N V UL    -
THE WORLD BANK
SECTOR POUCY AND RESEARCH STAFF
Environment Department
Energy Conservation Policy,
Energy Markets and the Environment
00
in Developing Countries'
Robin W Bates
May 1991
Environment Working Paper No. 45
This paper has been prepared for internal use. The views and interpretations herein are those of
the author(s) and should not be attributed to the World Bank. to its affiliated organizations or to
any individual acting on their behalf.


ACKNOWLElGEMENTS
The author is Principal Economist in the Environment Department of the World
Bank. The main work on the paper was carried out during two months study leave from the
World Bank, in 1990, at the Massachusetts Institute of Technology (MI). The views and
opinions expressed in the paper are the author's and do not necessarily reflect those of the World
Bank. The author acknowledges the support of MIT and the Center for Energy Policy Research,
especially the late David Wood, in making the work possible. A particular debt is owed to David
Wood for his continued support and encouragement for the work, following the author's return to
the World Bank, and his valuable comments on an earlier draft of the paper, which resulted in a
substantial improvement in the paper's structure and strengthening of the content. Thanks are
also due to Gunter Schramm of the World Bank for comments on the first draft of the paper.
The paper reports on findings that are particularly significant in the context of the overall work
program of the Environmental Policy and Research Division.
Departmental Working Papers are not formal publications of the World Bank. They
present preliminary and unpolished results of country analysis or research that are circulated to
encourage discussion and comment; citation and the use of such a paper should take account of
its provisional character. The findings interpretations, and conclusions expressed in this paper are
entirely those of the authors and should not be attributed in any manner to the World Bank, to
its affiliated organization, or to members of its Board of Executive Directors or the countries they
represent.
Because of the informality and to present the results of research with the least
possible delay, the typescript has not been prepared in accordance with the procedures
appropriate to formal printed texts, and the World Bank accepts no responsibility for errors.


Interest in energy conservation, although to some degree cyclical, has been stimulated
during the last twenty years by the rising cost of energy In a wide range of developing and developed
countries, especially following the oil price shocks of 1973-1974 and 1979-1980; by environmental
concerns, notably due to the impact of increasing energy consumption on global warming, pollution,
forests and natural habitats; and by national security considerations, as domestic energy supplies
continue to be vulnerable to political events in the Middle East. An active debate has ensued, In
which it is alleged that the existence of a variety of market fa Jures, imperfections and distortions
justifies government intervention in energy markets to promote expenditures on energy conservation.
It is the purpose of this paper to evaluate the validity and relevance of that debate to
developing countries, in terms of demand-side management, mainly where the public sector exerts
control over a significant portion of energy supply; and where that supply is sold predominantly in
markets subject to consumers acting competitively. The central tenet of the paper is that confusion In
the debate can only be avoided If a careful distinction is maintained between arguments related to the
proper functioning of energy markets, on the one hand; and externalities, on the other.
On the basis of a review of the literature on the sources of possible market failures and
the proposed remedies, much of which was inspired by circumstances in the U.S.A., the paper finds
that evidence of significant market failures sufficient to justify government intervention in energy
markets is not convincing; and that the proposed remedies In any case are inappropriate. The paper
concludes that, while environmental externalities may provide grounds for carefully-targeted public
sector intervention in energy markets, the national security argument is hard to sustain in developing
countries; and market imperfections and distortions generally do not justify non-price intervention,
with the possible eception of information provision and support for basic research and development


I.ntrductio
1.          Interest in energy conservation received a sharp impetus after the oil price increases of
1973-1974 and 1979-1980: in constant US$ of 1985, the average price of crude oil rose more than
threefold during 1974 and had Increased by a firther 50% by the end of 1980W. Consequently, the
real international  :e of crude oil was more than five times higher In 1980 than in 1973, with
serious implicatio..A for the balance of payments situation In key developed and a wide range of
developing countries. In the U.S.A., the value of net fuel Imports increased as a percentage of total
Imports, from 9.4% in 1973 to 21.7% in 1974 and 29.6% in 1980; while the corresponding figures
for Japan were 21.5% in 1973, 39.7% In 1974 and 49.7% in 1980. Also over the period 1973-1980,
the Philippines, Brazil, Tanzania and Kenya witnessed increases in the ratio of not fuel imports to
total imports by factors of roughly 2,3,4 and 5 respectively'.
2.          Predictably, the cost of thermal-based electric power generation has also risen, although
not solely as a result of higher oil prices. Data from the U.S.A. suggest that higher capital costs
were important, especially in connection with new nuclear- and coal-fired capacity'. While direct
information on trends in the typical long-run marginal cost of electricity generation Is difficult to
obtain, particularly for developing countries, recent movements in average financial costs are
probably reflected in the behavior of prices In the O.E.C.D. countries. The average electricity price
in the U.S.A., in constant US$ of 1984, increased steadily after 1973Y; figures for Japan and Europe
show average price increases of 25% and 11% respectively, for the period 1978-1981, following the
second oil price shood'; while, for the O.E.C.D. group as a whole, the real price of electricity
I/ World Bank, Price Prospet for Mqjor Primary Connoditl, Report No. 814/8, World Bank, Novemner
1988, Vol. I, p. 46.
Z/ World Bank, World Tabks 1987, World Bank, 1988.
I/ J. Goldemberg, T.B. Johamsson, A.K.N. Reddy and R.H. Williams, Energyor a Saainable World, Wiley
Eastern Limited, 1988, pp. 78-83.
g/ Ibid., p.80.
I/ Data provided by the Brazilian *Programa Nacional do Conservago do Energia Eldtricao (PROCEL),
Cnuis ldtricas Brasileras S.A. (ELETROBRAS).


-2-
lncreased by 1.4% p.a. (in constant US$ of 1986) for the decade 1979-1988 2'. Increases in the
long-run marginal costs of electricity supply are likely to have been even more pronounced.
3.          These developments In anergy prices and costs over the last twenty years, with their
adverse Implications for economic growth, stability and perceived living standards, were an important
driving force In stimulating a renewed Interest in the fnctioning of energy markets. In particular, an
active debate developed - especially in the U.S.A. - around the extent to which market forces,
operating through the price mechanism, could be relied upon to ensure the optimal lIplementation of
energy conservation measures; and whether or not Government intervention would be necessary to
promote energy conservation beyond the level which would be determined by market forces alone.
4.          In a roughly parallel development, environmental and national security concerns also
fuelled the debate over energy conservation policy. Joskow has described how the environmentl
movement became active in the reform of the electricity tariff structure in the United States In the
1970s, in order to siow down the rate of growth of electricity demand and mitigate *the adverse
environmental consequences associated with power plant construction"21; and how "environmentalists
were Interested in reducing the rate of growth in electricity consumption slnce they anticipated this
would reduce the need for more generating facilities and cut the Impact of electric power expansion
on the environment"'. Joskow then points out: "Several environmental groups, particularly the
Environmental Defense Fund, saw rate reform and especially the use of marginal cost pricing '
principles and peak-load pricing, as an important instrument for achieving these objectives*'. More
recently, the National Audubon Society developed *a comprehensive strategy to Iaprove US energy
efficiency and thereby reduce harmful impacts on wildlife and the general environment due to energy
production and consumption"tw; and the World Resources Institute reaffirmed Its conviction dat
wIncreased conservation in both developed and developing countries stands as perhaps the most
promising option for mitigating the various risks associated with the current global energy habits".
Most recent of all, the Gulf conflict, which took place at the beginning of 1991, again brought to the
fore concerns for national security, which stem from the high degree of dependence by oil-importing
countries on a single geographic source of supply for a vital energy source.
I/ World Bank, Review of Electrlety Tarfs in Dewoping Countrier during the 198fr, Industry and Energy
Department Working Paper, Energy Series Paper No. 32, World Bank, November 1990, p.25.
2/ P. Joskow, 'Electric Utility Rate Structures in the Unite States: Some Recent Developments, in W. Sichel
(ed.), Public Utility Rate Making in an Energy-Conscious nvironment, 1979.
I/ P. Joskow, "Public Utility Regulatory Policy Act of 1978: Electric Utility Rate Reform', Natwual Resourcur
Journal, Vol. 19, No. 3, October 1979.
2/ Ibid.
IQ/ P. Rollin and J. Beyea, 'US Appliance Efficiency Standards', Energy Policy, October 1985.
ILI World Resources Institute, World Resourcer 1990-91, Oxford University Press, 1990, p. 145.


-3-
5.          From the perspective of the developing countries, Goldemberg has been particularly
eloquent In linking environmental Issues and national self-reliance to the energy conservation
debatew. Among the developing countries, Brazil furnishes a striking example of the prominent
role which environmental concerns can play in discussions of energy policy, when plans to meet :he
rapid growth In electricity demand dep%ad, as they did until recently, upon the construction of large
hydroelectric projects in environmentally-sensitive areas of the Amazonw. Brazil's drive for greater
national self-sufficiercy In energy also gave impetus to efforts to reduce the consumption of petroleum
products'.
6.          It is the goal of this paper to examine the debate over Government Intervention in the
energy market, to promote commercial energy conservation. In doing so, we believe that it is crucial
to distinguish between arguments related to the proper functioning of energy markets, on he one
hand; and externalities, in the form of environmental and national security concerns, on the other.
While much of the literature reviewed was inspired by circumstances In the U.S.A., we are mainly
concerned with its application in developing countries, where the public sector exerts control over a
significant portion of energy supply; and where that supply is sold predominantly in markets subject
to private-sector consumers acting competitively. The major economies of Latin America would fit
this description, notably Argentina, Mexico, Chile and Brazil.
7.          We start by reviewing, in Section II, how a competitive market would ensure that energy
conservation measures are implemented optimally, relative to increases in energy supply, as part of a
least-cost solution, in the absence of externalities. In Section M, we classify the main types of
imperfection and distortion which have been alleged to interfere with the proper functioning of energy
markets, analyzing their effects on energy conservation, energy prices and energy supply in Section
IV. Section V considers the empirical literature on consumers' discount rates, which purports to
neasure the degree to which the existence of market imperfections interferes with the decision-taking
process of the competitive market; while Section VI discusses the principal measures which are
proposed in practice to ameliorate them, viz. information provision, Improved financing of energy
conservation investments, appliance efficiency standards, and building codes. Some of the empirical
evidence on the effectiveness of energy conservation mcasures Is rvicwed in Section VII, before we
return, in Section VIII, to the issue of externalities, notably with regard to the environment. Finally,
in Section IX, we conclude that, while environmental externalities may provide grounds for carefully-
targeted public sector intervention in energy markets, the national security argument is hard to
Il See, for example, J. Goldemberg, *Solving the Energy Problems in Developing Countries, Enery
Journal, Vol. 11, No. 1, January 1990; and J. Goldemberg et. aL, Eergy fbr a Sustainable World, op. cit..
11/ See, for example, Luiz Pinguelli Ross, O"delEtricas a Meio Ambient na Amaz6nia - AnAlise Critica do
Plano 2010, Ravista Brastleira de Energla. Vol. 1, No. 1, 1989; H.S. Geller, J. Goldemberg, J.R. Moreira,
P Kukai, C. Scarpinells and M. Yosiawa,  lectricity Conservation in Brazil: Potential and Progress",
Energy, Vol. 13, No. 6, 1988; and H.S. Geller, Electricity Consernation ti Brazik fatus Report and Analyb,
American Council for an Energy-Efficient Economy, August 1990.
I Antnio Aureliano Chaves do Mendonga, Brazilian Ewrgy Policy, Ministry of Mines and Energy,
September 1988.


