3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.

sgm LaTeX2e(2002/01/18) P1: GCE

10.1146/annurev.energy.27.122001.083429

Annu. Rev. Energy Environ. 2002. 27:57–81

doi: 10.1146/annurev.energy.27.122001.083429
Copyright °
c 2002 by Annual Reviews. All rights reserved

AN INTELLECTUAL HISTORY OF ENVIRONMENTAL

ECONOMICS
David Pearce
Department of Economics, University College London, Gower Street, London,
WC1E 6BT, and Department of Environmental Science and Technology, Imperial College
London, United Kingdom; e-mail: dpearce@ucl.ac.uk

Key Words history, valuation, policy instruments, discounting, sustainable

development
■ Abstract From modest beginnings in the 1960s, environmental economics has
grown to be a major subdiscipline of economics. It combines traditional work in the field
of welfare economics and the theory of economic growth with more recent perspectives
on the political economy of choosing policy instruments and the philosophy of sustain-
able development. The central tenets are that environmental problems have their roots
in the failure of economic systems to maximize human well-being, that environmental
quality matters for human well-being and for more traditionally oriented economic
growth objectives, and that efficient policy can be achieved through incentive design.

CONTENTS
INTRODUCTION: THE ORIGINS OF
ENVIRONMENTAL ECONOMICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
SUSTAINABLE DEVELOPMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
MEASURING SUSTAINABLE DEVELOPMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
ECONOMIC VALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
COST-BENEFIT ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
THE CHOICE OF POLICY INSTRUMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
ECOLOGICAL ECONOMICS: A NEW PARADIGM? . . . . . . . . . . . . . . . . . . . . . . . 75
CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

INTRODUCTION: THE ORIGINS OF

ENVIRONMENTAL ECONOMICS
The origins of environmental economics lie in the 1950s when, in the United States,
Resources for the Future (RFF) was established in Washington, DC. RFF is an
independent research organization that has both developed and applied economics
to a large array of environmental issues. The early focus at RFF was on the issue
of natural resource scarcity, prompted by the U.S. President’s Materials Policy
1056-3466/02/1121-0057$14.00 57
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

58 PEARCE

Commission (the Paley Commission). The Paley Commission was asked to review
the future supply of minerals, energy, and agricultural resources in light of the
formidable demands made on these resources by World War II. RFF’s early work
focused on the same issue, culminating in the influential and widely cited work
Scarcity and Growth, by Barnett & Morse (1) in 1963, together with Christy &
Potter’s statistical analysis of historical resource price trends (2) in 1962. But it
was in the 1960s that environmental economics truly came of age. The political
backdrop was the first environmental revolution initiated by Rachel Carson’s Silent
Spring (3) in 1962. Carson’s warnings about the effects of agrochemicals in the
environment were not new, but they were cogently put and had already gained
public attention earlier in the same year with a series of three articles in the New
Yorker. That economics should have a lot to do with environmental concerns of
the kind raised by Carson was not surprising. First, agrochemicals were and are
big business. Second, the use of chemicals such as DDT had done a lot to raise
agricultural productivity and protect human health. Third, economists were already
familiar with the idea that there are likely to be costs and benefits from any form
of economic activity. The costs take the form of “external effects,” in this case
the alleged loss of biological diversity about which people had begun to care far
more than previously. It is no surprise, then, that economists began to link the
theory of external effects with an economic interpretation of the rising tide of
environmentalism.
Like all subdisciplines of economics, environmental economics borrowed heav-
ily from thought of its precursors. The idea of an externality, a detrimental (or ben-
eficial) effect to a third party for which no price is exacted, was already familiar
from the work of Pigou (4) in the 1920s. Pollution damage fitted neatly into this
framework. Polluters cause damage to third parties but may not be required to pay
for that damage. Because market-oriented economic systems did not account for
externalities (any more than planned ones such as the former Soviet Union did in
practice), those systems could not be maximizing human well-being. Intervention
in some form to internalize the externality—to get the third-party effect included
in the internal costs of the polluter—was justified.
That policies could be evaluated in terms of their costs and benefits, with costs
and benefits defined in terms of human preferences and willingness to pay, was
established by Dupuit in the nineteenth century (5, 6). The body of modern-day
welfare economics was established by Hicks (7, 8), Kaldor (9), and others in the
1930s and 1940s. Practical guidelines for using welfare economics in the guise of
cost-benefit analysis were drawn up first for the water sector in the United States.
Considerable attention was also being devoted to the wider issue of efficiency in
government, especially military spending, by bodies such as the Rand Corpora-
tion. In 1958 three seminal works appeared: Eckstein’s Water Resource Devel-
opment (10), Krutilla & Eckstein’s Multipurpose River Development (11), and
McKean’s Efficiency in Government Through Systems Analysis (12). The feature
of these works was the synthesis of practical concerns with the theoretical welfare
economics literature. The essential step was the justification for the benefit-cost
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 59

principle: justifying projects or policies on the basis that benefits exceed costs
is wholly consistent with there being losers, i.e., those who suffer the costs. The
Kaldor-Hicks compensation criterion had established that projects were nonethe-
less justified because gainers could compensate losers, such that losers would be
no worse off, and gainers would still have a net benefit. This implies that, pro-
vided the compensation takes place, no one is actually worse off, thus meeting the
long-established Pareto criterion for an improvement in overall well-being. How-
ever, actual compensation need not occur: It is necessary only that it could take
place.
In a separate strand of intellectual development, the idea that any natural re-
source had some optimal rate of use had been established formally by Gray (13) in
the early twentieth century and later by Hotelling (14). Initially, these optimal use
theorems were confined to natural resource economics as opposed to environmen-
tal economics. The distinction between the two was that the former was mainly
concerned with rates of exhaustible resource depletion and the determination of
optimal harvest rates for renewable resources. Environmental economics, on the
other hand, focused on pollution. The distinction largely broke down once it was
recognized that theorems from the former were applicable to the latter contexts,
especially where pollutants were cumulative, and also in the context of the the-
ory of optimal economic growth. The growth theory contributions culminated in
elegant if demanding treatises in the 1970s, e.g., Dasgupta & Heal (15). Mathemat-
ical models of economies with single exhaustible natural resources were in turn
stimulated by real world issues. In 1973 the first Organization of the Petroleum
Exporting Countries’ (OPEC) oil price increase occurred, which prompted con-
cerns about the stability of fossil fuel–dependent economic systems. The optimal
use rate for a renewable resource, such as a fishery, was the subject of a separate lit-
erature dating mainly from Gordon’s 1954 paper on fisheries as a common property
resource (16). Gordon also explained why a fishery faced with open access, i.e.,
totally absent property rights as opposed to common property where rights exist
for a defined community, could be exploited to the point where all economic rents
were dissipated. By implication, if certain other conditions are present, open ac-
cess may be consistent with extinction of the resource. Interestingly, the paper that
commanded substantially more attention for saying the same thing (although con-
fusing common property and open access) was Hardin’s 1968 paper “The Tragedy
of the Commons” (17). Hardin is a human ecologist, and “Tragedy” has been one
of the most reprinted articles in the environmental literature.
Nonetheless, even today, textbooks tend to preserve the distinction between
natural resource and environmental economics. For reasons of space, the rest of
this paper adopts the distinction and focuses on environmental rather than natural
resource depletion issues. The history of natural resource economics has yet to be
written, but there is considerable historical perspective in Fisher’s 1981 text (18).
Another precursor literature relevant to the birth of environmental economics is
that relating to some notion of the ecological limits of economic activity. The envi-
ronmental movement of the 1960s had begun to focus on the apparently wasteful
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

