets. Once they have a clear financialstake, utilities can play the role envisioned forthem by Thomas Edison when he invented theincandescent bulb and proposed that utilitiessell illumination, thereby giving them a finan-cial interest in providing this service in themost cost-effective way.

In some cases, utilities can't or won't createneeded energy conservation programs. For ex-ample, electric or gas utilities may not wish tooffer retrofit services for oil-heated homes.Some may have so much excess capacity thatthey see no need to help customers use energymore efficiently. In such circumstances, govern-ment could stimulate the creation of new inde-pendent companies that would market energy-efficiency improvements by, for example,making loans or grants available to customersof such firms.

Where neither utility nor private sector ef-forts to market comprehensive energy conser-vation are feasible, local governments couldassume the responsibility.

Making Capital Available for Energy Effi-ciency Investments. Capital for energy-effi-ciency investments can be made more readilyavailable through general tax reform andthrough measures that direct capital resourcesto specific applications.

Because so much capital has been directed toenergy supply investments, eliminating sub-sidies for new supplies should free up morecapital for purposes other than energy produc-tion. The question is whether more should bedone to make capital available for investmentsin energy efficiency. For example, shouldgovernment direct capital to the steel industryto stimulate its modernization and thereby im-prove its energy efficiency, or should the mar-ket determine how much capital goes to the

steel industry? Different market-oriented indus-trialized countries have different answers tothese questions. Such questions of general in-dustrial policy are too complex to considerhere.

Without question, though, government inter-vention is needed to make capital more avail-able to individual consumers. Mortgage laws,for example, could be changed so that thefinancial institutions that determine the size ofthe allowable mortgage consider the energy-efficient household's increased ability to repaya loan. Special funds could be established by

Simply making more capital available atmarket interest rates in conjunction withimproved overall delivery of energyservices may open be more importantthan providing capital at subsidized rates.

government to finance qualifying energy-efficiency investments. And energy utilitiescould be encouraged to make capital availableto their customers for such purposes.

Simply making more capital available at mar-ket interest rates in conjunction with improvedoverall delivery of energy services may oftenbe more important than providing capital atsubsidized rates. Capital subsidies should beused only where less costly alternatives cannotwork—such as investments in energy efficiencyfor the poor.

Assisting the Poor. Middle- and upper-incomehouseholds in industrialized countries can ad-just to higher energy prices by insulating theirhomes, buying more fuel-efficient new cars,and the like. But many poor people with littleor no capital resources or credit live in poorly

95

insulated, drafty houses built before the oilcrises and depend for transportation on gas-guzzling cars discarded by the better-off.

The burdens of high energy costs would bereduced for the poor by implementation ofsome general energy-efficiency policies, such asenergy performance standards for appliancesand automobiles. But these measures alone arenot likely to prevent economic hardship. More-over, policies aimed at raising energy prices toreflect the long-run marginal costs of produc-tion will increase the suffering of the poor.

The needs of the poor have often been citedas reasons for keeping energy prices low.However, an economically efficient pricingsystem complemented by special programs tohelp the poor is far preferable to economicallyinefficient systems of price controls to protectthe poor.

In adjusting utility rates to reflect long-runmarginal costs, one way to ease the burden onthe poor would be to keep the overall revenuegeneration rate fixed, while allowing the ratecharged for energy to vary with consump-tion—with rates below the present price forlow levels of consumption to full marginalcosts for high consumption levels. Alternative-ly, rates could be raised across the board tomarginal cost levels, with compensating meas-ures to protect the poor. When rates are raisedto marginal costs, utilities would realize wind-fall profits proportional to the difference be-tween the marginal and average costs. Theseprofits should be taxed away or returned toconsumers in ways not directly related toenergy consumption. In either case, some ofthe excess revenues could be allocated to thepoor.

Similarly, if a new tax is levied on energy,the regressive nature of the tax might becountered by using the resulting revenuescreatively. They might be used to offset someregressive tax (such as the Social Security taxin the United States55) or to help the poordirectly.

Some kind of direct subsidy is generallyneeded to help poor households cope withhigh energy costs. But what kind? In theUnited States, the federal government providesmodest assistance to help the poor pay fuelbills and even more modest assistance to helpwinterize the homes of the poor. But more am-bitious programs are called for. Assistance inpaying fuel bills helps reduce immediate hard-ship, but the greatest need is for investments toreduce fuel bills, especially for space heating.Because many poor people live in inferiorhousing, energy-efficiency improvementsshould be coordinated with programs aimed atmore general improvement of this housing.56

Large one-time investment subsidies to thepoor for retrofitting their homes would beeconomically more efficient than continuallysubsidizing their fuel bills, and they would notbe as demoralizing as continued dependenceon assistance programs.

In the long run, changes in the structure ofemployment induced by implementing anenergy-efficiency strategy may be more impor-tant to low-income people in industrializedcountries than any changes they experience asconsumers. Almost all econometric studiesshow that labor will be substituted for energyas energy prices rise,57 so that energy taxesshould generate employment. In addition, eco-nomic production associated with improvingenergy efficiency and providing products lowin energy intensity tends to involve higheremployment levels per dollar of economic ac-tivity than the production that is replaced.58

Employment associated with end-use efficiencywould also tend to be less specialized, and in-creasing less specialized employment couldreduce structural unemployment—one of themost intractable poverty-related problems con-fronting industrialized countries.

Promoting Research and Development onEnergy End-Uses. Although major improve-ments in energy efficiency can be made withend-use technologies that are already commer-cially available, considerable further improve-ments could be realized with appropriate R&D.

96

The era of energy-demand consciousness isbarely a decade old, and energy-efficient end-use technology is still in its infancy. R&D ex-ploring opportunities for improvements inenergy efficiency will help make energy effi-ciency a design criterion in the process oftechnological innovation, in which the newprocesses and products chosen for commer-cialization tend to be those that offer simul-taneous improvement of several characteristics.

Research is also needed on the social aspectsof energy end uses—to improve techniques forevaluating conservation programs, to provide abetter understanding of how to motivate con-sumers to base energy decisions on life-cyclecosts, and to clarify the energy problems of thepoor and how to meet their needs effectively.

Government should help create an economicclimate conducive to such private sector R&Dand should sponsor promising R&D activitieswhen private efforts fall short. The policiesoutlined here for improving the climate for in-vestments in energy-efficiency improvement—especially eliminating subsidies to the energysupply industries and adjusting prices to reflectmarginal costs—would by themselves tend tocreate an economic climate conducive to thepursuit of energy-saving innovations by theprivate sector. In addition, tax laws might bemodified to give favorable treatment to R&D.

In general, strong government support forR&D is needed when risks are too high, bene-fits too diffuse, and payback too long to justifystrong private support. Specifically, public sup-port is needed for basic research (scientific in-vestigation aimed at expanding scientific under-standing), applied research (generic researchaimed at applying scientific methods to solvingtechnical problems), research on external costs(such as indoor air pollution), and technologyassessment (including monitoring and evalua-ting the field performance of end-usetechnologies).

Despite the importance to end-use energystrategies of research on the technical and

social aspects of energy end uses, suchresearch is only a minor part of government-sponsored R&D. In 1981, only about 6 percentof government R&D funds in the InternationalEnergy Agency countries were committed toenergy conservation—about one tenth theamount spent on nuclear energy. (See Table 23.)By 1983, overall government R&D expendituresand government expenditures on energy con-servation had declined 20 percent while expen-ditures on nuclear energy remained at the 1981level. In the United States, R&D expenditureson energy conservation were sharply reducedunder the Reagan administration while thosefor nuclear energy remained largely un-changed. Adopting end-use energy strategieswill thus require a major restructuring ofenergy R&D.

Creating and Maintaining Energy End-UseData Bases. To set priorities for end-useenergy strategies, policy-makers need to under-stand how the present end-use system worksand how trends in particular end-use activitiesshape aggregate energy demand. Planners' in-formation needs include the energy require-ments and consumer prices of various energyforms, the energy needs and expenditures ofeach economic sector and subsector, patterns ofenergy consumption disaggregated by end-uses, etc. Planners also need demographic andeconomic data. To meet these needs, detailedand highly disaggregated energy demand-supply data bases should be developed andmaintained.