4-
sustain In developing countries; and market Imperfections generally do not justify non-price
Intervenflon, with the possible exception of Information provision and support for basic energy
research and development (&D).
U. En   a C    amumn an ad Wl  tCM   Soupi
8.         In order to evaluate the question of market imperfections and public sector Intervention
more carefully, we remind ourselves how a perfectly competitive market would take care of energy
conservation as part of the least-cost solution. Throughout the discussion, we assume for convenience
the absence of externalities, which are treated separately and in detail In Section VII.
9.         Fig. I portrays a simplified model of the energy market: the curve SS', representing the
marginal cost of energy supply (MCES), is assumed to be upward sloping (or highly elastic In the
case of imports) over the relevant range of supply; and the demand curve Is shown by DD'V. For
any quantity under SS', the program of lowest Investment and operating costs for the energy sector
has been selected. Similarly, for any quantity under the curve DD', consumers have optimized
investment in energy conservation, i.e. to the point where fte incremental investment In energy
conseivation (A) equals the present value of the stream of incremental expected energy savings
produced by that investment. More precisely, the optimum condition for energy conservation
investment is:
A IS        t Q'       (1)
ir*i (1+rf
where r is the discount rate applied to energy consumption Investments and P, and AQ=, are
(respectively) the energy price and the expected additional (physical) energy savings in year t
attributable to the incremental investment in energy conservation Alec. The values of r, P, and Q,
will vary across consumers.
I The case of a faling MCHS, e.g. through economies of scale, and a downward-sloping SS', could also be
considered, but we take the position that the situation in Pig. 1 is more typical (see para. 20).


pt
10.         Fig. i treats pricm and quante as costant annual a==uns. With PPo Qa0  AQac
for ail t, equadrn (1) reduces to:
where ag, l the fator for the present valu of an anmity for n year at an nteret rate r, Le.
r
Dedning the marginal cost of energy conservadon (MCEC) as the annual equivalent of Alc per unit
of addidonal energy saved, then


MCEC     a   -ci,P          (4)
AQSC,
since cl,, the capital recovery factor, is the inverse of a,.
11.         Assuming diminishing returns to energy conservation investment, Alec8AQw will
increase as the level of energy conservation investment goes up. It follows that less energy
conservation will take place as we move down the demand crve DD' in Fig. 1: as P fails, with a
given discount rate r, MCEC is reduced by reducing conservation Investment, to maintain the
marginal optimizing condition. As a reference, curve DD'. In Fig. 1 represents the corresponding
demand curve with no energy conservation investment; and the horizontal gap between DD' and
D.D'. therefore measures the amount of energy conservation taking place. Similarly, an Increase in r
raises MCEC via the capital recovery factor c, and consumers respond by cutting energy
conservation investment at all prices to reduce Aac/AQgc to the point where equality between MCEC
and P is restored. It follows that the curve DD' in Fig. 1 moves to the right for hcreases in r; and,
with the MCES curve SS', price and supply Increase, while energy conservation decreases.
12.         In the case of Fig. 1, the market ensures an optimal allocation of resources at the
equilibrium price Po and quantity Qo. energy is supplied at least cost, with MCES=MCEC=Po,
compared with the higher price P. and supply Q. with zero energy conservation. No Intervention is
necessary to encourage investment in energy conservation measures and demand-side management is
handled entirely through economically-sound pricing policy.
II. Mat Imperfectlans andDistortio
13.         In contrast to the perfectly competitive market, it is often argued that public sector
intervention is necessary in energy markets, due to the existence of market imperfections and
g/ Mathematically, we defne: the relationship Q = f(1) between annual physical energy savings Q. and
energy conservation investment I, with its inverse I   g(Q); an MCEC as -6 8 A = cj,g'(Q.) > 0, with
g-(Q,) > 0 (diminishing returms). For Q. - 0, we define demand Q = (P). With a gien discount rate r and
price P, energy conservation Q will be pursued to minimize annual energy expenditures W * P(Q-Q) +
%1,g(Q.), i.e. until 8E/Q - P + c,g'(Q.) - 0, giving our optimizing condition P - MCBC in equation (4).
Q increases with P to maintain the optimizing condition: dQJdP - l/[cg(Q.)j > 0, as g*(Q) > 0. With a
given energy price P and an increase in r, Q. decreases with r as MCEC is adjusted to maintain the optimizing
condition: dQ./dr = -(6MCEC/6r)/(6MCEC/8Q,) < 0, since kj8br > 0, g'(Q) > 0 and g'(Q) > 0.


*7-
distortionse. The goal of such Intervention is to foster energy conservation beyond the level which
wold be determined by market forces alone. These Imperfections can be classified in a variety of
ways, but the main types Identified in the literature can be categorized in terms of: those which distort
the price signals or else place constraints on the consumer's ability to respond, even if the price
signals are rorrect; and those which stem from deficiencies in the basic process underlying energy
conservation decisions. However, the empirical evidence is unsatisfactory on the relative importance
of these imperfections and the actions best suited to deal with them.
) Disto     Nte e Signals and Rapns
14.         If energy prices are below the full economic costs, the pricing signals will be distorted.
There may also be distortions between energy production costs and prices, compared with energy
conservation costs, for example due to subsidies on certain forms of energy production and
macroeconomic distortions, e.g. taxes and duties on the Import of energy-efficient equipment. A
more subtle distortion in energy pricing signals can be traced to the high transaction.costs of efficient
energy pricing, arising from metering and other requirements of load management.
15.         Whereas the pricing signals in the energy market may be correct, there could be
constraints which limit the extent to which consumers are able to respond efficiently to those signals.
Imperfections In the capital market (along with inadequate internal cash generation in firms) may limit
the availability of finance; currency restrictions may curtail the supply of foreign exchange to
Ihwlement energy conservation measures; and there may be quotas or an outright ban on the Import
of certain energy-efficient equipment or devices to measure energy efficiency, which are not produced
locally.
(Il) Defiient Dion-Tadn Processes
16.         Various deficiencies have been identified in the way that energy consumers take
decisions, leading to inertia or under-investment in energy conservation. Consumers could be slow in
responding to changes in energy prices, especially where energy represents a small proportion of total
costs; they may face imperfect Information about the benefits and costs of energy conservation (partly
because the private sector may be unable to capture fully the benefits of research and development on
jZ/ See, for example, B. Hirst, W. Fulkerson, R. Carlsmith and T. Wilbanks, "Improving Energy Efficiency:
the Effectiveness of Government Action, Energy Polcy, June 1982: "Govemment programmes are justified on
the basis of the market barriers listed above and also on the basis of social benefits not ly captured by private
interests"; and L. Schipper, J.M. Hollander, M. Levine and P.P. Craig, "The National Energy Conservation
Policy Act: An Evaluation, Natural Raourew Journal, Vol 19, October 1979: "The need for government
participation is illustrated by the abundant set of market failures that deter conservation of energy in buildings".
A more qualified and accurate statement is provided by A.C. Fisher and M.H. Rothkopf, "Market Pailure and
Energy Policy", Energy Policy, August 1989: -To the extent that markets fail, government intervention may
improve the effciency of the allocation of resources - though it is, of course, not guaranteed to do so".


-
generic energy conservation technologies and therefore unwilling to carry It out); they may lack the
methodology to evaluate that inruimation (.e to handle the inter-temporal comparisons or trade-offs
which must be made between additional capital costs and reductions in energy costs); they may not
know how to design, initiate and implement energy conservation programs (notably due to a specific
lack of expertise In energy management at the plant level); and they may be ignorant about the
sources of energy conservation finance and the procedures for securing It.
17.         Writers have frequently identified institutional and regulatory barriers to efficient
decision-taking in the energy market, notably the divorce between those who take decisions about the
design and construction of energy-using facilities (architects, builders etc.) and those who will pay the
operating costs (owners and managers) or between the owners of a building and those who will
occupy and use it. Furthermore, firms which consume energy may not be motivated by profit
maximization; managers may not pursue energy efficiency or may attach a low priority to it, due to
the role of non-profit Incentives in managerial decision-takingly; and there could be a widespread
use of cost-plus pricing, particularly in the public sector, under which companies routinely try to pass
higher energy costs on to consumers.
(il) EmWlial vde
18.         Direct empirical evidence of the impact of these market imperfections is unsatisfactory,
especially in the second category. The literature descrilNs their existence, and there is a body of
circumstantial evidence for particular cases, but their relative importance and the actions best suited to
overcome them have not been demonstrated'. The argument hinges predominantly on Indirect
evidence, from econometric studies, that consumers reveal discount rates far In excess of the cost of
capital in their energy-conservation decisions; and engineering evidence of a significant difference
between the energy consumption that could be achieved theoretically from more efficient and known
technologies and actual consumptionv. This theoretical difference amounts typically to at least 10-
111 For example, there may be more prestige attached to the implementation of large-scale investment projects
ather than more mundane and low-prestige energy conservation measures or managers may be more
preoccupied with short-term problems.
2/ This position was taken by C. Blumstein, B. Krieg, L. Schipper and C. York, in "Overcoming Social and
lastitutional Barriers to Energy Conservation", Energy, Vol. 5, No. 4, April 1980; and more recently, nearly a
decade later, by R.S. Carlsmith, W.U. Chandler, LB. McMahon and D.J. Santini, Enery Effidency: How Far
Can We Go, ORNL Report TM-11441, January 1990.
2&/ A good example is B. Hirst e. aL, Improving Energy Efficiency: th Effectiveness of Govenment
Action, op. cit.. From the fact that energy audits of 48 hospitals in four States in 1981 proposed energy
conservation measures that would cut energy use by 20%, they infer that Govemment information programs
could resolve the issue; and that the high potential rates of return to investment in conservation measures in new
houses, to improve thermal performance, shown by a study in Kansas City, *suggest that the now housing
market is not economically efficient" and there is 'an important role for government programmes .


-9-
20% of total energy consumption in most countries: an even higher figure is sometimes quotedP.
For Brazil, a 1987 study estimated that the additional electricity savings available from laplementing
technologies in six major end-use areas "that are technically and economically feasible and, In many
cases, already available in Brazil" could amount to 20% by 2000 and that "additional savings may be
possible In other areas such as water heating and air conditicalng. A more comprehsive
review of energy savings potential in a sample of 25 developing countries by the Jont UNDP/World
Bank/Bilateral Aid Energy Sector Management Assistance Program (ESMAP) suggested that an
average of 20% of total end-use energy could be saved from the existing capital stock. By
showing that the cost of conservation measures is considerably below the value of the energy savings
and the cost of supply alternatives, the advocates of Government Intervention conclude that action by
the public sector can remove the institutional obstacles and market imperfections which prevent these
gains from being realized, or else introduce additional distortions In order to offset those caused by
market imperfections2y.
IV. The Effect of Imperfections and Distortions on the Energy Market
19.         Using the taxonomy developed In Section I, we now analyze the impact of market
imperfections and distortions on the workings of the energy market. Our discussion will te
conducted in terms of energy prices, energy supply and the amount of energy conservation investment
which is carried out and we address explicitly the implications for demand-side management.
() Distorted Price Signals and Resnonses
21/ E. Hirt er. al., 114, cite a 1978 study at the Mellon Institute, which estimated the hypothetical savings
from a "least-cost' energy strategy in 1978 of 25% of total energy actually consumed in that year. I. Geller,
J. Harris, M. Levine and A. Rosenfeld suggested in 1987 that the U.S.A. could have reduced its total energy
bill by 25%, simply by adopting available technologies which are economically justified. See -The Role of
Federal Research and Development in Advancing Energy Efficiency: A $50 Billion Contribution to the U.S.
Economy", Annual Review of Energy, Vol. 12, 1987.
2V  H.S. Geller e. al, Electricity Conservation in Brazil: Potential and Progress, op. cit..
ZY' World Bank, Energy fficlency &rrategy for Developing Countria: the Role of ESiAP, Background Paper
for Discussion at ESMAP's Annual Meeting, World Bank, 1989.
2A/ Aside from the U.S. literature, for developing countries see J. Goldemberg et. al., EnerVfor a
Sustainable World, op. cit.; M. Manasinghe, "Third World Energy Policies, Demand Management and
Conservation', Energy Policy, March 1983; M. Munasinghe and 0. Schramm, Energy Economs, DeMand
Management and Conserwaon Policy (1983); H. Kohli and E. Segura, "Industrial Energy Conservation in
Developing Countries*, Finance and Development, December 1983; and J.R. Gamba, D.A. Caplin and L.
Muakhuyse, Industrial Energy Radonaltation in Developing Countries, World Bank, 1986.