60 PEARCE

lifestyles of the occupants of modern economies, prompting a logic that said if these
lifestyles were endangering the planet, then the lifestyles must change. Gradually,
the unsustainable lifestyle issue became synonymous with the pursuit of economic
growth, and the antigrowth movement was born. But even this movement was not
new. The notion of absolute limits was central to Malthus’s concerns 160 years
before. Ricardo had developed a separate notion of scarcity arising from rising
marginal costs of resource extraction and use. John Stuart Mill had analyzed the
notion of a stationary state in which critical stocks of population and capital were
perpetually constant. These concerns about ultimate constraints to economic activ-
ity surfaced again in the now neglected work of Kapp, The Social Costs of Private
Enterprise (19), published in 1950. But probably the most celebrated paper to pro-
voke the many questions subsequently to be analyzed in environmental economics
was Boulding’s 1966 “spaceship Earth” essay (20). Boulding likened planet Earth
to a spaceship in which there is a finite supply of energy, which can only ultimately
be replaced with solar power, and a finite supply of water and materials, which
can only provide a sustainable future if they are reused and recycled. Viewed with
more than 30 years’ hindsight, Boulding’s essay sets the foundations for what
many today would still regard as a sustainable society, one that operates within the
limits set by finite supplies and finite flows of materials and energy. Production and
consumption cease to be good things and instead, attention has to be paid to the
maintenance of stocks of assets, including the stock of knowledge that Boulding
rightly foresaw as one of the means of improving the human lot without changes
in physical resources.
Boulding’s essay remains to this day the basis of ecological economics, where
the focus is still on physical limits and where technological change through hu-
man capital formation is not regarded as an obvious and viable means of escape
from those limits. Within environmental economics, however, Boulding’s work
prompted a different development. The theory of external effects had concluded
that, when present, externalities will produce suboptimal levels of human well-
being. But until Boulding’s notion of spaceship Earth, externalities were gener-
ally regarded as fairly minor and manageable deviations from the optimum. Silent
Spring had already suggested exactly the opposite—agrochemicals were pervasive
to economic systems. Externalities were showing up long distances from the sour-
ces of emissions and were cumulating through time as well. Spaceship Earth simi-
larly invoked the first law of thermodynamics to point out that whatever was taken
out of natural resource sectors must reappear in equal weight as waste, which will
likely affect the environment when disposed of: Matter and energy cannot be cre-
ated or destroyed. As economies expand in the economic sense, so they are likely
to expand in terms of physical resource extraction and hence in terms of physical
waste emissions to the environment. Worse, economic activity chemically trans-
forms materials and energy into waste gases: Carbon becomes carbon dioxide, for
example. These transformations become systematically more and more diffuse, or
entropic, and hence less and less easy to recapture in terms of materials reuse and
recycling. Moreover, energy cannot be recycled at all, although waste gases can
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 61

be captured. But if waste is pervasive to economic systems and if environmental

systems have finite capacities to receive waste, then externalities are also likely to
be pervasive, as Ayres & Kneese showed formally in the first materials balance
model of an economy (21).
A final important contribution was Coase’s “Problem of Social Cost” (22).
Coase observed that an externality context was conducive to two potential solu-
tions. The first, familiar from the work of Pigou, was a tax on the creator of the
externality (the polluter), or some form of regulation that imposed the burden of
action on the polluter. The second involved the sufferer paying the polluter not
to pollute. In the first case, the polluter pays and in the second the victim pays.
Coase argued that, in terms of efficiency (though not in terms of distributional bur-
dens), the solutions were equivalent. For those opposed to more state regulation,
in this case in the name of environmental quality, Coase’s argument opened up a
substantial potential for free market environmentalism. The state was redundant
except perhaps in terms of setting the conditions under which such market bar-
gains took place. While there is a tendency to dismiss Coase’s second solution as
being unfair, one real world context in which it operates occurs when the polluter
is a low-income agent and the sufferer a high-income agent. It is not uncommon
for states suffering transboundary pollution, for example, to provide grants and
technology for the upgrading of polluting technology in the polluting country.
The pieces of the economic jigsaw were now in place. Welfare economics pro-
vided the analytical foundations for determining optima in economic systems.
Within welfare economics, externalities were transformed from being fairly minor
deviations from the optimum to being pervasive, central, and potentially large. Eco-
nomic systems could therefore be significantly inefficient. The materials balance
principle, embodied in the notion of spaceship Earth, established that there could
be limits of a Malthusian kind unless technological change provided an escape.
Economic growth models in which resource endowments were explicitly modeled
came up with the same answer—an optimal world might require significant in-
tervention, but it might also be unsustainable without technological change. The
instruments for achieving optimality had also begun to be forged. Pollution taxes
and perhaps Coaseian bargains provided efficient solutions, and these instruments
were soon to be joined by perhaps one of the most ingenious policy instruments
of all, the tradable permit.

SUSTAINABLE DEVELOPMENT
Sustainable development has become a central notion in modern economic policy.
Its most celebrated formulation comes from the World Commission on Environ-
ment and Development (the Brundtland Commission) (23):
Sustainable development is development that meets the needs of the present
without compromising the ability of future generations to meet their own
needs. It contains within it two key concepts:
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

62 PEARCE

■ the concept of ‘needs’, in particular the essential needs of the world’s poor,
to which overriding priority should be given, and
■ the idea of limitations imposed by the state of technology and social orga-
nization on the environment’s ability to meet present and future needs [(23)
p. 43].

In economic terms, the definition involves the notion of an economic system in

which well-being per capita increases over time on a sustained basis. A similar, far
less researched notion, would apply to the current distribution of well-being. But
economic growth models traditionally adopted the same temporal perspective by
looking at growth paths of utility and consumption. Hence it seems appropriate to
reinterpret traditional growth theory in terms of sustainable development. Growth
models were concerned with what the optimal path of consumption, and hence
well-being (or utility) looked like, and also, incidentally, with whether such a path
is sustainable or not over long periods.
Defining sustainable development in terms of sustained per capita increases in
well-being needs to be distinguished from stating the conditions for the achieve-
ment of such increases. Like traditional growth theory, the modern theory of sus-
tainable development is based on notions of capital assets as the means of gen-
erating well-being. Along with developments in growth theory, e.g., endogenous
growth theory, capital is decomposed into human-made capital, human capital
(knowledge, skills), natural or environmental capital, and social capital. The con-
dition for sustainability (or potential sustainability) is that the sum of these assets,
i.e., total wealth, should increase in per capita terms over time. Intergenerational
rules of behavior follow immediately, e.g., that each generation should bequeath
to the next generation a capital endowment no less than the one it has now. This
interpretation of sustainable consumption was debated in the 1930s, with Hicks in
1939 setting out the notion that it involved maximum consumption consistent with
maintaining capital assets intact (24). The resulting Hicksian income measure can
then be interpreted as “the interest on total wealth.”
The links between sustainable development and traditional neoclassical growth
theory can be illustrated. Ignoring technical change and population growth, one
central equation describing an optimal path of consumption is given by the Ramsey
rule:

Ċ αR β /K M
1−α
− dM − ρ
= .
C µ

Here C is consumption, and hence the left-hand side of the equation is the per-
centage growth rate of consumption. The first expression on the right-hand side
is the marginal productivity of human-made capital, assuming a Cobb-Douglas
production function in human-made capital (KM) and natural resources (R). The
growth rate of consumption depends upon the utility discount rate (ρ), the marginal
productivity of human-made capital, the elasticity of the marginal utility of con-
sumption (µ), and the rate of depreciation on human-made capital (dM). Following
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 63

Dasgupta & Heal (15), consider what happens to the consumption path as R → 0.
Since ρ > 0, the rate of growth of consumption becomes negative. To avoid this, let
KM → 0 as well. But if KM → 0, then so does income and consumption. The impli-
cation is that the optimal path of consumption eventually goes to zero. Obviously,
this is not consistent with any notion of sustainable development.
The discussion is sufficient to establish that optimality, defined in terms of
maximizing the present value of future consumption streams, is not necessar-
ily consistent with sustainability. This nonsustainability comes about because of
positive utility discounting or because particular features of the depreciation of
human-made capital, and despite working with a Cobb-Douglas production func-
tion in which a fair degree of substitutability between human-made and natural
capital is assumed (in fact the elasticity of substitution is unity in this model).
The Ramsey growth path above is based on a highly simplistic model. The
potential nonsustainability of the model can be made worse by allowing population
to grow, so that the chances of bequeathing rising per capita wealth are less. But
one optimistic perspective on population growth is to allow its rate of growth to be a
determinant of technological change, which raises the productivity of capital assets
in the sense argued by Boserup (25). How far Boserupian technological change
describes real world contexts is debated in the literature. But more generally,
the omission of technological change from any growth model makes the model
seriously unrealistic.
One consumption path emerging from a model with technological change is:

Ċ
= g − βρ,
C
where g is the growth rate of the efficiency of production inputs (the growth rate of
total factor productivity), and it is interpreted here to reflect technological change.
The β is the exponent on R in the production function and is interpreted as the
elasticity of output with respect to the rate of resource use, R. The ρ is familiar
as the utility discount rate. The growth rate Ċ/C will be positive if g > βρ, i.e., if
the rate of technical progress exceeds the elasticity of output with respect to the
natural resource multiplied by the pure time preference rate. Results of this kind
are familiar from the 1970s literature on optimal growth (26–28). Technological
change has the potential to rescue optimal growth paths from being unsustainable.

