- i
Benefits Transfer: Procedures,
Problems, and Research Needs
1992 Association of Environmental and Resource
Economists Workshop
Snowbird, Utah
June 3 - 5,1992
U.S. EPA Headquarters Library
Mail code 3201
1200 Pennsylvania Avenue NW
Washington DC 20460
f
PA
'jO
•3
-J18
c.2
Tayler H. Blngham
Research Triangle Institute
Research Triangle Park, NC
Dr. Elizabeth David
State of Wisconsin
Department of Natural Resources
Madison, Wl
Dr. Theodore Graham-Tomassi
Michigan State University
East Lansing, Ml
1 Dr.MaryJoKealy
U.S. Environmental Protection Agency
Washington, DC
Dr. Michael UBIane
U.S. Department of Agriculture
Economic Research service
Washington, DC
Dr. Robert Leeworthy
National Oceanic and Atmospheric Administration
Rockvilie, MD
-------
I
•»
J
J
J
-------
•I
FOREWORD
The 1992 Association of Environmental and Resource Economists (AERE) Workshop
was die third in a series of important recent activities related to benefits transfer. In November
1992 die National Oceanic and Atmospheric Administration (NOAA) hosted a workshop
directed toward developing databases to support benefits transfers. The U. S. Environmental
Protection Agency (EPA) has taken a key first step in titis development by compiling a
bibliography of tiieir environmental benefits studies (see Appendix A). In March 1992 a special
section of Water Resources Research was dedicated to papers addressing issues related to
benefits transfer. The AERE workshop sought to expand on dus base by addressing questions
related to die adequacy of existing methods in4 valuation studies for performing benefits transfer
and by identifying die research needed to enhance benefits transfers.
Appreciation is extended to die workshop sponsors—EPA, NOAA, and U J. Department
of Agriculture's Economic Research Service—for tiieir continuing support of die workshop
series and for dus workshop hi particular.
"Benefits transfer" is die use of information from existing nonmarket valuation studies to
develop value estimates for anodier valuation problem. It n*i reduce bom die calendar tit^c and
resources needed to develop original r,srimatrs of values for environmental commodities. These
estimates are used to evaluate die attractiveness of potential governmental policies, to assess die
value of policies implemented in die past, and to identify die compensation required under
CERCLA when toxic substances, such as oil or PCBs, are released to die environment
Benefits transfer is not new. In any ex ante studies of policy options, researchers must
transfer information from other times and places to die present question. The policy researcher
straddles two points in time—die past and die rutnre—attempting to apply experience from die
past to a future situation. For example, die economist evaluating die likely effects of a possible
minimum wage increase on employment must draw on previously conducted research to forecast
die effects of die specific policy under consideration. This research may have analyzed a
"natural experiment" in a past setting. From dus research die economist may conclude tiiat a
10 percent increase in die miirftnym wage above die equilibrium wage caused employment to fall
5 percent in die affected labor markets. The economist offering advice on die specific policy
under consideration may use tiiis information to forecast tiiat die same (a lesser or greater) effect
is expected this time because die present situation is like (unlike) die past
For hypothesis testing purposes, frequently only die sign of die variable(s) of interest is
critical for supporting a tiieory. But for policy analyses die magnitude of die effect is critical.
111
WAKH'K'C'iUN, L-.C.
-------
Indeed in formal benefit-cost analysis, quantification is virtually the sine BOB qua. Benefits
transfer pmviftes a means of economically obtaining these magnitudes. However, the process of
benefits transfer is complex, and a "science" of benefits transfer does not now exist One
purpose of die workshop was to increase our awareness of the types of decisions involved in
performing benefits transfer and the research needed to close some of the gaps in our knowledge.
The workshop consisted of three formal papers, six benefits transfer case study protocob,
the concluding remarks of three discussants, and an after-dinner speaker who outlined the utility
of an information system to support benefits transfers. The case study protocols were selected to
provide a forum for evaluating the potential for conducting benefits transfer in specific
applications and to identify research needs. The case study groups comprised workshop
participants and a leaders) who provided the initial case study niaterials to the members of the
group, presented the results of the group's discussions to the entire workshop, and wrote the final
David Brookshire in his opening remarks to the workshop observed that the question we
face is not whether benefits transfers will be done but rather how. The imperative for such
studies is simply too strong to resist He highlights the complementary relationship between
many of the issues the researcher must address in benefits tnmsfer and in original nonmarket
valuation studies. Fonhemom, he raise* qii«triftn* regarding ifrp adequacy of me existing
research base to support benefits transfer applications.
Leland Deck and Lauraine Chestnut consider how good benefits estimates must be for
transfer purposes. Taking a value of infonnation approach, they look at the "market" for benefits
estimates and the costs of developing them. They identify several stages in the development
process, each of which represents a possible stopping point in developing benefits estimates.
Edward Morey investigates the rrJarionship between consumer's surplus and consumer's
surplus for a day of recreational use. Estimates of consumer's surphis for a day of use are
commonly used for benefits transfers. He shows mat compensating variation per day of use is a
well-defined concept for a change in the price of visiting a recreational site but is not, in general,
well-defined for a change in the characteristict of a tifr ffe idcnti*^f "rfpetent conditions for
when it M wdUdgfbied for ehataeterartiea change «nrf n*e* ritt^nUtin«ff tff JOBffniHfatP tfre bifttCS
from using approximations for the compensating variation.
In the first case study, John Bergstrom and Kevin Boyle develop a protocol for estimating
the value of protecting groundwater in a rural area dependent on it for its water supply. They
identify several studies mat provide information for ifaeir benefits transfer problem, provide a
.1
J
\
\
IV
-------
catalogue of fre duniEterfcrics of these studies, ind present a benefits value from the transfer
process. An actual policy rite study serves as a validity check on the estimates they derive
through die benefits transfer process. That study supports their benefits transfer value.
Bill Desvousges and bis coauthors examine the me of benefits transfer to value use
damages from the Arthur Kill oil spflL Their group evaluated me adequacy of existing studies
for several categories of water and wetlands use. They identify several important gaps in the
data, one of the more important of which is in wetlands value*—especially nonuse values for
wetlands preservation/restoration.
Carol Jones describes the Department of Interior's Type A oil spill model and evaluates
its adequacy for estimating the value of recreational fishing losses in a natural resources damages
The Type A model, which provides a computerized approach to predicting the fate and
effects of spills and to valuing injuries, is die major benefits transfer model for natural resource
damage assessments. Her paper identifies some of the improvements to the model, especially the
valuation component, that the case study participants thought would mate the results more valid.
Susan Kask examine* the potential for transferring healm benefits estimates to a study
site involving health risks from surface water contamination. She describes a theoretical model
and identifies a number of factors that may influence the value estimate. She finds mat a major
problem with healm benefits valuation is the absence of studies addressing both morbidity and
mortality in a comprehensive fashion.
Mary Jo Kealy and her coauthors develop a protocol for estimating the recreational
fishing benefits of reductions in acid deposition. This is an ex ante analysis frccaosf it examines
the expected benefits from implementing the Clean Air Amendments of 1990 (CAAA). They
tmyf their protocol on fit- Deck flftd Chestnut ittfiy^ process; f***^* jrtajr represents a decision
point at which the researcher asks if die expected value of the benefits of the information gained
from proceeding with the next stage exceeds its costs.
Lauraine Chestnut and Robert Rowe also conduct an ex ante study of the CAAA. They
ine the potential to transfer previous studies of the value of visibility improvements to a
study of the value of the expected reduction in regional haze in the Eastern United States. They
argue for a protocol mat incorporates all available infonnation, propedy weighted, and assesses
the uncertainty of the results. Their group was relatively comfortable with the availability of
infonnation for their benefits transfer problem; however, they aU felt the contingent valuation
method could be significantly improved.