-10-
20.         Evidently, if P<MCES, there will be cost-effective measures which are not being
Implemented, but that is precisely why economists recommend MC pricing. The situation in Fig. 1,
where the curve of energy supply SS' Is upward sloping, Is typical for the major sources of
commercial energy, such as petroleum products, coal and natural gasW. While a downward-sloping
supply curve may have been appropriate for electricity supply in the past, the era of declining long-
run marginal costs, when the industry enjoyed substantial economies of scale, is over In most
developing and developed countries.
21.         Suppose that price Is set at P,<Poin Fig. 1. There Is empirical evidence that this
situation is widespread in developing countries, at least for electricity supply: It may result because,
with rising MCES, the average cost of energy supply (ACES) is likely to be less than MCES and
excess profits would be generated; or simply from a desire to subsidize energy prices (P<ACES)N.
Then quantity Q, will be supplied, with MCES,> P1. Since energy conservation will be pursued only
to the point where MCEC=Pl < MCESI, energy conservation will be more economic at the margin
than energy supply. However, the argument that marginal-cost pricing does not exist In practice Is
insufficient where the Government has a direct Influence on energy prices. From an economic point
of view, it Is better public policy for the Government to use its influence to ensure that energy prices
reflect economic costs, before intervening In the market to create Arther distortions, since it has
direct control over at least some energy pricing (e.g. electricity) In most developing countries. Non-
price intervention should not be justified on the grounds that it may be politically easier, although it
frequently is; and if excess profits would be generated (ACES < MCES), this can be viewed as an
additional benefit in most developing countries, creating an opportunity for the Government to
supplement a generally inadequate fiscal base by levying taxes and yielding a valuable new source of
revenue'.
22.         A more sophisticated version of the energy model in Fig. I recognizes that energy
supplied at different times of the day, week and year has a different value and cost, so that different
demand and supply curves are needed, leading to different equilibrium prices and quantities for
different time-periods. The differential pricing of certain types of energy (notably electricity and gas)
and the related concept of load management, e.g. using interruptible supplies and time-of-day
metering, involves transaction costs, which may be significant. The issue of cost effectiveness must
be addressed, particularly with some (but by no means all) residential customers. If proper metering
is not employed, consumers may make choices about appliance use and energy conservation on the
basis of energy prices which do not reflect economic costs. However, a variety of approaches are
available to handle the problem, such as offering optional metered tariffs, where the Individual
21/ For tradeables, notably petroleum products, SS' is highly elastic and yields the se results as the upwad-
sloping curve in Fig. 1.
?&/ See World Bank, Review of Electricity Taro  in Developing CountriW durtg he 1980M, op. dit.; and
Enery Efficiency Swrategy for Developing Counries: the Role of EMAP, op. cit..
2,/ Or off-setting tax reductions can be made, if there is a constraint on increasing total tax receipts.


-11-
consumer pays for the meter and weighs the cost of metering against the benefits. The example
of the 'chuveiro', from Brazil, is a case where special metering is unlikely to be cost-effective. The
'chuveiro' is a relatively inexpensive shower fitting, which heats the water electrically as it Is used,
contributing significantly to the evening system peak. About 20 million are estimated to be in use in
Brazil,v. However, rather than, say, regulating the design and manufacture of the device, a tax
could be included in the purchase price to reflect its contribution to system capacity costs'.
23.         In the context of developing countries, a particular variation is heard on the argument
that energy prices and costs give inadequate signals to consumers or consumers may not be able to
respond appropriately to those signals, due to macroeconomic distortions, notably in trade policy.
Thus, the foreign exchange required to purchase energy-efficient items may be restricted'; there
may be taxes or controls on the import of energy-efficient equipment and devices to measure energy
efficiency; or some forms of energy supply may be subsidized, either directly by the Government or
indirectly, by failing to capture the high economic cost of the foreign exchange implicit in capital-
intensive supply solutions. It is then proposed that taxes and subsidies (which could be tax
deductions or other tax incentives, such as accelerated depreciation for energy-efficient investments)
are needed to offset distortions.
24.         On the demand side of the energy market, macroeconomic distortions have a range of
impacts on DD' in Fig. 1. A tax on energy conservation equipment (ad valorem or unit) will raise
MCEC throughout DD'. At each energy price, investment in energy conservation will decline, to
reduce MCEC: DD', in effect, shifts to the right, say from DOD'o to D1D', in Fig. 2 N. With the
supply curve SOS'O, the equilibrium price and supply increase from Po to P, and from Qo to Q,
respectively, with less investment in energy conservation than before. Limitations on the availability
of energy efficiency devices, either directly (through quotas and physical controls) or indirectly
(through a lack of financing sources, including foreign exchange) are more complex. If they are
effective, then (by definition) at some energy demand Q, > Qo on DoD'o in Fig. 2, the demand for
21/ See P.L. Joskow, "Public Utility Regulatory Policy Act of 1978: Electric Utility Rate Reform', op. cit..
Setting a tariff to reflect the weighted average cost of peak and off-peak use does not really address the issue,
since the allocation of conserved electricity between peak and off-peak hours at the margin may differ from the
average allocation used to calculate the weights.
2&/ OTestes Revelam Problemas com o Chuveiro Eltrico*, Sdo Paulo Energla, Ano VII, No. 63, April 1990.
,V/ The transaction costs of the tax need to be compared with the cost of regulations. There is also an issue
concerning the recipient of the tax proceeds. While the tax way not significantly affect the use of the
'chuveiro' and bence the cost of supply, the consumer will bear and be forced to recognize the costs imposed
on the system and an incentive provided to produce a more efficient shower fitting. Of course, we do not
address here regulations designed to improve the safety rather than the energy efficiency of the 'chiveiro'.
U/ Presumably, it would also be restricted for energy-producing items, e.g. imported capital for electricity
generating plant.
I/ Mathematically, an additional term Q, (a unit tax) or =Ivl (an ad valorem tax) is added to E in footnote
16, so that the condition becomes P = c ,g'(Q) + r or P - c ,g'(Q(l + r) respectively.


-12-
energy conservation devices is curtailed (equivalent to an artificially high MCtC at the curtailment
point); and the price elasticity of demand for energy falls with respect to price increases above P..
Hence, for prices above P,, the demand curve In effect again shifts to the right, say to DD'0D's, with
qualitatively similar impacts on energy price, supply and conservationw.
51
P---------.
z *
25.           Special treatment for energy conservation measures amounts to an attempt to shift
!ng or the relevant section of DD'2D'o in Fig. 2 to the left but represents an inefficient way to find
the optimum energy price and supply. Apart from the difficulties in implementing a program of
special treatment and predicting its consequences, taxes, liIted foreign exchange availability and
import restrictions are general distortions and are not best tackled through the energy sector by
manipulating the demand curve for energy: there is no presqpposition that special treatment for
energy conservation has greater net welfare benefits than, say, making more foreign exchange
U/ Mathematically, referring to footnote 16, a constraint I, on energy conservation investment sets an upper
limit f(I) on energy savings and determines energy demand Q -  (P) - f(L) for Q > Q,, P > P, The
optimizing condition becomes P - g'(Q)( + \), where the Lagrangean X is the shadow value of the
(constrained) energy conservation investment. Market forces would tend to establish a premium value for the
constrained investment equal to X.


-13-
avalable or subsidizing other constrained activities outside the energy sector. Broader Government
action is more appropriate on trade and exchange rate policy.
26.         On the supply side of the energy market, developed as well as developing countries have
subsidized energy productionw; and analytically, such subsidies will shift the MCES curve to the
right, say from SOS'o to SIS', in Fig. 2, by artificially depressing the real MCES, leading to a lower
equilibrium price P2 relative to the optimum P. an increased supply Q  > Q0 and again under-
investment in energy conservation. However, demand-oriented measures, based on special programs
for energy conservation, are theoretically questionable in dealing with supply distortions. In effect,
they try to shift DoD'o to the left to offset the rightward shift in SS'. The remedy is to discontinue
all subsidies, not to subsidize energy conservation measuresP.
(l) Diclt Decisiodakng rM
27.         It is sometimes claimed that consumers respond (at best) only slowly to price signals and
meanwhile costs will be incurred by the economy, due to the dynamics of price adjustment in the real
world, which are not recognized in Fig. 1N. For example, a given increase in ral income may
shift the demand curve outward from DoDO to DID', in Fig. 2 in the short run, but as more energy
conservation takes place in response to price increases, the curve will settle somewhere between
DoD'O and D1D'I. Producers supposedly, for some period of time, overestimate demand and during
the response lag, resources will be misallocated between production and conservation.
28.         While there is nothing intrinsically different about energy, as lags occur in the adoption
of other products, it is not usually argued that Government intervention is required. An interesting
approach has been elaborated by Shama, In which a parallel is drawn between the adoption of
conservation devices and the diffusion of innovationsW. Shama argues that the slow adoption of
JJ/ For example, Brazil has subsidized nuclear power and eAawl, to reduce its dependence on imported oil
(pars. 67). Widespread subsidies to U.S. energy development, for nuclear and hydroelectric power, coal, oil
and gas were described by R.L Bezdaek and B.W. Canm, TPederal Incentives for Energy Development,
Energy, Vol. 5, No. 5, May 1980.
35 At the level of individual project analysis, foreign exchange limitations can still be addressed by evaluating
supply alternatives with shadow exchange rates, refiecting (inte ala) JIe &&! ofroa nMcwd
the private sector used for energy conservation.
I See, for example, L. Schipper et. aL, -The National Energy Conservation Policy Act An Evaluatione,
op. cit.: 'When neither expensive new energy sources nor spontaneous changes in demand patterns coma about
quickly enough to satisfy political goals (such as the reduction of oil imports), it is necessary for governnt to
stimulate investments in conservation and supplies or to remove barriers that prevent a 'spontaneous ' response
of the marketplaceO. Also L. Schipper, 'Another Look at Energy Conservation!, American Economic
Assocladon (Papers and Proceedings), May 1979.
Il A. Shama, 'Energy Conservation in U.S. Buildings, Energy Policy, June 1983.


