96/12

June 1996 Documents

Statistics Norway
Research Department
Environmental Valuation, Cost
Benefit Analysis and Policy
Making: A Survey
1. Introduction
Monetary valuation of environmental goods has by now become the subject of numerous economic
books and articles. Interest in the topic seems to be increasing in the economics profession, and both
theoretical insight, methodological improvements and the number of empirical findings are expanding
rapidly.
The aim of such valuation is usually to incorporate environmental concerns into a cost-benefit-
analysis. Another purpose is to construct environmentally adjusted national income measures (UN,
1993). Environmental value estimates have also been combined with macroeconomic models, e.g. to
estimate welfare effects of a climate treaty (Brendemoen and Vennemo, 1994). Further, estimated
willingness to pay is now accepted in the USA as a basis for legal compensation claims for damages
to natural resources caused by spill of hazardous substances.
In this paper, I will review some of the literature on environmental valuation, with a particular view to
the relevance of such valuation for environmental policy making. In line with this purpose, much
emphasis will be put on interpretation issues and on the controversies in the valuation literature. No
extensive summary of the more technical parts of the literature on empirical estimation methods will
be provided. Moreover, I will concentrate my attention to the intended use of environmental value
estimates for inclusion in cost -benefit analysis!
Thus, I will start in sectore 2 with a look at cost-benefit analysis in general, and its relevance as
information background for policy making. In section 3, a brief overview of some currently used
methods to value environmental changes in monetary units is provided. Section 4 reviews the
professional debate following the Exxon Valdez accident in 1989, while section 5 discusses some
distributional issues relating to environmental valuation. Finally, section 6 points out some
philosophical objections to environmental valuation, which may be of importance for their use in a
policy making process.
2. Cost-benefit analysis
Cost-benefit analysis is a way of assessing the consequences of public projects and reforms, in which
the estimated benefits are weighed against the costs. For this purpose, all consequences must be
measured in the same unit, and the traditional choice of unit is money. To be explicitly included in a
cost-benefit analysis, then, environmental changes must be valued in monetary terms.
Dreze and Stern (1987) provide a standard reference for the theory of cost-benefit analysis. As they
point out (p. 911), the two basic ingredients of cost-benefit analysis are the ability to predict
consequences (a model) and the willingness to evaluate them (an objective function). The shadow
price
of a good is defined as the net impact on social welfare of a unit increase in the supply of that
good. Accordingly, the shadow price of wilderness, for example, is defined by the change in social
welfare when the amount of wilderness changes by one measurement unit (for example one m
2). The
terms value (or social value) and shadow price are frequently used interchangeably.
For policy purposes, an important consideration is whether normative premises are embedded in the
analysis, and if so, which ones. One can hardly expect a policy maker to be interested in the economic
analysis if one employs a priori
assumptions which run counter to the policy maker's political goals.
Dreze and Stern's theoretical approach is fairly general, allowing for various specifications of the
social welfare function. However, they emphasise that shadow prices cannot generally be estimated
1
For a discussion of environmentally adjusted national income measures, see Aaheim and Nyborg (1995).
1
without an explicit model and a specific choice of welfare function. For applied, numerical analysis,
this means that in addition to a proper description of how the economy works, normative premises for
the analysis must also be chosen.
Some economists have argued that this is not necessary, and that efficiency and distribution are two
distinct concerns which could and should be dealt with separately (Hicks, 1939, Frank 1992). Dreze
and Stern, on the other hand, discuss several arguments for the view that distributional concerns can
be disregarded in the analysis, and reject all of them. 2 They conclude that the analyst should present
results involving a selection of possible social welfare functions. Since the environment affect
different people in different ways, changing the specification of the social welfare functions will
generally imply that shadow prices for environmental goods change as well. Thus, the social value of
the environment cannot be defined in an objective way; it depends crucially on the chosen social
welfare function.
Dreze and Stern further suggest that one may use the inverse optimum methoth to assess welfare
weights which reflect the government's value judgement. This implies estimating implicit welfare
weights, either by calculating the welfare weights implied by previous policy decisions, or asking
decision makers more directly about their judgement concerning trade-offs between income to
different groups or individuals. The former version of this method requires the assumption that
previous decisions (or direct statements from politicians) originate from optimising behaviour.
In applied cost-benefit analysis, however, one is usually concerned with the maximisation of some
concept of aggregate income or well-being, disregarding its distribution entirely. Within the formal
framework of Dreze and Stern, then, this can be defended in two ways. The first is to assume that the
government is capable of redistributing income costlessly. In that case, the government can maximise
aggregate income when choosing public projects, and then subsequently distribute income among
individuals in any way it wishes. Distributional concerns need then not be addressed in the project
analysis.
Such an assumption clearly seems somewhat unrealistic. Dreze and Stern actually claim that this
represents a mistaken understanding of second-best welfare economics. It may be true that policy
instruments exist - e.g. income taxes - which allow a more direct influence on the distribution of
income than public projects; and the planner's model should include them. However, in general,
redistributing income using these instruments will have social costs, and therefore the implications of
projects for income distribution should not be ignore& (Dreze and Stern, 1987, p. 958). The
assumption that costless lump-sum redistribution is possible is common in economics, but the
frequency with which this assumption is made does not make it more plausible. Indeed, Hammond
(1979) shows that in the presence of private information, lump-sum redistribution based on anything
else than unalterable identifiable individual characteristics (such as age or sex) is generally infeasible
in large economies, due to incentive compatibility constraints.
2.1 Normative premises
The alternative justification of disregarding distributional concerns in the applied project evaluation is
to introduce a utilitarian social welfare function. In addition, one must make the assumption that every
individual has the same marginal utility of income, regardless of income or personal characteristics 3 .
This amounts to assigning an equal welfare weight to all individuals' income changes, which means
2 See also Auerbach (1985) and Quiggin (1995).
3 Alternatively, and equivalently, one might assume that the marginal utility of income varies, but let the weight
attached to any single individual's utility vary disproportionally with her marginal utility of income. Under the
assumption that rich people have a lower marginal utility of income than poor people, this is equivalent to
systematically giving rich people's utilities more weight. Such a normative premise may seem unethical, but is,
nevertheless, perhaps a fairly good description of actual policies in some countries.
2
that explicit welfare weights are not needed in the cost-benefit calculations. This is of course still a
quite explicit normative premise; it allows us to disregard distributional concerns simply because we
have chosen the normative view that distribution issues are not socially important. 4
The ethical aspects of utilitarianism has been the subject of much debate both in economics and
philosophy; see, for example, Sen and Williams (1982). John Harsanyi (1955) argued that individuals
who are maximising their own expected utility will prefer a social welfare function corresponding to
utilitarianism, provided that their evaluations are made in a situation where they do not know their
own position in society. On the other hand, others have argued strongly against the utilitarian moral
philosophy; for example on the grounds that it attaches no weight to concepts such as rights and
duties (Sen, 1985). Note, however, that even if one accepts utilitarianism, the assumption of equal
marginal utilities of income remains controversial.
This obviously represents a problem for cost-benefit analysis as information background for policy
decisions: The distributional preferences of the current government become irrelevant to the analysis;
distributional weights are determined a priori. The analyst has implicitly chosen a specific social
welfare function quite independently of the political views of the government. 5 This is clearly a quite
controversial point of departure for a policy analysis, and has indeed prompted a lot of dispute (e.g.
Kelman (1981), Blackorby (1990), Blackorby and Donaldson (1990), Hammond (1990), Bromley
(1990)). In fact, it poses two distinct problems: Firstly, whether equal weight on all income changes
can be defended ethically, and secondly, whether decision makers will find the analysis useful at all if
their answer to the first question is no.
2.2 Some views on cost-benefit analysis as a basis for policy recommendations
Authors differ a lot with respect to their confidence in the ability of cost-benefit analysis to identify
optimal provision levels of environmental goods (or, less ambitiously, welfare-increasing marginal
changes). Navrud (1992, p. 37) seems to be on the optimistic side:
There are two main arguments for putting a price on environmental goods such as water and air
quality, biodiversity and natural environments. First, we need to know the marginal value of
environmental goods to find the socially right (optimal) quantity/quality of different environmental
goods. Second, if environmental goods are not valued explicitly, they will still be valued implicitly
through policy decisions. Since the decision makers are often not aware of that they make these
valuations, this procedure produces an arbitrary and inconsistent set of prices.
Navrud advocates the idea that a socially optimal level of environmental goods can indeed be
identified, without discussing that the optimal solution will depend on which policy goals one is
aiming for, i.e. which social welfare function is chosen. Furthermore, he seems to be on the
pessimistic side regarding the quality of the political decision process, since he argues that the
implicit values deduced from such processes are likely to be not only inconsistent, but even arbitrary.
This latter view is clearly inconsistent with the inverse optimum method advocated by Dreze and
Stem, for example.
Freeman (1993, pp. 9-12) is somewhat more modest on the cost-benefit methodology's behalf,
regarding the extent to which it can provide final answers in a world of political disagreement. He
points out that decision makers may have other objectives besides economic efficiency, such as equity
considerations and intergenerational effects. Freeman does not regard advocating cost-benefit analysis
4
More precisely, utilitarianism implies that the distribution of utility is not important, only their sum; and the
additional assumption of equal marginal utility of money ensures that the income distribution is not important
either.