MEASURING SUSTAINABLE DEVELOPMENT

The Earth Summit at Rio de Janeiro in 1992 gave a strong impetus to the search
for measures of sustainable development. However, almost all of the resulting
indicators of sustainability are simply indicators of environmental and economic
change. A sustainability indicator should have the property of giving at least a first
approximation of whether a given economy is sustainable or not. The Rio Summit
called for the development of revised measures of gross national product (GNP) to
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

64 PEARCE

reflect this concern. In the 1990s, environmental economists developed indicators

based on the capital theory of sustainable development outlined above. The GNP is
a poor measure of well-being because it implicitly assumes there is only one capital
asset—human-made (or reproducible) capital. Revising the GNP to reflect changes
in other assets turned out to be intuitively simple, although debate continues on
the various adjustments, and measurement problems abound. An extensive guide
to modern national income theory is given in Hartwick (29).
Ignoring other forms of capital, the proper measure of well-being is net national
product (NNP) rather than GNP because capital depreciation does not contribute to
well-being—it merely replaces depreciated stocks. In 1976 Weitzman (30) showed
that current period consumption plus investment is, if held constant and the present
value taken, just equal to the present value of consumption along an optimal path.
The theoretical foundations for modifying the national accounts were subsequently
set out in various papers by Hartwick (31) and Mäler (32). Solow (33) also showed
that increases in national product from some baseline value are equal to the discount
rate multiplied by the cumulated capital wealth from the baseline period to the
current period. In other words, national product becomes interest on total wealth,
the Hicksian income concept introduced above. An expression that modifies NNP
to be an appropriate indicator of well-being is:

NNP = GNP − dM − dN − dH − dS .

In each case d refers to net depreciation and subscripts M, N, H, and S refer to

human-made, natural, human, and social capital, respectively. In turn GNP is made
up of consumption C and gross investment, I. Since I − dM = Inet, or net investment,
the revised measure of NNP becomes

NNP = C + Inet − dM − dN − dH − dS .

In terms of measurement, dM is estimated in conventional national income ac-

counts. Thus dN will be made up of any increases in environmental assets (e.g.,
growth of renewable natural resources) less any use rates (harvest), i.e., dN will be
positive if resources are used more than sustainably and are negative otherwise. By
definition, dN will be negative for exhaustible resources, but allowance needs to be
made for the discovery of exhaustible resources. The dH is likely to be negative,
i.e., human capital appreciates as knowledge and skills increase through time in a
cumulative fashion. The dS could be positive or negative with indicators such as
crime rates, family breakdown, etc., indicating depreciation of social capital.
For this measure of true or “green” NNP to be estimated, each depreciation
indicator needs to be measured in monetary terms. The issue of money valuation
thus arises (see below). The correct unit valuation in each case is the economic rent,
i.e., price minus marginal cost. While data on rents are difficult to come by, the
valuation issue for marketed products—fisheries, timber, energy, for example—is
not complex in principle. But where the asset being depreciated is not marketed, the
problem arises of finding individuals’ willingness to pay to prevent those changes
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 65

(or to secure a gain). Probably the largest, and most controversial, research effort
in environmental economics has in fact been devoted to this issue of valuing
nonmarketed asset change. The focus has been on environmental assets. Only now
is attention being paid to the valuation of changes in social assets. The development
of valuation procedures is discussed shortly.
Empirical attempts to estimate modified national accounts predate the interests
of environmental economists in green NNP [for a review see Eisner (34)]. The first
set of accounts to incorporate environmental depreciation was produced in 1989
for Indonesia by scholars at the World Resources Institute in Washington, DC
(35). The adjustments involved deductions for the rents foregone in the petroleum
and forestry sectors and for output losses from soil erosion. The resulting “net
domestic product” (a confusing term since it has a direct meaning in conventional
accounts) grew at 4% per annum from 1971–1984 compared to recorded gross
domestic product (GDP) growing at 7.1%; although after 1975, growth rates in the
two magnitudes were almost identical, which illustrates one of the problems of
interpreting modified national income measures. Essentially, a modified product
measure could grow less fast, just as fast, or faster than conventional GDP. It is
not intuitively obvious what this means for sustainability. Several dozen studies of
modified national income have been published since the World Resources Institute
study and are reviewed in Hamilton & Lutz (36).
The World Resources Institute study did in fact show an alternative illustration
in terms of net “true” investment, i.e., investment minus the depreciation on envi-
ronmental assets. Pearce & Atkinson (37) established that the correct measure of
sustainability could be expressed in terms of “genuine” savings. Savings can be
thought of as a fund set aside to cover depreciation on assets. A simple rule then
emerges to the effect that savings must be greater than depreciation on all assets
for an economy to qualify as potentially sustainable. Savings minus depreciation
equaled genuine savings. The notion of genuine savings was subsequently extended
and its theoretical underpinnings more rigorously established in a sequence of pa-
pers by Hamilton, for example (38). The concept was adopted by the World Bank,
and genuine savings indicators appear for over 100 countries (and across time) in
the Bank’s annual publication World Development Indicators, [see also (39)]. By
and large, economies with high ratios of savings to GNP have positive genuine
savings, i.e., are in principle sustainable. This result is as one would expect from
traditional development economics. That more developed economies emerge as
being sustainable is not therefore surprising. But high-consuming countries, such
as the United States, do not fare well on the genuine savings indicator; neither
do natural resource-dependent economies, e.g., Middle East oil economies, unless
depletion is compensated for by investment in other capital assets—bearing out
the “Hartwick rule” for sustainability (40).
One surprising feature of the various modifications to indicators of sustainability
is the omission of population change, i.e., measures have not been expressed in
per capita terms, even though this is an obvious requirement if the indicator is to
reflect sustained well-being. Recent efforts to measure changes in per capita total
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

66 PEARCE

wealth show that even modest rates of population change can put sustainability at
risk (41).
Although growth theorists can legitimately argue that the economic growth
models of the 1970s had already encompassed the issue of sustainable develop-
ment, it seems fair to say that the wider political embrace of the concept in the
1990s has stimulated the substantial progress in conceptual models of sustainabil-
ity and its measurement. This development contrasts starkly with the broader and
ever-growing literature on sustainable development, much of which lacks the rigor
brought to the subject by environmental economists.

ECONOMIC VALUATION

Although modified national accounting requires that asset depreciation be valued

in monetary terms, efforts to develop valuation techniques arose directly, and far
earlier, from project appraisal and, eventually, policy appraisal. Initial efforts to
broaden the coverage of environmental impacts in investment (project) appraisal
involved the addition of some form of environmental impact appraisal to state-
ments of costs and benefits. This remains the dominant form of appraisal. Its
obvious weaknesses are that environmental impacts are not fully integrated into
the appraisal process and that simply conducting an environmental assessment is
easily construed as having taken account of environmental concerns, without the
assessment affecting project design or even the final decision.
Welfare economics had established (partly through the rediscovery of Dupuit’s
work) that market prices are conceptually the correct measure of the economic
value of a marginal change in the supply of a marketed economy. But for non-
marginal changes, the relevant magnitude is the sum of consumers’ and producers’
surpluses, and to estimate these requires knowledge of demand and supply curves.
The same welfare economics also established that all such changes in well-being
arising from a project (later extended to policies also) should be included in a cost-
benefit appraisal. In practice, the first conceptual guidance was followed and the
latter ignored because environmental changes—usually referred to as intangibles—
were ignored. The basic assumption appeared to be that changes in surplus for
nonmarketed goods could not be estimated. Two of the triumphs of environmen-
tal economics have been to emphasize the incompleteness of appraisals that omit
environmental change and to develop the means of incorporating environmental
values into appraisal.
In fact, there had been early suggestions on how to value some environmental
benefits. Responding to a request from the U.S. National Parks Service in 1947
about the worth of national parks, Hotelling actually set out what today is known
as the travel cost method (42). The parks had no entrance fees so the problem was
similar to that analyzed by Dupuit in the 1840s—there was no market for roads and
bridges, but people obviously were willing to pay for them. In the park case there
were expenditures associated with the recreational use of the park, notably the costs
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 67

