Vous êtes sur la page 1sur 14

CONTINGENT VALUATION

SESSION 28

Objective: To describe and apply the Contingent Valuation Method to directly measure
WTP/WTA for valuing environment impacts.

Summary: The Contingent Valuation (CV) approach elicits directly, through surveys,
responses to valuation questions about environmental impacts. The method
may be used to measure both use and non-use values. This session discusses
issues that must be resolved in applying the techniquethe design of the CV
scenario, the technique that will be used to elicit willingness to pay (WTP),
methods for analyzing responses. Methods for validating CV responses are
also discussed.

TABLE OF CONTENTS

What is to be Measured? ...............................................................................2


Sample Selection............................................................................................3
Description of the Contingent Market. ............................................................3
Choice of Technique for Eliciting Willingness to Pay........................................5
Analysis of Contingent Valuation Responses ...................................................6
The Validity of the Contingent Valuation Method ............................................8
The Use of Contingent Valuation Studies ........................................................9
References...................................................................................................12

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washington, D.C.
Session 28 Contingent Valuation

28.1 Contingent valuation is a method of estimating the value that a person places on a good,
usually one that is not sold in markets, such as environmental quality or good health. The
approach asks people directly what they are willing to pay for the good, or what they are willing
to accept to give it up, rather than inferring this from observed behavior. The commodities
most often valued using this technique include public goods such as improvements in air and
water quality, and private non-market goods such as reductions in risk of death, days of illness
avoided or days spent hunting or fishing.

28.2 While other methods are available for estimating the values people place on some of
these goods, such as the travel cost method and hedonic pricing, contingent valuation is the only
method that can recover existence or non-use values--benefits from knowing that a good exists,
even though one may not experience it directly. Contingent valuation is also of use when the
policy option to be analyzed is outside of the range of available data. It is, for example, difficult
to use observed behavior to value proposed improvements in water quality at a lake if the lake
has always been polluted.

What is to be Measured?

28.3 The goal of a contingent valuation survey is to measure either the compensating or
equivalent variation for the good in question. If the individual must purchase the good, the
appropriate measure is compensating variation--the most the person will pay and keep his utility
constant. If an individual owns a good that may be taken away from him (e.g., if he may suffer
environmental damage) the appropriate measure is equivalent variation--the minimum
compensation the individual requires to keep his utility at its original level when he loses the
good. Compensating variation for an increase in a commodity from q0 to q1 or equivalent
variation for a decrease in the commodity from q1 to q0 can be defined using the individual's
expenditure function. Letting E denote the expenditure function, P the vector of prices for
market goods, q the quantity of the non-market good consumed, Q a vector of other non-
market goods and Ui the individuals utility when he consumes qi, compensating variation
(WTP) and equivalent variation (WTA) are given by:

(1) WTP = E(P,q0,Q,U0) - E(P,q1,Q,U0)

(2) WTA = E(P,q1,Q,U1) - E(P,q0,Q,U1).

28.4 In theory, the two measures may yield very different values for the same commodity
change, hence it is important to determine which valuation concept is the appropriate one. The
amount by which WTA exceeds WTP varies directly with the income elasticity of demand for q
and inversely with the elasticity of substitution between q and other goods (Hanemann 1991).
If the income elasticity of demand for q is zero, or if q is a perfect substitute for a private
good, WTP should equal WTA. If, however, the elasticity of substitution between q and
marketed goods is zero, the difference between WTA and WTP can be infinite.

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 2
Session 28 Contingent Valuation

28.5 Although WTA is the appropriate measure of value when a good that someone owns is
damaged, it is often difficult to measure WTA accurately in contingent valuation surveys.
Evidence of this is provided by the fact that willingness to accept compensation for quasi-private
goods (hunting licenses) in contingent valuation surveys has been found to exceed actual
willingness to accept compensation for the same goods (Bishop and Heberlein 1979; Bishop,
Heberlein and Kealy 1983). It is also the case that willingness to pay for a good is usually
many times lower than willingness to accept compensation to forego the same good (Bishop and
Heberlein 1979; Knetsch and Sinden 1984). When the good is not unique, this is interpreted as
evidence that WTA is more difficult to measure than WTP. For these reasons, researchers
have focused almost exclusively on WTP as the measure of value in contingent valuation
surveys.