5
Moreover, quite in contrast to the practice in most of modern economics, he is employing a cardinal utility
concept, based on the view that a dollar can be interpreted as a unit of utility.
3
as a simple decision rule as being particularly useful. Rather, he argues that the main value of cost-
benefit analysis lies in its ability to organise and simplify information.
Bromley (1990) takes a much more sceptical standpoint regarding the political usefulness of cost-
benefit analysis, arguing as follows: Economists who have persevered in this tradition seem content
to overlook the logical inconsistencies in welfare economics, this obduracy apparently being justified
on the grounds that a little economic analysis - even if indefensible on theoretical grounds, and
therefore bogus - is better than a political process left to its own devices. Vatn and Bromley (1994, p.
144), discussing environmental valuation, argue that ...the necessity claim for such valuation (...)
must rest on clear proof that the values (prices) derived from hypothetical valuation studies capture all
of the information pertinent to a particular environmental choice. In the absence of such proof, values
(prices) from hypothetical valuation studies carry no more normative significance than do competing
claims expressed by self-proclaimed interest groups on either side of any particular decision.
Evidence would suggest that a great many enlighteneth choices concerning the environment have
been taken in the absence of pricing.
2.3 Practical use of cost-benefit analysis in environmental decision making
While there is disagreement regarding whether environmental values and cost-benefit analysis should
be used for policy prescriptions, there also seem to be somewhat diverging opinions on whether such
analysis have actually been used in policy formation. The discussion below focuses on environmental
policy making, i.e. questions where the environmental issue is a major concern. Several of the
references below are collected in Navrud (1992).
The use of cost-benefit analysis as a part of environmental policy making differs quite a bit between
countries. In the United States, cost-benefit analysis is frequently carried out by federal agencies, and
this is at least partly due to political decisions. In 1981, President Reagan issued an executive order
requiring all federal agencies to prepare a Regulatory Impact Analysis before undertaking any major
regulatory action. Following this, the Environmental Protection Agency (EPA) developed its own
manual of cost-benefit analysis. Later, Executive Order 12291 issued by President Reagan and
Executive Order 12866 issued by President Clinton confirmed this (Hanemann, 1992, Portney, 1994).
However, the fact that cost-benefit analysis are frequently carried out in the USA does not necessarily
mean that they have had a large impact on policy. There seem to be diverging opinions on this point.
One reason why cost-benefit analysis may have had a limited influence on actual decisions is that
some of the laws governing the use of environmental regulations, such as the Clean Air Act, actually
do not permit balancing benefit and costs (Portney, 1990).
Another important development in the USA was the pass in Congress of CERCLA, the Compre-
hensive Environmental Response, Compensation and Liability Act, in 1980. It established a liability
on the potentially responsible parties to pay damages for the injuries to natural resources resulting
from the spill or release of hazardous substances, in addition to the costs of clean-up, removal,
remediation and any other necessary responsible costs (Hanemann, 1992). Although CERCLA was
not particularly concerned with cost-benefit analysis, it prompted a renewed interest in valuation of
natural resources and environmental goods. The interest reached a peak after the Exxon Valdez oil
spill in 1989. A summary of the debate following this event is provided in section 4 below.
Environmental values in cost-benefit analysis (and indeed, cost-benefit analysis altogether) seem to
have been less used in Europe than in the USA. Kuik et al. (1992) discuss environmental decision
making in Europe, and conclude that benefit estimates play a role in environmental decision making,
not only in the United States but also in Europe. (...) The extent and purpose of use of benefit
estimates varies across countries. The extent to which benefit estimates are used in Europe to
4
influence environmental policy decisions is negligible. However, they are used at the project level in
some sectors in a number of European countries. 6
Navrud and Strand (1992), discussing experiences from Norway, conclude as follows: So far the
benefit estimation studies seem to have been useful for providing support for decisions involving the
environmental goods they valued, but they have not played a crucial role in the decision making
process. Thus, the valuation exercises appear to be used as arguments for decisions which had
already been made, rather than serving as informational background for such decisions.
Similarly, Johansson and KristrOm (1992), commenting on the Swedish experience, conclude: The
general impression, however, is that these social cost-benefit evaluations have played a minor role in
the actual outcome of the decision making process. Even if a cost-benefit analysis shows that a project
is highly profitable to the entire society, the project is not necessarily (or even generally) undertaken,
and vice versa.
A discussion of the situation in the Netherlands is provided by Hoevenagel and Kuik (1992). They
maintain that the interest in environmental valuation is much larger among Dutch researchers than
among their sponsors, and that all four government departments which are directly involved in
environmental policy seem to show little interest in the subject. They explain this by a lack of
confidence in welfare-theoretic approaches based on the concept of willingness to pay.
In the United Kingdom, the Department of the Environment issued a specific set of guidelines of
economic appraisals in 1991, which openly embraced valuation techniques (Turner et al., (1992)).
Turner et al. claim that monetary valuation of environmental goods is currently enjoying a revival in
the UK. One of the explanations they provide for this is that a decade of Conservative government
brought with it a concern for efficiency in government, while politicians at the same time faced a
large demand for environmental improvements. Turner et al. do not, however, analyse the impact such
valuation might have had on actual decisions, and they maintain that monetary valuation remains
controversial in the UK.
3. Some methods for estimating environmental values
Individuals' willingness to pay for environmental goods is a central concept in environmental cost-
benefit analysis, but empirical data on this is not readily available from market prices. Several
estimation methods have been proposed, however. Some focus on market data, while others use
interview techniques. Some are direct in the sense that individuals are asked directly about their
valuation. Others are indirect, meaning that valuation of the environmental good is deduced from
other information, for example by using specific assumptions about the nature of the relationship
between demand for a private market good and the environmental good in question.'
In this section, I will give a brief overview of the most commonly used methods for estimating
environmental values. Special emphasis is put on the contingent valuation method, since this method
has been the focus of much debate lately. A thorough presentation of measurement of environmental
values can be found in Freeman (1993). See also Hanley and Spash (1993), Braden and Kolstad
(1991), and Johansson (1993, 1987).
6
It should be noted here that Kuik et al. seem to define <colic)/ decisions as choices between general policies on
a macro level, not including, for example, the adoption of a particular project.
7 The distinction between direct and indirect method is not quite equivalent to the distinction between market
based and interview-based techniques; for example, conjoint analysis (see the description below) is an indirect
method which is based on data from individual interviews.
5
3.1 Methods based on market prices
Although most environmental goods are not bought and sold in markets, the use of an environmental
goods is often connected to consumption of private, traded goods and/or services. The private good
may be complementary to the environmental good, as in the case of a cross-country skiing area and a
pair of skis, or they may be substitutes, for example clean water from the tap and bottled drinking
water. Information about the demand for such private goods (or services) can be used to estimate the
latent demand for the associated environmental good. Such estimation depends, of course, on the
plausibility of the assumptions made about the relationship between demand for the market good and
the environmental good.
The travel cost method has frequently been used to value recreational sites. A crucial assumption is
that consumers regard the journey to and from the site as a cost, not as part of the recreation. The
travel costs incurred may then be regarded as a price to visit the recreational site. Using information
about visitors' travel costs, and studying the relationship between the number of visits to the site and
individual' travel costs, one can estimate a demand function for the recreational services of the site.
For a discussion of the method, see Bockstael et al. (1991).
Hedonic methods are based on the fact that some goods or services, although serving approximately
the same purpose, are not perfectly homogeneous. For example, two living houses may differ with
respect to the number of rooms, the view from the house, and local environmental features, such as air
pollution. By comparing the prices of houses with varying characteristics, and using econometric
techniques such as multiple regression analysis, it is possible to estimate a market valuation of those
characteristics.
Similarly, hedonic methods have been used to estimate how much larger wages workers demand to
accept risky work, and this have been used to deduce valuations for a statistical life. However, this
application requires an assumption that individuals behave in accordance with expected utility theory
when faced with risk; an assumption which has been questioned by several scholars (Kahneman and
Tversky, 1979, Rubinstein, 1988).
A large number of studies have been conducted using hedonic methods, particularly in the USA
However, such methods are less reliable if the relevant markets are heavily regulated, since in such
cases, prices may not reflect marginal willingness to pay for the commodity or service in question. In
several European countries, this is the case, or has been so until recently, for the housing and/or
labour market. This is perhaps one reason why such methods have not been used much in Europe. A
closer description of hedonic methods can be found in Freeman (1993).
In some cases, environmental goods can be valued by using information about the demand for private
good substitutes. For example, a filtering device to purify tap water may diminish or even abolish the
problems connected to contaminated drinking water. In such cases, demand for the private substitute
can give an indication of the underlying demand for the environmental good. Usually, it is not
possible to find a perfect substitute for an environmental good. Purifying equipment does perhaps not
clean the tap water good enough, or consumers may prefer to know that their drinking water
originates from a pure water source.
3.2 The contingent valuation method
The methods mentioned above are based on market prices, while other methods involve, instead,
asking people directly how they evaluate environmental goods. The most prominent of these is the
contingent valuation method (CVM). This method has been extensively used in recent years, and a
huge body of empirical and methodological research has developed. A brief summary of the method is
provided below, while a review of the recent debate concerning the method is given in subsequent
sections. A detailed presentation of CVM can be found in Mitchell and Carson (1989).