-14-
some energy conservation measures is rational, if the innovations analytical framework is adopted.
The relative advantages of conservation devices must be demonstrated, compared with well-tried and
established supply alternatives; they must be perceived as compatible with existing life-styles or a
decision taken to change life-style; and they are more complex than generally supposed, requiring an
understanding of basic energy economics. Ie conclusion is that these factors should determine the
type of marketing which is necessary, for example using small-scale trials to reduce consumer risk or
targeting the segment of the population which is ready to take up the Innovation (e.g. the *early
adopter*)V. Price is less important in encouraging adoption at some stages than at others. Rather
than intervening in the market, authorities might do well to examine more carefully and publicize the
successful marketing strategies which were pursued for select Innovations, e.g. in the commancations
industry. Removal of barriers to competition and restraints on trade would also be appropriate, to
encourage the supply of cost-effective technologies.
29.         Imperfect information; ignorance surrounding the way in which to evaluate, design,
initiate, finance and implement energy conservation programs; and an institutional separation of the
designers, constructors, owners and operators of buildings have all been identified as distortions in the
decision-taking process regarding energy consumption, causing the energy demand curve in Fig. 1 to
lie between DM'. and DD'. MCES and MCEC would not then be brought into line through the price
mechanism; and, evidently, the market solution would entail a higher energy price and supply than PO
and Qo respectively.
30.          Firms consuming energy as an intermediate product may exhibit patterns of behavior
which are not motivated by profit maximization. For example, they may give low priority to energy
conservation and try to pass on higher energy costs to final consumers through cost-plus pricinge.
In terms of Fig. 1, the effect is complex to interpret, but one possibility is to reduce the price
elasticity of energy demand with respect to an increase in the prevailing energy price: the effect of an
energy price increase on energy demand and supply would then be less than otherwise, leaving
unexploited investment opportunities in energy conservatioO. Notice, however, that the magnitude
is hard to predict: cost-plus pricing raises the price of the final product by the full amount of the
incremental cost of the energy component, reducing dwmand for the final product and hence energy
consumption. Again, if these patterns are economy-wide, it is not obvious that action in the energy
sector alone will lead to a more efficient result, compared with better information and an improved
policy framework encouraging greater competition. Where the public sector itself, as an energy
consumer, is using deficient pricing rules, it is even more difficult to argue that intervention is
necessary: the Government should rectify such distortions through policy action.
i/ The type of adopter depends on the stage which has been reached in developing the innovation.
)2/ The behavior is essentially the same as in a firm where enery costs are a small proportion of total costs.
A/ C.f. the discussion of the impact of limitations on the availability of energy conservation devices in pam.
24.


-15-
31.         The possible deficiencies in the decision-taking process identified above, especially in
para. 29, have provided a fertile ground for the debate over non-price measures to promote energy
conservation. First, the most extensive and rigorous empirical work on market imperfections has
been devoted to the calculation of discount rates revealed by consumers in their decisions about
appliance choice. Second, the most popular energy conservation measures are directed at
ameliorating these particular distortions, notably through: nformation programs; attempts to increase
the funding available for energy conservation (e.g. through demand-side bidding); and energy
efficiency standards for appliances and buildings. We discuss these subjects in detail In Sections V
and VI.
V. Empirical Evidence on Market Imperfections: Consumer Discount Rates
32.         Empirical testing of the energy consumption model underlying Fig. 1 has concentrated
on comparing the cost of capital with consumers' implicit discount rates, the latter calculated from
quantitative studies of decisions about appliance choice. The hypothesis is that, where preferences
reveal discount rates substantially in excess of the relevant cost of capital, there Is primafacle
evidence of deficiencies in the decision-taking process and of market Imperfections.
33.         An appliance choice model requires a restatement of equation (1) in para. 9, to allow for
selection between mutually exclusive alternatives, with different investment outlays (first costs) and
lifetime energy running costs. The incremental streams for investment and energy savings are then
defined as differences in investment and operating costs for the (mutually exclusive) alternatives under
comparison, which provide the same energy service. Comparing appliances A and B, the consumer
selects A if its life-cycle cost is less than that of B, i.e. If:
"*PA(QM-Qk)<
(1 -1Q) +   M         0     (5)
si   (1+rY
Here, I and Is are the initial outlays and Qm and Q& the energy consumed in year t for appliances A
and B respectively; while r is the discount rate. For example, A may have a higher investment outlay
(IA> 1) but lower annual operatng costs than B (Q  < Q). As before, the values for the
variables are consumer-specific, due, Inter alia, to differences in expectations and transaction costs.
Facing given expectatios about energy prices and appliance investment and operating costs, an
increase in r will reduce the present value of A's energy savings stream (QA - Q.) and make
condition (5) less likely to hold: first costs become relatively more Important in the calculation.
Higher discount rates may represent, for example, increased uncertainty about ftture energy prices
and the physical savings to be expected from conservation investment or the higher cost of credit.


-16-
34.         The classic empirical work on consumers' discount rates appeared in an article written
by Hausman, based on an econometric model using the type of decision-taking process embodied in
condition (S)'. The model found that, in the purchase of room air conditioners, consumers
revealed discount rates of 15%-25%. It is Important to notice, however, that the discount rates were
a function of income: low-income consumers exhibited much higher discount rates than high-income
consumers, which Is entirely consistent with economic analysis. Findings by Dubin and McFadden,
for the choice of space and water heating, are consistent with Hausman: an implied discount rate of
20% at the sample mean Income, varying inversely with the level of incomeO'.
35.         The results of these and more than 20 other studies of consumer discount rates in
appliance choice were reviewed by Train in 19850. While the results varied widely, they were
generally consistent with those of Hausman and of Dubin and McFadden, being located within a
range of 10% to 30%, except for refrigerators, where discount rates in excess of 40% were typical.
Fourteen of the studies reviewed by Train estimated the influence of income, finding (ike Dubin and
McFadden) that in all -ses the discount rates fell as income increased; three contained evidence on
the effect of ownershij status, with lower discount rates for consumers who own their homes rather
than renting them; and two provided some limited evidence that discount rates increase with the age
of the household head.
36.         Work by Hartman and Doane on weatherization investments over 1982-1983, published
after Train's survey, also included an allowance for occupancy status, age of the head of household
and income. Most of the estimated implicit discount rates were high, although income was again an
important factor, with the average household revealing a discount rate of 27%. Renters and (unlike
the studies surveyed by Train) young heads of household showed much higher rates (45%-160%) than
owners and older householders (0%-90%)#. More recent investigations, by Ruderman et. al.,
similarly found high discount ratesf'.
411 LA. Hausman, -Individual Discount Rates and the Purchase and Utilization of Energy-Using Durables,
Bel Journal of &onomIa, Vol. 10, No. 1, Spring 1979.
4V L.A. Dubin and D.L McFadden, 'An Econometric Analysis of Residential Electric Appliance Holdings
and Consumption, Econonerca, Vol. 52, No. 2, March 1984.
41 K Train, 'Discount Rates in Consumers' Energy-Related Decisions: A Review of the Literature", Energy,
Vol. 10, No. 12, 1985.
!aI R.S. Hartman and MJ. Doane, *Household Discount Rates Revisited-, Energy Journal, Vol. 7, No, 1,
January 1986.
a/ H. Ruderinan, M.D. Levine, J.E. McMahon, 'The Behavior of the Market for Energy Efficiency in
Residential Appliances, including Heating and Coohrg Equipment, Energy Journal, Vol. 8, No. 1, January
1987. The study covered 1972-1981 data for gas and oil central space heating, room and central air
conditioners, electric and gas water heaters, refrigerators and freezers. The derived discount rates were in the
range 40%-825%, except for air conditioners (20%).


-17-
37.         Although high implicit discount rates are consistent with certai types of market
imperfection, they do not constitute proof that market Intervention is necessary. Indeed, bearing in
mind the inevitable uncertainties surrounding econometric estimates, we do not find the weight of
empirical evidence on discount rates to be surprising: consumers can be expected to regard the
benefits of energy conservation as highly uncertain, partly because Ature energy prices are uncertain
but also because they will vary widely from case to case; conservation Investments are much less
liquid than many financial assets; consumers may have good reasons for preferring more energy-
Intensive appliances in their individual circumstances; the estimated energy savings may be wrong; the
marginal cost of credit may be high, for example represented by credit cards; or some consumers may
not be creditworthy, etcqw.
VI. Eney cansfnrion masnure
38.         Advocates of Government intervention have identified a range of measures to reduce the
perceived market imperfections. These comprise various programs to provide Information, improve
financing opportunities for energy conservation investments, Introduce appliance efficiency standards
and establish codes requiring certain standards of energy efficiency in buildings. We examine these
in turn.
(Information ProvI
39.         Programs to improve the provision of information about energy conservation are targeted
at reducing the perceived costs and raising the perceived benefits of energy conservation measures.
In essence, they reduce the transaction costs associated with making the energy investment decision
and, at the conceptual level, they are readily fiterpreted in relation to equation (1) in para. 9.
Operating on the left-hand side, better information about the availability of energy conservation
measures and the associated investment costs reduce the perceived value of Ala. Consumers will
invest more in energy conservation at any given energy price, to restore equality with the right-hand
side, shifting the demand curve DD' In Fig. 1 to the left. Operating on the right-hand side of
equation (1), better and more reliable information about future energy prices and physical energy
savings may raise the benefits directly or (by reducing the risk premium included In the discount rate)
0/ Among the attempts to explain the phenomenon of high implicit consumer interest rates, see D.N. Dowees,
*Energy Conservation in Home Furnaces, Energy PoUcy, June 1979; R.W. Gates, Investing in Energy
Conservation: are Homeowners Passing up High Yields?, Energy PoUcy, Match 1983; H. Charnoff,
aIndividual Purchase Criteria for Energy-Related Durables: the Misuse of Lfe Cycle Costa, EnW  JourWal,
VoL 4, No. 4, 1983; H. Ruderman et. al, 'The Behavior of the Market for Energy Efficiency in Residential
Appliances, including Heating and Cooling Equipment', op. ct.; and R J. Sutherland, 'An Analysis of
Conservation Features in Commercial Buildings,' Eny Sytem and PoUcy, 1991 (forthcoming). A. Shama,
-Energy Conservation in U.S. Buildings', op. dt., expressed the view that high implicit discount rates may
refect rational behavior if the innovations framework is adopted.


-18-
Indirectly. Again, consumers would Invest more In energy conservation, reducing the right-hand side,
until equality Is restored with Als, and DD' shifts to the left. Thus, by reducing the transaction costs
of energy conservation, information programs can lower the equilibrium energy price and quantity
supplied.
40.         Government Intervention In the energy market, to provide basic Information, may be
justified if there are economies of scale or If property rights are hard to establish. For example, it
may not be efficient for large numbers of Individual agencies or firms to develop theit own
information programs and audit methodology, when much is generic. Also, Government can educate
or inform mortgage lenders and builders, Including speculative builders, about the additional value of
energy-efficient homes. Finally, Government initiatives to require the provision of certain :ypes of
Information and to promote standardization of the form o! information provision, through appliance
labelling, have been advocated, to increase the efficiency of the market. However, such initiatives
impose costs on suppliers and require a central decision by Government on exactly which type of
information to provide (for example, to make it broader than energy consumption) and how to display
it.
41.         A particular category of information which may merit Government support is
fundamental research and even limited development In areas where It is difficult to grant property
rights. Such support is no different from the general argument for public funding of certain research
activities. Basic research on combustion and materials and the environmental consequences of energy
production and use, as well as applied research on the testing and demonstration of generic new
technologies could fall into this category. However, the role of Government In energy R&D must
always be carefully scrutinized: apparent lack of Interest or resources on the part of the private sector,
the possible past success of Government programs and the long-term nature and sometimes high risk
of R&D are not sufficient arguments for Government involvementR; and the cost-effectiveness of
energy R&D efforts in contrast with other applications of scarce Government funds should be
establisheds
42.         It is frequently argued that an energy corporation or public utility is well-placed to
market conservation services. For example, it has detailed knowledge of the energy demand of its
consumers, well-established contacts, technical expertise and some degree of goodwill and credibility.
On the other hand, the utility usually has knowledge of only one supply source and limited experience
of selling conservation services rather than supplying energy. Interest may or may not exist,
depending on the relationship between price and the marginal cost of supply or non-supply and the
A/ See R.J. Sutherland, "An Analysis of the U.S. Department of Energy's Civilian R&D Budget*, nergy
Journal, Vol. 10, No. 1, January 1989.
M/ Some of the successes of U.S. Government-supported R&D are detailed in H. Geller e. A, 0The Role of
Federal Research and Development in Advancing Energy Efficiency: a $50 Billion Contribution to the U.S.
Economyn, op. cit.. A strong case for reorienting government R&D efforts more towards end-use energy
strategies, including those supported by aid agencies, has been made by J. Goldemberg at. aL, Enery for a
Swtainable World, op. cit..