Keeping Score on Conservation Programs. Asnoted, the complexity and ever-changing char-acter of energy-use patterns make it difficult topredict the best comprehensive end-use strat-egy. In implementing such strategies, policy-makers will have to try promising new ap-proaches, monitor and assess the efficacy ofnew programs, and make continual adjust-ments in light of successes and failures.

To this end, fast and reliable "scorekeeping"techniques are needed for measuring the effec-tiveness of particular conservation programs

97

Table 23. Distribution of 1981 andpercent).

ConservationOil and GasCoalNuclear (Nonbreeder)Advanced NuclearNew Energy Sources

(Solar, Wind, Ocean,Biomass, Geothermal)

Other Sources andNew Vectors

Supporting Technologies

Total

Total Expenditure atCurrent Prices (milliondollars)

Sources: For 1981: Organisation for

1983 Government Energy R&D Budgets in

1981

5.93.6

14.127.826.0

12.9

0.79.1

100

8,356

Economic Co-operation and Development,Research and Development and Demonstration in the IEA Countries:

IEA countries (in

1983

5.94.58.0

36.631.2

8.8

0.44.6

100

6,632

"Energy1981 Review of

National Programmes," Paris, 1982. For 1983: Organisation for Economic Co-operationand Development, "Energy Research and Development and Demonstration in the IEACountries: 1983 Review of National Programmes," Paris, 1984.

and understanding the reasons for successesand failures.59 Timely feedback on the energysavings actually realized in ongoing programswould enable planners to improve these pro-grams and would also help protect consumersagainst fraudulent or incompetent energy-service firms.

Energy and Economic DevelopmentIn developing countries, energy systems com-mand such large shares of developmentresources that energy policy cannot be con-sidered apart from development policy general-ly. Energy policy determines not only thekinds and amounts of energy sources devel-

oped but also the allocation of energy suppliesamong sectors and consumer groups. (SeeChapter IV and Energy for Development60.)

Energy Efficiency in the Modern Sectors ofDeveloping Countries. Most opportunities formore efficient energy use in industrializedcountries are also relevant to the modern sec-tors of developing countries. Because energyefficiency investments often lead to reducedoverall capital requirements for providingenergy services, such investments can be evenmore important to developing countries, wherecapital is scarce. (See Figure 25.)

However, energy-efficiency strategies may be

98

more difficult to implement in developingcountries. One reason for this judgment is thatnot all policy instruments that can be used inindustrialized countries are practical for thedeveloping country situation. For example, therebate programs used by some U.S. utilities topromote energy-efficient appliances in house-holds would not work where there are noeffective mechanisms for delivering rebates tothe poor.

In addition, price-reform efforts in develop-ing countries must be accompanied by compen-sating efforts to ease the burden of higherenergy prices on the poor. For example, incountries where the kerosene price is kept lowto protect the poor, the prices of kerosene anddiesel fuels should be increased only in con-junction with programs that give the pooralternatives to kerosene for lighting (for exam-ple, rural electrification) and cooking (say, bio-gas or producer gas).

Developing countries also face special policyissues in introducing new energy-efficienttechnologies. If these technologies must be im-ported, the country must decide whether tospend precious foreign exchange on them. Aproper evaluation of the foreign exchangeissue, however, should cover the entire energysystem—improved end-use technology andenergy-supply technology—because the in-crease in foreign exchange for a more efficientdevice is often more than offset by a reductionin foreign exchange requirements for newenergy supplies.

Ultimately, of course, the domestic manufac-ture of energy-efficient technology would bedesirable in many countries, and the invest-ment and infrastructure requirements forbuilding that capability must be understood.Some countries are already developing suchcapabilities. In India, for example, typical five-passenger cars in use get 21 to 24 mpg (10 to11 lhk), but typical new domestically manufac-tured cars have fuel economies of about 40mpg (6 lhk). There is strong evidence thatBrazilian manufacturers could produce energy-

efficient refrigerators, lighting systems, heatpumps, motors, and motor control devices injust a few years, if there were sufficient de-mand for such products.61

These examples, of course, are not surprisingin view of the fact that exotic technologies aretypically not required to improve energy-usingdevices dramatically. The key to introducingsuch new products is convincing manufacturersthat there would be adequate markets. Thus,utility and government programs to promotethe development of such markets through pro-curement, loan programs, and the like areespecially important.

Planners in developing countries face a newset of challenges in implementing "technolog-ical leapfrogging" strategies. Developing coun-tries should not be content to adopt energy-producing and energy-using technologies fromthe industrialized countries, which often willnot be matched to local human and naturalresources or be compatible with economic ex-pansion in the new era of higher world energyprices. Instead, developing countries shouldcontinually be seeking new technological op-portunities that could lead to improved produc-tivity, consistent with available resources, en-vironmental goals, and security concerns.Adopting such new technologies before theyare proven in the industrialized countries en-tails greater technological risk-taking than mostdeveloping countries are accustomed to.Although any action should be preceeded by acareful evaluation of the potential benefits, pru-dent risk-taking can enhance long-termdevelopment prospects.

Energy and Basic Human Needs. As pointedout here, the structure of the energy demand-and-supply system depends on the approachtaken to alleviating poverty, and an especiallypromising approach is to allocate energy andother resources directly to meeting basichuman needs for nutrition, shelter, sanitation,clothing, health, and education.

The energy policy implications of the basic

99

human needs approach to poverty alleviationwill vary from country to country. Ascertainingthese implications requires detailed data collec-tion and close analyses of how much energyvarious economic activities require, whichenergy resources are available, and which alter-native combinations of energy-supply andenergy end-use technologies could be used tomeet demand.

Some specific actions are widely applicable,however. Chief among them is providingenergy-efficient cooking stoves. Comprehensiveprograms are needed, including research anddevelopment on promising new designs, fieldtesting for actual energy savings and consumeracceptability, and promoting the diffusion anduse of stoves. (See Figure 8). Particular attentionmust be given to introducing stoves in house-holds outside the market economy.

Bringing electricity to all households shouldalso be given high priority. In rural areas,where access to centralized electrical grids isparticularly costly or impractical, decentralizedpower sources based on local resources e.g.,producer-gas generator sets using biomass fuelshould be examined closely. Because tech-nologies for decentralized power generation arenot nearly so well-established as those for cen-tralized power, R&D on energy-efficient small-scale power sources should receive highpriority.

A closely related but more general challengeis to modernize bioenergy resources by devel-oping efficient ways to convert raw biomassinto high-quality energy carriers—such asgases, liquids, processed solids, and elec-tricity—so that scarce biomass resources canprovide far more useful energy than is feasibleat present.62 Here too, major R&D efforts arerequired.

Finding the resources for the R&D will bechallenging. Support from international aidagencies may be needed in many instances.Cooperative R&D programs mounted bygroups of developing countries might some-

times be desirable. Collaboration with indus-trialized countries should also be considered; apromising division of labor might involve car-rying out basic research (e.g., combustion, heattransfer, and the fundamentals of gasification)in industrialized countries and more appliedresearch (aimed at designing and testing newtechnologies) in developing countries.63

Energy and Employment Generation. Inevaluating the employment implications of allalternative technologies and strategies fordevelopment, planners should give particularattention to the problems posed by over-invest-ment in basic materials-processing industriesthat generate few jobs. (See Chapter IV.) To theextent that such over-investment is the resultof subsidies to producers and consumers, ef-forts to bring energy prices into line with long-run marginal energy costs would be especiallyhelpful, as would tax and investment policiesthat do not favor such industries over others.

Assessing the appropriateness of new indus-trial technologies requires particular attentionto the potential for employment generation.Particularly desirable are modern competitivetechnologies that are also employment-intensive—the Brazilian alcohol and charcoalsteelmaking industries, for example.