of traveling to the park. Because people traveled different distances, the costs they
faced varied. In principle, treating the differential costs as prices meant that the
demand curve for recreational visits could be derived. The resulting area under the
demand curve gave an estimate of the total consumer surplus accruing to visitors
to the recreational site. Hotelling’s response was ignored by the National Parks
Service since other respondents had expressed a consensus view that the problem
could not be solved although his note appeared in the Parks Service report on the
issue (42). Ten years later it resurfaced, first in a study of recreational use of the
Feather River in California (43) and almost simultaneously in work from Resources
for the Future (44, 45). Since then, hundreds of travel cost studies have been carried
out, mainly, but far from exclusively, in the United States, where various pieces of
legislation have required that the benefits of sites be demonstrated. Methodologies
have been refined and techniques have been extended to cover benefit estimation
in the context of multiple recreational sites. Various issues remain debated: how to
deal with the value of time and the treatment of joint benefits where visits are part of
a broader travel experience. A significant feature of travel cost studies is that they
involve surveys of users to determine their mode of travel and the starting point
of the journey. Many issues of survey design and securing adequate and truthful
responses arise in travel cost, just as they do in the “stated preference” procedures
to be discussed below. A detailed survey and guide to travel cost techniques is
given in Ward & Beal (46).
The travel cost method is a “revealed preference” approach to valuation: Indi-
viduals’ preferences for a nonmarketed good are revealed through the inspection
of other markets. A second form of revealed preference relates to property—land
and housing—markets. The intuition is again simple. Property prices are capital
values that reflect the implicit rentals arising from the asset, rentals that in turn
reflect the value of the flows of services from the asset. If environmental charac-
teristics are one feature of those services then, in principle, it should be possible
to decompose the values of each of the characteristics, including the effect of the
environmental variable(s) on property prices. Proofs that these relationships bear
formal connection to measures of welfare change came after the first attempts
to use property prices. Ridker (47, 48) established statistical links between levels
of air pollution and property values in St. Louis, Missouri. Essentially, the value
obtained is the coefficient on air pollution in a regression of property prices on
determining factors. (His study is also of interest because it devotes one chapter to
the assessment of the avoided costs of cleaning materials if air quality is improved
and another to valuing the epidemiological links between air pollution and human
health using forgone output and burial costs).
The obvious question, however, was whether the resulting coefficient is actu-
ally a welfare-consistent measure. The regression equation linking property prices
to the bundle of price-determining characteristics became known as the “hedonic
price function.” Its derivative with respect to the environmental pollution is the
“implicit price function,” so that the value of the coefficient in the hedonic price
function can vary with the level of the environmental characteristic. Although most
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

68 PEARCE

hedonic property price studies stop at the estimation of this implicit function, once
the change in environmental quality is nonmarginal, implicit price estimation is
insufficient as a measure of the change in welfare, and a complex two-stage pro-
cedure is required to estimate an actual demand curve (49, 50). As a general rule,
the coefficient on pollution in the regression equation produces an implicit price,
which is a theoretically sound measure of welfare change provided the environ-
mental change is marginal, but not otherwise. Even where the estimation problem
is resolved, the resulting welfare measure is incomplete because it excludes (a)
changes in the welfare of any landlords, (b) effects of changes in environmental
quality on supply conditions generally, and (c) changes in welfare accruing to
nonhouseholders (e.g., visitors). Efforts to provide a more encompassing measure
of the change in welfare have defined the recent contributions to the hedonic price
literature (51, 52).
The notion of an hedonic price is general. Early work on hedonic models ten-
ded to focus on air pollution, as discussed above, and noise nuisance. Most studies
used regression approaches to estimate the relevant implicit price, but some at-
tempted direct estimation of the coefficient through questionnaires of estate agents
(realtors). A notable attempt in the United Kingdom using the latter approach in
a study of aircraft noise was the work of the Roskill Commission on the siting
of London’s third airport (53, 54). The early aircraft and road traffic noise studies
were reviewed by Nelson (55, 56).
One other major area of research using the hedonic approach has been on the
valuation of risks to life. The principles are the same as for hedonic property prices,
but the dependent variable is the labor wage, and the independent variables are
wage-determining factors such as skill level, unionization, and risk. Traditional
approaches to valuing life risks involved estimating the forgone output due to pre-
mature death, the idea being that society forfeits the (net) product of the individual
concerned. Apart from obvious anomalies, such as retired persons appearing to
have no value to their remaining lives, this approach has no basis in the theory of
welfare economics. Schelling appears to have been the first to note the relevant con-
cepts of willingness to pay to avoid a risk and willingness to accept compensation
to tolerate a risk (57). The resulting “value of a statistical life” is an aggregation of
individual risk valuations. Assume the probability of dying next year is 0.004 for
each person, and assume there are 1000 persons in the population. A hypothetical
risk reduction policy reduces the risk to 0.003, a change of 0.001. If each person
values this risk change at $1000, the aggregate willingness to pay is $1 million. The
change in risk will result in one statistical person being saved each year (1000 ×
0.001). Thus the value of a statistical life is $1 million in this example. What is
being valued is aggregate willingness to pay (or accept, in the hedonic wage case)
for the change in risk. The phrase “value of statistical life” has caused many prob-
lems for economists because it is easily shortened to “value of life,” which raises
moral concerns about procedures for placing money values on life itself. But the
central economic proposition remains: No society allocates all its resources to life-
saving, nor does any society treat lives worldwide as if they have equal monetary
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 69

value. Some of the sensitivity surrounding life valuation is avoided by noting that
the policy context is one of ex ante valuation not the valuation of identified ex
post deaths (58). The impossibility of avoiding monetary valuation was noted very
early on in a neglected classic paper by Thomas (59).
Travel cost and hedonic price methodologies are examples of revealed prefer-
ence approaches to valuation. The underlying theory of economic valuation was
eloquently brought together in Mäler’s 1974 classic treatise (60), which also set out
some of the theorems relating to the choice of policy instruments. In addition, Mäler
devoted some attention to what today would be called stated preference techniques,
basically the elicitation of the willingness to pay from the use of questionnaires.
Mäler was concerned with the classic Wicksell problem of misrepresentation of
preferences in such contexts, i.e., whether it is possible to design questionnaires
that avoid the incentives that respondents otherwise have not to tell the truth. He
concluded that it was impossible to guarantee truth telling, or incentive compat-
ibility as it has come to be known, but that directions of bias could be defined.
According to Hanemann (61), the idea of using questionnaires appears to date
from a suggestion by Ciriacy-Wantrup in 1947 (62) when addressing the issue of
how to derive a demand curve for soil conservation measures. As with the travel
cost method, however, the suggestion was not taken up, and the first, unrelated,
efforts to use what came to be known as contingent valuation occurred in 1958
and 1961. The 1958 exercise related to recreationists in the Delaware River Basin
(63), and the 1961 application related to recreationists in the Maine woods (64).
Soon after, contingent valuation studies multiplied rapidly. Only now, in the new
millennium, are there signs of shifts away from contingent valuation toward other
stated preference techniques that do not involve direct questions such as “what are
you willing to pay?” and “are you willing to pay $X?” This shift, which should not
be exaggerated, reflects concerns that direct questions may strain cognitive abili-
ties in respondents. Other stated preference techniques such as conjoint analysis,
choice experiments, and contingent ranking and rating (terminology, unfortunately,
varies) are now used on an increasing basis, but with the limitation that not all of
these approaches are consistent with the underlying theory of welfare economics.
Stated preference techniques have secured a major place in the valuation armory
of the environmental economist. Particular attractions are that, in principle, such
approaches can elicit all the kinds of economic value relevant to a policy or project
decision. In particular, they can elicit values placed on an asset by nonusers of the
asset, i.e., those who may want to have the asset preserved or improved, but who
do not make direct use of it. This notion of “nonuse” or “passive use” value came
to be extremely important, and controversial, in the practical applications of stated
preference techniques. A second reason for the use of stated preference techniques
is that questionnaires elicit far more information than stated willingness to pay.
For example, motivations for being willing to pay are a standard part of most
contingent-valuation exercises. Such studies have served to underline the richness
of human motivation. Motives such as altruism, stewardship, concern for future
generations, etc., have been shown to be important. Although there is nothing in
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