-------
I
Tmdy Cameron's icmaiks draw from the Uterttore and from experiences outside I
environmental and resource economict to suggest both inchmeal and institutional changes flat ^p
would improve benefits transfers. She shows that tbe benefits transfer issue is not an activity |
mittpie tn m; it has much in common with other efforts to develop more rigorous procedures for
ombining information. She addresses the issue of ample bias mat may be present in original - I
studies wben applied to a policy site and suggests a procedure f or rewtighting the original data
based on the policy rite variables. She also Ascribes a way to devdopestinutes from pooled 1
data using prior information that has potential application for benefits transfer. Finally, she •*
J
proposes some institutional changes mat would result in improved archiving and sharing of
original data sets for others to use in their research.
Alan Knronick discusses the demand for bencfitttnuisferstodies,mp«rticiiljf thcguse 1
for developing estimates of me external costs of electric power. He considers several types of
benefits and offen his opinions on those for which the research base is strong enough to support I
their transfer to oner contexts. He «!«*> raises some important issues regarding the value of *
standard protocols for documenting the choices benefits transfer practilionets •»•*»> so ***•* thfir •
choices and reasoning are clear to readers. He concludes h** icmarks with a suggestion for I
research that would improve the quality of bcncfHs transfers.
Tim Opalnch's and Marisa Mazzotta's concluding remarks argue mat a valid and reliable V
research base of original studies is complementary to benefits transfers. They also identify the %
need to provide empirical tests of benefits transfers and to develop better methods for I
transferring benefits estimates.
Martin TAavM pftinly fflif in Jfly BpmflHrg tfi^t, QfKg COllfiCtBdt data htW n*"W Of flip
characteristics of a public good. A system f or arcbivmg and sharing data would promote good \
science, learning, and better policy analysis. He provides some suggestions for an information '
system based on his experience wim other complex data sets. .
The papers, case studies, and discussants* remarks highlighted several cpncems
researchers have about performing benefits transfers. The wockshop participants* concerns mn^ 11
suggestions for a research program and some of my own are provided below.
ise benefits transfer begins wim original studies, many of the issues raised applied 'J
aswdltomem. Specifically.woikshoppartknpam^ were concerned mat bom U^SCOD^
studies and the reporting of data, methods, and findings m bom u^ published Uteiature and in •]
reports are not complete enough to perform good benefits transfers.
vi
-------
I
[
More original studies are needed dial address die human health effects of the
environment The available information on morbidity and mortality values is very limited and
tends to focus on adult health and life expectancy. Additional research is also needed on the
value of reducing infant and children's morbidity and mortality, including die value of reducing
die risks of reduced IQ and physical effects from botii die pregnant modier's exposure to
environmental pollutants and die infant's or child's subsequent direct exposure. Born parents
and prospective parents would probably place a high value on risk reductions in this area, but die
literature provides virtually no estimates of diese values.
More work is needed on specific services provided by die environment and die
fhafTi/t»rirarinn of hnw die valne of these services is affected by changes in die quality of die
environment Workshop participants specifically identified die need for additional studies of
water resource use, including beating and beach use and wetlands values, Abo die link between
injury to die environment and damages needs to be clearer in bom die original and benefits
transfer studies. Achieving this clarity may require an expanded role for economists in die
modeling of physical systems and their relationship to human activity.
Only one case study at die workshop touched on nonnse values, yet tiiey are die most
controversial component of benefits studies. Part of die reason for interest in nonuse values is
clear—even a small value when multiplied by a large number of affected individuals can result in
a number large enough to dwarf use values. The nonuse issue begins by asking when nonuse
values are relevant and extends to both technical and policy issues. Bodt die values elicitation
process and die extent of die "market!" for nonnse values are controversial issues. More research
is needed on die way nonuse values enter individuals* utility functions and on die values
elicitation process.
Systematic implementation of unproved benefits transfers is probably impossible wimout
better access to well-documented data, AERE should develop a standardized protocol for
documenting survey procedures used in original studies. For example, die protocol would
provide information on sample sizes and selection, data coding and checking, response rates and
steps tp1*^*1 to minimi^ nonresDonse bias, and die treatment of outiiers in dig estimation phase.
A completed protocol could be required for all Journal of Environmental Economics and
Management papers submitted for review. The completed form could be attached as an appendix
to papers accepted for publication; die journal editor could keep copies of die completed form, or
die authors could be required to express their willingness to provide diem to odier researchers
when so requested.
vii
-------
Related to the issues of documentation for both original and policy studies is the
importance of providing dear definitions of die baseline quantities of resource services, of
reporting value functions not just means, and of employing sensitivity analysis for key
parameters when performing benefits transfers.
Many participants discussed the lack of incentives to share original data such as survey
information. This problem will be hard to solve. One approach that may result in lower costs of
f»««nrig and ffffly r*™«ft«e enmnilinfrflH"" " *» devdnp • stmAmrMrrA approach to managing J
survey data. Again AERE could play an important role: it could design a univenal information
system to provifc a stamiariforauufor elec^^ I
would have to.be flexible enough to meet a wide range of researcher mterests yet structured
erough so that usenunfaniiliar with the d^ I
link need to mvolve the origmd researchers. Such a system would not solve the proprietary '
interest that developers of the data may have, but, given their willingness to share their data, it . \
would lower their and the recipient's cost of mat sharing. I
We should consider the value of studies that replicate die work of others. Too often I
existing editorial policies are opposed to replication; then when the "righT signs are found, the
opportunity for publishing papers confirming the results of others is very limited. But the 1
parameter estimates are at best just mat—estimates. The estimates are conditional on the 9*
institutional context and constraints impinging on the individual decision makers. A broader .
basg of empirical ffftv*fct ff Tf^fd to support benefits transfer. Further research may help to J
develop the preponderance of evidence needed for theories to have broad acceptance.
A benefits transfer must assess the extent to which the following arc "similar" between '
the study and policy site contexts: affected resomce(s)t damageCs), substitutes, and affected \
population. Studies will be similar in some features, different in Others. How should we weight I
studies for use in benefits transfer, and how should we communicate those weights to our
audience? Can this weighting be done objectively? Quantitatively? Meta-analysis and some of J
the literature cited by Trudy Cameron may be useful in addressing these questions.
Most benefits estimates have been developed in the United States and to a lesser extent in *
Europe. More research is needed to evaluate the extent to which these estintates are transferable -»
across societies where preferences, constraints, and institutions differ. Similarly, move work is J
needed to identify the carcnrnsnincrs for whicfa mlergeiierational benefits transfers are
appropriate and the procedures mat should be used to modify ciinent estimates to express the j
values and constraints appropriate for future t
*
-------
Changes in environmental quality are likely to affect both the intensity and quality of
resource use. For example, a beach with oil on it may experience reduced visits by beachgoers;
however, some people may still frequent it In both original and policy studies we should
explicitly value the losses in utility for the foregone visits as well as the reduction in the value of
the remaining visits.
Benefits transfers will not have the elegance of pure theory or the rigor of hypothesis
testing. This method seems likely to emerge as a different science, one that uses the results from
original research but is based on interpreting economic history and applying it to current
problems. Itcan provide useful input to policy issues mat directly affect resource allocation and
to compensation questions that may indirectly change resource allocation as liability rules are
internalized into future choices.
TaylerH. Bingham
Research Triangle Park, NC
-i
-------
I
j
J
'\
-------
David S. Brookshire
LdandB.Deck
Lamaifle G. Chestnut
Edward R. Morey
John C. Bergstrom
Kevin I. Boyle
William H. Desvousges
Richard W. Dunfbrd
Kristy E. Mathcws
Carol Adairc Jones
Susan B.Kask
MaryJoKealy
Susan Herrod
George Parsons
Maxk Montgomery
Lauraine G. Chestnut
Robert D. Rowe
Trady Ann Cameron
Alan J. Krupnick
James J. Opaluch
MarisaJ.Mazzotta
Martin R David
CONTENTS
Issues Regarding Benefits Transfer
Benefits Transfer: How Good Is Good Enough?
What is Consumer's Surplus Jbr a Day of Use And What Does It Tell Us
About Consumer's Surplus?