Energy for Agriculture. Few energy needs indeveloping countries are as crucial as the needsfor expanding agricultural production to feed agrowing population. Although agriculturalenergy needs will expand rapidly, the absoluteamounts of energy required would not neces-sarily be formidable with an end-use energystrategy. (See Chapter IV.)

Efforts to improve energy efficiency in othersectors, especially efforts to save oil in trans-portation, would go a long way toward makingoil import requirements for agriculture afford-able. Efforts to save oil in the market-orientedindustrialized countries would also help byhelping to keep the world oil price from rising.(See Figure 10.)

100

Efforts to modernize bioenergy would alsohelp. Possibilities include using producer gasengines for running pump sets, vehicles, andother farm equipment and using biomass-derived liquid fuels (methanol and ethanol) invehicles.

Political Feasibility. The energy strategydescribed here for developing countries is tech-nically and economically feasible. However,adoption of this strategy would require amarked departure from the status quo in manycountries.

Logically, energy policy should promote sus-tainable development, but not all those inpower in developing countries hark to thislogic. Not the least of required changes wouldbe a major shift in the allocation of capital.Although the proposed effort would be partial-ly "self-funded" in the sense that the pursuitof energy-efficiency improvements would freeeconomic resources for other purposes, somefunds would have to come from taxing theelites.

Will the elite abide such changes? Eliteminorities control virtually every aspect ofpolitical and economic life in developingcountries—in marked contrast to the poor andpolitically weak majority, dispersed in villagesand crowded in metropolitan slums. And whatabout foreign interests? When the elite buy im-ported technologies, export commodities, andborrow large sums of capital, they are oper-ating in an international economic arena domi-nated by the interests of the industrializedcountries. These twin realities shape decision-making in developing countries.

Generally, the elite put great emphasis onbig projects, such as dams and airports, and onpatterns of industrialization and trade thatbring those in power the material comfortsavailable in industrialized market countries-automobiles, air conditioners, airplanes, refrig-erators, and the like. As it happens, these in-terests coincide nicely with the interests of themajor international credit and aid agencies,

most of whose loans or grants cover expensesinvolving foreign currency. Aid money is spentprimarily on technologies and consulting andengineering services from the aid-giving coun-tries. Thus, much of the aid ends up back inthe industrialized countries, increasing demandfor their bulldozers, turbines, irrigation equip-ment, nuclear reactors, etc.

Energy policy in the developing countries hasstressed large centralized energy supplyprojects—especially for electricity generationand transmission—and imported oil. Too oftenthe energy needs of the poor, especially inrural areas, have been neglected. Hydroelectricprojects, nuclear power plants, and refinerieshave attracted far more attention and capitalthan, say, the humble wood stove on which somany depend for cooking and heating and inwhich so much energy is wasted.

Yet, it is not far-fetched to expect the elitesin the developing world to support develop-ment along the lines described here. Poverty isso dire and pervasive that neglecting it or rely-ing on some variation of the "trickle down"approach is to risk social upheaval. There is, ofcourse, a political risk in providing poor peopleaccess to basic economic and educationalresources, but this risk pales beside that of thepolitical instability inherent in festering pover-ty. Thus, it is in the long-term self-interest ofthe elites to identify and support effective pro-grams for eradicating poverty.

Moreover, the outlook for sustained eco-nomic growth in many developing countriesthrough business as usual is not promising.Continued heavy emphasis on exports, whichhas largely shaped the present mix of economicproduction in the developing countries, maycause serious problems in the years ahead.Although the shift from producing com-modities for export to producing processedbasic materials has helped protect theeconomies of developing countries from thevicissitudes of world commodity markets, theoutlook for the export of manufactured goodsto the rich industrialized countries is clouded.

101

The developing countries are already beginningto feel the protectionist measures by industrial-ized countries bent on preventing the furtherloss of jobs in basic industries. Unfortunately,the problem is not a transient one but, rather,an indicator of a long-term trend—a responseto the continuing shift in the industrializedcountries toward less materials-intensiveeconomies.64

Given this long-term prospect of saturatedmarkets for many basic products in the indus-trial countries, shifting the mix of productionto exploit mass markets within developingcountries is a more promising way to realizesustained long-term economic growth and,thus, is in the long-term self-interest of the elite.Of course, no country can have mass marketsfor manufactured goods if its people don'thave income—hence, the necessity of meetingthe basic human needs of the poor, of gener-ating gainful employment, and of improvingthe productivity of the agricultural sector, inwhich so many of the poor labor.

A major challenge for the enlightened elite isto induce their fellow elite to reinvest their cap-ital at home instead of in the industrializedcountries, where it is safe from political over-throw and it might earn more in the shortterm. But, of course, the flight of capital is aproblem whatever energy strategy is chosen.

A major advantage of end-use energy strat-egies in developing countries is that they couldhelp resolve important social problems stronglylinked to energy use. The reduced require-ments for imported oil and the net reduction incapital requirements for the energy systemwould mean smaller expenditures of foreignexchange for such purposes, making it easierfor the indebted developing countries to servicetheir debts. Emphasis on efficient use of bio-mass would also help reduce deforestation,desertification, and soil erosion. Modernizationof bioenergy would not only help reduce oilimports and increase self-reliance but wouldalso generate employment and stimulatedomestic technological development.

International ActionsIf implemented, the national energy policiesdiscussed here would go a long way towardmaking ours a more equitable, more environ-mentally sound, more self-reliant, and morepeaceful world. But providing support fornational programs in developing countries thathelp those countries become more self-reliantor dealing with international or global prob-lems may require international cooperation.

Helping Developing Countries Become MoreSelf-Reliant. Most developing countries havebeen forced to depend on international aid fortheir energy activities because they lack capital,technical resources, and adequate energy in-frastructures. The energy expenditures of themultilateral and bilateral aid agencies totaledabout $14 million between 1972 and 1980. (SeeTable 24.) The bulk of this expenditure camefrom the large development banks, but bilateralaid accounted for about 30 percent of the total.

Energy supply has dominated all energy aidprograms. More than 90 percent of the expen-ditures have gone into large systems forgenerating, transmitting, and distributing elec-tricity (mainly hydroelectric power). Fossil fuelexploration has accounted for about 5 percentand new and renewable energy sources forabout 3 percent. Efforts on the demand side,directed mainly at industrial conservation, haveaccounted for less than 1 percent of the total.

It remains unclear how much energy-relatedaid is needed to implement end-use energystrategies in developing countries. "Front-end"costs of many of the technologies (improvedcooking stoves, biogas plants, producer gasgenerators and engines, biomass-fired gas tur-bine cogeneration systems, more efficient lightbulbs, variable-speed drives for motors, etc.)are relatively modest.

The challenge of implementing end-useenergy strategies is not so much in the amountof capital required as in the institutionalhurdles. Once a decision has been made tobuild a $2 billion nuclear power plant, con-

102

Table 24. Expenditures of Multilateral andof current dollars)

MULTILATERAL AID

World Bank(FY 1972-December 1978)

ConventionalPower Gener-ation (Hydro,

Nuclear,Thermal),

Transmission,Distribution;Power Sector

Studies

5,210

Inter-American DevelopmentBank(FY 1972-FY 1978) 2,596

Asian Development Bank(FY 1972-FY 1978) 1,183

European Development Fund(to May 1978) 141

U.N. Development Programme(to Jan. 1979) 72

U.N. Center for NaturalResources, Energy andTransport(to Jan. 1979)

Subtotal

3

9,205

Bilateral Aid Aj

Fossil FuelsRecovery(includes

Studies andTraining)

305

158

21

—

23

5

512

*encies in the

New andRenewables

(includesGeothermal,Fuelwood)

170

4

0

9

29

4

216

Energy Area (millions

TechnicalAssistance,

EnergyPlanning,

Other

—

—

—

13

5

18

TotalEnergy

Aid

5,686

2,758

1,204

150

137

17

9,952

struction is relatively straightforward, and it isachievable by a fairly small, disciplined team.On the other hand, the number of people in-volved in spending $2 billion on end-use tech-nologies—say, on the construction and distribu-tion of efficient cooking stoves—is likely to belarge and the task far more complicated.