70 PEARCE

economics that requires motivations to be of the self-interest variety, it remains the

case that economists have themselves contributed to the fallacy that economic man
has only selfish motives. This does not mean that a multiplicity of motives avoids
all problems—the extent to which values based on altruistic motives, for example,
should be included in aggregated values has been the subject of some controversy.
So important did contingent valuation, in particular, become that it resulted
in several sets of guidance from practitioners. Most applications occurred in the
United States because various pieces of legislation and court assessments on envi-
ronmental damages allowed such methodologies to be used. In contrast, European
environmental liability legislation is far more limited, and economic valuation
methodologies have rarely been used in liability cases. Nonetheless, government
use of contingent valuation has grown in Europe, although the United Kingdom is
probably the only example where contingent valuation has been used to measure
damages as the basis of environmental taxation. In the United States, Mitchell &
Carson (65) produced a detailed guide to contingent valuation. Coincidentally, the
same year, 1989, saw the running aground of the Exxon oil tanker, Exxon Valdez, in
a remote part of Alaska. A contingent valuation study by Carson & colleagues (66)
focused on the nonuse values of the general public and found that total damages,
as measured by willingness to pay to avoid an accident, ranged from $2.8 billion to
$9.3 billion. In the event, Alaska State and the U.S. Federal Government settled for
around $1 billion in damages. Exxon, however, initiated an attack on contingent
valuation, and the resulting controversy was partly responsible for the creation in
1992 of a distinguished panel of experts to appraise the credibility of contingent
valuation. The National Oceanographic and Atmospheric Administration (NOAA)
Panel itself set out an extensive set of guidelines for the implementation of a contin-
gent valuation study. An example of one of the more important recommendations
was that values must be “sensitive to scope,” i.e., values must vary with the scale of
the environmental effect (67). While not themselves uncontroversial, the NOAA
guidelines are widely used in modern contingent valuation studies. European guid-
ance on contingent valuation and other stated preference techniques is contained in
Pearce et al. (68), and detailed guidance on choice modeling (i.e., stated preference
techniques other than contingent valuation) can be found in Louviere et al. (69).
All stated preference studies have tended to be controversial, because the issue,
identified as early as Wicksell, of whether respondents are telling the truth remains
central to the debate. While a few efforts have been made to look at charitable
donations as a means of eliciting nonuse values, it remains the case that stated
preference techniques are the only means of uncovering these values. Nonuse
value had been identified first by Krutilla in a seminal paper in 1967 (70). The
main feature of nonuse value is that it has no “behavioral trail” and so will not
appear in conventional estimates of demand curves for environmental assets. The
same is true for another notion of economic value, option value, first identified
by Weisbrod in 1964 (71). Although nonuse value relates to a positive valuation
independently of use now or in the future, option value relates to a willingness to
pay now for the option to make use of the asset in the future, regardless of whether
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 71

the option is exercised or not. Substantial debate arose over both concepts, with
early lack of consensus fostered partly by different definitions and terminology. At
some time in the 1980s, the notion of “total economic value” (TEV) emerged, with
TEV defined as the sum of use values (the usual measure of consumer surplus),
option values (seen as a premium over and above consumer surplus), and nonuse
values. By and large, nonuse values have proved to be important for environmental
assets with a degree of uniqueness about them.

COST-BENEFIT ANALYSIS

Although most of the conceptual foundations for cost-benefit analysis developed

independently of environmental economics, the integration of environmental im-
pacts into cost-benefit analysis did lead to some significant changes. The most
obvious was that, through the techniques for measuring the economic value of
changes in environmental assets, cost-benefit analysis became more complete in
its coverage. Given that cost-benefit had always claimed to measure all changes
in well-being with a project or policy, and given that, via the materials balance
principle, all projects and policies do have environmental impacts, this process of
extending cost-benefit analysis was important.
A second contribution to change came from the new focus that environmental
issues gave to the choice of the discount rate. The problems with positive dis-
counting had always been known—far-distant effects could be substantial but,
expressed in present value terms, could quickly become insignificant. This raised
issues of fairness through time, and later it also appeared inconsistent with sus-
tainable development because the latter appeared to imply that current generations
should not impose significant costs on future generations. An assembly in 1977
of the leading analysts on discounting, once again arranged by Resources for the
Future, produced an impressive, technical volume on all the arguments about se-
lecting discount rates (72). After nearly 500 pages of exposition, a policy analyst
was not much wiser about what rate to select despite a masterly (and still widely
recommended) attempt at consensus by Lind in the introduction to the volume. A
more recent volume, produced as a deliberate follow-up to the first set of essays,
offers more guidance, especially because of a short but incisive essay by Weitzman
(73). Weitzman suggests that the apparent trivialization of large future costs by
discounting arises from a confusion between discount rates and discount factors.
Treating the latter rather than the former as a random variable can be shown fairly
simply to produce the result that the discount rate declines over time. Weitzman
even offers some rules of thumb about what appropriate short- and long-run dis-
count rates are. No one would pretend that the issue is resolved, but there is an
elegance about Weitzman’s solution.
A third area in which environmental economics modified cost-benefit analy-
sis arose from the contribution of Krutilla & Fisher (74). The problem developed
from the use of cost-benefit procedures to evaluate projects that had irreversible
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

72 PEARCE

effects, e.g., the flooding for hydroelectricity of an environmentally important val-

ley. It did not seem justified to commit resources to a project that yielded benefits
(electricity) for a possibly very short period of time (given rates of technological
change that would displace such technologies), while incurring a perpetual loss
of natural amenity. One major adjustment to cost-benefit procedures had already
been suggested in Krutilla’s 1967 essay (70), i.e., the inclusion of nonuse values
of the environmental amenity as part of the opportunity cost of the development.
Option values in the sense of Weisbrod were also relevant. The second adjustment
was to build in the rate of technological change as a decay factor in the benefits
of development, which lowers the present value of the benefits of the electricity
by effectively raising the discount rate applied to development benefits. The third
adjustment involved raising the present value of the benefits of the amenity by
lowering the effective discount rate applied to those forgone benefits. The ratio-
nale here was already familiar: If environmental assets are fixed in supply and
demand for them grows through time, there will be a relative price effect that will
inflate the conservation benefits. The end result was a modified cost-benefit algo-
rithm that appreciated the benefits of conservation and depreciated the benefits of
development, which biased the rule more in favor of conservation.
A final consideration involved the value of delaying irreversible action. Delay
generates more information, and more information provides for better decisions.
Because the context of most decisions is one of uncertainty and because uncertainty
pervades our knowledge of the natural environment and how it provides for the life
forms it supports, there is a value to be attributed to conservation over and above
all the previous considerations. Essentially, conservation generates information.
Also initially called option value, generating confusion with Weisbrod’s concept,
this notion subsequently became known as quasi-option value. Quasi-option value
reflects the value of information gained from the delay in making an irreversible
decision. The notion also underlies options theory, which, originally formulated in
the context of financial markets, results in a modification to cost-benefit analysis as
shown by Dixit & Pindyck (75). The same idea can also be utilized to make some
sense of “the precautionary principle,” widely adopted in national and international
legislation, but which has little formal meaning. Quasi-option value suggests there
is a value to caution in a world of irreversibility and uncertainty.
While the basic infrastructure of cost-benefit analysis remained intact during
these developments, the modifications have proved to be very important. Environ-
mental economics has therefore improved and extended cost-benefit analysis.

THE CHOICE OF POLICY INSTRUMENTS

The last major theme in the development of environmental economics relates

to the choice of the means of achieving an environmental goal. Traditionally,
environmental policy has been based on what came to be known as “command-
and-control” policies. Although widely used as a piece of terminology, it is not
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 73