Groundwater Valuation: Dougherty County, Georgia
Natural Resource Damages Valuation: Arthur Kill Oil Spill
Recreational Fishing Valuation: Application of me Type A Model
Long-Term Health Risks Valuation: Pigeon River. North Carolina
Acid Rain Provisions of the Clean
Air Act Amendments
Visibility Valuation: Add Rain Provisions of the Clean Air Act
Issues in Benefits Transfer
Benefit Transfer and Social Costing
Fundamental Issues in Benefit Transfer and Natural Resource
Benefiting Benefits Transfer. Information Systems for Complex Scientific
Data
Appendix A: Economic Analysis and Resource Branch Environmental Benefits Database
Appendix B: Attendees List
XI
-------
J
J
1
-------
ISSUES REGARDING BENEFITS TRANSFER
David S.Brookshire*
ABSTRACT
Although benefits transfers are not new, many issues remain unresolved. In this paper I
make three arguments: most, if not all, of the issues regarding nonmaiket valuation tie also
relevant to developing protocols for benefits transfer; considering the required level of
accuracy for different uses of nonmarket values is central to the benefits transfer process; the
existing set of nonmarket studies does not form an TK^qiratf base for benefits transfer.
•I
Benefits transfer has been a widely used methodology in policy analysis and natural
resources decision making for decades. The process involves
focus[ing] on measuring On dollars) how much die
gain from it They are not forecasts, and they usually
exogenous influences on people's behavior. Instead, a
i affected by some policy will
• not attempt to predict other
Ined set of conditions is
assumed to characterize the nonpolicy variables. Then benefit estimates are derived by
focusing on the effects of die conditions «Mmn«ii to be changed by die policy. (Smith,
1992, p. 686)
Viewed simplistically, the benefits transfer process applies a data set that was developed for a
unique purpose to an application for a different purpose.
The use of benefits transfer has increased recently and is thus receiving renewed
attention. The renewed interest stems from various sources, including recent court decisions
(State of Ohio, 1989), increased federal agency interest, and financial pressures due to increased
costs and limited funding for primary studies. As recently as fall 1991, an environmental
database workshop held in Washington, DC assessed the availability of existing nonmarket
valuation studies and considered means to enhance the availability of these studies for purposes
of benefits transfer.*
My renewed interest in benefits transfer was rekindled by two papers on the topic that I
received approximately two years ago for possible publication in Water Resources Research
(Luken. Johnson, and Kibler, 1992; Desvousges, Naughton. and Parsons, 1992). The review
process raised a relatively unique problem for an editor. The reviews ranged from "publish this
paper, it is great, timely etc.** to "you cannot do this, benefits transfer make no sense." The
distribution of recommendations was so highly bimodal that I decided to edit a special issue of
'University of New Mexico. Department of Economics.
l
-------
^flferJfwoimrtf/^*^^ to direcdy address the benefits transfer process. I believed the Luken
and Desvousges papers were controversial and represented a challenging contributJon to the
literature. Too ofta controversial papers never make it into the literitorebecwse they ire, in
fret, controversial Further, after reflecting about the notion of benefits trinsfcr. I believed we
had already become committed to this method: the issue was not whether, bat how benefits J
transfer should be condocted. We needed a forom for discussing possible issues and protocols
to benefits traiisfer. Hopefully, itespecid 1
direction.*
This paper builds on the issues already identified in the extant literature (see, for instance, J
Smith and KIOTO. 1990; Walsh, 1992) and those brought forth in the special Water Resources
.3 and raises additinirt I argue
that
£•'• i
• most, if not afl, of me issues regarding nonutttet valuation ate also relevant to I
developing protocols for benefits transfer;
1
is central to the benefits transfer process; and J
fee eatisting set of
transfer protocols;
• fee eatisting set of Mnniaiket studies d^ I
|
PROTOCOLS FOR BENEFITS TRANSFER4
frffi^iiMMMe henefit* tnnsfer undtes and guidelines have assigned values through using
expcrtopinion as wefl as results from observed behavior and direct elicitation models.5 Why J
attempt to develop benefits transfer protocols? Why not just conduct a primary study? Two
reasons justify developing these protocols: primary studies can be time consuming and costly. 1
• •• '* • •
Studies based on original data require developing survey tastnraeats, selecting and
drawing a sampfc, administering the instnment, and analy^ J
In some castss the caleixlartmMreqmred is simply n^ f
^I^mfegK)g)pCffMSOf ••lyCOBttilJBlOflMdiBWBWBB,MB SpBCiaiSBMOOWiSJMBWHM•§IHE> JO^WB «77*•*••» .^
-Jp BAlSl^M' S'AMftStfW^Pff tf^^WMVWA •
>SeeAtktean,QDCte.aBdSao|ma992XBi«ftste 'I
. .. aa*IIklcr(1992);MeOoBBdl
(1992); SaiMh (1992); ^Wrf^JoiMW».^McBJ«a992XWa»«ao«r»
1992.
^lieBnokdilie sod Neffl (1992), Looori«a992) and McCai^ 0992)
-fl
CooncU, 1983). J
-------
cases do not always have the
L
I
•i
governmental policy makers and litigants in damage a
time to conduct primary studies.
Financial resources are also limited Recognizing this, the U.S. Environmental Protection
Agency (EPA), National Oceanic and Atmospheric Administration (NOAA), and the U.S. Forest
Service (USFS) are actively addressing data and protocols for benefits transfer.
Local and state governments also have t growing need for nonmarket valuation
information. Many environmental matters are addressed at the state level, yet financial resources
are very limited. For instance, New Mexico has a large natural resource portfolios
competing needs. Developing countries, including eastern Europe, also lack the financial
resources for primary studies. Mexico is a case in point: it has a great need to obtain an
understanding of the overall environmental problems (Mclntosh, 1991). In general, applying
nonmarket valuations across differing national economies is a relatively unexplored area. An
obvious place to start would be with benefits transfer rather man costly primary studies.
METHODS FOR ASSIGNING NONMARKET VALUES
An overview for benefits transfer must begin with a consideration of the methods
available for initially assigning nonmarket values, the accuracy of the methods, die diversity of
nonmarket commodities of interest, and the existing databases. Additional issues might include
the existing form of research agendas, the availability of data, and die rote of judgment
Observed behavior methods (direct or indirect), such as the travel cost and hedonic
methods, and/or hypothetical behavior, such as the contingent valuation approach, form the core
of desirable methods. The literature reveals that many variants *n<* much discussion of die
robustness of each exist on nonmarket methods.6
The available, primary nonmarket valuation methods are not completely reliable or
accurate. In my opinion, accuracy concerns preclude a cookbook approach to benefits transfer,
as is true for ncnunarket valuation efforts in general. Further, not an applications of nonmarket
techniques are created equal The more recent studies are not necessarily superior, more
accurate, or more useful as some would seem to imply. We have not reached a consensus about
die correct procedures with which to conduct all of our nonmarket valuation investigations.
judgment issues not withstanding. Nor have we reached a consensus on valuing various types of
component values. For instance, is the appropriate valuation framework a total value framework,
tbe contingent valmdao method
Bnx)kshin.M^
-------
I
or arc specific nonuse values the appropriate focus, or both? Thus, we should not expect all »
nonmarket valuation tedmkpTff to ^ equally useful in all cages of benefits transfer V
In addition to nonmarket valuation methods, the nature of the commodity is central to the
reliability of die benefits transfer process. When itacarchcrs think of a study site (the primary I
study) and policy site (the site for which benefits are being transferred), they immediately must '
consider questions of uniqueness and substitute and complement availability. For example, if not »
all groundwater is just grouwlwater, then tte specific nature of the commodity at the study and J
policy sites is important
The quality of existmgnoninarlBetd^ also is important. Given that primary studies are
far from perfect, benefits transfer studies more than likely compound the accuracy problems of i
primary studies. The accuracy problems that exist in the primary studies do not disappear when •
die benefits transfer process is undertaken. This general theme has been with us for years (see »
Morgenstem, 1973). I
Although a large number Of SttKPO ^fo, *^>g nmnh^r «vai1aM<» fnr gppcjfic nf>nmgrlp*»i I
might h^ limitfd, At best the current valuation database is A collection of studies •
that represent a serendipity of perceived needs. To some extent, funding agencies find
coordinadne research agendas difficult. Further, data have been lost through the process of
changing affiliations of researchers as well as through changing computer technology. Perhaps
not an of the raw data actually exist mow ever-expanding bibtiography of studies. ^
Researchers should be concerned about extending the base of available studies. A |
systematic and coordinated research program is needed as well as a changeinhowwe *
characterize productive research. Recently the legal community has become a significant source .
for the funding of new studies. In some sense die legal community is pacing our research efforts. I
This pacing and direction of efforts with ynMiifie agfMlflft in *»fad may outstrip our actual
abilities to assign sufficiently accurate values.7 Further, replication is often viewed as not I
productive to journal editors and reviewers and thus not rewarded by the prof esnon.