6
In contrast to the previously mentioned valuation methods, CVM is capable of measuring so-called
non-use values or existence values. Use value is the value someone attaches to his or her actual use of
the good in question. In addition to the use value, a person might be willing to pay something to
secure someone else's use of the environment, for instance members of future generations. A person
may also be willing to pay something just to know that an amenity is not being explored, or that a
species is being preserved. Several CVM studies have found that a large proportion of reported
willingness to pay relates to other motives than one's own use. For example, in Stevens et al. (1991),
only 7 percent of respondents' willingness to pay for the protection of four wildlife species was
assigned to a current use or option category. Respondents allocated 34 percent of their willingness to
pay to a bequest value category, while 48 percent was reported to be motivated by the view that
animals have a right to exist independent of any benefit or harm to people. 8
The distinction between use values and non-use values is not clear-cut, since some authors include the
value of others' use (for example future generations) in the concept of use value, while others regard
this as part of the non-use value. Some authors also maintain that e.g. the pleasure obtained by
looking at pictures of a beautiful scenery in a book is a kind of use value. In this latter case, estimated
values obtained by using e.g. the travel cost method clearly do not provide estimates of all kinds of
use value associated with that site.
Indeed, some critics have claimed that the concept of existence value cannot be defined in an
operationally meaningful way (Larson, 1991, Cummings and Harrison, 1995). The distinction
between use valueo and existence value further touches upon a fundamental issue raised by Sagoff
(1988) among others; namely, whether willingness to pay for environmental goods can reasonably be
interpreted as expressions of individuals' personal, exogenous preferences at all. I will return to this
discussion in section 6.
In a contingent valuation survey, one asks people directly about their willingness to pay to secure
(avoid) a change in the provision of some specified good, or, alternatively, what compensation they
would require in order to give it up (accept it). The questions are usually of a hypothetical character,
such that no actual payments are made. Responses to such questions are naturally contingent upon the
hypothetical circumstances under which respondents are told that the good is to be provided, which
means that one must be careful to use the estimated values in another context that the one in which
they were originally obtained. 9
The hypothetical nature of the questions also imply that misunderstandings between interviewer and
respondent may easily occur. Much methodological research has recently focused on how to avoid
such misunderstandings. Further, the unfamiliar situation of being asked to trade personal income
against a public good may create ambiguities; respondents may, for example, never have considered
such a trade-off before, so that their responses may not be thoroughly considered. Mitchell and
Carson (1989, pp. 236-237) provide a list of potential sources of bias in contingent valuation studies.
They also suggest several procedures which can be used to reduce the occurrence of such bias. See
also the summary of the NOAA Panel's recommendations in section 4.4 below.
3.3 Other methods
3.3.1 Conjoint analysis
Conjoint analysis is quite similar to the CVM in that the analyst asks people about their judgements,
rather than using market data or other revealed choices. Like CVM, conjoint analysis is based on
hypothetical choices. However, people are not asked directly about their willingness to pay or
8 See also the discussion in Mitchell and Carson (1989), p. 291-292.
9 This is nevertheless frequently done. A discussion of such benefit transfers is found in Navrud (1994).
7
willingness to accept, but are instead asked to compare situations which vary in several respects, and
state which one they prefer. For example, an individual could be asked to choose between travelling
by train from A to B in 40 minutes, with a view of the sea, paying 10 USD, or travelling the same
distance in 30 minutes with no view, paying 12 USD. A series of such questions is asked. The analyst
can then apply econometric estimation techniques and infer consumers' implicit valuation of time and
of a nice sea view. Conjoint analysis thus has similarities with hedonic methods as well as contingent
valuation. For a further discussion of this method, see Louviere (1988).
3.3.2 Implicit valuation
The term implicit valuation is frequently used to describe a method which is very similar to the
inverse optimum method recommended by Dreze and Stern (1987) (see section 2) to obtain welfare
weights. Rather than estimating distributional weights, however, one is concerned with decision
makers' implicit revealed willingness to pay (on society's behalf) for environmental goods. If, for
example, Parliament decides to choose the most expensive of two alternative projects because it has
more desirable environmental effects, and this is the only difference between the two projects, the
cost difference provides a minimum estimate of Parliament's implicit willingness to pay for the
corresponding change in environmental quality. Note that this method does not measure consumers'
preferences, so that values estimated by this method have quite another interpretation than value
estimates obtained by the methods mentioned above. Carlsen et al. (1993), who estimated implicit
environmental values in connection with the Norwegian Master plan for Water Resources, provide
one example of applied implicit valuation.
The method rests on an assumption that previous policy decisions are the result of optimising
behaviour on the government's part. Chase (1968) pointed out that there is an element of circular
reasoning involved in using this method as input into a cost-benefit analysis, and then use the
resulting net benefits estimates to guide policy makers about optimal policy decisions. This argument
may be less relevant if the cost-benefit analysis is intended as a device of guiding purely
administrative decision processes, not processes involving the government itself.
However, in a democracy, implicit values cannot be expected to be constant, neither over time, nor
between different decisions made at approximately the same time. Discrepancy between values
estimated at different occasions may well occur even if decision makers are rational, well-informed,
and do not change their preferences: When decisions are made through democratic procedures of
collective decision-making, there is of course no central planner with dictatorial power. Generally,
one cannot expect collective decisions made by a set of rational individuals to follow the same
patterns as decisions made by a single rational individual. For example, it is well-known that majority
voting can produce intransitive choices (see Feldman, 1980).
In politics, it may vary from case to case which interests are given priority by decision makers.
Consequently, implicit values will vary too. Requiring constancy of such values would imply that
those groups or interests who succeeded in gaining political priority last year should succeed in doing
so this year, and next year as well; while those who failed last year, should continue to fail. Probably,
an important feature of a stable democratic system is precisely that there usually comes another
chance for losers; it is not the same interests that win every time (Buchanan, 1954, Miller, 1983).
3.3.3 Expert panels
An obstacle for reasonable estimation of people's preferences for environmental goods is that the
population may be poorly informed about the nature of the good in question. For instance, a project
may have effects on a rare biotope, in which case the willingness to pay is perhaps not really
connected with the area as such, but with preserving biodiversity. Most people, however, have little
knowledge of the probability that different species will survive under various circumstances, the
prevalence of species which are very similar to the ones which are threatened, the probability that
those species will become commercially useful in the future, and so forth. In such cases, willingness
8
to pay is not just based on individual preferences, but also subjective probability judgements and
judgements concerning highly complicated cause/effect-relationships. These judgements may
correspond poorly to the opinion of better informed experts.
Some researchers have proposed using so-called focus groups to investigate the public's under-
standing and knowledge of the issue prior to performing a contingent valuation study (Hutchinson et
al., 1995). In a focus group, people discuss the issue in an informal manner, while the analyst ob-
serves the discussion. If there are common misunderstandings, ambiguities or lack of knowledge of a
subject, this may become apparent during the discussion, and the analyst can make sure to provide
information about these things to the respondents in the ensuing contingent valuation survey.
However, providing such information in a satisfactorily way may sometimes be exceedingly difficult
or time-consuming. m It has therefore been proposed to study the preferences or views of a group of
experts, rather than estimating the population's preferences. For a discussion of this method, see
Keeney and Raiffa (1976). For such expert values to be interpreted as representing the population's
values (if they were better informed), one must assume that the chosen experts' preferences do not
differ from the preferences of the rest of the population, at least not in a systematic way. However, it
seems quite plausible that the process of recruitment to a profession involves some selection of people
according to interests and preferences; also, working with a particular issue for a long period of time,
as experts do, could presumably change their judgements of the importance of their own field
compared to other issues.
The use of expert panels in environmental valuation stem from the view that consumers' judgements
are not made on a well-informed basis, and thus cannot be trusted. This is to some extent inconsistent
with the theoretical foundations of cost-benefit analysis, where consumer sovereignty is a central
concept.
A general argument for using market prices or monetary evaluations in project evaluations is that
prices provide highly aggregated information. If expert panels are used, the evaluation of the good in
question is made by a few experts, in which case it might be fairly easy to sum up the judgements
anyhow. The usefulness of monetary values simply as a means of aggregating information then
becomes less relevant. Moreover, the process of transforming professional scientific judgements into
monetary estimates is probably difficult, particularly since most natural scientists are not educated
economists as well. One might thus question whether translating experts' views into monetary values
is really worth the exercise, compared to the alternative of simply presenting expert views as part of
an impact assessment (without measuring them in monetary units).
4. Controversies: The debate following Exxon Valdez
Economists have been discussing valuation of environmental goods for a long time, but in recent
years, the debate has intensified. The single event contributing most to this was probably the
Exxon
Valdez accident in 1989, when the tanker Exxon Valdez ran aground in Prince William Sound in
Alaska, spilling vast amounts of oil and contaminating a huge area. According to CERCLA (see
section 2.3), Exxon was liable to pay damages for the resulting injuries to natural resources. Since the
physical damages were enormous, large economic interests were clearly at stake; and since there was
no consensus on how to measure the damages in monetary terms, both Exxon, the State of Alaska and
federal authorities spent a lot of resources trying to establish new arguments and knowledge about
valuation methods.