-19-
extent to which excess capacity existsO. While there is no reason why an energy supplier should
not be permitted to supply conservation services, provided that it does not take advantage of a
monopoly position and it operates profitably, energy service companies may be more suitablev.
43.         The concept of National Energy Conservation Centers or Energy Extension Services
(akin to agricultural extension services) has been proposed for developing countries, to conduct
promotional and technology demonstration campaigns, give technical advice, conduct audits or train
firms to do their own auditsk1. Such a strategy may be effective where Imperfect Information
makes conservation investments look risky or innovative to consumers unfamiliar with the technology.
As in the case of the utility (which might in practice be the same entity), such Centers should charge
for the cost of their services wherever feasible.
01) Eina     Eneg     r   vaonnkInvemets
44.         A wedge may be driven between the two sides of equation (1) In para. 9, if
imperfections in the capital market limit consumers' access to capital. Also, high labor mobility, or
the various types of owner/renter problem discussed in para. 55, may lead to an emphasis on first-
costs rather than operating costs, equivalent to high implicit discount rates in equation (1), as
discussed earlier. In these situations, financing opportunities can be designed and geared to Individual
circumstances, without resorting to subsidies (e.g on interest rates) or other special actions favoring
the energy market. Subsidies inevitably have an opportunity cost, in terms of the reduced availability
of capital to non-energy uses.
4/ The conditions under which a U.S. public utility would be interested in energy conservation, i.e. reducing
demand rather than increasing supply, are discussed in A.M. Marino and J. Sicilian, The Incentive for
Conservation Investment in Regulated Utilities', Jowal of Envkvmental Economier and Managment, Vol.
15, No. 2, June 1988. Itportant possibilities would be P < MCES, deferment of lumpy investments and
capital rationing.
,g/ The issue of supplying energy conservation services in tbe U.S. context is discussed by D. Norland and J.
Wolf, -Utility Conservation Programs: Opportunities and Strategies-, Public UtiltW Forlghty, August 8,
1985; LM. Stelzr, What Role Can Utilities Play in Energy Conservation Programs, Eer y Jow"ral, Vol. 3,
No. 1, January 1982; and M.C. Whittaker, Conservation and Unregulated Utility Profits: Redefining the
Conservation Market, Public Udilar Fornghdly, July 7, 1988. A good discussion of the comparative
advantages of technology-oriented energy service compeanies is in R.H. Williams, lrnovative Approaches to
Maketing Blectric Efficiency*, in T.B. Johansson, B. Bodlund and R.H. Williams, Electriefty: ficient End-
Use and New Generation Technologles and their PlannAg INpicatiow, Lund, Sweden, 1989.
I J.R. Gamba et. aL, Indwtri Energy Ratlonalization in Developing Countries, op. ce.; and J. Goldemberg
t. aL, Energy for a Sustainable World, op. cit..


-20-
45.         Various types of shared savings contracts are possible to deal with the problem of
consumer finance for energy conservation. Essentially, an energy service company (which could be
the public utility itself) agrees to install the energy conservation measure at its own expense, while the
consumer contracts to reimburse the company at regular intervals, say monthly, from the estimated
energy savings. If these savings are as large as claimed, there should be ample scope for both the
consumer and the company to make a profit. Appliance rental and lease options are similar ways to
deal with the perceived aversion to high first-cost, as well as the problem of those who move
frequently. The public utility or a private rental/lease company finances the initial appliance cost and
hires or leases the appliance on, say, a monthly basis.
46.         A particular form of shared savings contract has been proposed by Cicchetti and Hogan,
who point out that it is possible to *unbundle* energy supply, such that the energy conservation
services which are sold to the utility are dWinguished from the energy services retained by the
customerI. The Punbundling* makes It possible to differentiate the costs of investing in energy
conservation (demand-side alternatives) from the benefits of erjoying the same level of services, at a
lower level of energy consumption. The former reduce the quantity of energy which the utility needs
to supply and the costs would therefore be reimbursed by the utility. The latter, In contrast, are
enjoyed by the consumer, who (under the Cicchetti-Hogan scheme) continues to pay the supplier for
an equivalent amount of energy supplied. In effect, the consumer pays for the energy which is
conserved (not consumed), while the supplier pays for the cost of the investment which was necessary
to obtain that conservation.
47.         There are serious practical problems involved with the Cicchetti-Hogan approach, e.g.
because the energy service relates not to real energy but to estimates of energy not supplied and the
scope for misrepresentation on the part of the energy service company will be clearly; but Joskow
has demonstrated that, from an analytical point of view, the "unbundling* approach is entirely
consistent with the standard competitive model, as described In Fig. 1, because the consumer pays the
full cost of the conservation service' and the utility is merely acting as an intermediary between the
conservation supplier and the energy consumer. However, Joskow correctly underlines two critical
components in the approach: first, the utility pays no more than the equivalent MCES for the
conservation project; and second, the consumer who benefits from the conservation project continues
to pay the utility for the energy which would have been consumed without the project. Joskow then
concludes that the scope for shared savings is limited to the difference between MCES and P, which
of course is zero in the properly-functioning competitive model.
I/ C.J. Cicchetti and W. Hogan, -Includig Unbindled Demand-side Options in Electric Utility Bidding
Programs, Public Utites Fortnightly, June 8, 1989.
&I CJ. Cicchetti and W. Hogan, Ibt. Some of the practical problems are discussed fi*ther in M.R.
Hoover, J.S. Garces and LS. Ridge, wDemand Side Bidding: A Practical View, Public Ustiles Fortalghtly,
June 21, 1990.
M1 P.L. Joskow, 'Understanding the 'Unbundled' Utility Conservation Bidding Proposal", Public Utiliti
Fortnightly, January 4, 1990. See also L. Ruff, "Environmental Protection though Energy Conservation: a
Free Lunch at Last?', Energy and the Environment In the 21st Centwy, M.LT. Press, Cambridge, MA, 1990.


-21-
48.         In terms of the basic energy conservation model underlying Fig. 1, the maximum
payment v which the energy supplier can make to the conservation supplier Is MCES. Under the
Cicchetti-Iogan scheme, the maximum savings to be shared between the conservation supplier and
the consumer (who could be the same person) are correspondingly w-MCEC = MCES-MCEC =
MCES-P. Evidently, at the equilibrium PO,QO in Fig. 1, there is no incentive for this type of shared
savings contract, as MCES=P. Nevertheless, Joskw suggests that the energy supplier has a useful
role to play by, In effect, acting as a broker between the energy consumer and the energy
conservation supplier, e.g. providing a convenient way to amortize the investment cost of the
conservation measure, through the monthly energy f110; and to some extent ameliorating
informational barriers by giving its "seal of approval" to the conservation measure. He also suggests
that the energy supplier "can offer subsidies to correct pricing distortions*, if marginal cost pricing
does not exist, up to a limit of MCES-P, although we have argued that it Is better public policy to
attack pricing distortions directly rather than resort to subsidies or other market Interventions.
49.         Evidently, not all forms of shared savings contracts and rental/lease agreements meet the
test of covering costs. Where the Government or public utility is Involved, there is the danger that
some subsidy will be made for the rented appliance or energy conservation service'. A particular
case in point is a version of the shared savings contract which has recently appeared in the literature
and generated a lively debate in the context of the U.S. electricity Industry. It focusses on the
proposition that it is more economic to conserve energy than to produce it; and the perceived need, in
consequence, to include energy conservation alternatives along with supply alternatives in an "all-
source" bidding program, to find the least-cost way to supply future electricity demands. One of the
strongest proponents of buying "negawatts" on an equal footing with kilowatts has been Lovins'.
The argument is deceptively simple: since the public utility would pay for extra kilowatts of capacity,
why should it not be willing to pay an equivalent amount to reduce consumption, i.e. to buy
"negawatts"?. "After all, a kilowatt-hour saved is just like a kilowatt-hour generated, only cheaper,
and can be resold to someone else, so they should be treated alike-. The payment could be made
directly to consumers or to energy service companies on their behalf, to cover some or all of the
investments necessary to implement energy conservation measures. In effect, the utility would be
a/ In effect, being the vehicle to amortize the investment shown in equation (4), par. 10.
I/ Some interesting examples from U.S. utilities in Pennsylvania and New Jersey are given by M.A. Brown
and D.L. White, 0Stimulating Energy Conservation by Sharing the Savings: a Community-based Approach",
Environmen and Planning A, Vol. 20, 1988. These shared savings contracts are three-way agreements,
involving the public utility, an energy service company and the consumer. They are unfortunately not
transparent with regard to the distribution of benefits and costs.
E1 See A.B. Lovins, "Saving aigabucks with Negawatts", Puble Utlkes Foanlghly, March 21, 198S.
B/ A.B. Lovins, Ibid, p. 24.


-22-
paying the consumer to invest in, for example, more efficient appliances, higher thermal efficiency
and autogeneration .
50.         At first sight, the Lovins proposal appears similar to the Cicchetti-Hogan scheme.
However, the problem with the "negawatt" auction is that it eliminates the second of the two critical
components identified by Joskow (see para. 47). Turning to our basic energy conservation model:
while the payment for a supply-side alternative covers its cost (MCES), the cost of an economical
energy conservation measure (MCEC) is already compensated through the reduction in the customer's
bill on the kilowatthour savings (P). If the payment for the energy conservation measure is
w=MCES and there is no off-set in the consumer's bill, the subsidy received by the energy
conservation supplier and energy consumer jointly becomes (in contrast to par  48) v+P-
MCEC=MCES+P-MCEC=2P-MCEC: there is a windfall gain, which would encourage inefficient
energy conservation measuresf. Furthermore, if the payments for energy conservation measures
are directed only or mainly at one form of energy (such as electricity), which Is likely to be the case,
consumers' inter-fuel substitution decisions will be distorted.
51.         Special credit programs are sometimes proposed for low-licome consumers on the
grounds that they have insufficient access to financial markets. Yet there is no obvious reason why
such arrangements should be made for energy marketr as opposed to other markets. The financial
market is normally assumed to assess adequately the credit-worthiness of borrowers and charge
appropriate Interest rates to reflect the risk of default. Similarly, direct Income support for poorer
consumers, for example through tax credits, free audits and grants, as has been provided under U.S.
legislation, is hard to justify, without evidence that they increase the welfare of poor consumers more
than other types of intervention.
§2/ The flexibility benefits of demand-side alternatives are also emphasized in the literature, from a system
planning perspective, because they can be adjusted in small increments. See E. Hirst, Tlexibility Benefits of
Demand-Side Programs in Electric Utility Planningl, EneWy Jowal, Vol. 11, No. 1, January 1990; and A.B.
Lovins, "Saving Gigabucks with Negawatts, op. it. However, there is nothing special about such programs:
marginal cost pricing, for example through interruptible loads, and small-scale supply options, such as gas
turbines, have the same feature.
(/ The point is well-put by C.J. Cicchetti and W. Hogan, Including Unbundled Demand-Side Options in
Electric Utility Bidding Progamst, op. c.: In the cas in which the utilfy pays the customer to make the
conservation investment .... the utility buys the 'supply' of conservation, but the customer receives the energy
service produced by that investment free! The incentive given the customer is wrong. He is paid twice: once by
giving him the dollars for the conservation device, again by providing a free service. This misplaced incentive
could result in inefficient over-investment in conservation. In practice, the value of the subsidy depends not
only upon the relationship between MCES, MCEC and P: it also reflects the way in which a utility's payments
for conservation investments translate into a higher average revenue requirement, through the inclusion of the
cost of those investments in the rate base. With P=MCES, inefficient energy conservation measures will raise
the average revenue requirement and hence P but the consumer's bill will still fal so that he is indeed
compensated wice for his expenditure on the (subsidized) energy conservation measure. For a numerical
example, see LE. Ruff, OEnviroamental Protection through Energy Conservation: a Free Lunch at Last?", op.
dt., pp. 21-22.