Ag t hf ^fl^^ y» frpy nHc *nA K»iv»fitg transfer bfoome morft pfpvatent, another «ffw
-------
•i
i
article directly reported on be made available, or should the complete dan set be made available
even if the authors intend further work?8
Finally, what is the role of judgment in the benefits transfer process, both in the original
studies as well as in the process itself? Can we be judgment free and purely scientific?
ADEQUACY OF EXISTING SET OF STUDIES
A wide range of measurement issues is associated with all nonmarket valuation
techniques. Consider the contingent valuation method, in part, because of the recent court
rulings (State of Ohio, 1989). Let us ask a version of the question put forth by Burness,
Cummings, and Ganderton (1991): "Which households place what value (or types of values) on
which nonmarket goods?" (p. 432, emphasis added). We might add: Are households accurate in
revealing their preferences and how do households form values and how should these values be
elicited?
In light of the recent court rulings potentially leading to compensable damages as well as
efforts to include externality costing for utilities, these issues are becoming increasingly
important. These issues implicitly include the aggregation issue and the scope of the market
question. To my knowledge we have never completely agreed on how to designate the market
area nor agreed on appropriate aggregation procedures. This issue becomes especially difficult
when we move from considering use values to existence values. For instance, what is the
appropriate population to aggregate over for the Grand Canyon or El Mono National
Monument? Further, one might include design concerns such as the level of specificity of the
commodity that is described and issues of embedding.
In listening to the exchanges at the recent AERE1992 sessions in New Orleans, I can
only conclude that a serious dffratp continues over the accuracy levels thyt we can tolerate in
primary nonmarket studies and how these studies should be designed Thus, we might ask: Can
we tolerate die additional accuracy concerns ***** are necessarily involved in die benefits transfer
process?
One answer is to assert that die stability of die foundation for die frcnf-fhs transfer process
dgremff on die intended use of a particular implication of die benefits transfer process. As an
illustration, consider a juxtaposition of perceived needs and purpose of a benefits transfer
made available.
povisknftifcfar/EEtfitlhttttkatt
-------
ise. Figure 1 illustrates a stylized continuum of uses representing alternative applications of
I
*
Rtqulrad Accuracy .
High
Gain* in Senming Pofcy
Knowtodgf Dcdsfcna DwmgM/Utifty
compensate damages vri real econoite
have been mstii^ real ecooomk
valoatkn.
i
Flgur* 1* A Continuum of Decision Settings from Uast Required Accuracy to I
Most Required Accuracy 1
Viewing the required level of accuracy for a benefits transfer within t conditional j
framework provides two insights. The use of the inf onnarion uiherentiy deteniuiies the . .
underlying accuracy requirements. Tmsmsight applies m both the primary data collection case I
and the benefits transfer cases. For instance, gains in knowledge might be represented by •'
benefits transfer uses such as the scope of the U.S. Geological mapping program. Screening
efforts might be represented by die CERCLAType A analysis. Policy decisions might involve
regulatory rule making, and compensable damages might involve cases associated with large-
scale natural resource damage assessments and externality costing for electric utilities. A
difference between the policy decisions and the compensable damage cases is mat, whfle in bom
cases real dollars are exchanged, we do not know precisely to whom in the poEcycase, That is. I
the policy case includes a hidden distributional issue, fr fte compentaHe case, real dollars arc
exchanged and the parties are relativdy more easUy identifiable.9 i
A continuum such as mis suggests accuracy tempered by the use of the valuation
information. For instance, in the case of gains from knowledge we might argue mat some ]
decisions, if incorrect, will not result in too high t cost to society. In the cases where large dollar
amounts are involved the response is sometimes different Often one hears that, as real dollars 1
become involved, the information (either from a primary or benefits tiansier study) is not precise
enough. That is, as me real ecooomkconiinitnient becomes more real, we sho^ i
information for decision making. The argument ttiwt we cannot ™*At*t*ir+ 1 policy i^ynopge *
without knowing the exact nature of the functional rp.litionships echoes the earlier implication of
'
-------
i
•i
me theory of second best Reaction to nihilism of the theory of second best was swift Several
researchers argued that piecemeal welfare policies coiild be pursued for those sectois satisfy
separability from die original distorted sector. Relevant to the benefits transfer issue is the work
of Yew-Kwang Ng (1977 and 1979) regarding a third best allocation. Ng demonstrates that, in
the absence of perfect information, correcting a distortion wiU always improve social welfare in
m expected value sense. Decisions based on even imperfect information, as from a benefits
transfer, are superior to no decisions.
SOME SAMPLE GENERIC GUIDELINES FOR BENEFITS TRANSFER PROTOCOLS
The original site study must be scientifically sound in the use of conceptually correct
economic methods, experimental design, and implementation procedures. The original site study
should report, maybe in an appendix, the empirical procedures, including details regarding all of
the information collected, and whether the information was useful in the empirical process.
The commodity of value should be similar between the study and policy site. Assessing
mis similarity might include quality and quantity considerations as well as the property right
structure. In addition the study and policy site markets should be similar, an assessment that
could include innumerable considerations. The overall issue becomes, what is similar enough?
To address the degree of similarity, consider a simplified benefits transfer framework.
For the study site (A) the results of a study enable one to estimate die following:
where
X.A
individual valuation regarding site A;
vector of socioeconomic characteristics
ace, cultural);
characteristics of the com
economic relevant notions [such as complement s, substitute's uniqueness]);
XmA * market conditions (size and composition).
The P's are the regression coefficients and are instrumental in the benefits transfer process.
For the policy site we estimate the following:
-------
1
where .
VB • individual valuation at site B, based on the p's of site A;
I
XCB - vector of sodo-economic characteristics at B;
Xj8 B vector of characteristics of the commodity at B; and ' |
XmB * vector of market conditions at rite B.
We are interested in die B coefficients. That is, is it acceptable to use the coefficients and
implicitly the underlying distributions from the study site to estimate the value for the policy
site? The research question is characterized m Figure 2. Given die array of information used in
valuation studies, what conditions are necessary for us to rely on the VB estimated from the VA •
equation? Thatis,forrite Aeachoftbesubelements(e.g.,forXs,asubelem |
rite B will have a corresponding distribution. The diagonal squares represent identical variables.
If this woe to occur then the benefits transfer process would not be of concern because the study I
and policy rites would be essentially identical. However, this condition is highly unlikely. The
question then bccomcst How yfrniiy are tftesf- dfotributfwy? How «ii»n«r must they be for I
different uses of die benefits transfer process f or alternative uses? •*
Can die issues raised in mis paper be answered by the existing base of nonmarket studies?