10
In addition to making respondents better informed, such discussions may change respondents' attitudes
towards the issue at hand; thus influencing the value estimates in a way which was not intended.
9
The debate which followed was particularly concerned with the contingent valuation method and its
ability to provide reliable estimates of existence value. Some of the controversies which emerged are
discussed below. Portney (1994), Hanemann (1994) and Diamond and Hausman (1994) provide
(together) a good and considerably more detailed overview of this debate.
4.1 What does contingent valuation really measure?
Many economists have expressed doubts as to whether responses in a contingent valuation study can
actually be interpreted as measuring preferences for the environmental good in question. Diamond
and Hausman (1994), for example, conclude that contingent valuation is a deeply flawed
methodology for measuring non-use values, one that does not estimate what its proponents claim to be
estimating. Among the arguments used for this is that CVM values seem to be very sensitive to
whether the valuation question is asked by itself or with other questions, and, if several questions are
asked, the sequence they are asked in. Another phenomenon is the finding in some studies that
willingness to pay is more or less constant even when the valued good is changed in quite significant
ways. The phenomenon that values lack responsiveness to changes in the valued good, but are
sensitive to the sequencing of questions, is frequently called embedding. Further, sceptics have used
as an argument the fact that willingness to accept (or the compensation demanded by consumers to
accept a change) is frequently several times higher than reported willingness to pay.
4.1.1 Warm glow>>
Kahneman and Knetsch (1992) have suggested that responses to CVM studies measure the purchase
of moral satisfaction rather than economic value in the usual sense. This is frequently called the
warm glow hypothesis, since according to this idea consumers are actually reporting their
willingness to pay for a private good: The warm glow one feels inside when giving money to a
worthy purpose.
This hypothesis has been supported by some empirical analyses, demonstrating that large changes in
the quantity or inclusiveness of the environmental good may be associated with very small (or no)
changes in willingness to pay. This is to be expected if people are not concerned about the particular
environmental good, but are rather looking for good causes to allocate their purchase of moral
satisfaction-contribution to. Kahneman and Knetsch (1992) themselves found that when each sample
was asked only one valuation question, willingness to pay varied surprisingly little with the
inclusiveness of the valued good: Mean willingness to pay for improved availability of equipment and
trained personnel for rescue operations did not differ significantly from mean willingness to pay for
improved preparedness for disasters more generally; which did not, in turn, differ significantly from
mean willingness to pay for improving all environmental services.
A much cited study is Desvouges et al. (1993), where respondents were asked about their willingness
to pay to avoid the death of a certain number of migratory waterfowl. Respondents who were asked
about their willingness to pay to avoid the deaths of, respectively, 2,000, 20,000 or 200,000 birds did
not give significantly different answers.
The warm glow hypothesis is not the only plausible explanation of this phenomenon, however. It is
possible that respondents have a tendency to misconceive the valuation question, interpreting it as
being more general than was intended by the researcher. For example, respondents to the study by
Desvouges et al. may have reported their willingness to pay to save animals' lives in general, not
specifically the birds affected by the project they were asked to value.
Other studies, however, have found that willingness to pay does vary with the scope or scale of the
good in question in the expected way. Smith (1995) provides one example. Hanemann (1994) gives
references to a number of such studies. He concludes that the results of Kahneman and Knetsch and
Desvouges et al. do not represent the majority finding in the CVM literature regarding the variation of
10
willingness to pay with scope. Further, he is critical to several methodological aspects of these two
studies.
Many studies have demonstrated that the number of valuation questions asked in one interview, and
the order the questions are asked in, are important factors to determine individuals' willingness to
pay for each of the goods. This observation seems to be much less controversial than the claim that
values lack sensitivity to scope. One example of values' sensitivity to the sequencing of questions is
the study of Kahneman and Knetsch (1992) referred to above: When people were asked about the
most inclusive category (all environmental services) first, and then asked what part of this
contribution should be allocated to the more specific purposes, mean willingness to pay for the latter
dropped quite dramatically, as compared to the case when each sample of respondents were asked
only one valuation question.
This phenomenon can partly be explained as an income effect: When you have paid for one
environmental good, you have less money left to pay for another (although these payments are, of
course, hypothetical). However, Diamond and Hausman (1994) claim that the income elasticities
which can be estimated from such data cannot reasonably be explained as a result from income
effects.
If two goods to be valued are close substitutes, the value of one of them may depend crucially on
whether the other is already available. Under such circumstances, one would expect value estimates to
be sensitive to the sequencing of questions, and aggregation of independently derived willingness to
pay estimates would overstate total willingness to pay (see Mitchell and Carson, 1989, p. 44-47).
Thus, the observation that values are indeed sensitive to the sequence of questions may indicate that
different environmental goods are regarded as close substitutes, and should hence not necessarily be
interpreted as evidence for the warm glow hypothesis. On the other hand, the warm glow
hypothesis provides one explanation why environmental goods may be close substitutes: If
respondents pay for the pleasure of giving to a good cause, rather than their own benefit from this
particular environmental change, one environmental purpose may be just as good as another to obtain
this goal.
Under the warm glow hypothesis, extensive use of contingent valuation as background information
for political decisions is hardly a good idea. If one conducts a CVM study each time a new project
proposal is to be evaluated, only asking about the valuation of one good at a time, respondents might
report all or most of their planned contribution to good causes every time, since the contributions
are hypothetical. Thus, the warm glow hypothesis is, as Arrow et al. (1993) note, votentially a
very damaging criticism of the method. Under this hypothesis, willingness to pay cannot be
interpreted as a welfare measure for the specific good to be valued; which brings up the question of
why the reported values ought to be aggregated in a cost-benefit analysis at all.
4.1.2 4,<Reasonableo payment
When asked to report their willingness to pay for an environmental good, it is also possible that
people report an amount of money which they perceive as a reasonable or fair payment, rather that
the maximum amount of money they would be willing to pay to achieve the good. When making such
judgements, they take into account such elements as the total costs of supplying the good, how much
each would have to pay if everybody paid their share, whether it is reasonable that they should pay
anything at all if the blame is really on someone else, and so on (Schkade and Payne, 1993). Such
value statements can be regarded as the results from a kind of intuitive cost-benefit analyses, rather
than an individual marginal benefit.
A somewhat similar critique is that of Kahneman et al. (1993), who claim that willingness to pay for
public goods measures attitudes rather than economic values. They maintain that if one primarily
wants to investigate people's attitudes, money is hardly the most natural unit to use; one could
11
instead, for instance, simply ask people more directly what they think.
The proponents of the contingent valuation method have responded to these critiques by pointing out
methodological weaknesses in their critics' analyses (Hanemann 1994, Smith 1992). Due to the many
possible biases and sources of error, CVM surveys are very expensive to conduct if one adheres to the
most strict standards. Thus, most studies have methodological weaknesses to some extent. Further
research is probably needed to draw more firm conclusions about the nature of CVM value estimates.
4.2 Willingness to pay or willingness to accept?
In CVM studies, respondents are asked about their willingness to pay (WTP) or willingness to accept
(WTA) for a given change in the quantity or quality of an environmental good (or bad). The WTP and
WTA format correspond to the two theoretical welfare measures of the Hicksian compensating
variation and equivalent variationn . The correspondence between the two question formats and the
two theoretical welfare measures depends on whether the environmental change to be valued is a good
or a bad.
WTP is the maximum amount of money the respondent is willing to pay to secure, or avoid, the
proposed change. If the change is an improvement, WTP to secure the change measures the
compensating variation. If the change is considered as a deterioration, however, WTP to avoid the
change is the equivalent variation.
WTA is the amount of money a person must at least have to be willing to accept a change, or to be
willing to forego it. If the change is an improvement, WTA measures the equivalent variation. If, on
the other hand, the change makes things worse from the respondent's point of view, WTA measures
the compensating variation.
Until recently, it was believed that equivalent and compensating variation measures ought to be fairly
close, due to Willig (1976) and Randall and Stoll (1980). However, a lot of CVM studies have
reported large discrepancies between WTA and WTP values (references to several such studies are
provided in Knetsch, 1994). This has been used as one argument that CVM studies do not measure
economic values in a reliable way. However, Hanemann (1991) demonstrated that when the analysis
is concerned with the welfare effects of quantity changes rather than price changes, the difference
between the compensating and equivalent variations may actually be infinitely large. This means that
it may matter a lot which of these measures one employs.
It is not always obvious whether the compensating and equivalent variation measures should be
chosen in a cost-benefit analysis. The equivalent variation of a single consumer will rank alternatives
in accordance with his utility function even when more than two alternatives are compared; the
compensating variation, on the other hand, may fail to do so (Johansson, 1993, p.30). However,
frequently one is really only interested in comparing only two alternatives, namely the situations with
and without the implementation of a certain project. Apart from such technical features of the
measures, they do however have different interpretations in terms of rights.
When asking someone of her willingness to pay to achieve something, there is an implicit assumption
that she does not really have a property right to it; she might have to pay to get it. On the other hand,
if one asks her of the compensation she would need to give it up, there is an implicit assumption that
she does have some kind of property right to this good. Now, CVM is frequently applied to public
goods, where property rights are usually not clearly defined, or at least not assigned to individuals.