Sample Selection

28.6 After determining the population whose WTP is to be measured, which depends on the
nature and location of the commodity to be valued, the researcher ideally uses probability
sampling procedures to select a sample who will receive the contingent valuation questionnaire.
This involves constructing a sampling frame and selecting from the sampling frame in such a way
that the probability of each person being selected is known. Cost considerations have
sometimes led to the use of convenience samples--interviewees selected at shopping malls or
through newspaper advertisements. Such samples are usually adequate when the goal is testing
how certain treatments (e.g., changes in scenario) affect responses, but do not allow WTP
values to be generalized to the general population.

28.7 The researcher must also determine the method of administering the questionnaire.
WTP may be elicited through in-person interviews, or in mail or telephone surveys. The choice
of administration method depends in part on the nature of survey. Visual aids can be used only
in in-person and mail surveys. Some elicitation techniques (e.g., questions whose wording
depends on the response to previous questions) are difficult to administer in mail surveys.
Which method is chosen will affect both the cost of conducting the survey and the response
rate. Response rates are generally lower for mail surveys and may lead to non-response or
sample selection bias. This occurs when people who refuse to answer the survey (e.g., due to
lack of interest in the topic) have WTP values that are systematically different from (e.g., lower
than) those of respondents.

Description of the Contingent Market

28.8 To elicit WTP in a questionnaire requires a description of the commodity to be valued


(e.g., a reduction in particulate matter of 10 micrograms per cubic centimeter), the mechanism
by which the good will be provided (controls on a nearby power plant) and the method of
payment (an increase in the respondent's utility bill).

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 3
Session 28 Contingent Valuation

28.9 It is essential that the commodity to be valued be described in a scenario that is


meaningful to the respondent. Goods that are unfamiliar to respondents--programs to clean up
ground water or to preserve a rare insect species--are inherently more difficult to describe than
more familiar goods--catching two striped bass on a fishing trip. Regardless of the familiarity of
the commodity, it is essential that the respondent understand the commodity as the researcher
intends it. If the researcher wishes the respondent to value an improvement in water quality in a
single lake, it is important that respondent do not value improvements in water quality in all lakes
in a region, or clean water in general. It is also important that the scenario be plausible. If
respondents feel that nuclear power is inherently unsafe, they will not believe a scenario that
asks them what they would pay to reduce health risks from a nuclear power plant to zero.

28.10 One important dimension of the scenario design is describing the availability of
substitutes for the commodity valued. The value of a malaria vaccine, for instance, should
depend on the existence of other methods for preventing the disease or mitigating its effects.
Such methods should be clearly spelled out to respondents in the course of the survey;
otherwise people may make different assumptions about the commodity being valued.

28.11 To determine if the scenario has been communicated successfully usually requires that
the respondent be debriefed at the end of the questionnaire. Failure to communicate a realistic
scenario to the respondent is likely to result in the respondent not taking the scenario seriously
or replacing it with something of his own invention, either of which may bias WTP.

28.12 How the good will be provided and how the respondent will pay for it must also be
described. In the case of a private good, when the good will be purchased and on what terms
must be clearly specified. For a public good it is common to frame the survey as a referendum.
That is, the respondent is asked whether he would vote in favor of a proposition on the ballot
that, if approved by the majority of voters, would provide the public good at a given cost to the
respondent household. The choice of payment vehicle is an important element of the market
scenario. Respondents are likely to react differently to a pollution regulation whose cost is paid
for by firms (even if firms pass the cost on to stock holders and consumers) than to pollution
abatement financed by tax dollars.

28.13 Before asking the individual what he would pay for the commodity valued, he should be
reminded that his expenditures on other goods must be reduced by this amount; i.e., he should
be reminded of his budget constraint. It is also desirable to follow questions that elicit WTP
with questions to determine whether a respondent who says he will pay nothing for the good is
stating his true valuation or a protest bid. A respondent may say that he will pay nothing for a
good that he in fact values if he believes that he should not have to pay for it or otherwise
disagrees with the scenario presented in the survey.

Choice of Technique for Eliciting Willingness to Pay

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 4
Session 28 Contingent Valuation

28.14 In constructing the contingent valuation survey a researcher must also decide what
technique to use to elicit willingness to pay. Open-ended elicitation methods, which include
bidding games and payment cards, provide a point estimate of the respondents maximum WTP
for the commodity in question. Closed-ended methods, which ask the respondent whether he
would pay at least a stated amount, provide intervals in which the respondents WTP lies.