If the international aid community is commit-ted to helping developing countries implementend-use energy strategies, aid programs mustbe restructured. To begin, the aid agencies anddevelopment banks will have to give less pro-ject and more program support. Typically, bigenergy supply efforts, such as the construction

103

Table 24. Continued

BILATERAL AID

French Aid(1976-1979)

Canadian InternationalDevelopment Agency(1978-1979, 1979-1980)

German Aid(1970-present)

Kuwait Fund(FY 1973-FY 1978)

Netherlands—Dutch

ConventionalPower Gener-ation (Hydro,

Nuclear,Thermal),

Transmission,Distribution;Power Sector

Studies

229

88

1,925

437

Development Cooperation(1970-present)

U.K. Overseas Devel.Admin.(1973-present)

U.S. AID(FY 1978-FY 1980)

Grand Total

Percentage in Each Sector

Source: T. Hoffman and B.inger, 1981).

119

146

403

12,71991

Fossil FuelsRecovery(includes

Studies andTraining)

16

0

41

99

71

1

2

7575

New andRenewables

(includesGeothermal,Fuelwood)

30

2

81

1

7

3

96

4373

fohnson, The World Energy Triangle (Cambridge,

TechnicalAssistance,

EnergyPlanning,

Other

5

1

48

—

2

—

46

1211

TotalEnergy

Aid

280

91

2,095

536

198

149

546

14,033100

Massachusetts: Ball-

of hydroelectric facilities, lend themselves to aproject-oriented approach. But end-use energystrategies require broad program supportbecause they involve diverse and often small-scale technologies tailored to regional and localconditions.

One objection often raised about such a shiftis that many developing countries lack thetechnological and management institutions andexpertise to plan and administer such pro-grams. In fact, one reason aid flows to projectsinstead of programs is that aid agencies that

104

support projects need not rely much on localinstitutions and capabilities.

The only way to overcome this weakness isto build institutions and strengthen indigenouscapabilities. Admittedly, this task is time-consuming and often frustrating, but the long-term payoffs would be well worth the effort.By helping developing countries implementend-use energy strategies, aid agencies wouldmake it easier for them to discharge theirdebts; continuing to emphasize increasinglyunaffordable energy supply projects wouldhave the opposite effect. Accordingly, aidagencies would do well to resist the temptationto achieve quick successes with big projectsthat undermine self-reliance.

Aid agencies should direct a portion of theseexpenditures to energy institution-building indeveloping countries or to modernizing existinginstitutions, such as the large utility com-panies. A possible model is the RockefellerFoundation's 20th-century contributions tobuilding medical institutions.

Further, indigenous energy-related technicalcapabilities should be strengthened. Typically,aid has not effectively fostered indigenoustechnical capability, partly because large pro-jects require highly specialized support ser-vices. As a result, project procurement andconsulting arrangements in developing coun-tries are frequently left to foreign countries,which become better and better at providingthese services. Perhaps more important, mostlarge loans and grants managed by interna-tional or bilateral organizations are madespecifically to cover expenses requiring foreigncurrency. Local expenditures are rarely coveredby the loans. Because a typical loan coversabout one-third the overall project cost, mostaid money therefore is spent on consulting andengineering services and on imported machi-nery. These practices, which recycle the aidback to the donor country, do not encourageself-reliant development. Instead indigenoustechnical capability would be strengthened if itis stipulated that: (1) foreign consultants cannot

be recruited unless it is shown that they areessential, (2) when foreign consultants arehired, measures be taken to associate localgroups with the programs, and (3) a significantfraction of the aid must be spent in the reci-pient country so that it helps build localtechnical capability.

Such changes will be difficult to sell political-ly in the donor countries, where support for"foreign aid" often hinges on the purchase ofproducts and services from the donor country.But a technically self-reliant developing countrywill be better able to buy the high-technologyexports of industrialized countries, such ascomputers, lasers, fiber optics, pharmaceu-ticals, and biotechnology products.

Finally, some aid should be used to supportfirst-of-a-kind commercial demonstrations topromote technological leapfrogging in rapidlyindustrializing developing countries instead oftechnological hand-me-downs from industrial-ized countries.

Coping with Global Problems. Global insecuri-ty owing to overdependence on Middle Eastoil, the prospect of global climate changeassociated with excessive use of fossil fuels,and the risk of nuclear weapons proliferationassociated with large-scale use of nuclearpower—all are global problems, the resolutionof which may require collective global actions.

BRINGING STABILITY TO THE WORLD OILMARKET We argue here for oil taxes or tariffsthat would stabilize consumer oil prices, thusstimulating investments in improved energyefficiency. However, a country might be con-cerned that its taxes or tariffs would make itsindustries less competitive in world markets.

This problem could be overcome if oil-importing countries cooperatively levied tariffsor taxes on oil. A cooperative effort would in-hibit such shifts in industrial market shareamong the participating countries, and eachwould benefit from the lower world oil pricesresulting from the oil-saving efforts of others.

105

Even oil exporters would benefit in the longrun from such a cooperative effort because theresulting stable consumer oil prices wouldreduce uncertainties about the future demandfor oil, thus making planning for investmentsin new oil production capacity less risky.

As in the case of a national tax or tariff,there are many questions about how and atwhat levels cooperative taxes or tariffs shouldbe administered. Achieving such an agreementwould be a heroic political accomplishment,but the mutual benefits of achieving some kindof agreement would be enormous. How muchlonger can the world economy endure theroller-coaster trend in the world oil price? (SeeFigure 1.)

The challenge of implementing end-useenergy strategies is not so much in theamount of capital required as in theinstitutional hurdles.

LIMITING THE ATMOSPHERIC CARBONDIOXIDE LEVEL The gradual phasing out ofcoal use proposed here to manage the carbondioxide problem over the next several decadesis perhaps the most formidable politicalchallenge of an end-use energy strategy. Somekind of CO2 control treaty, limiting coal usethrough taxes or other mechanisms, may beneeded. Such a treaty might involve just threecountries—the United States, the USSR, andChina—which together account for nearly 90percent of the world's estimated coal resources.

Limiting CO2 emissions would be far morepolitically difficult than cooperatively adminis-tering an oil tariff or tax because the benefits toindividual countries are less clear. In a globalwarming, some countries would be worse off,

but others might be better off. To complicatematters, it cannot yet be ascertained who thewinners and losers would be.

With emphasis on efficient end-use technol-ogy, however, a CO2 control treaty may not benecessary. Coal is cheap, but using it is not.Coal is a dirty fuel, requiring much morecapital investment than either oil or natural gasto use in environmentally acceptable ways. Inan energy-efficient future, with global energydemand growing hardly at all, coal would be amuch less desirable fuel than it would be ifenergy demand were growing rapidly.

Pursuit for energy efficiency is probably thebest single strategy for coping with the CO2

problem over the next several decades. So do-ing would both limit the CO2 build-up in thisperiod and buy time for developing alternativeenergy sources, such as hydrogen based on theuse of amorphous silicon solar cells.65

CONTROLLING NUCLEAR WEAPONS PRO-LIFERATION The main institutional deterrenttoday to the "horizontal proliferation" ofnuclear weapons is the Non-Proliferation Trea-ty (NPT). Despite its merits, the NPT has failedto attract several countries as signatories, and itdoes not forbid the nuclear fuel-reprocessingand plutonium-recycling activities that couldbring nations dangerously close to nuclearweapons capability.

The countries most closely allied to theUnited States and the Soviet Union appear tohave accepted the two-caste system underlyingthe NPT, in which the world is formally dividedinto countries with weapons and those with-out. But many taking a more independentpath—among them, Argentina, Brazil, India,Iran, Israel, Pakistan, South Africa, SouthKorea, Taiwan, and Yugoslavia—have not for-mally rejected the nuclear weapons option, andmost have failed to ratify the treaty. Most alsohave both a technology base and experience innuclear energy, enabling them to build nuclearweapons if they want to, and thus are con-sidered "threshold" countries.