11 See, for example, Johansson (1993). Some authors use the term surplus instead of variation when a
quantity change is concerned, while the term variation is used when welfare effects of a price change is
analysed.
12
This means that it is sometimes not obvious whether the respondent actually has a right to the good to
be valued. One possible approach is to take the status quo as a point of departure concerning rights;
regarding as the consumer's what she has got presently, not what she would have had if the project
was implemented. For a discussion of variation measures and rights, see Bromley (1995).
Whether one chooses to use the compensating or equivalent variation as the relevant individual
welfare measure, it is hard to find a theoretical justification for using different welfare measures when
the change is welfare-improving and when it is not. Nevertheless, this is exactly what is recommended
by several CVM theorists (and also by the NOAA Panel (Arrow et al., 1993), see section 4.4). The
reason is that when faced with WTA questions, a large number of respondents usually refuse to co-
operate, or they report values which are regarded by the researchers as implausibly large. This has
lead many researchers to prefer the willingness to pay-format, regardless of property right
considerations. Since WTP corresponds to the compensating variation in the case of an improvement,
but to the equivalent variation in the case of a deterioration, this amounts to changing the preferred
underlying theoretical welfare measure, depending on whether the change is welfare improving or not.
4.3 Non-use values and altruism
In empirical studies, non-use values frequently account for large parts of the estimated total values.
Some scholars have questioned whether non-use values should actually be included in cost-benefit
analyses at all. This question is closely related to the issue of altruism, since non-use values may to a
large extent be motivated by concern for others, either present or future people 12.
Milgrom (1993) maintain that willingness to pay which is motivated by altruistic concerns should not
be included in a cost-benefit analysis. This reasoning is discussed in more detail by Johansson (1992,
1994). The main argument is as follows. If person A is concerned about person
B' s utility, A will be
willing to pay something to secure B' s access to the public good. But A will also be concerned about
B
having to pay his share of the costs. Thus, counting A's altruistic concerns on the benefit side, but
not on the cost side of the calculations, gives rise to an inconsistency. However, if A is only concerned
about B's access to the public good, not his utility in general, the inconsistency vanishes.
Johansson (1994) concludes that the problem cannot be treated satisfactorily by simply ignoring
altruistically motivated willingness to pay. He points out, however, that if the question about
willingness to pay is posed in a particular way, the problem can be avoided. The idea is to ask
respondents to indicate their willingness to pay,
provided that everybody else pay just enough to be
exactly as well off as before.
In practice, however, such a scenario is likely to seem implausible to
respondents. It might also make valuation questions exceedingly difficult to understand for some
respondents.
4.4 The NOAA Panel
As a result of the large professional disagreement on this issue, and the huge economic interests
involved in the legal case following the Exxon Valdez accident, the American Department of
Commerce's National Oceanic & Atmospheric Administration (NOAA) convened an expert panel in
1992. The Panel was headed by the two Nobel laureates Kenneth Arrow and Robert Solow, and its
mandate was to evaluate the use of CVM in determining non-use values. The Panel's report was
released in January 1993 (Arrow et al., 1993), concluding that CV studies can produce estimates
reliable enough to be the starting point of a judicial process of damage assessment, including lost
passive-use values. Further, they put forward a set of guidelines for conducting CVM studies in a
methodologically satisfactory way, and argued that to be acceptable for the purpose referred to above,
these guidelines should be adhered to.
The NOAA Panel's guidelines have been much referred to, and now serves as a best practice
12
That is, at least if existence valueo is defined to include others' use of the environmental good.
13
standard by which the quality of CVM surveys can be compared, although several of the
recommendations have been criticised by other economists (see, for example, Bromley, 1995). One
must also take into account that the NOAA Panel was concerned with the situation in the USA,
meaning that the reasoning behind the recommendation may not be valid elsewhere. The guidelines
include the following recommendations, among several others:
One should generally choose a conservative design; that is, when alternative instruments of
eliciting willingness to pay are available, one should choose the alternative which is likely to yield
the lowest estimate.
One should ask for willingness to pay, not willingness to accept measures.
One should use a referendum format: The questions should be asked yes/no-questions (would
you be willing to pay x USD to secure the provision of...).
The environmental program or policy that is offered must be described accurately.
Both yes and no responses should be followed up by questions of why.
The questionnaire should include questions about features such as income, prior knowledge and
interest in the site, attitudes towards the environment and big business, and belief in the
scenarios.
The instrument must not be so complex that it poses tasks which many respondents are not able to
or willing to perform.
Respondents must be reminded of alternative expenditure possibilities.
The survey should be designed to minimise warm gloww-effects, as well as views such as it's big
business' fault, so they should pay, not me.
Note that the Panel started from the premise that passive use-loss - interim or permanent - is a
meaningful component of the total damage resulting from environmental accidents. The main
concern of the Panel was CVM as a means of estimating non-use values in the context of a judicial
process; hence, it did not, for instance, discuss whether non-use values (or passive-use values)
should be included in a cost-benefit analysis, or whether aggregating willingness to pay is generally a
desirable way to estimate changes in social welfare.
5. Distributional effects and interest group bias
Whether CVM or other methods are used to measure environmental effects in a cost-benefit analysis,
the disregard of distributional effects in such analyses is controversial. This is common to all cost-
benefit analyses, whether they are primarily concerned with environmental issues or not. Below, I will
discuss some distribution-related issues which I think are particularly relevant to cost-benefit analysis
of environmental change.
5.1 Consumers with different preferences
Public goods differ from most private goods in that the available amount of public goods must be
regarded as exogenously given for the individual consumer. In competitive markets with private
goods, individuals will generally consume different amounts of the private goods, while marginal
rates of substitution between different goods will be equal across individuals. The opposite is true for
pure public goods: The amount available is equal for everybody, while marginal rates of substitution
are generally different. Thus, consumers' marginal willingness to pay for public goods will differ
between individuals.
Brekke (1993) demonstrates that the accounting unit matters when willingness to pay for a public
good is aggregated across individuals. The problem arises because individuals have different
willingness to pay, which can happen even if they have exactly the same income, provided that they
have different tastes. Imagine, for example, two persons with the same income, where one (the
14
materialist) is relatively more concerned about money, while the other (the environmentalist) is more
concerned about the environment. If an environmental change is measured in monetary units (i.e.
willingness to pay in dollars), the environmentalist is favoured; he does not care much about money,
and can afford to give away much of it without losing too much utility. If the change is measured in
units of the environmental good, however (for example m2 of wilderness), the materialist is favoured;
he does not care much about wilderness, and so he can afford to give up much of it without getting a
lot worse off.
Aggregation using the two different units puts different emphasis on the interests of these two
persons. The two methods correspond to assuming, respectively, that everybody has the same
marginal utility of income (monetary units) or that everybody has the same marginal utility of the
environmental good (environmental units). It is difficult to see that one of these assumptions is
generally more plausible than the other. Brekke applied both methods to calculate aggregate
willingness to pay for a change in air in Norway, using data from Strand (1985). After converting both
aggregate numbers to monetary units, he found that aggregate WTP was 22 times (!) higher when
monetary units were employed in the aggregation process.
The implication of Brekke's finding is that the choice of measurement unit determines which interest
groups will be favoured by the analysis, a result which must be said to be quite troublesome. This is
really just an illustration of a more general problem. If lump-sum transfers are not feasible, cost-
benefit analysis requires a cardinal and interpersonally comparable utility concept. Since we do not
know much about marginal utilities, one usually adopts the assumption that everyone has the same
marginal utility of income, which simplifies the analysis considerably. But this introduces a kind of
arbitrariness into the analysis, since we do not know anything about the plausibility of this
assumption. Brekke's result demonstrates that this may have significant implications, since another,
equally plausible assumption yields benefit estimates of quite a different magnitude.
Many people seem to think that poor people have a larger marginal utility of income than rich people
do. If this is actually so, cost-benefit analyses will generally tend to favour rich people's interests. If
public goods have income elasticities above one, the effect described by Brekke will also contribute to
a such bias against the poor>>.
5.2 Sustainability issues and intergenerational equity
Another dimension of the income distribution issue is the question of intergenerational equity. Many
environmentalists seem to think that putting a monetary value on the environment, hence making it
visible in the economic analyses and political decision process, is a way to ensure that the interests of
future generations are properly taken care of.
However, this is not necessarily the case. The value that is attached to environmental goods will itself
depend on whether the economy is following a sustainable path or not. For example, if the current
generation regards environmental resources as theirs to spend freely, without any regards for future
generations, environmental resources will be regarded as less scarce than if they had to be shared with
future people. Hence, a lower value will be attached to them, and inclusion of these low values in
cost-benefit analysis may not be of much help for future generations.
More generally, using aggregate willingness to pay as the central indicator of social welfare pre-
supposes that the initial distribution of income (or wealth) is socially acceptable. This is true whether
one is concerned with the intra- or intergenerational distribution. Thus, Howarth and Norgaard (1992)
point out that incorporating environmental valuation per se in decision-making will not bring about
sustainability unless each generation is committed to transferring to the next sufficient natural re-
sources and capital assets to make development sustainable.