always clear what distinguishes command and control measures from other policy
instruments. One strong form of command and control combines setting a target and
telling the regulated party how to achieve the target. The most obvious example
here is the technology-based environmental standard that works by telling the
polluter what technology to use either in production of the good he is producing
or in terms of abatement equipment. The standard (e.g., level of emissions) is
whatever the technology achieves, and the means of achieving the standard is the
technology itself. In other cases, standards may be prescribed with the polluter left
alone to decide how best to achieve the standard. Most literature would classify this
approach also as command and control, but it is clearly primarily a command with
the control being left to the polluter. In contrast to these approaches, environmental
economists have been concerned to advance the use of economic instruments
or market-based instruments. Examples would be pollution taxes, deposit-refund
schemes, and tradable pollution or resource permits. The focus of these instruments
is on sending a price signal to polluters or resource users. The price can be either
explicit, as with an environmental tax, or derived from a quantity control, as with
the price of tradable permits.
The virtues of these approaches lie in the theoretical expectation that they will
(a) minimize the costs of complying with regulations, and (b) stimulate techno-
logical change because the tax (or need to buy permits) is avoided if pollution is
reduced. The notion of an environmental tax related to the money value of environ-
mental damage dates from Pigou (4). Pigou argued that, where “marginal private
net product” deviated from “marginal net social product” (i.e., when there is an
externality) intervention through a tax would be justified as a means of maximiz-
ing the “national dividend” (economic welfare). While the idea of setting taxes
equal to the marginal externality remains the cornerstone of the environmental tax
literature, Baumol & Oates early on demonstrated that taxes still have the desir-
able feature of minimizing compliance costs even in contexts where an arbitrary
standard is set (arbitrary from the standpoint of welfare economics, that is) (76).
The intuition is simple. For any tax rate, each polluter will abate pollution up to the
point where his marginal abatement costs just equal the tax. Hence the marginal
abatement costs of all polluters are equal, and this is the condition for minimizing
the sum of all abatement costs. If firms recognize this cost minimizing theorem, it
becomes difficult to explain why they are usually hostile to environmental taxes.
Some of the reasons suggested include (a) that they pay the tax on all pollution,
optimal and nonoptimal, whereas command and control procedures impose costs
only on the levels of pollution above the standard, and (b) the suspicion that what
starts as an environmental tax can quickly become a general revenue-raising tax.
Advocacy of environmental taxation has been one of the hallmarks of environ-
mental economics. A dominant figure in the early advocacy was Allen Kneese
who, along with John Krutilla, was widely regarded as one of the fathers of envi-
ronmental economics. Both were long associated with Resources for the Future.
Kneese wrote one of the first textbooks on the subject in 1977 (77), just as he also
contributed substantially to communicating research results on the economic value
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

74 PEARCE

of the environment (78). In 1964 Kneese authored an important book on regional

water quality management (79), which was extended and reissued four years later
(80). The volume set out to show that water management in the United States had
not been informed by any balancing of costs and benefits nor by attention to cost-
minimizing market-based incentives. Although critical of the management system
in the Ruhr basin in Germany, Kneese & Bower reported favorably on the pioneer-
ing efforts made there to bring rational economic concepts to bear on that system.
The more recent debate on environmental taxation has centered on the so-called
double dividend debate. Traditionally, economists did not concern themselves with
what happened to tax revenues (although some of Pigou’s work did suggest that
he favored earmarking of taxes). The idea of the double dividend is that a tax on
pollution yields revenues that can then be used to finance a reduction in some
other distortion in the economy, a reduction in labor taxes for example. In this
sense a single instrument, the tax, secures two goals or dividends—the reduction
in pollution and the reduction in labor market distortion. Another way of looking
at environmental taxation is that it always yields two dividends—the pollution
reduction and the benefit of whatever the tax revenues are spent on. The new interest
in double dividends arises because of the presence of existing distortions, i.e., the
analysis concerns what happens in a second-best world. It is usual to distinguish at
least two forms of the double dividend hypothesis. The weak form accepts that the
environment will be better with the tax than without and that well-being will also
improve relative to a situation in which tax revenues are redistributed as lump sums.
The strong form says that there will be two improvements in well-being, which
ignores the lump sum transfer baseline. But whether either double dividend exists
is open to debate. Much depends on the interactions between the environmental
tax and other distortions. For example, the environmental tax may raise the cost
of polluting goods, which reduces consumers’ real incomes. They may or may not
be better off with the tax than without it. What is required is a general equilibrium
modeling of the tax. The double dividend is thus far from assured (81, 82).
In terms of the practical relevance of environmental taxation to policy, an im-
portant event was the creation of the Environment Directorate at the Organisation
for Economic Co-operation and Development (OECD) in Paris. This Directorate
has been a persistent advocate of economic instruments generally; its early work
focused on taxation. It formulated the “Polluter Pays Principle” in the early 1970s
very much along Pigouvian lines, although, by the time of publication, the formu-
lation had been watered down to make almost any regulation imposing a cost on
polluters consistent with the principle (83).
The origins of marketable permits lies with the work of Dales (84), a Canadian
economist. Again, the basic notion is a simple one. Given the need to meet some
target, say X tonnes of pollution, and the source of the pollution is Y emitters,
distribute permits equal to X tonnes to the emitters, and then allow them to buy and
sell the permits. Because pollution without a permit is not allowed, each emitter
will reduce pollution so long as the cost of doing so is less than the price that would
have to be paid for a permit. High abatement–cost polluters will therefore tend to
buy permits, and low-cost polluters will sell permits. The market in permits will
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 75

determine an equilibrium price for the permits. That such a system will minimize
compliance costs was formally demonstrated by Montgomery (85), but the intu-
ition is the same as that for the Baumol-Oates cost minimization theorem for taxes.
While much younger in concept than environmental taxes, tradable permits have
developed rapidly in practice. They have been most widely used in the United States
for the control of acidic pollutants, but they have also been used to control overfish-
ing in several countries. Tradable water rights also exist in various countries, the
goal again being to ensure that scarce water supplies are allocated to those who have
the highest use value for the water. Farmers with water on (or below) their land,
for example, can then sell the rights to extract the water to other users with higher
willingness to pay for the water than that of the farmer. Interestingly, tradable quota
systems are not confined to high-income economies—tradable water rights have
been in place in Chile for two decades, and Chile also has an incipient tradable air
pollution permit market. Significantly, the continuing discussions on how best to
tackle global warming have led to a reasonably widespread consensus that tradable
greenhouse gas permits must be an integral part of the system of control. Trading
will take place not only within national borders, but also under the provisions for
the “flexibility mechanisms” in the Kyoto Protocol, trading will also occur inter-
nationally. Whether a fully fledged international emissions trading scheme will
develop remains to be seen, but there has already been over a decade’s experience
with bilateral joint implementation schemes whereby an emitting nation can offset
some of its target by reducing greenhouse gas emissions in other countries.
In the cases of environmental taxes and tradable permits, what was essentially
a theoretical literature a few decades ago has now become practical policy. Envi-
ronmental taxes are extensive in OECD countries and are emerging fairly fast in
middle-income and even low-income countries. Tradable permits extend across tra-
ditional air pollutants, fisheries, and water, and they are beginning to emerge in the
solid waste sector through tradable recycling obligations and tradable landfill quo-
tas. At one level, then, academic environmental economists have been enormously
successful in getting their ideas adopted. On another level, it remains surprisingly
unclear just how successful the instruments have been themselves. There is a great
need for more ex post study of the effectiveness of economic instruments.

ECOLOGICAL ECONOMICS: A NEW PARADIGM?

No history of environmental economics, however brief, would be complete without

reference to the dissension that pervades the subject. Probably the most substantial
difference of view at the moment concerns ecological economics and the extent to
which it defines a new paradigm when compared to environmental economics. Any
brief survey is also bound to do a disservice to the group claiming the title of ecolog-
ical economics because the literature is growing rapidly and it is far from clear that
the subject is a coherent one. This is as one would expect if indeed it is a new subject.
Ecological economists probably regard environmental problems as being far
more serious than do environmental economists. This shows up in fairly urgent
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

76 PEARCE

clarion calls for action. If so, the issue is an empirical one, and not one that can
be resolved by theorizing. There is far greater emphasis on limits in the ecolog-
ical economics literature; limits set by the carrying capacity of the Earth and its
environments. This is evident in the work of Herman Daly who has long been an
advocate of antigrowth as a means of keeping world economic systems within the
biogeophysical limits of the Earth [see (86)]. Daly’s work is very much in the spirit
of Boulding’s spaceship Earth essay and has also been heavily influenced by the
work of Georgescu-Roegen (87), which has stressed the role of the second law
of thermodynamics, increasing entropy, in making an economic system consistent
with maximum recycling and maximum use of renewable energy. Little seems to be
said by ecological economists about technological change—an introductory text
by the leading advocates of ecological economics does not discuss technological
change at all (88).
A second feature of ecological economics, which follows on from the greater
focus on limits, is its rejection of the substitutability assumption implicit in the
use of neoclassical production functions. This shows up most clearly in the de-
bate over sustainable development. Measures such as green NNP and genuine
savings assume substitutability among all forms of capital. Thus, while ecologi-
cal economists would accept the notion of Hicksian income as income that can
be sustained without running capital assets down, they would add a strong sus-
tainability constraint to the effect that the existing stock of natural capital should
not depreciate. Although not formally demonstrated in the ecological economics
literature, it is possible to show that this added constraint further lowers the level
of consumption per capita than can be sustained through time. In turn, this appears
to be consistent with the ecological economists’ view that overconsumption ac-
counts for much environmental degradation and that lifestyle change is required.
The notion of strong sustainability is attractive in many respects. However, it begs
the question of how the optimal stock of natural capital is determined, unless what
is intended is that what there is is optimal.
A third feature of ecological economics, also following from the previous points,
is a rejection of the smoothness of the various production functions in neoclassical
economics. If there are discontinuities in ecological systems, then comparatively
small variations in economic variables may bring about collapse of ecological
systems. Although this is an extension of the rejection of substitutability, it also has
implications for the choice of environmental policy instruments. It would appear to
suggest a preference for quantity-based command and control regulations because
instruments such as environmental taxes are not certain in their effect. Equally,
such a view may be consistent with the adoption of tradable permit schemes.
Fourth, ecological economists are far more suspicious of the practice of dis-
counting future costs and benefits and of monetizing environmental damage. The
stance against monetization follows logically from the view that the environment
has no substitutes. However, the extent to which the literature is consistent in this
respect is debatable. Some ecological economists appear, for example, to embrace
modified green national accounting, which assumes substitution and monetization.
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 77