106

These "threshold" countries are not likely tooppose further proliferation of nuclear weaponsor to support far-reaching international controlsover nuclear power as long as the current non-proliferation system remains so discriminatory.A country that formally renounces nuclearweapons can be seen as accepting a fundamen-tal restriction on its political independence,condemned to neocolonial status with respectto the superpowers. As a result, under thepresent ground rules, there seems no prospectof widening support for the NPT.

Even more serious, the groundwork forreprocessing and recycling plutonium in the in-dustrialized countries has already created de-mand for these technologies in Argentina,Brazil, Pakistan, South Korea, Taiwan, andelsewhere, including several countries thathave ratified the NPT.66 Such a demand is in-spired by a combination of technical considera-tions, desires for prestige, and militarymotives. Their many-sided character is whatmakes the plutonium fuel cycle technologies sotroublesome: nations can move step-by-steptoward a weapons capability without having todecide or announce their ultimate intentions inadvance. This "latent proliferation" under-mines the effectiveness of present nuclearsafeguards.

Latent proliferation is not significantly con-strained by the NPT because the treaty permitsthe development of all types of civilian nuclearpower facilities without discrimination. In fact,under Article IV, parties to the treaty agree tofacilitate the "fullest possible" exchange ofequipment, materials, and information for thepeaceful use of nuclear energy. Even countriesthat have ratified the NPT are moving closer tothe technical capability to produce nuclearweapons. Under extraordinary circumstances,they could withdraw from the treaty onrelatively short notice.

End-use energy strategies can provide thebasis for formulating a more hopeful non-proliferation policy that also makes economicsense—as is indicated by considerations of the

proliferation risks and economic aspects of dif-ferent parts of the nuclear fuel cycle.

As long as plutonium remains in spent fuel,it is protected against diversion to weaponspurposes by the intense radiation from the fuelelements. Plutonium can be separated only byreprocessing the spent fuel, and there are nosound reasons for doing so at present.67 Repro-cessing spent fuel to recover the plutonium forrecycling in today's reactors is uneconomical,and there is no need to reprocess spent fuel forbreeder reactors (which require plutonium asfuel).

Even if nuclear power grows moderatelyrapidly, the world's nuclear power systemswould not be constrained by limited suppliesof uranium for at least 50 years.68 The need forplutonium recycling and breeder reactorswould be far less with the nuclear powerscenario we have described. (See Chapter III.)

Finally, the argument that the disposal ofnuclear power wastes requires reprocessingdoes not stand up to critical analysis. There ap-pear to be no inherent problems with directspent-fuel disposal, though no satisfactorylong-term waste disposal scheme has yet beendeveloped for either spent fuel or reprocessedfuel.69

An important policy option for reducing therisk of proliferation would be to avoid repro-cessing spent fuel. Imposing such a constrainton non-nuclear weapons countries, whetherNPT signatories or not, would be possible onlyif the nuclear weapons countries engaged inthe vertical proliferation of nuclear weaponryaccepted parallel obligations—certainly on theircivilian power programs but perhaps also ontheir weapons programs, because the problemsof horizontal and vertical proliferation ofnuclear weapons are inevitably, intimately, andinextricably linked. A global policy to avoidreprocessing spent fuel from civilian powerprograms (implemented through an interna-tional agreement) might have to be supple-mented by a policy (agreed to by the nuclear

107

weapons countries) not to reprocess spent fuelto produce plutonium for weapons. To securethe threshold countries' support for effectivenon-proliferation conditions, internationalsafeguards should discriminate as little aspossible between nuclear and non-nuclearweapons countries. This symmetry of obliga-tions might even be formalized in a new non-proliferation treaty to replace the presentN P T 7 0

It is possible to entertain such possibilities forlimiting the dangers of proliferation becauseplutonium recycling technologies are not eco-nomic and may never be needed, largelybecause of progress in and future prospects forthe more efficient use of energy.

ConclusionPrecisely what combination of policies isneeded to implement end-use energy strategiesand how best to carry out these policies cannotbe known a priori. Different approaches willhave to be tried and modified in light of ex-perience. But what is clear is that the requiredeffort probably involves only the coordinateduse of familiar policy instruments. The creationof a new world order does not appear to be a

The creation of a new world order doesnot appear to be a precondition forbringing about a global energy futureradically different from what is usuallyprojected.

precondition for bringing about a global energyfuture radically different from what is usuallyprojected.

The energy future we have outlined here isnot the ultimate answer to the world's energyproblems. Eventually, the world will need eco-nomical and environmentally benign renewableenergy sources—the development of which willtake time and ingenuity. But this future wouldgive our children and grandchildren a worldfree of draconian energy-production regimesand, we hope, a world sufficiently prosperousand peaceful to allow them to work out long-term solutions to energy problems. It shouldgive them a little breathing space and someroom to maneuver.

Jose Goldemberg is President of the University of Sao Paulo in Sao Paulo, Brazil. Thomas B. Johanssonis Professor of Energy Systems Analysis, Environmental Studies Program at the University of Lundin Lund, Sweden. Amulya K. N. Reddy is Chairman, Department of Management Studies at theIndian Institute of Science in Bangalore, India. Robert H. Williams is Senior Research Scientist atthe Center for Energy and Environmental Studies at Princeton University, Princeton, New Jersey,USA.

108

Notes

1. The World Bank, World Development Report1984 (Oxford, New York, and Toronto: Ox-ford University Press, 1984).

2. R.H. Williams, "A Low Energy Future forthe United States," Energy, the InternationalJournal, in press.

3. The World Bank, The Energy Transition inDeveloping Countries (Washington, D.C.,1983).

4. H. Tsuchiya, "Energy Efficiency of Refrig-erators in Japan," Research Institute forSystems Technologies, Tokyo, 1982.

5. C.A. Berg, "Energy Conservation in In-dustry: the Present Approach, the FutureOpportunities," report to the President'sCouncil on Environmental Quality, Wash-ington, D.C., 1979; and R.M. Solow,"Technical Change and the Aggregate Pro-duction Function," The Review of Economicsand Statistics, 39(1957): 312-320.

6. S. Baldwin, et al., "Improved Woodburn-ing Cookstoves: Signs of Success," Ambio,vol. 14, no. 4-5(1985): 288-292; and S. Bald-win, Biotnass Stoves: Engineering Design, De-velopment, and Dissemination (Washington,D.C.: World Resources Institute,forthcoming).

7. U.S. Energy Information Administration,"International Energy Outlook 1985, with

Projections to 1995," U.S. Department ofEnergy, Washington, D.C., March 1986.

8. The World Bank, The Energy Transition inDeveloping Countries.

9. B. Bolin, et al., eds., The Greenhouse Effect,Climatic Change, and Ecosystems, SCOPE 29(Chichester, New York, Brisbane, Toronto,and Singapore: John Wiley & Sons, 1986).

10. D. Albright and H.A. Feiveson, "WhyRecycle Plutonium?" Science 235(March 27,1987): 1555-1556; and D. Albright and H.A.Feiveson, "Plutonium Recycle and theProblem of Nuclear Proliferation," PU/CEES 206, Center for Energy and Environ-mental Studies, Princeton University,Princeton, New Jersey, April 1986.

11. Energy Policy Project, A Time to Choose:America's Energy Future (Cambridge: Ball-inger, 1974).

12. W. Haefele, et al., Energy in a FiniteWorld—A Global Systems Analysis (Cam-bridge, Massachusetts: Ballinger, 1981).

13. World Energy Conference. Energy 2000-2020: World Prospects and Regional Stresses,J.R. Frisch, ed. (London: Graham & Trot-man, 1983).

14. J. Ogden and R.K. Williams, "The Pros-pects for Solar Hydrogen in an Energy-

109

Efficient World," Center for Energy andEnvironmental Studies, Princeton Universi-ty, Princeton, New Jersey, forthcoming.