-23-
52. The argument is sometimes heard that the market interest rate, which appears in equation (1) In
para. 9, exceeds the social rate of discount, causing a bias towards first-cost. If true, it requires
action throughout the economy, not confined within the context of energy conservation. It is not a
solution to treat the decisions of individual consumers, concerning the disposition of their own
Incomes In energy markets, as if they were public investment decisions, by maklhg calculations on
their behalf at discount rates determined by central planners4'.
(111) Applanc E=dln Standard
53.         In terms of condition (5) In para. 33, appliance efficiency standards shift decisions away
from certain appliances with lower investment costs and higher operating costs by, In effect, attaching
a very high or Infinite value to the investment cost of the appliances excluded under the regulations.
As pointed out by Hausman and Joskow, the ensuing reduction In consumer choice can be presumed
to decrease welfareV. For example, there are welfare losses due to consumers being forced to
incur higher investment outlays to *enjoy* lower annual energy costs, even assuming that the
engineering calculations turn out to be more correct than those which have Implicitly been made by
consumers. Referring to condition (5), mandatory standards preclude consumers with values of r
above a critical value from maximizing their welfare'. Second, appliances may be removed from
the market which have features perceived to be desirable by consumers, but which are not understood
by the central planners. The assumption underlying the "life-cycle* cost comparisons in condition (5)
is that A and B provide an equivalent quality of energy service. In reality, additional terms are
needed to attach monetary values to differences in the services provided by A and B. Third,
standards by definition must cover wide areas, although conditions of climate, energy prices, tastes
and habits can vary widely, thereby underlining the consumer-specific nature of the P, Qu and Qa in
condition (5)W. Even on a voluntary basis, appliance efficiency standards may serve to limit
jl It is not inconsistent to use a discount rate for public investment decisions which is different from the ones
used in the private sector.
2/ J.A. Hausman and P.L. Joskow, 'Evaluating the Costs and Benefits of Appliance Efficiency Standards-,
American Sconomic Assoclation (Papes and Proceedings), Vol. 72, No. 2, May 1982.
gi/ These welfare losses include income distribution effects, as the higher first-cost of more energy-efficient
appliances will impose a larger burden on lower-income consumers, who in any case tend to have hisomr
discount rates (see paras. 34, 35, 36). It is not sufficient to argue that the higher first-cost will be repaid in
energy savings: lower-income consumers have limited access to capital and the funds may yield higher benefits
to them spent in other ways. Writers frequently overlook this point (c.f. P. Rollin and J. Beyea, *US Appliance
Efficiency Standards", Energy Policy, October 1985).
f./ For example, an individual's tolerance to heat or use of specific rooms may lead to only occasional use of
certain room air conditioners; and with a move to peak load pricing, less efficient appliances may be appropriate
outside peak hours, because the costs during those hours are lower than average. Appliance standards cannot
accomnodate this type of economic effect. See J.A. Hausman, 'Individual Discount Rates and the Purchase
and Utilization of Bnergy-Using Durables*, op. ck.; and J.A. Hausman and P.L Joskow, 'Evaluating the Costs
(continued...)


冷｛’哗，…，：………｛


-25-
empirical evidence that ownership status afficts die Implementation of energy conswadon
measuresN. Rental units can be metered individually for electricity consumption or theta may be a
master meter, with the owner responsible for payment. In the former case,, theta Is no direct Incentive
for the owner to install energy conservation devices; and individual renters have only limited ability,
beyond behavioral changes, to affect energy consumption, as they are unlikely to invest In energy
conservation devices without lonkterm, tenure. Where a master mew Is provided. then is no
incentive for individual renters to conserve, as all energy costs are absorbed and avenged in overall
rents, so that die effective marginal price is zero. Hence, households In midtifamily dwellings with
central system (e.g. for water and space beating and for air-conditioninj) treat conservation as a
collective consumption good-&. While the role of energy costs In the reatal market Is complex,
individual metering of housing units would appear to resolve mmy Issues of energy conservation. In
the final analysis, the level of rents (where there Is master metering) and the level of energy bills for
individual rental units will affect the demand for those units; W competition between property
owners should therefore encourage investment in conservation devices, if they are economically
justified. Of course, the need for prices to reflect accurately the economic costs of supply at the
master or individual meters must again be underlined. 7be potential impact of information campaigns
should not be underestimated: if retrofit benefits exis4 and renters and owners are made aware, it a
not difficult to find ways to act Jointly to realize the benefits to their mutual advantage.
56.         More generally, even in the non-rental residential market, it is sometimes claimed that
owner-occupiers as a group choose energy-Inefficient buildings, just as they buy appliances which are
inefficient. The same arguments which were proposed against appliance efficiency standards In the
last section can be levelled against building codes and ordinances In the non-rental residential market,
but there Is also empirical evidence that home buyers are willing to pay higher market prices for
energy-efficient homes, i.e. they capitalize the energy savings in the sales price, so that a raw of
return can be earned on residential Investment In energy-conservationP . Competition among
builders can be made more effective by informing them and extending the concept of labelling to now
homesw.
U/ Sw K. Train, -Discount Raw in Consumers' Energy-Related Decisionsp, op. ck.; ad R.S. Hartman sad
MJ. Dom, 'Household Discount Rates Revisited% ov. dt..
(& E.T Fujii and 1. Ma1c, 'A Model of Household Mectric* Conservation BdkrAoe, Land Emnonda, VoL
6C,, No.          .1984.
2Q/ M.J. Horowitz and ff. Haeri, "Economic Efficimey v E=V Efficiency: Do Model Conservation
Standards Make Good Sem?", En&V Eeonomiw, April 1M, estimated a red tab of return of 8% an die
PC
resale of energy efficient homes. M. Longstreth, A& Coveney and J.S. Bowers, 'Onservation
Characteristics Among Determinants of Residential Property Value% JOW7W Of COWWW AW6arCh, VOL llo
june 1984, demonstrate dat basic enwV conservation measures to improve a home"s thwaW integrity, eg.
adding wall and ceiling insulation, increase reWe value. K. Train, op. dt., vqmu an ORNL study (R.
Johnson, 'Housing Market Capitalization of Energy-&ving Durable Good Investmento, Report ORNUCON-74,
1981) wbich calculated an implicit discount rate of about 4% applied to energy efficimey in the pufcbm of
homes.
U/ Recognizing that labelling involves transaction costs, as mentioned in pars. 40.


-26-
57.         In the non-residential market, it has not been explained why business investment in
buildings should be less rational than in other sectors. If there is widespread failure to give sufficient
attention to energy conservation, much more needs to be known about the reasons, as a basis for
tackling them in a more comprehensive fashion. With regard to the builder/owner/occupler d49emma,
Sutherland's recent study of energy conservation features in commercial buildings, based on a 1986
nationwide survey conducted by the Energy Information Administration of the DOE, concluded that
there is no significant difference between the number of conservation features installed by owners of
commercial buildings compared with renters, "which Implies that conservation strategies are not
impeded by the renting or leasing of buildings"N.
VII. The ffecivnesof Eey Consevation Measures
58.         Quite apart from the conceptual problems in justifying Government intervention in the
energy market, evidence on the effectiveness of energy conservation measures is not conclusive.
Researchers at the Oak Ridge National Laboratory (ORNL) cite significant savings from a program of
free residential audits and interest-free residential weatherization programs for the Tennessee Valley
Authority, amounting to 2.5 GWh and 1000 MW for 1987. They refer also to ORNL studies of the
corporate average fuel economy (CAFE) standards, which attributed savings of 230,000 barrels per
day to CAFE by 1980; and the Department of Energy conservation programs in effect in 1980, to
which savings of 3 million barrels per day by 2000 were imputed, over and above what the market
would yield-. More recently, ORNL stated that an additional Improvement In energy efficiency of
14% could be achieved by 2010 through cost-effective conservation measures, compared with the
gains that could be achieved through the market alone-; and studies by the staff at the Lawrence
Berkeley Laboratory of the impact of the appliance standards in the National Appliance Energy
Conservation Act (NAECA) of 1987 and the 1988 Amendment, yield impressive results. Also in
a recent study, Greene compared the effects of CAFE and fuel prices in terms of a manufacturer
decision-taking model, utilizing a penalty function to reflect the impact of deviations from CAFE
.U/ R.J. Sutherland, "An Analysis of Conservation Features in Commercial Buildings", op. ds.
71/ E. Hirst et. aL, Improving Energy Efficiency: the Effectiveness of Government Action", op. cit..
24/ R.S. Carsmith er. aL, "Energy Efficiency: How Far Can We Go?", Op. cit..
21/J.E. McMahon, D. Berman, P. Chan, T. Chan, J. Koomey, B. Labot, M. Levine, S. Stoft and I. Turiel,
*Economic and Energy Demand Impacts of U.S. Appliance Efficiency Regulations on Consumers,
Manufacturers, and Electric Utilities', Paper Presented at the 2drtwenth Annual International Couftence of the
International Association for Energy Economicr: Integrated Energy Mar*ers and Energy Synems, Lessons and
Perspecives, Copenhagen, Denmark, June 19-21, 1990.