28.15 Among open-ended approaches the simplest is to ask the respondent What is the most
you would pay . . .? for the good in question. This does not provide the respondent with
implied value cues, but may be a difficult question to answer, especially if the good is an
unfamiliar one. For this reason WTP data collected using this approach are sometimes
characterized by gross outliers, or refusal to answer the question. An alternative is to allow the
respondent to choose his maximum WTP from a payment card that lists possible dollar values.
This is likely to be an easier task, but may encourage the respondent to limit his announced
WTP to the values on the card.

28.16 In early applications of the contingent valuation method, WTP amounts were sometimes
elicited using iterative bidding protocols (Randall, Ives and Eastman 1974). In a bidding game
respondents were queried about a certain WTP figure. Those who agreed to pay were offered
higher and higher amounts, until they finally switched to a no answer. Respondents who
declined to pay the initially offered amount were offered lower amounts, until they said yes.
This approach has been virtually abandoned, because it was found to produce starting point
biases, i.e., the final WTP amount was significantly related to the choice of the initial figure.

28.17 Closed-ended approaches, which ask respondents whether they would pay a stated
amount for the good in question, require only a yes-no answer. Their take-it-or-leave-it format
mimics the choice facing buyers in an actual market, or, in the case of a public good, the choice
facing citizens in a referendum. In the case of public goods, closed-ended questions have been
shown to be incentive-compatible (Hoehn and Randall 1987). For these reasons closed-ended
(dichotomous choice) questions have become the standard approach to eliciting WTP.

28.18 The simplest dichotomous choice question asks each respondent whether he would pay
at least $z for the good in question, where z is varied randomly across respondents from a list
of values chosen by the researcher. An affirmative answer to this question, however, indicates
only that the respondents WTP lies in the open interval ($z,). For this reason, an initial
closed-ended question is often followed by a question designed to more narrowly bracket the
respondent's WTP. If the individual says he will pay at least $z for the commodity, he is then
asked if he will pay a higher value, $y, in hopes that he will pay less than $y. Accordingly,
tighter intervals may be formed in which the subjects WTP amount lies, resulting in improved
efficiency of the estimates of mean/median WTP and of the coefficients entering in the WTP
function (Hanemann et al. 1991).

28.19 Recent articles have, however, detected starting point biases in the pairs of answers to
the initial and follow-up payment questions (Herriges and Shogren 1996), and raised the issue

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 5
Session 28 Contingent Valuation

that such follow-up questions may induce rejection of the scenario, especially if respondents
interpret the amount suggested in the dichotomous choice payment question as the cost at which
the government is capable of providing a public good or program (Alberini et al. 1997).

28.20 Another important issue is how the values assigned to respondents in a dichotomous
choice survey should be selected if the goal is to obtain statistically efficient estimates of WTP.
The judicious selection of these values (henceforth referred to as bids) requires some initial
information about the distribution of WTP, which is usually obtained from pretest surveys.

28.21 If WTP is normally distributed with mean and standard deviation , and the primary
goal is to estimate (mean WTP) efficiently, the fiducial design principle from the statistical
literature suggests that only two bid values need be used, and that respondents should be evenly
divided between those values. Specifically, the bid levels should be set equal to d,
where d depends on the sample size, but is generally small. For instance, if n = 5000, d =
0.2411 (Alberini 1995). If the researcher wishes to estimate both and with reasonable
precision, the bid amounts should be further away from the center of the distribution, but in
practice no further than the 15th and 85th percentiles of the distribution (Kanninen 1995).

28.22 If the researcher is uncertain about the shape of the distribution, it is wiser to use a
wider range of values. One approach (Boyle et al. 1988) is to draw points randomly from the
distribution of WTP (estimated in a pretest) and then use the corresponding bid values in the
final survey. Alternatively, one can place bids at equally distant quantiles, where the number of
quantiles chosen depends on the number of subsamples possible given total sample size
(Alberini 1995).

Analysis of Contingent Valuation Responses

28.23 One goal of a contingent valuation survey is to estimate the population distribution of
WTP for the good in question. This is the (weighted) frequency distribution of responses when
the WTP question is open-ended. In the case of a closed-ended question one can compute the
percent of the sample that is willing to pay at least a given value, which provides an estimate of
one minus the cumulative distribution of WTP.