The base of studies from which a benefits transfer study can build is quite thin, at least for
contingent valuation applications. This paucity stems, in part, from the existing incentive 1
structure to publish and obtain research funds. The funding environment and the publishing
en vironmem have encouraged, if not required, studies n^ are unique. Often this uniqueness can 1
be found in the nature of the good valued. As such not enough studies address the same issue. '
The overall number of studies mat are not replications is large; thus, the number of off- I
diagonal studies is large. That is, we wffl be typically off the diagonal Unlike the more
tnun^cmalscience-oiienteddiscqdines. replicatiocmeconoinics and the publication of data are 1
not viewed as worthwhile. This attitude is not bad. However, we might need to consider other
fomis of research acceptable and piibfishabfes* cm j
discipline that contributes so heavily to the p()licy arena. Editors amf reviewers must confront *
this issue. Essentially, for the case of benefits transfer, we might want to consider what «
constitutes a substantive contribution to the literature. \
I
-------
B
B
*
Figure 2. Distribution tesut*
DESIGNING A META-ANALYSIS
Further research is necessary if we arc to toon fully understand the reliability
requirements of the benefits transfer process. Using meta-analysis can further our understanding
of the importance of various components of existing studies and help focus our research efforts.
The laboratory and field settings also offer the opportunity to explore various protocol guidelines
under varying degrees of control and realism. As such, I suggest a combined effort involving all
three settings. We have at least two ways of conducting a benefits transfer. Researchers could
simply take the value elicited at die study site and apply it directly to the policy site. For
example, a value for a change in clean air in Los Angeles may be applied directly to a similar
change in air quality in Denver. This application is clearly simplistic, and most researchers
would not wish to follow such an approach. A more technically valid strategy is to employ the
coefficients estimated with the study rite data to the variables describing the policy site. We
. need a protocol to judge sufficiently similar pairs of policy and study sites to employ benefits
transfer. To this end I offer a first hypothesis.
HI: Benefits transfer are robust to differences in site characteristics—whether X,,
Xg, or Xm or a combination of differences.
-------
If HI is not refuted then we arc able to conduct defensible benefits transfer although the »
policy and field sites may have substantially different characteristics. Researchers may conduct fK
the following tests to evaluate this hypothesis: I
• examine previous value elicitation (CVM, TCM, or HPM) studies to determine the ' .
elasticity of estimated values with respect to the independent variables. Lower I
elasticities imply that we may employ the obtained values across sites that are different *
in terms of those variables for which the elasticities are low;
• conduct laboratory investigations in which values are elicited in different institutions . J
where X., X., and Xm are varied individuaUy to determine the impact of these
difference Apt*". «*»« «fll indicate the characteristics critical to successful "1
application of benefits transfer, and J
• investigate the linearity of the valuation relationship obtained at the study site. The .
more linear this relationship the more critical are similarities in site characteristics I
between the study and policy site to successful benefits transfer.
Hypothesis 2 relates to the need to conduct and publish studies replicating previous work. J
H2: The values generated with die coefficients from Ae study rite applied to the
policy site characteristics are identical to me values that would be obtained from a
primary study at the policy site.
H3: The values from the study site are robust over time if underlying site
characteristics have not changed.
10
1
A test of diis hypothesis lequiies conducting at feast a p^ *
F-cM-.nri*iiy, we wnnlri then have original rite estimates at both the study and policy sites. The I
v«iii«y stained via benefit* transfer, vB given $A and XP. would be compared wife the primary
estimates, VB given p8 and XB. If mis hypothesis is not refuted through repeated investigations, J
the validity of benefits transfer would be supported for settings similar to those studied.
If values for a particular good obtained at a singfe site are not consistent across time, *
preferences are not stable and imply that benefits transfer is a questionable practice because it .
depends on the stability of preferences over bom time and location. This characteristic gives rise J
to a third hypothesis:
>.JP
:1
Robustness might be viewed as representing stable preferences. Whittington et al (1992) |
has addressed the effects of "time to think" and Kealy, Montgomery, and Dovido (1990) the
stability of wUlingness-to-pay values over time. Here we are interested m the shelf life of any
I
-------
given set of studies. What are die limits? Are recreation values sufficiently stable over a 10-year
period? We might consider replicating some of the earlier applications to address this question.
Repeated work with a fixed pool of subjects could also possibly give us some insights. The
time-to»think issue is relevant here bfcausfi it implies the primary estimates are themselves
subject to accuracy problems; most contingent valuation method studies do not provide much of
a thinking period between the presentation of information and the elicitation of values.
If we argue that the institutional setting is not important in individual valuations, then we
should not observe interactive effects between the vector components, X,, Xf, and Xm, of our
valuation studies. This characteristic suggests a fourth hypothesis:
H4: No interaction effects occur between Xt, Xg, and Xm. Thus differences in
some of these variables between die study and policy site do not imply that we are
unable to use the coefficients estimated for the remaining variables hi a benefits
transfer.
Out possible test of this hypothesis would involve econometrically checking for
interaction effects with the primary data from the study site. Another test would involve using
die meta-analysis technique as suggested by Smith and Kaorn (1990) and Walsh, Johnson, and
McKean (1989). A series of laboratory experiments could also be designed to investigate the
interactions of the components of the institutional setting with the values elicited from
individuals.
The more significant the interaction effects the more similar we will require settings to be
if we are to employ benefits transfer. An investigation of these (and possibly other) hypotheses
generated by a systematic investigation of benefits transfer applications will move us toward
protocols for benefits transfer.
CONCLUSION
In sum, no matter how well developed the benefits transfer process becomes, it will still
have the accuracy problems of the original studies. The accuracy needs of various types of
benefits transfer studies will vary. Overall accuracy can only be expected to deteriorate. The
current collection of original studies is not sufficient for fine tuning the benefits transfer process.
We may need to conduct additional primary studies in various settings such as die laboratory and
die field and to continue using meta analyses to improve our understanding and die accuracy of
benefits transfer.
11
-------
*
I
Atkinson, S.E..T.D. Crocker, and J-F.Shogrcn. 1992. "Bayesian Exchangeability, Benefit Q
Transfer, and Research Efficiency." Water Resources Research 28:715-722. - I
REFERENCES
Brookshire,D.B.,andRNeill. 1992. "Benefit Transfers: Conceptual and Empirical Is
Water Resources Research 28:651-655. - I
Boyle, KJ., and J.C Bergstrom. 1992. "Benefit Transfer Studies: Myths, Pragmatism, and
Idealism." Water Resources Research 28:657-663. .
Bumess, RS., R.G. Cummings, and P.T. Oandenon. 1991. "yarning Environmental Goods: A -*
Critical Appraisal of me Stale of die Alt'* In Economics and Management cf Water
Drainage in Agriculture, A. Dau»adI3.ZN)ermantcd&*t Kluwer Academy Press. *1
Qmunings,R.O.,D.S. Bn>okshire,andWJ3. Schulze(eds). 1986. "Valuing Environmental
Goods: An AtmBHimtt of the CV»»«i"g«"* Valuation Method." Totowa,NJ: Rowmanft -*
AUanheld. I
»*
Desvousges, WJL, M.C. Naughton, and GJL Parsons. 1992. "Benefits Transfer Conceptual
Problems in Estimating Water Quality Benefits Using Existing Studies." Water I
Resources Research 28:675-683. *
Kealy, MJ., M. Montgomery, and JJ. Dovido. 1990. "Reliability and Predictive Validity of )
Contingent Values: Does the Nature of me Good Matter." Journal of Environmental |
Economic Management 193A4-2&.
Loomis^U. 1992. *TTie Evolution of a More Rigorous Approach to Benefit Transfer Benefit
Function Transfer." Water Resources Research 28:701-705.
Luken, RJLtFiL Johnson, and V.Kibler. 1992. "Benefits and Costs of Pnlp and PanerEfflnent 1
Controls Under the Clean Water Act" Water Resources Research 28:665-674.