15. D.O. Hall, G.W. Barnard, and P.A. Moss,Biomass for Energy in Developing Countries(Oxford: Pergamon Press, 1982).

16. M. Ross, E.D. Larson, and R.H. Williams,"Energy Demand and Materials Flow in theEconomy," Energy, the International Journal,in press. See also E.D. Larson, M. Ross,and R.H. Williams, "Beyond the Era ofMaterials," Scientific American 254(June1986): 34-41; and R.H. Williams, E.D. Lar-son, and M.H. Ross, "Materials, Affluence,and Industrial Energy Use," The AnnualReview of Energy, 12(1987): 99-144.

17. D.P. Ghai, et al., The Basic Needs Approachto Development (Geneva: InternationalLabour Organization, 1977).

18. P. Streeten, et al., First Things First: MeetingBasic Needs in Developing Countries (NewYork: Oxford University Press, 1981).

19. M.Q. Quibria, "An Analytical Defense ofBasic Needs: The Optimal Savings Perspec-tive," World Development 10(1982): 285-291.

20. Government of Karnataka, Report of theWorking Group Constituted for Advance Plan-ning for Utilisation of Power in Karnataka(Bangalore, India: Government Press, 1982).

21. H.S. Geller, "Ethanol Fuel from SugarCane in Brazil," Annual Review of Energy,10(1985): 135-164.

22. Ibid.

23. United Nations Food and AgricultureOrganization, Agriculture: Toward 2000(Rome: 1981).

24. A.K.N. Reddy, "The Energy and EconomicImplications of Agricultural Technologies:An Approach Based on the Technical Op-

tions for the Operations of Crop Produc-tion," PU/CEES 182, Center for Energy andEnvironmental Studies, Princeton Universi-ty, Princeton, New Jersey, 1985.

25. United Nations Food and AgricultureOrganization, Agriculture Toward 2000.

26. R.H. Williams, G.S. Dutt, and H.S. Geller,"Future Energy Savings in US Housing,"Annual Review of Energy, 8(1983): 269-332.

27. Ibid.

28. G.S. Dutt, "House Doctor Visits-Optimizing Energy Conservation withoutSide Effects," in Proceedings of the Interna-tional Energy Agency's Conference on NewEnergy Conservation Technologies and TheirCommercialization (West Berlin: SpringerVerlag, 1981).

29. G.S. Dutt, et al., "The Modular Retrofit Ex-periment: Exploring the House Doctor Con-cept," PU/CEES 130, Center for Energyand Environmental Studies, PrincetonUniversity, Princeton, New Jersey, 1982.

30. Williams, Dutt, and Geller,Savings in US Housing."

'Future Energy

31. F.W. Sinden, "A Two-Thirds Reduction inthe Space Heating Requirements of a TwinRivers Townhouse," Energy and Buildings,vol. 1, no. 3(1978): 243-260.

32. Williams, Dutt, and Geller,Savings in US Housing."

'Future Energy

33. Ibid.; and H.S. Geller, Energy Efficient Ap-pliances (Washington, D.C.: AmericanCouncil for an Energy Efficient Economy,1983).

34. Williams, Dutt, and Geller, "Future EnergySavings in US Housing"; and Geller,Energy Efficient Appliances.

110

35. H.S. Geller, The Potential for Electricity Con-servation in Brazil (Sao Paulo, Brazil: Com-panhia Energetica de Sao Paulo, 1985).

36. Ibid.

37. Demand and Conservation Panel to theCommittee on Nuclear and AlternativeEnergy Systems of the National ResearchCouncil, Alternative Energy Demand Futuresto 2010 (Washington, D.C.: National Aca-demy of Sciences, 1979).

38. Ibid.

39. D. Bleviss, Preparing for the 1990s: The WorldAutomotive Industry and Prospects for FutureFuel Economy Innovation in Light Vehicles(Washington, D.C.: Federation of AmericanScientists, January 1987).

40. Ibid.

41. Ibid.

42. F. von Hippel and B.G. Levi, "AutomotiveFuel Efficiency: The Opportunity andWeakness of Existing Market Incentives,Resources and Conservation 10(1983): 103-124.

43. Demand and Conservation Panel, Alter-native Energy Demand Futures to 2010.

44. Blevis, Preparing for the 1990s.

45. Ibid.

46. United Nations Food and AgricultureOrganization, Agriculture: Toward 2000.

47. G.J. Hane, et al., "A Preliminary Overviewof Innovative Industrial Materials Pro-cesses," prepared for the U.S. Departmentof Energy by the Pacific Northwest Labora-tory, September 1983.

48. T.B. Johansson, "Energy End-Use in Indus-try: The Case of Steel," paper presented at

the OLADE/UNDP-DTCD Workshop on Ra-tional Use of Energy, Companhia Energeticade Sao Paulo (CESP), Sao Paulo, Brazil,November 24-30, 1985.

49. P. Steen, et al., Energy—For What and HowMuch? (Stockholm: Liber Forlag, 1981). InSwedish. Summarized in T.B. Johansson, etal., "Sweden Beyond Oil—The EfficientUse of Energy," Science vol. 219(1983):355-361; R.H. Williams, "A Low EnergyFuture for the United States," Energy, theInternational Journal, and PU/CEES No. 186(revised), Center for Energy and Environ-mental Studies, Princeton University,Princeton, New Jersey, 1?87; J. Goldemberget al., Energy for a Sustainable World (WileyEastern: New Delhi, in press).

50. Space heating, which is not needed in mostdeveloping countries, is not included in theenergy budget for this hypothetical country.

51. Ross, Larson, and Williams, "Energy De-mand and Materials Flow in the Economy."

52. W.J. Baumol and E.N. Wolff, "Subsidies toNew Energy Sources: Do They Add toEnergy Stocks?" journal of Political Economy,vol. 89, no. 5(1981): 891-913.

53. A.K.N. Reddy, "A Strategy for ResolvingIndia's Oil Crisis," Current Science 50(1981):50-53.

54. A.K. Meier and J. Whittier, "ConsumerDiscount Rates Implied by Consumer Pur-chases of Energy-Efficient Refrigerators,"Energy, The International Journal, vol. 8, no.12(1983): 957-962; and J.E. McMahon andM.D. Levine, "Cost/Efficiency Tradeoffs inthe Residential Appliance Marketplace," inWhat Works: Documenting Energy Conservationin Buildings, J. Harris and C. Blumstein,eds., proceedings of the Second SummerStudy on Energy Efficient Buildings (Wash-ington, D.C.: American Council for anEnergy Efficient Economy, 1983), 526-527.

I l l

55. See M.H. Ross and R.H. Williams, OurEnergy: Regaining Control (New York:McGraw-Hill, 1981), Chapter 15.

56. Williams, Dutt, and Geller, "Future EnergySavings in US Housing."

57. E.R. Berndt, "Aggregate Energy, Efficiency,and Productivity Measurement," AnnualReview of Energy 3(1978): 225-273.

58. C.W. Bullard, "Energy and EmploymentImpacts of Policy Alternatives," in EnergyAnalysis, A New Public Policy Tool, MarthaW. Guilliland, ed. (Boulder, Colo.: West-view Press, 1978).

59. M.F. Fels, ed., "Measuring Energy Savings:The Scorekeeping Approach," Energy andBuildings, special issue, vol. 9, no. 1-2February (1986).

60. J. Goldemberg, et al., "Basic Needs andMuch More with 1 kW per Capita," Ambio,vol. 14(1985): no. 4-5 190-200. This articleis based on a more extensive analysis bythe same authors, Energy for Development(Washington, D.C.: World Resources In-stitute, 1987).

61. Geller, The Potential for Electricity Conserva-tion in Brazil.

62. R.H. Williams, "Potential Roles for Bio-energy in an Energy-Efficient World,"Ambio, vol. 14, nos. 4-5(1985): 201-209;A.S. Miller, I.M. Mintzer, and S.H. Hoag-land, Growing Power: Bioenergy for Develop-ment and Industry (Washington, D.C.: WorldResources Institute, 1986); and Reddy, "AStrategy for Resolving India's Oil Crisis."