28.24 To compute the total benefits of a policy one typically computes the mean of the WTP
distribution and then multiplies mean WTP by the size of the population affected by the policy.
As an alternative, researchers frequently estimate median WTP, which provides a robust lower
bound for mean WTP and is less sensitive to skewed distributions and the presence of gross
outliers. It is also possible to compute -trimmed means to reduce the influence of extremely
large and extremely small reported WTP amounts. The -trimmed mean is computed after
deleting the 100 percent smallest and 100 percent largest WTP observations.

28.25 In addition to estimating the mean of the distribution of WTP, it is of interest to see how
mean WTP varies with characteristics of the respondent (income, education). This provides a

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 6
Session 28 Contingent Valuation

test of the internal validity of responses and is useful if one wishes to transfer WTP estimates
from one population to another. Equations (1) and (2) above imply that WTP can be written as
a function of income, Y, prices, other non-market goods, taste variables, T, and the original and
final levels of the good in question,

(3) WTP = f(Y,P,Q,T,q0,q1) = X .

28.26 Estimation of the coefficient vector allows one to test the internal consistency of
responses and predict mean WTP as a function of respondent characteristics. If one has a point
estimate of WTP for each individual in the sample, then it is a simple matter to append an error
term to (3) and estimate a WTP function.

28.27 When a closed-ended question is used to elicit WTP, it is assumed that the respondent
answers yes to the stated amount, zi, if Xi + i > zi , where i is an error term representing
variables that affect WTP but are not observed by the researcher. If { i} are independently
and identically normally distributed with zero mean and standard deviation , the coefficient
vector may be estimated using a variant of the probit model (Cameron and James 1987).
Specifically, the probability that the respondent will not pay zi is given by ( zi/ - Xi /)
where is the standard normal distribution function. By estimating a probit model with z and X
as independent variables one obtains estimates of 1/ and /. Mean WTP is then obtained
as Xi .

28.28 When a follow-up dichotomous choice question is introduced, more efficient estimates
of are possible for a given sample size. To illustrate, assume that a respondent is initially
queried about his WTP $30, and assume that the response is yes. The same respondent
would then be asked whether he would pay a greater amount, say $50. If the answer is no,
the respondents WTP lies between $30 and $50. If the answer is yes, the respondents
WTP lies in the interval ($50, ). Denoting the lower and upper bounds of the interval
bracketing the respondent is WTP by WTPiL and WTPiU, and assuming for the sake of
simplicity that i is normally distributed with mean zero and standard deviation , the log
likelihood function of the sample data is given by:
n WTPi U x i WTPi L x i
log L = log
i =1

28.29 Estimates of the parameters and are obtained by the method of maximum
likelihood. If the researcher is reluctant to make distributional assumptions about { i}, non-
parametric techniques can be used to estimate (Carson et al. 1994; Haab and McConnell
1997).

The Validity of the Contingent Valuation Method

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 7
Session 28 Contingent Valuation

28.30 Does the contingent valuation method yield reliable estimates of WTP for the
commodities to which it has been applied? Some economists have expressed skepticism of the
method because contingent valuation questions are hypothetical and may therefore yield
hypothetical answers. Skepticism of the method has been also generated by the implausible
answers observed in some contingent valuation studies (Hausman 1993).

28.28 Contingent valuation studies sometimes produce many zero WTP values and/or a few
implausibly large values. Zero values may be evidence of strategic behavior, of attempts by the
respondent to misrepresent WTP in order to influence the outcome of the survey. Or, zero
values may represent protest bids, indicating that the respondent does not think he should have
to pay for the good, even thought he has a positive WTP for it. WTP values that are large
relative to the respondents income and/or the availability of substitute commodities may indicate
strategic behavior, or that the respondent has failed to consider carefully his budget constraint.

28.32 In some studies WTP fails to increase very much with the quantity of the good valued.
The value of a reduction in risk of death may not increase with the size of the risk reduction
(Jones-Lee et al. 1985) or the WTP to save 1,000 birds may not differ significantly from the
WTP to save 100,000 (Desvousges et al. 1993). This suggests either that the good that is
valued

is not well understood or that it is a category of programs that is being valued; i.e., that the value
attached to a large class of goods is embedded in the good valued (Kahnemann and Knetsch
1992).

28.33 To test whether contingent valuation provides accurate estimates of value, researchers
have compared contingent valuation estimates with the results of actual or simulated market
transactions. These tests assume that actual WTP measures true compensating variation, and is
therefore an appropriate criterion to which to compare the contingent valuation method. This is
perhaps reasonable for private goods, but is problematic for public goods, where actual WTP
may understate compensating variation due to the free rider problem.