-27-
standards and market equilibrium demand on automobile manufacturers' costs. Greene concluded that
CAFE standards did have a significant impact on fuel efficiency In new cars.W
59.         These results seem impressive. However, other studies have concluded that the expected
and realized energy savings attributable to energy conservation programs may be over-estimated; or
market solutions might give similar results. Some of these studies are reviewed In paras.
60, 61, 62, 63, 64. Section VII then concludes (para. 65) with some observations on the
measurement problems faced by all empirical studies on the effectiveness of energy conservation
measures.
0) The Rebound Effect
60.         An important reason why both anticipated and realized energy savings may be over-
estimated is the "feedback" or "rebound" effect. The basic energy conservation model in equation (1)
in para. 9 and condition (5) in para. 33 operates as if the expected additional physical energy savings
in year t - A or (Q - QW) - are independent of the energy price P. In fact, we can expect a
"rebound" from carrying out an energy conservation measure. The demand for energy is derived
from the demand for the services which it yields and an increase in the efficiency of an energy-
utilizing device is equivalent to a reduction in the price of the energy which it uses. The benefit
which occurs as a result of this effective decline in the price of energy can be enjoyed as: an increase
in disposable income (i.e. reducing physical energy consumption but maintaining the same level of
energy service at lower cost); an increase in the level of energy service (I.e. increasing physical
energy consumption and enjoying a higher level of energy service for the same cost); or a
combination of both. This holds true not only when the efficiency of a motor vehicle engine or air
conditioner is increased. The application of weatherization in a house, for example, has a similar
effect to an increase in furnace efficiency. The more elastic the utilization of a given appliance with
respect to changes in the price of the energy which it uses, the more significant the "rebound" effect
and increases in efficiency could in theory increase total energy consumption. While the "rebound"
effect is clearly welfare-Increasing - as in the standard case of a price decrease - the "rebound"
analysts are emphasizing that quantitative studies which try to predict the results of energy
conservation programs should take into account the ensuing adjustment in consumption.
61.         A variation of the "rebound" effect has been ncted with regard to time-of-day pricing In
the electricity sector'. High peak-load prices combined with low off-peak prices will change the
pattern of appliance use. If off-peak electricity becomes sufficiently cheap, the consumer might
choose a less efficient appliance for use predominantly off-peak; similarly, if use during peak drops
7A/ D.L. Greene, 'CAFE or Price?: An Analysis of the Effects of Federal Fuel Economy Regulations and
Gasoline Price on New Car MPG, 197-89", Energy Journal, Vol. 11, No. 3, July 1990.
2/ J. Hansman and J. Trimble, "Appliance Purchase and Usage Adaptation to a Permanmnt Time-of-day
Electricity Rate Schedule, op. ct..


-28-
sharply, due to higher electricity rates, it may again be more economic to choose a less efficient
appliance.
62.         The theoretical existence of the "rebound" phenomenon is clear; the issue concerns its
numerical ImportanceW. Khazzoom has argued that the 'rebound" effect Is significanty. For
example, he has shown, on the basis of 1969-1980 data In Sacramento, California, that half the
energy savings which resulted from the increased efficiency of electrical appliances was lost after four
years; in the long run, less than 30% remained. In relation to CAFE standards, he also cites work
by Blair, Kaserman and Tepel, to the effect that, in the short run, 30% of the potential reduction in
gasoline consumption ia offset by increased travelfs. Dubin, Miedema and Chandran concluded
thAt engineering estimates of reductions in electricity consumption for heating and cooling, caused by
improved technologies, could be over-estimated by 8%-12% and 13% respectivelyV. An
evaluation of the energy-conservation services provided in Connecticut, under the 1978 National
Energy Conservation Policy Act (NECPA), found that estimates of the program savings were
statistically Insignificant and well below the estimates shown in the audits (2 million BTUs per
participant for the former against 16 million for the latter), partly due to households taking some of
the gains in the form of increased comfortg. The unreliability of engineering estimates, due to the
*rebound* effect, has also been indicated in work on weatherization measures in the Pacific
NorthwestW. A particularly cautious view on the effect of energy audits on energy consumption
2V/ Some discussions of the theoretical aspects of the 'rebounda effect can be found in M. Enhorn, Economic
Implications of Mandated Efficiency Standards for Household Appliances: an Extension', Energy Journal, Vol.
3, No. 1, January 1982 and "PEA Efficiency Standards and Energy Usage: Resolving Khazzoom versus Beson
and Johnson", Enery Journal, Vol. 5, No. 1, January 1984; and F.C. Bold, 'Responses to Energy Efficiency
Regulations', Energy Journal, Vol. 8, No. 2, April 1987.
72/ See, for example, D. Khazzoom, 'Economic Implications of Mandated Efficiency in Standards for
Household Appliances', Energy Journal, Vol. 1, No. 4, October 1980; 'The Demand for Insulation - A Study
in the Household Demand for Conservation*, Enerjy Journal, Vol. 8, No. 3, July 1987; 'Energy Savings
Resulting from the Adoption of More Efficient Applianceso, Energy Journal, Vol. 8, No. 4, October 1987;
Energy Savings from More itticient Appliances', znery Journal, Vol. 10, No. 1, Januw:y 1989; and An
Eonometric Model Integrating Conerwtion Measur in the Residential Demand for Electriciy, JAI Press,
1986.
jg/ R. Blair, D. Kaserman and R. Tepel, 'The Impact of Improved Mileage on Gasoline Consumptions,
Economic Enquiry, April 1984, cited by D. Khazzoom, 'The Demand for Insulation - A Study in the Household
Demand for Conservation', op. dt.
J/ J.A.Dubin, A.K. Miedems and R.V. Chandran, 'Price Effects of Energy-Efficient Technologies: a Study
of Residential Demand for Heating and Cooling', Rand Journal of Economics, Vol. 17, No. 3, Autumn 1986.
JV E. Hirst, P.S. Hu, E.F. Taylor and LM. Thayer, 'Connecticut's Residential Conservation Service',
Enery Pollcy, February 1985. Another explanation given by the authors was that the audits assumed that
conservation measures would be implemented simultaneously and not in isolation.
12/ G.A. Leaf, 'The Effectiveness of Weatherization Measures in the Pacific Northwest', Public Utiliter
Fortnightly, April 12, 1984. The failure to implement all the measur  assumed in the audit was an additional
explanation.


-29-
was given by Frankel and Duberg, who maintained that the savings in consumers' energy bills do not
match the investment costs required to produce them, as homeowners respond by raising thermostats
after implementing the audits' recommendations&. Finally, a study of changes in the energy
consumption of consumers who have installed energy-conserving devices, by Longstrath and Topliff,
throws additional light on the complex issue of consumers' response to energy conservation
programsW. Using the DOE's Residential Energy Consumption Surveys, it was found that 49% of
consumers who installed retrofit measures in 1982 did not reduce energy consumption between 1981
and 1983. These results were attributed partly to consumers increasing their non-heating and non-
cooling uses, to compensate for the savings that they enjoyed in their heating and cooling bills. The
authors also argued that there were differences between the groups of energy savers and non-savers,
in terms of the price changes they faced, house size, household size etc. We shall return to some of
these statistical issues shortly.
(ii) Mart-Based Istrumeints
63.         The possibility that market solutions can yield the same results as efficiency standards
was broached by Mayo and Mathis, who concluded that the CAFE standards had no Independent
statistically significant impact on the efficiency of the passenger vehicle fleet or the demand for
gasoline from 1977 to 1983W. Khazzoom also concluded, on the basis of his work on home
insulation in the Sacramento area, that conservation measures were implemented in response to
economic incentives (notably energy prices), without the need for Government regulationsu. In its
comments on the amendment to the NAECA of 1988, the U.S. Administration pointed.to the
significant gains made by manufacturers of energy appliances in the 1970s, without Federal
standards,W.
64.   In a 1981 study of the cost-effectiveness of energy conservation measures, Hartman produced
evidence that market solutions can lead to gains in appliance efficiency which are commensurate with
the gains achievable through standards, perhaps as a result of information campaigns and higher
p/ M.L. Frankel and J.A. Duberg, 'Energy Audits as an Investmnent: the Residential Conservation Service
Program Analyzeda, Public Utilita Fortlghtly, April 12, 1984. Furthermore, higher energy prices may
already have induced homeowners, including those who were not audited, to implement conservation measures
(the methodology was based on comparisons between audited and non-audited homeowners); while the
engineering estimates of the potential savings could have been too optimistic, even aside from the "rebound"
effect.
g/ M. Longstreth and M. Topliff, "Determinants of Energy Savings and Increases after Installing Energy-
Conserving Deviceso, Energy, Vol. 15, No. 6, June 1990.
/ J.W. Mayo and J.E. Mathis, -The Effectiveness of Mandatory Fuel Efficiency Standards in Reducing the
Demand for Gasoline", Applied Economics, Vol. 20, No. 2, February 1988.
SI 'The Demand for Insulation - A Study in the Household Demand for Conservation', op. ci.
ja/ Senate Report (Energy and Natural Resources Committee) No. 100-345, May 13, 1988, p. 6.


-30-
pricesw. Using the ORNL model, he predicted the effect of DOE efficiency- standards for 1980 on
average appliance efficiency for space heating (electric, gas and oil), room air-conditioning, water
heating (electric, gas and oil), refrigerators, freezers and cookers (electric and gas). The results were
compared with estimates for three alternative scenarios: reliance only on market prices; a 100% fuel
tax; and information campaigns to reduce perceived market imperfections. He concluded that "only
for refrigerators, electric ranges and electric water heaters do the standards generate greater efficiency
gains than all 'market' scenarios". A study by Corum and Van Dyke also looked at the effect of
alternative policy instruments, this time to achieve a given level of energy efficiency in commercial
buildings by the year 2000, viz., mandated building efficiency standards, tax credits and fuel
taxes2w. It was found that a fuel tax of 15% or a tax credit of 40% could be expected to give the
same results as the building efficiency standards required by the 1978 NECPA. Each of these
measures has different welfare implications, with building standards reducing consumer choice. On
the other hand, econometric testing by Walsh produced no evidence that tax credits had any Impact on
energy conservation investments, as a result of ignorance about the availability of the credits, failure
to understand them, their insufficient size to make a difference and too much paperworkW.
(III) Measurement Problems
65.         The inconclusive nature of the empirical evidence on the impact of energy conservation
measures may not be surprising, in view of the measurement obstacles involved. An excellent
discussion of the econometric problems has been provided by HartmanW. He points especially to
the need to establish appropriate control groups, since often the impact of the measures is estimated
by the difference between the energy use of those who took part in a conservation program and thrue
who did not. Self-selection bias arises, because certain types of household have a predisposition
towards taking part in the program, for example due to their level of energy consumption, appliance
holdings, and family characteristics (notably education, income, number of residents, and age of
householder). Non-participants may be "early adopters" who have already started to implement
conservation measures. Indeed, Hartman and Doane confirmed the importance of a wide range of
factors when they looked at the Residential Conservation Service Programs of the Portland General
Electric Company (PGE), following introduction of the 1978 NECPA2. Using 1977-1981 data,
during which PGE offered audits and zero-interest loan programs for weatherization investments, they
found that only a small percentage (8%-36% or 25% for the average household) of the observed
§2/ R.S. Hartman, "An Analysis of Department of Energy Residential Appliance Efficiency Standards", Enery
Journal, Vol. 2, No. 3, July 1981.
2/ K.P. Corum and J. Van Dyke, "Stimulating Energy Conservation in Buildings', Enet Polley, Mar&c
1983. The ORNL model was used to simulate the results.
91/ M.J. Walsh, "Energy Tax Credits and Housing Improvement", Energy Economic, October 1989.
2Z/ R.S. Hartman, "Energy Conservation Programs: the Analysis and Measurement of their Effects', Energy
Policy, October 1986.
21/ R.S. Hartman and M.J. Doane, "The Estimation of the Effects of Utility-Sponsored Conservation
Programmes", Applied Economicr, Vol. 18, No. 1, January 1986.