63. E.D. Larson, "Producer Gas, EconomicDevelopment, and the Role of Research,"PU/CEES 187, Center for Energy and En-vironmental Studies, Princeton University,Princeton, New Jersey, April 1985.

64. Ross, Larson, and Williams, "Energy De-mand and Materials Flow in the Economy."

65. Ogden and Williams, "The Prospects forSolar Hydrogen in an Energy-EfficientWorld."

66. Albright and Feiveson, "Plutonium Recycleand the Problem of Nuclear Proliferation."

67. Ibid.

68. H.A. Feiveson, F. von Hippel, and R.H.Williams, "Fission Power: An EvolutionaryStrategy," Science 23 (January 26, 1979):330-337.

69. H. Krugmann and F. von Hippel, "Radio-active Waste: The Problem of Plutonium,"Science 210 (October 17, 1981), 319-321; andAlbright and Feiveson, "Plutonium Recycleand the Problem of Nuclear Proliferation."

70. H.A. Feiveson and J. Goldemberg, "Denu-clearization," Economic and Political Weekly,vol. 15, no. 37 (1980): 1546-1548.

112

Appendix

A Top-Down Representation of OurBottom-Up Global Energy DemandScenario

The "bottom-up" or end-use construction ofthe global energy demand scenario presentedin Chapter IV and summarized in Figure 9 willbe unfamiliar to those more accustomed to"top-down" model representations of theenergy future.

To express our global energy demandscenario in terms more familiar to most energymodelers, we have constructed a simple modelrelating commercial final energy demand percapita (FE/P) to gross domestic product percapita (GDP/P), the average price of finalenergy (Pe), a rate of energy efficiency im-provement (c) that is not price-induced, an in-come elasticity (a), and a long-run final energyprice elasticity (-b):

FE/P (t) = A x [GDP/P (t)]a x [Pe (t)]'b /(I + c)t,

where "A" is a constant. This is the aggregateenergy demand equation underlying the Insti-tute for Energy Analysis/Oak Ridge AssociatedUniversities (IEA/ORAU) global energy-econ-omy model.1 Here the equation is appliedseparately to industrialized and developingcountries, relating FE/P, GDP/P, and P e valuesin 2020 to those in 1972 for illustrative valuesof the parameters (a, b, and c).

Figures Al and A2 show per capita GDP andenergy-price parameters consistent with this

study's energy demand scenarios for industrial-ized countries and for developing countries,respectively, for alternative assumptions aboutincome and price elasticities and the non-price-induced energy-efficiency improvement rate.

The year 1972 is chosen as the base year forthis modeling exercise because it is the lastyear before the first oil price shock, sopresumably the economic system was then inequilibrium with the existing energy prices(unlike the situation in 1980, say). For this baseyear, the values of FE/P were 4.7 kW and 0.38kW, compared to the 2020 scenario values of2.5 kW and 1.0 kW for industrialized anddeveloping countries, respectively. (Note: thecommercial energy use values needed for thisanalysis account for only about half the totalfinal energy use in developing countries atpresent.)

For the income elasticity a value of 0.8 wasselected to capture the effects of the ongoingshift to less energy-intensive economic activityin industrialized countries. The value of unityassumed in many modeling efforts is includedfor comparison.

For developing countries, income elasticitiesof 1.4 and 1.1 are assumed for these displays.The value of 1.4 was used for developing coun-tries in the 1983 IEA/ORAU study2 and may beroughly characteristic of the historical situationin developing countries. However, as develop-ing countries modernize in the decades ahead,

113

the income elasticity can be expected todecline. The two assumed values may span therange of uncertainty for the income elasticity indeveloping countries for the period of interesthere.

As for the long-run price elasticity, Nordhaushas reviewed various studies and has con-cluded that the range of plausible values isfrom -0.66 to -1.15,3 with -0.8 a "best-guess"value based on a judgmental weighting ofvalues from various studies.4 (These elasticitiesappear high but they are not. Long-run priceelasticities are much larger than short-runelasticities. Likewise, final demand elasticitiesare greater than secondary demand elasticities,which in turn are greater than primary demandelasticities.5) The illustrative values chosen here(-0.7 and -1.0) span most of this range.

The assumed non-price induced energy-effi-ciency improvement rates are 1.0 and 0.5 per-cent per year. The higher value is the oneassumed for the "CO2-benign" global energyscenarios developed in a 1983 MassachusettsInstitute of Technology Energy Laboratorystudy;6 the lower value approximately reflectsthe contribution from such energy-efficiencyimprovements in a 1984 IEA/ORAU analysis.7

The higher energy-efficiency improvement rateis coupled with the lower price elasticity andthe lower rate with the higher price elasticity toreflect the tendency of non-price-inducedenergy-efficiency improvement policies todiminish the efficacy of prices in curbingenergy demand.8

Although we did not make explicit assump-tions about energy prices in the construction ofour global energy demand scenario (most end-use technologies underlying our analysis wouldbe economic at or near present prices on a life-cycle cost basis, with future costs discounted atmarket interest rates), there may well be con-tinuing final energy price increases to reflectrising marginal production costs, the expectedcontinuing shift to electricity, and the levy ofsome energy taxes to take externalities into ac-count. Prices have already risen substantially

above 1972 values: in West Germany andFrance, average final energy prices in 1980were 1.5 and 1.6 times the 1972 values in realterms, respectively,9 and in the United States,the average price in 1981 was 2.3 times the1972 price.10 Looking to the future, the IIASAstudy11 projects that by 2030, final energyprices will be 3 times the 1972 value in allregions except the WE/JANZ region (WesternEurope, Japan, Australia, and New Zealand),for which a 2.4-fold increase is projected in-stead. A 1983 projection by the U.S. Depart-ment of Energy is for much larger (3.6-fold to5.7-fold) average final energy price increasesfor the United States between 1972 and 2010,associated with an 11- to 17-percent reductionin final energy use per capita in this period.12

It is reasonable to associate an average increasein the energy price by 2020 somewhere in therange 2 to 3 times the 1972 value with theglobal energy demand scenario.

For industrialized countries and the cases (a,b, c) = (0.8, 0.7, 1.0) and (0.8, 1.0, 0.5), ourenergy demand projection is consistent with a50- to 100-percent increase in per capita GDP(comparable to the values assumed in theIIASA and WEC low scenarios) and 2020energy prices 2 to 3 times the 1972 values. (SeeFigure Al.) If there were no ongoing structuralshift (a = 1.0) to less energy-intensiveeconomic activities, somewhat higher energyprices would be required.

For developing countries, our scenario with2020 prices in the range of 2 to 3 times 1972prices and the high-income elasticity (a = 1.4)would be compatible with the per capita GDPgrowth rates assumed in the IIASA and WEChigh scenarios. With the lower-income elasticity(a = 1.1), the base case projection would beconsistent with much more rapid GDP growth.(See Figure Al.)

This modeling exercise shows that althoughour global energy demand projection for 2020is far outside the range of most other projec-tions, it appears to be consistent with plausiblevalues of income and price elasticities and with

114

Figure A-l

/P(1

972)

O

020)

]P(

2

QO

4 0

3.5.

3.0.

2.5.

2.0.