28.34 The results of such comparisons are mixed. In the case of private goods, several
experiments comparing hypothetical and actual WTP for hunting permits have found no
statistically significant difference between mean values of hypothetical and actual willingness to
pay (Mitchell and Carson 1989). Similar results were obtained by Dickie et al. (1987) in the
case of strawberries. Comparisons of stated v. actual WTP for public goods have generally
found that the former overestimates the latter (Kealy et al. 1987, Arrow et al. 1993).

28.35 A second test of validity compares contingent valuation estimates of WTP with
estimates produced by indirect methods, such as hedonic pricing or the travel cost approach,
that infer WTP from actual behavior. To the extent that the two are similar, it is unlikely that
people are overstating their WTP in a contingent valuation survey simply because they do not
have to pay for the good. A meta analysis of 83 studies that estimate WTP using both direct

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 8
Session 28 Contingent Valuation

and indirect methods found that the mean ratio of the contingent valuation estimate to the
estimate produced by indirect methods was 0.9 (Carson et al. 1996). The mean ratio of the
two estimates was not significantly different from 1.0 at the 0.05 level.

28.36 Comparing contingent valuation to other methods of valuing a good provides important
evidence of the reliability of the contingent valuation method, but clearly cannot be applied in all
cases. Tests of internal validity attempt to determine whether responses to a given contingent
valuation survey are internally consistent. Tests of content validity ask whether the questionnaire
describes the commodity to be valued in a meaningful way and avoids biases introduced by
failing to mention substitute commodities or encouraging the respondent to consider his budget
constraint.

28.37 Tests of theoretical validity ask whether WTP behaves as theory predicts: whether it
increases with income and with the quantity of the good valued. To see whether WTP increases
with income and is influenced in an expected way by education and taste variables, it is essential
that such covariates be included in the questionnaire and that a WTP function be estimated as

part of the data analysis. Tests of scope, which measure whether WTP increases with the
quantity of the good valued, should also be included in the questionnaire (Arrow et al. 1993).
By having each respondent value different quantities of the good in question or different
respondents vary different quantities of the good, one can see whether WTP increases with the
quantity of the good.

The Use of Contingent Valuation Studies

28.38 Despite the criticisms of the method, contingent valuation has been used to value the
benefits of environmental regulations, especially in the areas of air and water pollution control.
Contingent valuation studies have served as inputs to investment decisions by the U. S. Forest
Service and the World Bank, and have been used in litigation to value damages to natural
resources caused by oil spills and other toxic releases.

28.39 In the area of environmental regulation, the USEPAs cost-benefit analysis of the Clean
Air Act, 1970-1990 (USEPA 1997) used contingent valuation studies to value both reductions
in acute and chronic illness and reductions in premature mortality associated with air quality
improvements. Specifically, contingent valuation studies provided estimates of WTP to avoid a
symptom day (Loehman et al. 1979; Tolley et al. 1986), as well as WTP to reduce ones risk
of contracting chronic bronchitis (Krupnick and Cropper 1992). Contingent valuation studies of
WTP for reduced risk of death were used together with compensating wage studies to value
reductions in premature mortality (Jones-Lee et al. 1985). In evaluating proposed water quality
regulations for the iron and steel industry, EPA relied upon contingent valuation studies of the
value of a fishing day (Desvousges et al. 1983) to estimate the value of water quality
improvements (USEPA 1982).

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 9
Session 28 Contingent Valuation

28.40 Contingent valuation studies have been used to compute the benefits and costs of
infrastructure investments. The World Bank has used contingent valuation studies to estimate
WTP for piped water connections and sewerage services in developing countries (Griffin et al.
1995) and has based investment decisions on these results. Comparisons of stated and actual
willingness to pay for piped water connections in Kerala, India found that contingent valuation
studies correctly predicted 91% of actual decisions to connect to piped water. In estimating the
returns from electricity generation in the United States, contingent valuation studies have been
used to compute the social costs of electricity generation for use in investment decisions by
power plants. Social costing studies (Krupnick and Burtraw 1997) have relied on contingent
valuation to value visibility impairment (Chestnut and Rowe 1990) as well as health damages.

28.41 Resource planning decisions by the U.S. Forest Service have also employed contingent
valuation to estimate unit values for various recreation activities (Walsh, Johnson and McKean
1992). These include WTP for a day spent fishing, hunting or hiking. Contingent valuation
studies, which ask respondents what they would pay rather than forego a day spent on the
activity in question, have complemented travel cost studies as a method of valuing recreation
resources.