Others reject monetized national accounts in favor of satellite systems where mone-
tized accounts have attached to them physical statements about changes in natural
resource endowments. Similarly, the stance on discounting is ambiguous. Like
many environmental economists, ecological economists express a concern that
high discount rates discriminate against future generations and hence against sus-
tainability. It was observed earlier that this result is not new and emerges clearly
from the optimal growth literature of the 1970s. But how far discount rates should
be lowered, even to zero, is not clear, and the literature sometimes reads as if one
should not discount at all. But zero is a discount rate, and zero discounting would
have profound implications for the distribution of well-being between current and
future generations, lowering well-being now to very low levels in the name of
future generations. Again, the role of technology appears to be ignored, and there
is perhaps an implicit substitution of a survivability criterion for a sustainable
welfare criterion.
A final issue is perhaps less significant in dividing ecological from environmen-
tal economists, but it seems fair to say that ecological economists want to revisit
the well-known fact in welfare economics that what is economically efficient is not
necessarily optimal from the standpoint of a social welfare function. Essentially,
social welfare functions are concerned with procedures for aggregating the well-
being of individuals. Rules for aggregation will depend on what constitutes a fair or
deserving distribution of benefits and costs, an issue that may reflect assumed sets
of rights. As is well known, there are many such rules, and there is no unique rule
by which just distributions can be judged. The issue of how to weight individual
gains or losses between people now and, for that matter, people across generations
is not resolved in the philosophical literature nor in the economics literature. Nor,
many would argue, can it be. There are as many forms of cost-benefit analysis, for
example, as there are social welfare functions. Ecological economists tend to em-
phasize one set of social welfare functions, for example those stressing the rights of
future generations. Environmental economists may not be so motivated, although
many are. The issues are complex because it is not clear whether nonexistent beings
can be said to have any rights at all, an issue much debated in the moral philoso-
phy literature. Although these differences between ecological and environmental
economists should not be exaggerated, they are not resolvable by logic.
It may be no surprise if ecological economics has contradictions and ambi-
guities; it is a new subject, and it appears to be a broad church, embracing many
different persuasions. What is true is that it provides a refreshing challenge to envi-
ronmental economists who might otherwise not challenge some of the fundamental
assumptions underlying their own subject.

CONCLUSION

Environmental economics is remarkably young as a subdiscipline of economics.

It can claim at most a 50-year heritage dating from the formation of Resources
for the Future in 1952, and more probably, a 40-year duration dating from some
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

78 PEARCE

of the seminal works of the 1960s by the likes of Boulding, Krutilla, Kneese,
Dales, and Weisbrod. The 1970s onward saw a lengthy period of theoretical con-
solidation in terms of valuation theory and policy instrument design, together with
the most important advances in optimal growth theory in the presence of envi-
ronmental constraints. The last decade has witnessed a greater effort to produce
an environmental macroeconomics through the focus on sustainable development
and, not covered in this essay, environment, and international trade. From humble
beginnings in the 1960s and 1970s, the discipline now boasts two major inter-
national federations of environmental economists, the (North American–based)
Association of Environmental and Resource Economists and the European Asso-
ciation of Environmental and Resource Economists. Ecological economists have
formed their own association, the International Society for Ecological Economics.
Perhaps more important, theoretical exercises, in what to the public are obscure
journals, a few decades ago have been transformed into practical policy measures
that now figure in the armory of environmental policy in the political world. Even
if major obstacles remain, it is no longer necessary to explain the language of
environmental economics to decision makers. That situation has been changed for
ever.

The Annual Review of Energy and the Environment is online at

http://energy.annualreviews.org

LITERATURE CITED
1. Barnett H, Morse C. 1963. Scarcity and 8. Hicks JR. 1943. The four consumer’s sur-
Growth: The Economics of Natural Re- pluses. Rev. Econ. Stud. 11:31–41
source Availability, Baltimore: Johns Hop- 9. Kaldor N. 1939. Welfare propositions of
kins Univ. Press economics and interpersonal comparisons
2. Christy N, Potter FT. 1961. Trends in Natu- of utility. Econ. J. 49:549–52
ral Resource Commodity Statistics of Pri- 10. Eckstein O. 1958. Water Resource Devel-
ces, Output, Consumption, Foreign Trade opment: the Economics of Project Evalua-
and Employment in the United States, tion. Cambridge, MA: Harvard Univ. Press
1860–1957, Baltimore: Johns Hopkins 11. Krutilla J, Eckstein O. 1958. Multipurpose
Univ. Press River Development. Baltimore: Johns Hop-
3. Carson R. 1962. Silent Spring, Cambridge, kins Univ. Press
MA: Riverside 12. McKean R. 1958. Efficiency in Govern-
4. Pigou A. 1920. The Economics of Welfare. ment Through Systems Analysis. New York:
London: Macmillan Wiley
5. Dupuit AJ. 1844. On the measurement of 13. Gray L. 1914. Rent under the assumption
the utility of public works. Transl. R Bar- of exhaustibility. Q. J. Econ. 28:466–89
back, 1952, in Int. Econ. Pap. 2 14. Hotelling H. 1931. The economics of ex-
6. Dupuit AJ. 1853. On utility and its haustible resources. J. Polit. Econ. 39:137–
measure-on public utility. J. Econ. 36:1– 75
27 15. Dasgupta P, Heal J. 1979. Economic The-
7. Hicks JR. 1939. Foundations of welfare ory and Exhaustible Resources. Cambridge
economics. Econ. J. 49:696–712 UK: Cambridge Univ. Press
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 79