15
Environmental valuation and intergenerational equity have been discussed in a large number of
papers. Most of these concentrate on discussions of the discount rate (for some recent examples, see
Hanley, 1992, Azar and Sterner, 1995, Beckerman, 1993). Although discounting and intergenerational
equity are obviously important when it comes to the applicability of environmental cost-benefit
analysis for policy purposes, I will not go further into those issues here, since that would clearly
require a separate paper. The reader can be referred to Dasgupta and Heal (1979), Lind (1982) and
Broome (1992) for discussions of discounting.
6. Environmental values and philosophy
Most of the recent debate about valuation of environmental goods, and contingent valuation in parti-
cular, has rested on the premise that the sum of individuals' willingness to pay is a relevant and
acceptable measure of changes in social welfare. This premise has been the subject of much debate in
other branches of economics (for example, see Sen, 1985, Hammond, 1990, Pauwels, 1993), but in the
valuation literature, it seems to be accepted without further questions to a quite surprising extent. The
debate following the Exxon Valdez accident, for instance, was largely concerned with CVM' s ability
to measure non-use values in a sufficiently accurate way, while few (although there were some) posed
the question of whether willingness to pay should really be a major concern in the determination of
environmental policy.
Recently, however, several scholars has focused on ethical and philosophical aspects of environ-
mental valuation and/or its use in cost-benefit analysis (Vatn and Bromley, 1994, Sagoff, 1994,
Berrens and Polasky, 1995, among others). Concerning the political aspects of the issue, I believe that
much still remains to be done; but some of the philosophically oriented critique seems quite relevant
for political purposes as well.
6.1 Does valuation reduce the value?
One early, and harsh, critic of both cost-benefit analysis and environmental valuation is Kelman
(1981). Kelman claims that there are good reasons to oppose efforts to put dollar values on non-
marketed benefits and costs. According to Kelman, putting a price on a benefit may reduce the value
of that benefit: The disgust that accompanies the idea of buying and selling human beings is based
on the sense that this would dramatically diminish human worth. If your beloved estimated a
monetary value of his/her relationship to you, this valuation exercise would probably itself reduce the
value of the relationship.
This line of thought is developed more formally in Frey (1992). Frey assumes that behaviour is
motivated by two factors; intrinsic (environmental ethics) and extrinsic (external factors, such as
economic incentives). He demonstrates that under certain conditions, pricing of environmental goods
(or bads) may actually be counterproductive; the intrinsic motivation to behave in an environmentally
friendly way is crowded out. This may perhaps be explained by a hypothesis of the following kind:
Individuals are raised to adhere to certain moral norms regarding their attitudes to public goods. These
norms differ from the norms governing market behaviour, where looking after one's own interests is
allowed to a larger extent. When an environmental good is priced, people may begin to regard it more
in line with a market good, and thus place less strict moral restriction on their behaviour towards it.
6.2 Social welfare judgements or individual utility?
Kelman (1981) further points out that using willingness to pay-figures to provide guidance for public
decisions implies an assumption of no difference between how people value certain things in private
individual transactions and how they would wish those same things to be valued in public collective
decisions. He argues that people might have what he calls Aigher and <dower preferences, the
latter coming to fore in private decisions, while people want the former to come to the fore in public
16
decisions. For example, precisely because we fail to give live-saving the value in everyday personal
decisions that we in some general terms think we should, social decisions may be regarded as an
opportunity to correct this.
Sagoff (1988) similarly argues that when environmental policy issues are concerned, people behave as
citizens, which means that they consider the benefits to public interests when evaluating a proposal.
In contrast, when acting as consumers, individuals pursue their own, personal interests. Sagoff argues
against using aggregate willingness to pay as a central criterion for environmental policy decisions,
since a simple aggregation of people's judgements as citizens does not have room for the important
process of public discussion and deliberation: Political decisions ought to be guided through search
for the best argument, not the highest willingness to pay.
Kelman's idea of Aigher and lower preferences does perhaps bear some resemblance to Amartya
Sen's notion of an agency aspect of individual utility (Sen, 1985). Sen points out that the action taken
by an individual is not necessarily motivated only by his self-interest. This is not really the same as
altruism (see section 4.3): Altruism is usually modelled such that other people's utilities enter the
utility function of the altruistic individual. Then, other people's utility contributes to one's own, so
that a person acting out of altruistic motivations is in one sense also acting out of self-interest. Sen
does not, however, accept the idea that any individual choice can ultimately be explained by self-
interest. There are cases where individuals, out of duty, go to personal sacrifices to fulfil their
perceived duty towards someone else to a degree which is difficult to explained as self-interest; the
extreme example being the wartime soldier who is willing to sacrifice his life. Sen (1977) maintains
that the traditional description of individuals as utility-maximisers is too simple; he argues that every
individual could, alternatively, be seen as having several preference orderings, corresponding to
different motives.
When people report their willingness to pay for an environmental good, the motivation behind the
answers is usually not known to the researcher. Perhaps willingness to pay-statements are mixtures of
purely selfish interests, altruistic concerns, and agency behaviour. If we do not know how to interpret
the responses, however, simply summing them up and interpreting the result as an indicator of social
welfare may yield a social welfare concept of which no-one understands the ethical implications
properly.
6.3 Intrinsic values
Many authors have been concerned about the question of whether or not the environment has some
kind of intrinsic value; that is, a value distinct from the individual utility human beings can derive
from it (Nelson, 1995, Norton, 1986, 1995).
A claim that cost-benefit analysis is not capable of taking nature's own interestso into account at all
would not be true: When willingness to pay for environmental goods is included in the analysis, such
considerations are taken into account to the extent that individuals are prepared to pay for it. After all,
willingness to pay for non-use values is what much of the debate about CVM has been about.
However, this does not necessarily imply that aggregating individual willingness to pay accounts for
nature's interests in a satisfactory way. The manner in which such values are included in cost-
benefit analysis depends, among other things, on whether individuals respond to WTP questions as
consumers or citizens in the sense of Sagoff (1988). It is quite possible that the procedure of
aggregating individual willingness to pay, regardless of the underlying motives, implies taking
nature's intrinsic value into account in a way which no respondent would regard as appropriate.
Common et al. (1993) comment as follows: We find it difficult to see how value can exist
independently of a valuer, and it seems more reasonable to regard claims for such intrinsic value as
claims that altruism be extended to embrace non-humans. However, the fact that it is hard to define
17
value without a valuer does certainly not imply that individual well-being is the only relevant concern
for environmental policy making. Considering Sen's argument that people may be acting out of duty,
it seems quite possible that people may be acting out of duty to a cause rather than another person.
Taking care of the environment can be a such cause. This means, firstly, that that willingness to pay
may not necessarily reflect individual well-being. Secondly, however, it is quite possible that
individuals want environmental policy to take such moral duties into account.
For example, some people's religion may say that it is wrong to treat nature as if it belonged to the
human race, because this is to put humans in the place of God (or the gods). The view that nature has
an intrinsic value seems quite consistent with a pantheistic religious view, for example, and could
probably be justified by several other religious views as wel1. 13 It seems to me that such views do not
need any other justifications than belief.
Goodin (1994) suggests a view of the value of nature which does not involve a value without a
valuer, but which nevertheless seems to imply that the value of the environment can hardly be
expressed in monetary units. He argues as follows: The value of natural processes is to provide a
context, outside of ourselves (individually, or even collectively), in which to set our lives (...) What is
wrong with environmental despoliation is that it deprives us of that context; it makes the external
world more and more one of own (and perverse) creation. That is ultimately a wrong to humans,
rather than to nature as such, to be sure. It is, nonetheless, a wrong that cannot be recompensated by
cash payments.
7. Concluding remarks
Some economists have argued that environmental valuation, combined with cost-benefit calculations,
can indicate the socially optimal level of environmental goods provision. According to this view,
analyses of the value of specific environmental changes ought to have a larger impact on policy
decisions than is presently the case.
However, what is best for society is a controversial issue. Individuals with different political and
ethical views would presumably not agree on this, even if they managed to agree on a description of
the situation. The sources of such disagreement may be distributional issues, differing views on the
importance of non-utility information such as rights and duties, and different judgements on which
factors are important for a good life.
To identify the preferred policy alternative from a social point of view, then, one must make
controversial choices regarding which interests and values are to count the most. This means that the
preferred alternative will not be the same for different policy makers. Cost-benefit analysis is one way
of weighing various kinds of concerns against each other, but not the only one. Hence, environmental
valuation cannot by itself provide any objective answer to the question of how much we should
protect the environment.
Well-founded environmental policy decisions requires that accurate and understandable information
about the issue is available for policy makers. However, it is not obvious that monetary valuation is
always the most suitable format for such information. One alternative is to present policy makers with
environmental indicators measured in physical units, verbal descriptions, or pictures. Contingent
valuation practitioners have put much emphasis on developing methods for providing survey
respondents with accurate and understandable information about the goods they are asked to value.
Such methods could, of course, be used to provide information to policy makers as well.
13 Discussions of various religions' views on nature can be found in Bratton (1993), Hallman (1994), and Tucker
and Grim (1994).