28.42 Finally, contingent valuation remains one of the most important techniques used to
assess natural resource damages. When an oil spill or other toxic release results in injuries to
natural resources, such as wildlife or groundwater, trustees of the resource typically sue
potentially responsible parties for damages. Trustees seeking compensation for injuries have
often relied on contingent valuation studies, as in the cases of the Nestucca and Exxon Valdez
oil spills (Rowe, Shaw and Schulze 1992; Carson et al. 1994). In other cases, such as the
Eagle Mine in Colorado, the plaintiff employed a combination of techniques, including a
contingent valuation survey, travel cost and hedonic price methods (Kopp and Smith 1993).
While Department of Interior rules once gave preference to methods of assessing damages
other than contingent valuation, the District of Columbia Circuit Court of Appeals ruled in Ohio
v. United States Department of the Interior (880 F.2d 432 1989) that contingent valuation was
not inferior to other methods of estimating natural resource damages.

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 10
Session 28 Contingent Valuation

References

Alberini, A. 1995. Optimal designs for discrete choice contingent valuation surveys: single-
bound, double-bound and bivariate models. Journal of Environmental Economics and
Management 28: 187-306.

Alberini, A., Kanninen, B. and Carson, R. T. 1997. Modeling response incentive effects in
dichotomous choice contingent valuation data. Land Economics 73:

Arrow, K., Solow, R., Portney, P. R., Leamer, E. E., Radner, R. and Schuman, H. 1993.
Report of the NOAA panel on contingent valuation. Federal Register 58: 4601-4614.

Bishop, R. C. and Heberlein, T. A. 1979. Measuring values of extramarket goods: are indirect
measures biased? American Journal of Agriculture Economics 61: 926-30.

Bishop, R. C., Heberlein, T. A. and Kealy, M. J. 1983. Contingent valuation of environmental


assets: comparisons with a simulated market. Natural Resources Journal 23: 619-34.

Boyle, K., Welsh, M. P. and Bishop, R. C. 1988. Validation of empirical measures of welfare
change: comment and extension. Land Economics 64: 94-98.

Cameron, T. A. and James, M. D. 1987. Efficient estimation methods for closed-ended


contingent valuation surveys. Review of Economic Statistics 69: 269-76.

Carson, R. T., Mitchell, R. C., Hanemann, M. W., Kopp, R. J., Presser, S. and Ruud, P. A.
1994. Contingent valuation and lost passive use: damages from the Exxon Valdez. RFF
Discussion Paper QE 94-18, Resources for the Future, Washington D.C.

Carson, R. T., Flores, N. E., Martin, K. M., and Wright, J. L. 1996. Contingent valuation and
revealed preference methodologies: comparing the estimates for quasi-public goods. Land
Economics 72: 80-99.

Chestnut, L. G. and Rowe, R. D. 1990. Economic valuation of changes in visibility: a state of


the science assessment for NAPAP. In Acidic Deposition: State of Science and Technology,
Report 27 eds. G. Brown and J. M. Callaway, Washington, D. C.: National Acid Precipitation
Assessment Program.

Desvousges W. H., Johnson, F. R., Dunford, R. W., Hudson, S. P., Wilson, K. N. and Boyle,
K. P. 1993. Measuring natural resource damages with contingent valuation: tests of validity and
reliability. In Contingent Valuation. A Critical Assessment ed. J. A. Hausman, Amsterdam:
North Holland Publishers.

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 11
Session 28 Contingent Valuation

Desvousges, W. H., Smith, V. K. and McGivney, M. P. 1983. A Comparison of Alternative


Approaches for Estimating Recreation and Related Benefits of Water Quality
Improvements. Washington, D. C.: U. S. Environmental Protection Agency.

Dickie, M., Fisher, A., and Gerking, S. 1987. Market transactions and hypothetical demand
data: a comparative study. Journal of American Statistical Association 82: 69-75.

Griffin, C. C., Briscoe, J., Singh, B., Ramasubban, R. and Bhatia, R. 1995. Contingent
valuation and actual behaviour: predicting connections to new water systems in the state of
Kerala, India. World Bank Economic Review 9: 373-395.

Haab, T. C. and McConnell, K. E. 1997. Referendum models and negative willingness to pay:
alternative solutions. Journal of Environmental Economics and Management 32: 251-70.