16. Gordon HS. 1954. Economic theory of a environmental resources. Environ. Resour.
common property resource—the fishery. J. Econ. 1(1):1–15
Polit. Econ. 62:124–62 33. Solow R. 1986. On the intergenerational al-
17. Hardin G. 1968. The tragedy of the com- location of exhaustible resources. Scand. J.
mons. Science 162:1243–48 Econ. 88:141–49
18. Fisher AC. 1981. Resource and Environ- 34. Eisner R. 1988. Extended accounts for na-
mental Economics. Cambridge UK: Cam- tional income and product. J. Econ. Lit. 26:
bridge Univ. Press 1611–84
19. Kapp K. 1950. The Social Costs of Private 35. Repetto R, McGrath W, Wells M, Beer C,
Enterprise. London: Spokesman Press Rossini F. 1989. Wasting Assets: Natural
20. Boulding K. 1966. The economics of the Resources in the National Income Acco-
coming spaceship Earth. In Environmen- unts, Washington, DC: World Resour. Inst.
tal Quality in a Growing Economy, ed. H 36. Hamilton K, Lutz E. 1996. Green National
Jarett, pp. 3–14. Baltimore: Johns Hopkins Accounts: Policy Uses and Empirical Ex-
Univ. Press perience. Pap. 039. Environ. Dep. Wash-
21. Ayres R, Kneese A. 1969. Production, con- ington, DC: World Bank
sumption and externalities. Am. Econ. Rev. 37. Pearce DW, Atkinson G. 1993. Capital the-
59(3):282–97 ory and the measurement of sustainable de-
22. Coase R. 1960. The problem of social cost. velopment. Ecol. Econ. 8:103–8
J. Law Econ. 3:1–44 38. Hamilton K. 1996. Pollution and pollution
23. World Commission on Environment and abatement in the national accounts. Rev. In-
Development (WCED). 1987. Our Com- come Wealth. 42(1):13–33
mon Future, Oxford: Oxford Univ. Press 39. Hamilton K, Clemens M. 1999. Genuine
24. Hicks JR. 1939. Value and Capital. Oxford: savings rates in developing countries.
Oxford Univ. Press World Bank Econ. Rev. 13:333–56
25. Boserup E. 1980. Population and Techno- 40. Hartwick J. 1977. Intergenerational equity
logical Change, Chicago: Univ. Chicago and the investing of rents from exhaustible
Press resources. Am. Econ. Rev. 67:972–74
26. Dasgupta P, Heal G. 1974. The optimal de- 41. Hamilton K. 2000. Sustaining Economic
pletion of exhaustible resources. Rev. Econ. Welfare: Estimating Changes in Per Capita
Stud. Symp. pp. 3–28 Wealth. Policy Res. Work. Pap. 2498. Wash-
27. Solow R. 1974. The economics of resour- ington, DC: World Bank
ces or the resources of economics. Am. 42. Hotelling H. 1947. The economics of pub-
Econ. Rev. 64:1–14 lic recreation. In The Prewitt Report. Wash-
28. Stiglitz J. 1974. Growth with exhaustible ington, DC: Natl. Parks Serv.
natural resources: efficient and optimal 43. Trice AH, Wood SE. 1958. Measurement of
growth paths. Rev. Econ. Stud. Symp., pp. recreation benefits. Land Econ. 34(3):195–
123–38 207
29. Hartwick J. 2000. National Accounting and 44. Clawson M. 1959. Methods of Measuring
Capital. Cheltenham, UK: Elgar the Demand for and Value of Outdoor
30. Weitzman M. 1976. On the welfare sig- Recreation, Repr. No. 10. Washington, DC:
nificance of national product in a dynamic Resour. Future
economy. Q. J. Econ. 90:156–62 45. Clawson M, Knetsch JL. 1966. Economics
31. Hartwick J. 1990. Natural resources, of Outdoor Recreation. Baltimore: Johns
national accounting and economic de- Hopkins Univ. Press
preciation. J. Public Econ. 43(3):291– 46. Ward FA, Beal D. 2000. Valuing Nature
304 with Travel Cost Models. Cheltenham, UK:
32. Mäler K-G. 1991. National accounts and Elgar
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

80 PEARCE

47. Ridker R. 1967. Economic Costs of Air Pol- 62. Ciriacy-Wantrup SV. 1947. Capital returns
lution: Studies in Measurement, New York: from soil conservation practices. J. Farm
Praeger Econ. 29:1181–96
48. Ridker R, Henning JA. 1967. The deter- 63. Mack RP, Myers S. 1965. Outdoor recre-
minants of residential property values with ation. In Measuring Benefits of Govern-
special reference to air pollution. Rev. ment Investments, ed. R Dorfman. Wash-
Econ. Stat. 49(2):246–57 ington, DC: Brookings Inst.
49. Freeman AM. 1974. On estimating air pol- 64. Davis R. 1963. The value of outdoor recre-
lution control benefits from land value stud- ation: an economic study of the Maine
ies. J. Environ. Econ. Manag. 1(1):74–83 woods. PhD diss., Harvard Univ., Cam-
50. Rosen S. 1974. Hedonic prices and implicit bridge, MA
markets: product differentiation in perfect 65. Mitchell RC, Carson RT. 1989. Using Sur-
competition. J. Polit. Econ. 82(1):34–55 veys to Value Public Goods: The Contin-
51. Cheshire P, Sheppard S. 1998. Estimating gent Valuation Method. Washington, DC:
the demand for housing, land and neigh- Resour. Future
bourhood characteristics. Oxford Bull. 66. Carson RT, Mitchell RC, Hanemann WM,
Econ. Stat. 60(3):357–82 Kopp RJ, Presser S, Ruud P. 1994. Contin-
52. Palmquist R, Isangkura A. 1999. Valuing gent Valuation and Lost Passive Use: Dam-
air quality with hedonic and discrete choice ages from the Exxon Valdez. Discuss. Pap.
models. Am. J. Agric. Econ. 81(5):1128–33 94–18. Washington, DC: Resour. Future
53. Commission on the Third London Airport. 67. Arrow K, Solow R, Portney PR, Leamer
1970. Papers and Proceedings, Vol. 7. Lon- EE, Radner R, Schuman H. 1993. Report
don: HMSO of the NOAA panel on contingent valu-
54. Walter A. 1976. Noise and Prices. Oxford: ation. Fed. Regist. Jan.15. 58(10):4601–
Oxford Univ. Press 14
55. Nelson J. 1982. Highway noise and prop- 68. Pearce DW, Ozdemiroglu E, Bateman I,
erty values. J. Transp. Econ. Policy. 16: Day B, Hanemann WM, et al. 2002. Stated
117–38 Preference Techniques for Estimating Eco-
56. Nelson J. 1980. Airports and property val- nomic Value. Cheltenham, UK: Elgar
ues: a survey of recent evidence. J. Transp. 69. Louviere JJ, Hensher DA, Swait JD. 2000.
Econ. Policy. 14(1):37–52 Stated Choice Methods: Analysis and App-
57. Schelling T. 1968. The life you save may lication. Cambridge UK: Cambridge Univ.
be your own. In Problems in Public Expen- Press
diture Analysis, ed. SB Chase. Washington, 70. Krutilla JV. 1967. Conservation reconsid-
DC: Brookings Inst. ered. Am. Econ. Rev. 57(4):777–86
58. Ulph A. 1982. The role of ex ante and ex 71. Weisbrod B. 1964. Collective consumption
post decisions in the value of life. J. Public services of individual consumption goods.
Econ. 18:265–76 Q. J. Econ. 78:471–77
59. Thomas HA. 1963. The animal farm: a 72. Lind R, ed. 1982. Discounting for Time
mathematical model for the discussion of and Risk in Energy Policy. Baltimore: Johns
social standards for control of the environ- Hopkins Univ. Press
ment. Q. J. Econ. 77:143–48 73. Portney P, Weyant J, eds. 1999. Discount-
60. Mäler K-G. 1974. Environmental Eco- ing and Intergenerational Equity. Wash-
nomics: A Theoretical Inquiry. Baltimore: ington, DC: Resour. Future
Johns Hopkins Univ. Press 74. Krutilla JV, Fisher A. 1985. The Economics
61. Hanemann M. 1992. Preface. In Pricing the of Natural Environments. Washington, DC:
European Environment, ed. S Navrud, pp. Resour. Future. 2nd ed.
9–35. Oxford: Oxford Univ. Press 75. Dixit A, Pindyck R. 1994. Investment
3 Oct 2002 9:52 AR AR171-EG27-03.tex AR171-EG27-03.sgm LaTeX2e(2002/01/18) P1: GCE

HISTORY OF ENVIRONMENTAL ECONOMICS 81

Under Uncertainty. Princeton, NJ: Prince- tect the environment and improve the tax
ton Univ. Press system at the same time? South. Econ. J.
76. Baumol W, Oates W. 1971. The use of stan- 61:915–22
dards and prices for protection of the envi- 83. OECD (Organisation for Economic Co-
ronment. Swed. J. Econ. 73:42–54 operation and Development). 1975. The
77. Kneese AV. 1977. Economics and the En- Polluter Pays Principle: Definition, Anal-
vironment. London: Penguin ysis, Implementation. Paris: OECD
78. Kneese AV. 1984. Measuring the Benefits 84. Dales JH. 1968. Pollution, Property and
of Clean Air and Water. Washington, DC: Prices. Toronto: Univ. Toronto Press
Resour. Future 85. Montgomery WE. 1972. Markets in li-
79. Kneese AV. 1964. The Economics of Re- cences and efficient pollution control pro-
gional Water Quality Management. Balti- grams. J. Econ. Theory 5:395–418
more: Johns Hopkins Univ. Press 86. Daly H. 1977. Steady State Economics. San
80. Kneese AV, Bower B. 1968. Managing Wa- Francisco: Freeman
ter Quality: Economics, Technology, Insti- 87. Georgescu-Roegen N. 1971. The Entropy
tutions. Baltimore: Johns Hopkins Univ. Law and the Economic Process. Cam-
Press bridge, MA: Harvard Univ. Press
81. Bovenberg AL, de Mooij RA. 1994. Envi- 88. Costanza R, Cumberland J, Daly H, Good-
ronmental levies and distortionary taxation. land R, Norgaard R. 1997. An Introduction
Am. Econ. Rev. 94:1085–89 to Ecological Economics. Boca Raton, FL:
82. Oates WE. 1995. Green taxes: Can we pro- St. Lucie