1.5.

x

Each

A Top-Down Representation of Our End-Use EnergyIndustrialized Countries

Key For (a,b,c)a = Income Elasticityb = - Energy Price Elasticityc = Annual Energy Efficiency

Improvement Rate (%)

In n n T-I -i n\

(0.8,1.0,0.5)

(1.0, 1.0, 0.5)

/

/

/

/

/ / /

/ / /

/ / /

Price of Energy (2020) Divided by Price of Energy (1972)

i i

2 3

f ]Federal

r t r n M A P .rrance j iir\c>/\ i roj

Demand Scenario for

8i O

/ O/ ^ _

DUu

•£,

// S• • 03

%o

O

inua

ler

a

<

i

4

ection1980 U.S., 1981; For NA, 2030

Republic IIASA Projectionof Germany For WE/JANZ, 20301980

ine represents the combinations of per capita gross domestic

2 9

< IIASA HighScenario

.2.6

^ Sweden, 1960-82

.2.3

-+— U.S., 1960-82WEC High Scenario

^ U.K., 1960-82

.1 .9

-<— IIASA Low Scenario

_1.5

^ WEC Low Scenario

.0 .8

product and energy price in the year2020 (relative to 1972 values) consistent with the energy demand scenario developeddustrialized countries, for the indicated values of income elasticities,induced energy efficiency improvement rate.

in Chapter 4 for in-price elasticities, and the nonprice-

115

Figure A-2. A Top-Down Representation of OurCountries

in

9 _

8 _

g 7-rH

d,

D£ 6_

2020

)],

6 5"

4 _

3 _

2 _

Key For (a,b,c)a = Income Elasticityb = - Energy Price Elasticityc = Annual Energy Efficiency

Improvement Rate (%)

^ — — » (1.1, 0.7, 1.0)(1.1, l.o' 0.5)(1.4, 0.7, 1.0) /

——.—.- (1.4, 1.0, 0.5) /

/

/ y/ /

/

// y/ /' *v

r

Price of Energy (2020) Divided by Price

1

1 2

End-Use Energy

/

//

/

/ y/ y

r

of Energy (1972)

13

tIIASA ProjectionFor 2030

Each line represents the combinations of per capita gross domestic2020 (relative to 1972 values) consistent with the

Demand Scenario

y/

o

sINRrH

D(J

For

y 1

te (

Pei

row

t

O

3Cr-

rage

Ave

1

4

product and

4.9

_ 4.7

„_ 4.4

-*—

_ 4.1

- 3.8

- 3.4

M

"*—

_ 2.3

-«—-•—

_ 1.5

for Developing

Brazil, 1960-82

Thailand, 1960-82

Malaysia, 1960-82

Middle IncomeCountries, 1960-82

WEC High Scenario

Low IncomeCountries, 1960-82

IIASA HighScenario

WEC Low ScenarioIIASA Low Scenario

energy price in the yearenergy demand scenario developed in

developing countries, for the indicated values of income elasticities,induced energy efficiency improvement rate.

Chapter 4 forprice elasticities, and the nonprice-

116

plausible expectations about energy price andGDP growth, if the non-price-induced energy-efficiency improvement rate can be in therange 0.5 to 1.0 percent per year.

Although the non-price-induced energy-efficiency improvement rate is a measure of thepublic policy effort that would be required tobring about the energy future described in thisstudy, not all this efficiency improvementwould have to be public-policy-induced.Energy-efficiency improvements associated withgeneral technological innovations have oftenbeen made even when energy prices have been

declining,13 a phenomenon that led IEA/ORAUanalysts to include the non-price-induced tech-nological improvement factor in their model forthe industrial sector in the first place.14

At the same time, however, the energy-efficiency improvement factor may not repre-sent the full extent of the needed public policyeffort if low energy-demand levels were tocause energy prices to stabilize or even fall. Insuch a case, energy taxes may also be neededto keep gradual upward pressure on final (con-sumer) prices.

117

Notes for Appendix

1. J. Edmonds and J. Reilly, "A Long-TermGlobal Energy-Economic Model of CarbonDioxide Release from Fossil Fuel Use,"Energy Economics 5(1983): 74-88.

2. J. Edmonds and J. Reilly, "Global EnergyProduction and Use to the Year 2050,"Energy, the International Journal 8(1983):419-432.

3. W.D. Nordhaus, "The Demand for Energy:An International Perspective," in Interna-tional Studies of the Demand for Energy, W.D.Nordhaus, ed. (North-Holland, Amster-dam, 1977).

4. W.D. Nordhaus and G. W. Yohe, "FuturePaths of Energy and Carbon Dioxide Emis-sions," in Changing Climate, report of theCarbon Dioxide Assessment Committee(Washington, D.C.: National AcademyPress, 1983).

5. Energy Modeling Forum, "Aggregate Elasti-city of Energy Demand," Vol. 1, EMF Rep.No. 4, Stanford University, Stanford,California, 1980.

6. D.J. Rose, M.M. Miller, and C. Agnew,"Global Energy Futures and CO2-InducedClimate Change," MITEL 83-115, Massa-chusetts Institute of Technology, Cam-bridge, Massachusetts, 1983.

7. J. Edmonds et al., "An Analysis of PossibleFuture Atmospheric Retention of Fossil Fuel

CO2," DOE/OR/21400-1, Washington D.C.,1984.

8. Energy Modeling Forum, "AggregateElasticity of Energy Demand."

9. C.P. Doblin, "The Growth of Energy Con-sumption and Prices in the USA, FRG,France, and the UK, 1950-1980," RR-82-18,International Institute for Applied SystemsAnalysis, Laxenburg, Austria, 1982.

10. U.S. Energy Information Administration,U.S. Department of Energy, "State EnergyPrice and Expenditure Report 1970-1981,"Washington, D.C., 1984.

11. W. Haefele, et al., Energy in a FiniteWorld—A Global Systems Analysis (Cam-bridge, Massachusetts: Ballinger, 1981).

12. Office of Policy, Planning, and Analysis,U.S. Department of Energy, "Energy Pro-jections to the Year 2010: A TechnicalReport in Support of the National EnergyPolicy Plan," Washington, D.C., October1983.

13. R.H. Williams, E.D. Larson, and M.H.Ross, "Materials, Affluence, and IndustrialEnergy Use," The Annual Review of Energy,12(1987): 99-144.

14. J. Edmonds and J. Reilly, "A Long-TermGlobal Energy-Economic Model of CarbonDioxide Release from Fossil Fuel Use."

119

World Resources Institute

1735 New York Avenue, N.W.Washington, D.C. 20006 U.S.A.

WRI's Board of Directors:Matthew NimetzChairmanJohnE. CantlonVice ChairmanJohn H. AdamsRobert O.BlakeJohnE. BrysonRichard M. ClarkeMarc J. DourojeanniAlice F. EmersonJohn FirorCurtis A. HesslerMartin HoldgateJames A. JosephIan K. MacGregorAlan R. McFarlandRobert S. McNamaraPaulo Nogueira-NetoRuth PatrickJames Gustave SpethM.S. SwaminathanMostafa K. TolbaRussell E. TrainArthur C. UptonGeorge M. Woodwell

James Gustave SpethPresidentJessica T. MathewsVice President and Research DirectorMohamed T. El-AshryVice President for Policy AffairsWallace BowmanSecretary-Treasurer

The World Resources Institute (WRI) is a policyresearch center created in late 1982 to helpgovernments, international organizations, the privatesector, and others address a fundamental question:How can societies meet basic human needs and nurtureeconomic growth without undermining the naturalresources and environmental integrity on which life,economic vitality, and international security depend?

The Institute's current program areas include tropicalforests, biological diversity, sustainable agriculture,global energy futures, climate change, pollution andhealth, economic incentives for sustainabledevelopment, and resource and environmentalinformation. Within these broad areas, two dominantconcerns influence WRI's choice of projects and otheractivities:

The destructive effects of poor resourcemanagement on economic development and thealleviation of poverty in developing countries; and

The new generation of globally importantenvironmental and resource problems that threatenthe economic and environmental interests of theUnited States and other industrial countries andthat have not been addressed with authority intheir laws.

Independent and nonpartisan, the World ResourcesInstitute approaches the development and analysis ofresource policy options objectively, with a stronggrounding in the sciences. Its research is aimed atproviding accurate information about global resourcesand population, identifying emerging issues anddeveloping politically and economically workableproposals. WRI's work is carried out by aninterdisciplinary staff of scientists and policy expertsaugmented by a network of formal advisors,collaborators, and affiliated institutions in 30 countries.

WRI is funded by private foundations, UnitedNations and governmental agencies, corporations, andconcerned individuals.

W O R L D K E S O U R C K S\7.\5 New York Avenue, NVVWashington, DC 20006USA

I \ S 1 I T U T I.

$10.00 US