McConnell, ICE. 1992. "Model Building and Judgement: Implications for Benefits Transfers t
with Travel Cost Models." Water Resources Research 28:695-700. ]
McIntosh,M. 1991. "Doing Business in Mexico: The Evolving Legal Framework
(Environmental Considerations Regaidmg Waste Disposal)." International 1
Transboundary Resources Center. «
Mitchell, R.C., and R.T. Carson. 1989. "Using Surveys to Value Public Goods: The Contingent 1
Valuation Method." Washington, DC: Resources for the Future. J
Morgenstem, O, 1973. "On the Accuracy of Economk Observations." 2nd Ed. Princeton: ••»
Princeton University Press. I
Ng, Yew-Kwang. 1979. "Welfare Economics: Introduction and Development of Basic
Concepts." London: Macmillan. -1
Ng, Yew-Kwang. 1977. Towards a Theory of Third Best" Public Finance/Finances
PubUques 32:1-15. 1
12
I
-------
Smith, VJL, and Y. Kaora. 1990. "Signals or Noise? Explaining the Variation in
Environmental Benefits Estimates." American Journal of Agricultural Economics
72(2):419-433.
Smith, VJL 1992. "On Separating Defensible Benefits Transfers from 'Smoke and Mirrors'."
Water Resources Research 28:685-694.
State of Ohio vs. U.S. Department of the Interior. 1989. 880 E2d 432, DC Or.
U.S. Water Resources Council 1983. Economic and Environmental Principles and Guidelines
far Water and Related Land Resources Implementation Studies. Washington, DC
Walsh, R.G., D.M. Johnson, and J.R. McKean. 1992. "Benefits Transfer of Outdoor Recreation
Demand Studies (1968 - 88)." Water Resources Research 28:707-713.
Walsh, R.G., D.M. Johnson, and J.R. McKean. 1989. Issues hi Nonmarket Valuation and
Policy Application: A Retrospective Glance." Western Journal of Agricultural
Economics 14(1):178~188.
Whittington, D., V.K. Smith, A. Okorafor, A. Okore. J.L. Liu, and A. McPhaiL 1992. "Giving
Respondents Time to Think in Contingent Valuation Studies: A Developing Country
Application." Journal of Environmental Economic Management 22:205-225.
13
-------
J
J
•I
-------
BENEFITS TRANSFER: HOW GOOD IS GOOD ENOUGH?
Leland B. Deck and Launine G. Chestnut*
ABSTRACT
Transferring benefits estimates developed in one context to other contexts to analyze
related valuation questions is appealing because it can save time and resources. However,
fundamental questions regarding the accuracy of the transfer must be addressed to determine
first whether the transfer should be done at aft, second how it should be done, and third how
much confidence to place in the transferred results. The answers to these questions wUl
depend on die purpose of die analysis. Assessing die baste purpose of the analysis is a value
ceai in benefits estimate reuires tim an
of information question. Reducing
money. The benefits of reduced
ly in
requres
margin. The political/institutional context
tainty are finite and probably diminishing at the
text for the benefit analsis is an imortant fact
factor in
net benefits may be r
benefits fall within a
nefit analysis is an importa
how much accuracy is BffV^ In some cases a clear demonstration of positive
uired, but in other cases evidence supporting the- likelihood that
range may be sufficient. Judging whether the uncertainty
involved in a transfer is acceptable requires considering the decision-making context, as well
as the economic valuation questions involved. This paper raises and discusses the following
questions in ft"* context!
• Is die likely direction of potential error in the transferred results clear?
* Is a benefits transfer analysis better than no benefit analysis at aH?
• Does the regulatory or other deciipftn-m>|i"n£ context require fo^t benefits be
demonstrated to exceed costs or are other factors more central to the decision?
• What is the actual feasibility of conducting a new study?
• How much might a new study be realistically expected to reduce uncertainty?
• What are the chances of being so far wrong that a different decision would result?
The markets for benefit practitioners' analytical products are people making decisions.
Decision makers obviously prefer to obtain fafansAVk benefits information for as little
expenditure as possible. Benefits transfers offer quicker and less expensive results than
undertaking original benefit analyses, but they extract a price in terms of reduced accuracy,
validity, and acceptability of the results. The question then becomes, umpgr what circumstances
does a benefits transfer provide adequate information for decision making?
Sufficient accuracy cannot be objectively defined independent of die context Adequacy
•is not a matter of simply defining acceptable confidence intervals on the estimates and assuring
the estimates meet that standard. The institutional context motivates the need for benefit analysis
AbtAiiociaw,lnc.»dRCG/H«gkr,BiiIly,lnc.re«pccti«ly. Iteopbikxtfapreucdinthis paper are solely fee
auhon'and do not Decessarfly represent tbevfews of the U£.&^ Alko Batata.
Tayler Bingham. Bill Desvoosges. and Bob Rowe provided many helpful oonmeott •Klnggettioiii.bat die
aatbarsarenipoosibleforanyshoctoofningsarinaccmcia.
-------
and must be roiuidered^^ndeieniiiiiing a sufficient level of accuracy. "Good enough" can be I
determined only by considering the role of benefit analysis in the deciskra-making process and ^t
the tolerance for uncertainty in the benefits estimates in that setting. |
In some situations, more sophisticated analysis, even if the results ate statistically I
o*ifleientfromi»revioiislyavaUabte •
at hand bother cases more detailed study and analysis, even if it meter/ confirms previous -.
results with a greater degree of certainty, may alter the decision. A judgment mat more analysis J
won't make any difference to the decision at hand provides a clear stopping point Given the
time and money required for additional benefit analysis, the question is best staled as a value- I
added question: Win it make enough of a difference to justify the cost?
In this paper we explore several analytical and institutional issues in deciding how good I
is good enough. The following section describes a spectrum of benefit analysis choices into .
which benefits transfer fits and discusses factors to consider when determining the appropriate |
level of benefit analysis to meet the need of the decision maker. Next, we describe some of the
institutional letting* where benefit analysis is VK4 and how these vvff differ. Finally, we nmVf> I
a few comments about strategic considerations mat can play a role m the benefit analysis process,
THE BENEFIT ANALYSIS SPECTRUM W
Discussions about benefits transfer have tended to fociuoo conducting a benefits transfer 1
versus conducting an original study, as if these are the only two options. In reality, a wide range
of options for benefit analysis can be matched to each setting, depending on how much new t
information needs to be generated, how much can be borrowed, and how much detail is needed -J
in the results. A benefit analysis speuium may be defined as follows, wim detail and effort ^
increasing from first to last: J
• qualitative benefit analysis I
•transfer scoping analysis J
• foH benefits transfer
• original pilot study J
• full original study
AguofoaaVefcen^anafyriristhelowert -'
Qualitative analysis presents as much information as possible on the physical, social, and .
: impacts of the policy option, as well as information on the demand for the policy's J
J
effects, but it does not attempt to estimate the monetary benefits. The next level of the spectrum
-------
is transfer scoping analysis, which locates and examines existing relevant benefit studies for
method, results, and relationship to the decision option in question. Transfer scoping includes
analyzing the possibility of preparing a benefits transfer but stops short of adopting the existing
results to the current situation. The next level is a/u# benefits transfer, including designing an
approach for applying die information from existing studies to the current decision, obtaining
additional necessary information on the current question, preparing a quantitative benefits
estimate, and assessing the quality of information hi that pff'|Mtift The level of effort for a
benefits transfer can vary considerably: it can range from a simple threshold or bounding
analysis to detailed procedures to adjust and interpret results from previous studies and analysis
of the sensitivity of results to specific transfer assumptions. The fourth level is an original pilot
study. Pilot studies involve method and instrument development with a small-scale application.
An original pilot study can address some of the questions raised by a benefits transfer, such as
the degree to which changes in the specific scenarios affect the willingness-to-pay (WTP)
estimates, and provide preliminary new benefits estimates. A pilot study riso r?" address the
feasibility of conducting a full benefit analysis. The final level is a. fall original benefit analysis,
involving extensive data collection among a representative sample of the affected population.