-31-
reductions in electricity use could be attributed to the programs. An Interpretation in terms of the
underlying psychology of consumers, rather than econometric specification, Is given by Stern, Berry
and Hirst. The authors show how the size, type and uon-financial features of incentives may have
an Impact on a conservation measure beyond changing its effective cost. For example, financial
incentives may serve to attract attention, in which case increases In size beyond a certain point may
have little or no Impact; or they may be noticed mainly by those looking for them, so that the effect
would be to speed the rate of adoption rather than the levelW. Different types of financial incentive
may be perceived quite differently, although the financial effect may be Identical (e.g. manuacturers
rebates, grants, subsidized loans, tax rebates, etc.). Finally, a range of features which are part of the
financial package are critical: its promotion, simplicity of procedures, reliability of the associated
work and trust in Its outcome.
VII. Exenlt
66.         In examining the debate over Government intervention in the energy market, to promote
energy conservation, we have so far focussed strictly on the functioning of the energy market itself, in
order to keep separate the effects of externalities (see para. 8). Evidently, market solutions will not
be socially optimal, If external costs in the production or consumption of energy are not reflected In
MCES and price. Two externalities have been a particular preoccupation in the debate over energy
conservation: the production and consumption of energy may have deleterious environmental impacts,
which impose costs on the national or trans-national community; and the consumption of imported
energy may involve a reduction in national security, due to vulnerability to supply interruptions, again
representing a welfare loss to society. In this argument, environment and national security are viewed
as collective goods .
(1) Natonal Scrit
67.         The argument that energy conservation increases national security, by reducing energy
imports, requires close scrutiny. First, energy is only one commodity in a complex system of
interational tr2de which would be vulnerable to disruption in a time of crisis. Second, intervention
j/ P.C. Sterm, LO. Berry and E. Hirst, "Residential Conservation Incentives", Energy Policy, April 1985.
M/ As the authors point out, if non-participants are averse to borrowing, zero-interest loans which emphasize
the indebtedness aspect of the conservation measure are not likely to be successful.
2/ For example, see A.C. Fisher and M.H. Rothkopf, "Market Failure and Energy Policy*, op. cit.; E. Hint
et. aL., "Improving Energy Efficiency: the Effectiveness of Government Action", op. cit.; L. Schipper et. al.,
"The National Energy Policy Conservation Act: An Evaluation", op. ck.; and J.H. Gibbons and W.U.
Chandler, Energy: te Conserw~ton Revoltion, Plenum Press, 1981.


-32-
has to be targeted at the specific energy products which are Imported at the mArgin. Generic energy
conservation is not selective, affecting indigenous as well as imported enerry. A selective tax, e.g on
gasoline, could reduce imports, but may not be as cost-effective as other types of action. A public
stockpile or Strategic Petroleum Reserve, flly financed by a levy on imported oil may be a preferred
alternatives'. Third, past experience shows that a policy of greater self-sufficiency can carry high
economic costs, which are particularly hard to justify in the case of developing countries. Brazil's
experience is instructive. Following a determined drive for greater energy Independence, energy
investment increased from 12% of all investment and 2.5% of GDP in 1973 to nearly 20% of all
investment and more than 4% of GDP in the early 1980sW. The nuclear power program has
produced virtually no electricity, despite investments of several billion dollars'; and while the
alcohol program has been highly successful in displacing gasoline consumption, the program's total
investment had reached US$7 billion by the end of 19870. Furthermore, the costs of greater self-
sufficiency were not always transparent: most of the fiscal transfers to the ethanol program failed to
appear in the public budget, since they were indirect, represented by the loss of tax revenues through
reduced gasoline sales.
(II) Envrnment
68.         Of all the arguments for Government intervention in the energy market, environmental
externalities constitute the most convincing and there is scope for making more effective use of
economic incentives to reduce adverse environmental externalities. For this reason, it is crucial to
distinguish between environmental harm caused by energy consumption, on the one hand, or by
production, on the other, so that the remedies can be chosen accordinglym'.
69.         In general, natural gas and petroleum products are of more concern than electricity in
terms of consumption; the opposite is true for production, even granted the increasing problems raised
by oil spills. Coal has harmful production and consumption effects. Much of the world's air
pollution, acid rain and global warming are attributable to the consumption of hydrocarbons, notably
carbon monoxide and dioxide, sulfur dioxide and nitrogen oxide1'. The production of electricity
imposes environmental costs, partly because it is a consumer of harmful energy products as
intermediate goods; but nuclear plants also pose environmental issues, in use and in the final disposal
of radioactive materials. Hydroelectric plants are associated with a range of environmental problems:
22/ See M.A. Adelman, `Coping with Supply Insecurity", Energy Journal, Vol. 3, No. 2, 1982.
2A1 Based on data from the Ministfrio des Mines e Energia, Comisslo Nacional do Energia, Subgrupo do
ReferOncias Bdsicas, Reladrlo Final, March 1989.
22/ Ministdrio des Minas e Energa, Balanfo Energdrico Nacional 1988, Table 1.3, Brasiia, 1988.
j1l Coislo Executiva Nacional do Alcool (CENAL), The National Alcohol Program, May 1988.
10a/ C.f. L Ruff, "Environmental Protection through Energy Conservation: a Free Lunch at Last?*, op. cit..
j/ See, for example, World Resources Institute, World Raourea 1990-91, op. cit..


-33-
the risks of catastrophic floods to people and property situated downstream; the flooding of land for
reservoirs, which may displace people, create health hazards from water-borne diseases, such as
achistosomsis, and threaten human cultures, the ecological balance, cultural treasures and wildlife;
and the Impact on people and ecology which occurs during the development of the affected areas as
construction proceeds.
70.         Since these consumption and production effects vary between energy products and
production plants, generalized reductions in energy consumption are not a cost-effective way to deal
with them. It would be more efficient to tackle the externality directly, either through the use of
market-based incentives (MBIs) or command and control (C&C) measures. MBIs cover economic
instruments, such as taxes on emissions, subsidies to abatement measures and tradeable pollution
permits. C&C measures are typically embodied in regulations governing activities which inflict
environmental damage. There is a considerable body of theoretical and empirical literature to support
the view that MBIs are generally more cost effective than C&C measures, as the former encourage
consumers and producers to seek least-cost ways of meeting environmental goals, through adjustments
in the level of consumption and production and through inter-fuel substitution'. For example, in
the case of pollution, they help to equalize the marginal costs of pollution control across all sources:
firms with more expensive pollution control costs will bid up the price of tradeable permits, until (at
the margin) permission to pollute costs the same as pollution control or paying a pollution tax. It is
extremely difficult to achieve the same result through C&C, because regulation requires a
considerable amount of information about the pollution costs of the different emitting sources.
Pollution taxes may also be less distortionary than other taxes in meeting a public sector revenue
constraintkv. As Tietenberg has noted, there has been widespread recognition among leading
environmental groups "that the power of the market can be harnessed and channelled toward the
achievement of environmental goals, through an economic incentives approach to regulation"N.
Even when MBIs may be difficult to Implement, e.g. because of high monitoring and enforcement
costs, it will be more efficient to use indirect instruments, such as the taxation of polluting 'tels,
rather than energy conservation to handle environmental externalities.
PY-1.      T" *!!       I-trate th ises, it is  ctv tolk somc of the public policy alternatives
in the case of Brazil. Growth in electricity demand in Brazil is expected to be met mainly from
hydroelectric power and the potential environmental impacts will vary from the very serious to
relatively benign. However, increasing Interest is being shown in developing non-hydroelectric
alternatives, utilizing bagasse, fuel oil, coal and natural gas. The result will be a complex structure
of marginal costs, including environmental costs. Natural gas and bagasse are environmentally the
I/ See, for example, G. Eskeland and E. Jimenez, "Curbing Pollution in Developing Countries', Finawe
and Development, March 1991; and Choosing Pole Instranents for Pollution Control - A Review, Policy,
Research and External Affairs Working Paper Series, WPS 624, Country Economics Department, World Bank,
March 1991.
1AI C.f. G. Eskeland and E. Jimenez, op. cit..
M/ T.H. Tietenberg, "Economic Instrments for Environmental Regulation', Ofont Review of Econonde
Policy, Vol. 6, No. 1, 1990, p. 17.


-34-
least harmful thermal alternatives. Not all hydroelectric projects will have adverse environmental
effects; and those that do can often be mitigated with foresight and appropriate action. Similarly,
with coal and fuel oil, the negative environmental effects can be reduced with appropriate
technologies, at a cost. Demand-side measures are not good instruments to deal with this situation.
Given the high degree of central control In the Brazilian power system, a better approach in that case
Is to tackle the environmental problems directly in supply-side planning and to Internalize the
environmental costs wherever possible, so that they can be reflected in marginal cost and price.
72.         The production of natural gas and petroleum products in Brazil, on the other hand, has
had fewer adverse environmental impacts. To date, the main problems arose from pollution from
refineries, which Is being tackled directly, again because the system Is subject to central control;
although, in the future, the possible exploitation of natural gas and oil in the Amazonian region will
have to be approached with great caution. Environmental problems are more relevant In the
consumption of gasoline, diesel and fuel oil. Taxation of these products, based on identifiable
environmental costs, would provide an economic basis for inter-fuel substitution, including the
development of ethanol, in place of gasoline, and natural gas, to replace fuel oil, gasoline and diesel.
IX. Conclu
73.         Our major conclusion is that, where the public sector exerts control over a significant
portion of energy supply and sells it in markets with an active private sector, market imperfections do
not provide an adequate justification for the public sector to intervene in the energy market, given the
welfare arguments against intervention and the potential for resource misallocation. We have also
tried to show that our present understanding of the nature and extent of these imperfections Is
inadequate to provide a realistic and practical basis for Government non-price intervention to promote
energy conservation, and we lack clear-cut evidence that intervention has worked in practice. Far
from "clearing away the institutional obstacles and market imperfections which stand in the way of
achieving the full range of energy efficiencies which are technically and economically feasibleMS,
the conservation measures usually advocated would create additional distortionsim'. While
recognizing fully the political obstacles which must be surmounted, the economist is on much sounder
ground, qua economist, in trying to ensure that an overall competitive framework exists for the
economy and that prices in the public sector are based on economic costs. Including allowances for
environmental costs. However, we acknowledge that the public sector may need to complement these
pricing signals with limited and well-conceived action, to ensure an adequate flow of information to
MI J. Goldemberg et. aL, Energyor a Swainable World, op. cit., p. 372.
M/ Trying to target our actons precisely enough and choose our delivery mechanism to bring about the right
kind and amount of conservation requires a great deal of additional information about consumer behavior.
Some of these questions are pursued further in M.A. Einhorn, "Measuring the Costs and Benefits of
Conservation Programs, and D. Norland and J.L Wolf, -Utility Conservation Programs: A Regulatory and
Design Framework", both in Pubic Ultks Fornighrly, July 25, 1985.


.35-
the market and to support basic R&D. A wider consumer choice, absent regulations and standards,
and more Informatic may raise transaction costs for consumers but are normally regarded as Integral
elements to the market solution.
74.         Externalities create the only major qualification to our conclusion, although not In the
sense that they provide a general rationale for the public sector to implement energy conservation
measures. Rather, the environmental externality needs to be addressed directly, for example through
tradeable permits in the case of pollution. Alternatively, the prices and costs of the specific energy
products that create the externalities could be adjusted: through output prices (e.g with pollution
taxes), where externalities are caused by consumption; and through input prices, where they are
caused by production. If adjustments in prices are impractical, the externalities may have to be
handled directly, through regulations. In the case of developing countries, the revenue implications of
our recommendations for energy supply companies should be beneficial, to the extent that marginal
cost exceeds average costN; and while environmental issues are a legitimate public policy concern,
we believe that it is hard to justify the use of energy policy to increase national security in a
developing country, given the typically high costs of greater energy self-sufficiency.
J19I Where economic pricing leads to excess profits, the Govemment can take advantae of a valuable new
source of revem by levying taxes.