Hanemann, W. M. 1991. Willingness to pay and willingness to accept: how much can they
differ? American Economic Review 81: 635-47.

Hanemann, M., Loomis, J. and Kanninen, B. 1991. Statistical efficiency of double-bounded


dichotomous choice contingent valuation. American Journal of Agricultural Economics 73:
1256-63.

Hausman, J. A. (ed.) 1993. Contingent Valuation. A Critical Assessment. Amsterdam:


North Holland Publishers.

Herriges, J. A. and Shogren, J. F. 1996. Starting point bias in dichotomous choice valuation
with follow-up questioning. Journal of Environmental Economics and Management 30:
112-28.

Hoehn, J. P. and Randall, A. 1987. A satisfactory benefit cost indicator from contingent
valuation. Journal of Environmental Economics and Management 14: 226-47.

Jones-Lee, M. W., Hammerton, M. and Philips, P. R. 1985. The value of safety: results of a
national sample survey. Economic Journal 95: 49-72.

Kahneman, D. and Knetsch, J. 1992. Valuing public goods: the purchase of moral satisfaction.
Journal of Environmental Economics and Management 22: 57-70.

Kanninen, B. J. 1995. Bias in discrete response contingent valuation. Journal of


Environmental Economics and Management 28: 114-25.

Kealy, M. J., Dovidio, J. and Rockel, M. L. 1987. Willingness to pay to prevent additional
damage to the Adirondacks from acid rain. Regional Science Review 15: 118-41.

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 12
Session 28 Contingent Valuation

Knetsch, J. L. and Sinden, J. A. 1984. Willingness to pay and compensation demanded:


experimental evidence of an unexpected disparity in measures of value. Quarterly Journal of
Economics 99: 507-21.

Kopp, R. J. and Smith, V. K. (eds.) 1993. Valuing Natural Assets, Washington, D. C. :


Resources for the Future.

Krupnick, A. J. and Burtraw, D. 1997. The social costs of electricity: do the numbers add up?
Resources and Energy 18: 423-466.

Krupnick, A. J. and Cropper, M. L. 1992. The effect of information on health risk valuations.
Journal of Risk and Uncertainty 5: 29-48.

Mitchell, R. C. and Carson, R. T. 1989. Using Surveys to Value Public Goods: The
Contingent Valuation Method. Washington, D. C.: Resources for the Future.

Loehman, E. T., Berg, S. V., Arroyo, A. A., Hedinger, R. A., Schwartz, J. M., Shaw, M. E.,
Fahien, R. W., De, V. H., Fishe, R. P., Rio, D. E., Rossley, W. F. and Green, A. E. S. 1979.
Distributional analysis of regional benefits and costs of air quality control. Journal of
Environmental Analysis and Management 6: 222-243.

Randall, A., Ives, B. C. and Eastman, C. 1974. Bidding games for valuation of aesthetic
environmental improvements. Journal of Environmental Economic Management 1: 132-49.

Rowe, R. D., Shaw, W. D. and Schulze, W. D. 1992. Nestucca oil spill. In Natural Resource
Damages: Law and Economics, eds. K. M. Ward and J. W. Duffield, New York: John
Wiley.

Tolley, G. S., Babcock, L., Berger, M., Bilotti, A., Blomquist, G., Brien, M., Fabian, R.,
Fishelson, G., Kahn, C., Kelly, A., Kenkel, D., Krumm, R., Miller, T., Ohsfeldt, R., Rosen, S.,
Webb, W., Wilson, W., and Zelder, M. 1986. Valuation of reductions in human health
symptoms and risks. University of Chicago. Final report to the U. S. Environmental Protection
Agency.

United States Environmental Protection Agency. 1997. The Benefits and Costs of the Clean
Air Act, 1970 to 1990. Washington, D. C.

United States Environmental Protection Agency. 1982. An Economic Assessment of the


Benefits of Final Effluent Limitation Guidelines for Iron and Steel Manufacturers. EPA-
230-04-82-005. Washington, D. C.

Walsh, R. G. , Johnson, D. M. and McKean, J. R. 1992. Benefits transfer of outdoor and


recreation demand studies, 1968-1988. Water Resources Research 28: 707-713.

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 13
Session 28 Contingent Valuation

Environmental Economics and Development Policy Course


World Bank Institute, July 15-26, 2002, Washingto n, D.C. Page 14