This spectrum is laid out in the same order as die steps an researcher typically takes in
preparing a benefit analysis. In most cases, a full-blown original benefit analysis is not prepared,
but even when it is, some amount of transfer scoping is usually done first Researchers usually
decide that somewhere along the spectrum a study short of a full original study is adequate for
the current purposes and that additional steps are either impossible (usually because of time or
money constraints), infeasible (e.g., focus groups indicate tremendous difficulty with evaluating
the decision in a monetary context), or mat the value of potential improvements in the quality of
information from the next level of analysis is linriiH for the decision at hand.
Despite the level of effort judged adequate for a given benefit analysis, researchers will
help ensure me professional credibility of the analysis and results by including the following
steps:
• carefully reviewing and reporting the underlying studies;
• providing the underlying studies and data as part of the administrative record;
• discussing and documenting all transfer assumptions, omissions, and known biases;
• supporting assumptions with data and literature;
• characterizing uncertainty in the results;
* providing other supporting data/literature;
-------
• ensuring consistency with scientific and economic theory; and
* providing specific transfer algorithms or programs.
This kind of quality control and full reporting is required for any benefits transfer to be good
enough; bad analysis is never good enough no matter now tangential to the decision at hand.
When two parties come up with different benefit estimates, this kind of reporting of the analysis
allows third parties to sort out the sources of the differences.
MEETING THE NEED: WILL MORE ACCURATE BENEFITS INFORMATION
MAKE A DIFFERENCE?
Everyone faced with an option to expend greater effort to obtain more precise benefits
information must confront the fundamental question: Will more accurate benefits information
make a difference to me decision at hand? Although moving up the benefits analysis spectrum
can provide additional information and diminishing uncertainty, it is not necessarily me better
option. Whether it is better to move up die benefits analysis spectrum is a judgment that can be
made only in the specific context of the situation. If a more complex study would be costly and
delay the decision but could not influence the outcome because of institutional factors, it is not a
better study for that situation. The decision iptf"ff faces a constrained optimization problem
where the optimal solution is rarely the unconstrained global maximum.
In many situations a benefits transfer may provide
information for the derision
at hand and. therefore, be the preferred level of analysis even though an original study might
provide more precise benefit estimates. For example, a benefits transfer win likely provide a
range of plausible benefit estimates (maybe even a probability distribution). If die entire range of
plausible estimates fail* above or below the costs of the action muff*1 consideration, and if the
decision criterion is based on positive net benefits of any magnitnd^ then increased precision in
benefit estimates cannot change die decision.
Value of Information Considerations
Deciding how far to go along the benefits analysis specti
of information perspective. Each step along the spectra
be analyzed from a value
represents a greater level of effort that,
hopefully, will provide more information about the benefits of the program under consideration
but at a cost of a greater investment of resources, including time. The value of information
analysis says that additional information garnering (in mis case benefits analysis) should be
undertaken as long as the benefits of the additional information exceed the CM* of nhtaining fr
(Freeman, 1984). Estimating die expected costs of additional levels of effort required for mother
J
J
•i
I
J
J
3
•I
J
-------
I
f
step along die continuum is probably fairly straightforward, but estimating the expected benefits
of additional effort is probably not so straightforward.
What are the expected benefits of the additional information obtained when additional
effort is put into benefits analysis? Let's focus on the additional effort required for an original
benefits analysis relative to a benefits transfer. A value of information analysis suggests that an
Original gtwly wpnld f Eliminate (nr tmhiee) the nncerttinty in me benefits estimates. We expect
mat a benefits transfer might, at best, provide a probability distribution of benefits estimates due
to various sources of uncertainty. If sonic part of the benefits distribution falls below the costs of
the program, while me expected value of the benefits exceeds costs, then mere is some risk that a
wrong decision is being made if the program is undertaken. The reverse situation of expected
benefits falling below costs, while part of the distribution exceeds costs, might also occur. (As
noted above, if the fall benefits distribution fan* entirely above or below the estimated costs of
the program then there is little*- risk of making a wrong decision, and additional benefits analysis
is not needed.) The benefit of additional information is a function of the probability of a wrong
decision and the magnitude of the negative net benefits that win be incurred if a wrong decision
is made.
If the researcher fc«» the following information, the value of information framework can
provide a clear direction about whether an original benefits study should be undertaken:
• a probability distribution far expected benefits so that the probability of making a
wrong decision rfln be reasonably estimated,
• cm estimate of negative net benefits if a wrong decision is made,
• an estimate of the reduction in uncertainty in fh? benefits efftimatfs *fr*r could be
obtained with an original benefits study, and
• an estimate of the cost of an original benefits study.
Clearly, in many situations much of mis information win be unknown or highly uncertain.
Designing this decision framework and filling m plausible ranges for unknown elements may be
useful in judging the sensitivity to the different elements of the decision to do an original study.
Reductions in Uncertainty EKpttttd Fran in Original Study
An interesting link in the value of information framework is the question of how much an
original benefits study can be expected to reduce the uncertainty in the benefits estimates relative
to a benefits transfer. Murh nf thy ^iy^ifffjrm nf tha pros «xi COBS of benefits transfer has
presumed that an original study could be conducted and would provide reliable benefits
-------
timates.
Given mat availablp txonmnic tgjchnkpK* fffr fflmmfrg fr
goods are subject to considerabk uncertainty, and in some cases considerable controversy, this
presumption may not be appropriate in many instances. The state of the art in benefits estimation
is such that uncertainty in the benefits estimates remains high even if an original benefits study is
undertaken
iig how much an prif1"1 *t"tiy "^y ""hiee uncertainty in die benefits «stinMHgf
before the study is completed and the results thoroughly evaluated is very difficult If we knew
enough to accurately predict the effects of the specific ciicnnifTinrrtof the original study on the
uncertainty in the benefits estimates, we would probably ato know enough to predict ho w the
benefits estimates would change as wen, so a new study would not be needed. Mote likely , the
researcher begins with a set of benefits estimates for similar, but not exactly the same,
circumstances. Some evidence may exist about how certain characteristics of the site or good m
question affect the benefits estimates, at least in terms of the direction of me effect (positive or
negative), out this evidence is often not very precise. For example, consider the benefits of
protecting a recreational fishing spot Predicting that benefits are higher at a site where the
avenge size of the catch is higher may be possible, but uncertainty about how much higher the
benefits are may also exist The researcher may assume mat a site with a prettier view would
have greater benefits, but perhaps no evidence on mis is available. A new study may find mat
the view has no significant effect on what fishermen are wining to pay to protect a given fishing
spot More studies over time might actually result in greater admowk^ged uncertainty in the
lf th* »ttimfltMi effeen of certain charmeteriittiea on me benefits estimates am
not consfatgnt and if thf amCTPT fff unfwpfr*"'*1 y*"*^^ «" **fan** »rm**
The designers of an original study can count on having more information when me study is -'
completed,, but having rextoced ymx<*tmmty in the benefit* estimate* for the apecJfic question at M
hand is only one of several possible outcomes. ,'.
In most cases an original study should not be treated as supplanting all previous work but "1
as adding to the body of available information. The evaluation of the results of an original •'
benefits study should consider the results of previous similar studies. Atkinson. Crocker, and -*
Shogren (1992) conclude that given uncertainty in all the benefits estimates, the best estimate of J
benefits consider, in some appropriately weighted fashion, rstimatrm from past studies as well as
from an original study designed specifically for circumstances at hand. This conclusion, based - J
on an empirical Bayestan approach, appeals to common sense. Most available benefits
fftinmtfon frchf"
-------
benefits study in light of previous related studies to help determine how much weight to place on
the original study results for the decision at hand.
Different Sources of Uncertainty
Discussions about uncertainty in benefits transfers often focus primarily on the
uncertainty in the average benefit to the affected individual or party and on the characteristics
that determine this average benefit This focus implies that the ideal benefits transfer can be
undertaken if we have a value function that includes all relevant individual and site or good
characteristics. For example, the focus for visibility benefits transfer has been the WTP function
for the household, which might incorporate income, education, and other characteristics of the
household, and location, use patterns, and other characteristics of the site where visibility is
expected to change. As Smith (1992) notes, uncertainty about the size of the market (Le,,
number of households affected) may have a greater impact on results man uncertainty in the
average WTP per household.