18
When choosing how to provide information, it is important to take into account how policy makers
are likely to use this information. If one believes that environmental issues will not be taken into
consideration by policy makers at all without including them in a cost-benefit analysis, this is clearly
an argument for valuing the environment in monetary terms. On the other hand, it is possible that
inclusion of too many considerations makes the cost-benefit analysis so complicated, and so
controversial, that policy makers are not liable to use it anyway. Hence, it does not seem possible to
draw unambiguous conclusions about the desirability of environmental valuation for policy purposes.
References
Aaheim, A., and K. Nyborg (1995): On the Interpretation and Applicability of a 'Green National
Product', The Review of Income and Wealth 41, 57-71.
Arrow, K., R. Solow, E. Leamer, P. Portney, R. Radner, and H. Schuman (1993): Report of the
NOAA Panel on Contingent Valuation. National Oceanic and Atmospheric Administration, Federal
Register/Vol. 58, no.10, 4602-4614.
Auerbach, A. J. (1985): The Theory of Excess Burden and Optimal Taxation, in A. J. Auerbach and
M. Feldstein (eds.): Handbook of Public Economics, vol. I, Amsterdam: Elsevier Science Publishers
B.V. (North-Holland), 61-128.
Azar, C., and T. Sterner (1995): Discounting and Distributional Considerations in the Context of
Global Warming. Paper presented at the Sixth Annual Conference of the European Association of
Environmental and Resource Economists, Umea, Sweden, June 1995.
Beckerman, W. (1993): Environmental Policy and the Discount Rate: Philosophy or Economics?
CSERGE Working Paper 93-12.
Blackorby, C. (1990): Economic Policy in a Second-Best Environment,
Canadian Journal of
Economics XXIII, No. 4, 748-771.
Blackorby, C., and D. Donaldson (1990): A Review Article: The Case Against the Use of the Sum of
Compensating Variations in Cost-Benefit Analysis,
Canadian Journal of Economics XXIII, No. 3,
471-494.
Bockstael, N. E., K. E. McConnel, I.E. Strand (1991): Recreation, in Braden, J. B., and C.D.
Kolstad (eds.) (1991):
Measuring the Demand for Environmental Quality. Amsterdam: Elsevier
Science Publishers B.V. (North -Holland), 227-270.
Braden, J. B., and C.D. Kolstad (eds.) (1991):
Measuring the Demand for Environmental Quality.
Amsterdam: Elsevier Science Publishers B.V. (North-Holland).
Bratton, S. P. (1993):
Christianity, Wilderness, and Wildlife: The Original Desert Solitaire,
Scranton:
University of Scranton Press.
Brake, K. A. (1993): Do Cost-Benefit Analysis Favor Environmentalists? Discussion Paper 84, Oslo:
Statistics Norway.
Brendemoen, A., and H. Vennemo (1994): A Climate Treaty and the Norwegian Economy: A CGE
Assessment, The Energy Journal 15, 77- 93.
Bromley. D. W. (1990): The Ideology of Efficiency: Searching for a Theory of Policy Analysis,
Journal of Environmental Economics and Management
19, 86- 107.
19
Bromley, D. W. (1995): Property Rights and Natural Resource Damage Assessment, Ecological
Economics 14, 129-135.
Broome, J. (1992): Counting the Cost of Global Warming, Cambridge, UK: White Horse Press.
Buchanan, J. M. (1954): Social Choice, Democracy and Free Markets, Journal of Political Economy,
114-123.
Carlsen, A. J., J. Strand and F. WenstOp (1993): Implicit Environmental Costs in Hydroelectric
Development: An Analysis of the Norwegian Master Plan for Water Resources, Journal of
Environmental Economics and Management 25, 201-211.
Chase, S. B. (1968): Problems in Public Expenditure Analysis, Washington D.C.: Brookings
Institution.
Common, M. S., R. K. Blarney and T. W. Norton (1993): Sustainability and Environmental Valuation,
Environmental Values 2, 299-334.
Cummings, R. G., and G. W. Harrison (1995): The Measurement and Decomposition of Non-use
Values: A Critical Review, Environmental and Resource Economics 5, 225-247.
Dasgupta, P. S., and G. M. Heal (1979): Economic Theory and Exhaustible Resources, Cambridge:
Cambridge University Press.
Desvouges, W. H., F. R. Johnson, R. W. Dunford, S. P. Hudson, K. N. Wilson (1993): Measuring
Natural Resource Damages With Contingent Valuation: Tests of Validity and Reliability, in
Hausman, J. A. (ed.): Contingent Valuation. A Critical Assessment, Elsevier Science Publishers B.V.,
91-164.
Diamond, P. A. and J. A. Hausman (1994): Contingent Valuation: Is Some Number Better than No
Number? The Journal of Economic Perspectives 8, 45-64.
Dreze, J., and N. Stem (1987): The Theory of Cost-Benefit Analysis, in A. J. Auerbach and M.
Feldstein (eds.): Handbook of Public Economics, vol. II, Amsterdam: Elsevier Science Publishers
B.V. (North-Holland), 909-990.
Feldman, A. M. (1980): Welfare Economics and Social Choice Theory, Boston: Kluwer-Nijhoff
Publishing.
Frank, R. H. (1992): Melding Sociology and Economics: James Coleman's Foundations of Social
Theory, Journal of Economic Literature XXX, 147-170.
Freeman, A. M. (1993): The Measurement of Environmental and Resource Values. Theory and
Methods, Washington D.C.: Resources for the Future.
Frey, B. S. (1992): Pricing and Regulating Affect Environmental Ethics, Environmental and Resource
Economics 2, 399-414.
Hallman, D. C. (ed.) (1994): Ecotheology: Voices From South and North, Maryknoll: Orbis Books.
20
Goodin, R. (1994): Selling Environmental Indulgences, K Y K LOS 47, 573-596.
Hammond, P. J. (1979): Straightforward Individual Incentive Compatibility in Large Economies,
Review of Economic Studies 46, 263-282.
Hammond, P. (1990): Theoretical Progress in Public Economics: A Provocative Assessment, Oxford
Economic Papers 42, 6-33.
Hanemann, M. W. (1991): Willingness To Pay and Willingness To Accept: How Much Can They
Differ? The American Economic Review 81, 635-647.
Hanemann, M. W. (1992): Preface in S. Navrud (ed.): Pricing the European Environment. Oslo:
Scandinavian University Press, 9-35.
Hanemann, M. W. (1994): Valuing the Environment through Contingent Valuation, The Journal of
Economic Perspectives 8, 19-44.
Hanley, N. (1992): Are There Environmental Limits to Cost Benefit Analysis? Environmental and
Resource Economics 2, 33-59.
Hanley, N., and C. L. Spash (1993): Cost-Benefit Analysis and the Environment, Edward Elgar.
Harsanyi, J. C. (1955): Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of
Utility, Journal of Political Economy 63, 309-21.
Hicks, J. (1939): The Foundations of Welfare Economics, The Economic Journal 49, 696-712.
Hoevenagel, R., and 0. Kuik (1992): The Netherlands, in S. Navrud (ed.): Pricing the European
Environment, Oslo: Scandinavian University Press, 100-107.
Howarth, R. B., and R. B. Nordgaard (1993): Intergenerational Transfers and the Social Discount
Rate, Environmental and Resource Economics 3, 337-358.
Hutchinson, W. G., S. M. Chilton, J. Davis (1995): Measuring Non-use Value of Environmental
Goods Using the Contingent Valuation Method: Problems of Information and Cognition and the
Application of Cognitive Questionnaire Design Methods, Journal of Agricultural Economics 46, 97-
112.
Johansson, Per-Olov (1987): The Economic Theory and Measurement of Environmental Benefits,
Cambridge, UK: Cambridge University Press.
Johansson, Per-Olov (1992): Altruism in Cost-Benefit Analysis, Environmental and Resource
Economics 2, 605-613.
Johansson, Per-Olov (1993): Cost-Benefit Analysis of Environmental Change. Cambridge: Cambridge
University Press.
Johansson, Per-Olov (1994): Altruism and the Value of Statistical Life: Empirical Implications.
Journal of Health Economics 13, 111-118.
Johansson, P.-O., and B. KristrOm (1992): Sweden, in S. Navrud (ed.): Pricing the European
Environment, Oslo: Scandinavian University Press, 136-149.
21
Kahneman, D. and J. Knetsch (1992): Valuing Public Goods: The Purchase of Moral Satisfaction,
Journal of Environmental Economics and Management 22, 57-70.
Kahneman, D., I. Ritov, K. E. Jacowitz and P. Grant (1993): Stated Willingness to Pay for Public
Goods: A Psychological Perspective, Psychological Science.
Kahneman, D., and A. Tversky (1979): Prospect Theory: An Analysis of Decision Under Risk,
Econometrica 47, 263-291.
Keeney, R. L., and H. F. Raiffa (1976): Decisions with Multiple Objectives: Preferences and Value
Tradeoffs, New York: Wiley.
Kelman, Steven (1981): Cost-Benefit Analysis. An Ethical Critique, Regulation, January/February.
Knetsch, J. L. (1994): Environmental Valuation: Some Problems of Wrong Questions and Misleading
Answers, Environmental Values 3, 351-368.
Kuik, 0., S. Navrud, and D. W. Pearce (1992): Benefit Estimation and Environmental Decision
Making, in S. Navrud (ed.): Pricing the European Environment. Oslo: Scandinavian University
Press, 274-287.
Larson, D. M. (1991): On Measuring Existence Value, Mimeo, Dept. of Agricultural Economics,
University of California, Davis.
Lind, R. C. (1982): Discounting for Time and Risk in Energy Policy, Washington D. C.: Resources for
the Future.
Louviere, J. J. (1988): Analyzing Decision Making; Metric Conjoint Analysis, Beverly Hills: Sage.
Milgrom, P. (1993): Is Sympathy an Economic Value? Philosophy, Economics, and the Contingent
Valuation Method, in J. A. Hausman (ed.): Contingent Valuation: A Critical Assessment. New York:
North-Holland, 417-435.
Miller, N. R. (1983): Pluralism and Social Choice, The American Political Science Review 77, 734-
747.
Mitchell, R. C., and R. T. Carson (1989): Using Surveys to Value Public Goods. Washington:
Resources for the Future.
Navrud, S., and J. Strand (1992): Norway, in Navrud, S. (ed.): Pricing the European Environment.
Oslo: Scandinavian University Press, 106-135.
Navrud, S. (ed.) (1992): Pricing the European Environment. Oslo: Scandinavian University Press.
Navrud, S. (1994): Economic Valuation of External Costs of Fuel Cycles. Testing the Benefit
Transfer Approach, in A. T. de Almeida et al. (eds.): Integrated Electricity Planning, Kluwer
Academic Publishers, 49-66.
Nelson, J. L. (1995): Health and Disease as 'Thick' Concepts in Ecosystemic Contexts,
Environmental Values 4.
22
Norton, B. (1986): On the Inherent Danger of Undervaluing Species, in B. Norton (ed.): The
Preservation of Species, Princeton: Princeton University Press.
Norton, B. (1995): Objectivity, Intrinsicality and Sustainability: Comment on Nelson's 'Health and
Disease as Thick Concepts in Ecosystemic Contexts', Environmental Values 4, 323-332.
Palmquist, R. B. (1991): Hedonic Methods, in Braden, J. B., and C.D. Kolstad (eds.): Measuring
the Demand for Environmental Quality, Amsterdam: Elsevier Science Publishers B.V. (North-
Holland).
Pauwels, W. (1993): The Implicit Welfare Weights Used When Maximising Aggregate Surplus,
Journal of Economics 57, 261- 277.
Portney, P. R. (1990): Public Policies for Environmental Protection, Washington D.C.: Resources for
the Future.
Portney, P. R. (1994): The Contingent Valuation Debate: Why Economists Should Care,
The Journal
of Economic Perspectives 8, 3- 18.
Quiggin, J. (1995): The Suboptimality of Efficiency, Economic Letters 47, 389-392.
Randall, A., and J. R. Stoll (1980): Consumer's Surplus in Commodity Space,
The American
Economic Review 71, 449-457.
Rubinstein, A. (1988): Similarity and Decision-Making Under Risk (Is There a Utility Theory
Resolution to the Allais Paradox?), Journal of Economic Theory 46, 145-153.
Sagoff, M. (1988): The Economy of the Earth, Cambridge: Cambridge University Press.
Schkade, D. A., and J. W. Payne (1993): Where Do the Numbers Come From?, in J. A. Hausman
(ed.): Contingent Valuation: A Critical Assessment. New York: North-Holland, 271- 303.
Schulze, W. D., D. S. Brookshire, E. G. Walther, K. K. MacFarland, M. A. Thayer, R. L. Whitworth,
S. Ben-David, W. Malm and J. Molenar (1983): The Economic Benefits of Preserving Visibility in the
National Parklands of the Southwest, Natural Resources Journal 23, 149-173.
Sen, A. K. (1977): Rational Fools: A Critique of the Behavioural Foundations of Economic Theory,
Philosophy and Public Affairs 6, 317- 344.
Sen, A. K. (1985): Well -being, Agency and Freedom. The Dewey Lectures 1984,
The Journal of
Philosophy LXXXII, No. 4.
Sen, A. K., and B. Williams (eds.) (1982):
Utilitarianism and Beyond, Cambridge: Cambridge
University Press.
Smith, V. K. (1992): Comment: Arbitrary Values, Good Causes, and Premature Verdicts,
Journal of
Environmental Economics and Management 22, 71- 89.
Smith, V. K. (1995): Can Contingent Valuation Distinguish Significant from Trivial Public Goods?
Discussion Paper 95-24, Washington: Resources for the Future.
23
Stevens, T. H., J. Echeverria, R. J. Glass, T. Hager, and T. A. More (1991): Measuring the Existence
Value of Wildlife: What Do CVM Estimates Really Show? Land Economics 67, 390-400.
Strand, J. (1985): Verdsetting av reduserte luftforurensninger fra biler i Norge (Valuation of Reduced
Air Pollution from Cars In Norway), Memorandum 1, Department of Economics, University of Oslo.
Tucker, M. E., and J. A. Grim (eds.) (1994): Worldviews and Ecology: Religion, Philosophy and the
Environment, Maryknoll: Orbis Books.
Turner, R. K., I. Bateman, and D. W. Pearce (1992): United Kingdom>, in S. Navrud (ed.): Pricing
the European Environment, Oslo: Scandinavian University Press, 150-176.
United Nations (1993): Integrated Environmental and Economic Accounting. Interim Version. Hand-
book of National Accounting, Series F, No.61. New York: UN Department of Economic and Social
Development, Statistical Division.
Willig, R. (1976): Consumer's Surplus Without Apology, The American Economic Review 66, 589-
597.
24
Issued in the series Documents
94/1 H. Vennemo (1994): Welfare and the Environ-
ment. Implications of a Recent Tax Reform in
Norway
94/2 K.H. Alfsen (1994): Natural Resource Accounting
and Analysis in Norway
94/3 0. Bjerkholt (1994): Ragnar Frisch 1895-1995
95/1 A.R. Swensen (1995): Simple Examples on
Smoothing Macroeconomic Time Series
95/2 E.Gjelsvik, T. Johnsen, H.T. Mysen and
A. Valdimarsson (1995): Energy Demand in
Iceland
95/3 C. Zhao, 0. Bjerkholt, T. Halvorsen and
Y. Zhu (1995): The Flow of Funds Accounts in
China
95/4 Nordic Indicator Group (1995): Nordic
Environmental Indicators. Draft document.
English version with main points from comments
received
95/5 H.A. Gravningsmyhr (1995): Analysing Effects of
Removing Survivors' Pensions, Using the
Microsimulation Model LOTTE
95/6 P. Boug (1995): User's Guide. The SEEM-model
Version 2.0
95/7 E. Bowitz, N.O. Mwhle, V.S. Sasmitawidjaja and
S.B. Widoyono (1995): MEMLI An
Environmental model for Indonesia. Technical
Documentation of data programs and procedures
95/8 K. H. Alfsen, T. Bye, S. GlomsrOd and H. Wiig
(1995): Integrated Assessment of Soil Degradation
and Economic Growth in Ghana
95/9 0. Bjerkholt (1995): Ragnar Frisch and the
Foundation of the Econometric Society and
Econometrica
95/10 P.J. Bjerve (1995): The Influence of Ragnar
Frisch on Macroeconomic Planning and Policy in
Norway
96/1 D. Kolsrud (1996): Documentation of Computer
Programs that Extend the SEEM Model and
Provide a Link to the RAINS Model
96/2 E. Bowitz, V.S. Sasmitawidjaja and G. Sugiarto
(1996): The Indonesian Economy and Emission of
CO2: An Analysis Based on the Environmental-
Macroeconomic-Model MEMLI, 1990-2020
96/3 H.V. Sm130 (1996): Use of Geographical
Information Systems (GIS) in Statistics Norway
96/4 E. Berg (1996): Some Results from the Literature
on the Impact of Carbon Taxes on the Petroleum
Wealth
96/5 E.J. Flottum (1996): Norwegian National
Accounts Documentation of the Compilation
and Methods Applied. I General Description. II
GDP From the Output Approach
96/6 E.J. Flottum (1996): Norwegian National
Accounts Documentation of the Compilation
and Methods Applied. III GDP by Expenditure
Approach. IV GDP by Income Approach
96/7 0. Bjerkholt, K.A. Brekke and R. Choudhury
(1996): The Century Model - on the Long Term
Sustainability of the Saudi Arabian Economy
96/8 R. Choudhury (1996): The Century Model.
Technical Documentation of Computer Programs
and Procedures
96/9 R. Choudhury (1996): The Implementation
Model. Technical Documentation of Computer
Programs and Procedures
96/10 R. Choudhury (1996): The Selection Model.
Technical Documentation of Computer Programs
and Procedures
96/11 R. Choudhury (1996): The 0M95 An Oil Model
for the Kingdom of Saudi Arabia. Technical
Documentation of Computer Programs and
Procedures
96/12 K. Nyborg (1996): Environmental Valuation,
Cost-Benefit Analysis and Policy Making: A
Survey
25
Documents
Statistics Norway
Research Department
P.O.B. 8131 Dep.
N-0033 Oslo
Tel.: + 47-22 86 45 00
Fax: + 47-22 11 12 38
ISSN 0805-9411
O W Statistics Norway