Sources of Uncertainty
Benefits analysis related to a change in environmental quality typically involves a
physical science component as well as an economic component because characterizing the
environmental impact in physical terms is usually necessary before the economic value of the
change can be estimated. For example, for the Best Available Retrofit Technology (BART)
analysis of the Navajo Generating Station, the environmental impact of concern for die benefits
analysis was die change in visibility that might be expected at the Grand Canyon as a result of
reduced emissions from the power plant Therefore, estimating the predicted change in visibility
conditions at the park was necessary before the value of this change could be estimated.
Considering die level of uncertainty ft*at CTJBS in fl*^ physical- science component of the
benefits assessment may be important tor the economist when determining the appropriate level
of benefits analysis to undertake (Smith, 1992). The decision maker may have little advantage in
having fine-tuned economic estimates if the physical science component is associated with a
wide range of uncertainty.
THE MARKET FOR
! ANALYSIS: WHAT IS THE NEED?
Benefit analysis has both formal and informal roles in many decision-making pn
in the private and public sectors, including the judicial, executive, and legislative branches of the
government The roles of benefit analysis vary substantially among and within the different
-------
government branches. Our basic contention in this paper is that the appropriate level of benefit I
analysis must be determined within the context of the specific institutional setting. The potential ^fe
influence on the outcome, legal limitations on using benefit analysis, time and money J
CTHJHra*"**! amount and quality of available benefits information from previous studies, -
propensity of individuals and institutions to consider benefits information, and even strategic -J
considerations are all factors in determining what is "good enough** for each situation.
Various institutional settings ask very different questions of benefit analysis. Clearly, the J
question being asked influences the appropriate level of effort At one extreme are situations
pltimaftfly umpiring a. angle dollar amount such as efforts to incorporate environmental I
lalities in utility planning. Here a direct link can be made between the magnitude of the
benefits estimates and
-------
administrative requirements such as E.O.12291, but demonstrating that benefits exceed costs is
not required under this legislation, which creates a low threshold for sufficient accuracy.
A fifth situation is where benefit analysis is not directly tied to an immediate decision but
is part of an «"forr"fl^pp gathering or rtiffff*ll"nl'*it•»?>tain public consensus among affected parties about
the desirability of the policy. Benefit analysis can help focus public and private attention on the
reasons for undertaking costly or burdensome activities. "Selling a program" does not end when
a decision is made but must be continually pursued as long as the decision is reversible. Benefit
analyses may be useful, and very simple benefits transfers may be good enough.
Judicial Branch
Judicial proceedings are one setting where the outcome is potentially directly tied to the
magnitude of the estimated benefits. A familiar example is monetary damages in a natural
resource damage assessment under the Comprehensive Environmental Response, Compensation
and Liability Act of 1980 (CERCLA or "the Superfund l»w^. Of aU the settings for benefit
analysis, uncertainty tolerance may be lowest in litigated damage assessment cases. When a
CERCLA case is decided in court, die judge, the trustees, and the potentially responsible parties
have a keen interest in die $"3 of tfif final jnHgm«wt Mflrpp«l changes in t*K? benefit £s*»»"ate
can affect the marginal size of the damage judgment, increasing the need for reducing benefit
uncertainty. The U.S Department of the Interior published guidelines (federal Register, 1980)
on the CERCLA benefit analysis, describing acceptable approaches for benefit analysis. Benefits
transfers are permissible hi the Type A" model but only between fairly well-matched situations
because the cost of new studies is presumed to be large relative to the potential damages
associated with a relatively small pollution incident The incentive for original work is fairly
high.
-------
Another role of benefit assessment in damage assessment cases can tolerate more
ungainly and hence often relies on benefits transfer. Either party in a potential litigation must
establish a broad strategy they will follow. Each party may want to have the issues decided in
court based on a detailed presentation of the evidence. Conversely, the case could be settled out
of court When considermg whether to setue, each par^ I
likelihood of winning in court, and the likely size, of the court's decision. In doing so, each party
may consider die magnitude of available benefit estimates from either transfers or original "1
•»__ *••
Executive Branch J
The executive branch's responsibilities to prepare, implement, and enforce regulations *
have a number of very different instimtional settings that indiide benefit analysis as one J
consideration. Each potential application of benefit analysis has to own set of legal, procedural,
practical and pottticalissties that a£rect the p^ The specific |
fr»mewtuk atsts either upper or lower limits (or both) on die influence of benefits estimates even
before considering the quality of the potentially available benefit information. 1
Relatively few regulatory situations legally or procedurally allow using benefit analysis
as a central tod in the decision process. One notable exception is the Toxic Substance Control
Act (TSCA), which explicitly allows the use of benefit analysis in setting ***mig>i compound
exposure regulations. The basic TSCA objective is to prevent *taieasoiiabfe risk," and benefit J
analysis is one way of assessing reasonableness. Although no legal or procedural impediments
exist to using benefit analysis, many actual TSCA regulations have relied on health risk or cost- I
effectiveness criteria, rather than monetized benefits. However, in 1991 the federal court
overturned EPA's ban on asbestos under TSCA and found mat EPA msufficiently examined
alternatives to an outright ban. The court ruled mat although a strict quantified benefit-cost
criterion is not required, unsubstantiated statements that benefits deady exceed costs are not a ^
sufficient rationale to justify a very costly program. How EPA win respond to the court ruling in J
the future use of benefit analysis under TSCA remains to be seen.
The best known use of benefit analysis in the federal executive branch is Regulatory
Impact Analysis (RIA) documents. E.O.12291 requires benefit-cost analysis of an "major rules'* "j
(regulations or requirements with annual costs over $100 nullionnta cause a major increase in -I
prices or have significant impact on competition, employment, investment, or international
I
y
»
-------
\
I
(
competition).1 E.0.12291 charges government agencies with the role of the neoclassical
benevolent social planner in traditional economic models by directing the agencies to select, as
permitted by the law, the policy option with the least net cost to society. The guidance issued by
the Office of Management and Budget (OMB) on the benefit analysis required in an RIA
encourages selecting the highest level of benefit analysis by stating **[A]n attempt should be
made to quantify all potential real incremental benefits to society in monetary terms to the
maximum extent possible" ("Regulatory Impact Analysis Guidance," 1989, p. S68). But the
12291 guidance also recognizes the choice depends on (he situation, by saying, "The amount of
analysis (whether scientific, statistical, or economic) that a particular issue requires depends on
how crucial that issue is to determining the best alternative and on the complexity of the issue.
Regulatory analysis inevitably involves uncertainties and requires informed professional
judgments" ("Regulatory Impact Analysis Guidance," 1989, p. 561).
The E.0.12291 guidance recognizes that in some regulatory situations the law prohibits
considering the monetary benefits, or any other economic factors, in determining the best
regulatory decision. This principle is commonly embedded in many of the United States' health-
based statutes and has been upheld in federal court. Thus, the potential for a dichotomy exists:
an Executive Order requires preparing a benefit cost analysis, but the implementing legislation
prohibits considering such information in the regulatory process. Even when more accurate
benefits information could be obtained, the legal barriers to using benefit analysis often
discourage the government from committing significant resources to preparing benefits analysis
for E.O.12291. The legal status, combined with chronically short budgets and pressing time
constraints, often limits the federal government to relatively quick and low-cost forms of benefit
analysis. The legal limitations and budget constraints result in relatively greater uncertainty in
manyRIAs.
On the other hand, any benefit information included in an RIA is not ignored. The OMB
examines the benefits information presented when fulfilling their duties under E.O.12291 to
examine the economic efficiency of all proposed regulations. Federal agencies, aware of the rote
.1
1Tbe role of benefit analysis has been reaffirmed and expanded to additional Executive Orders and several policy
statements from tbe President's Council on Competitiveness. On Marco IS, 1991, men Vice President (,_,_
wrote to EPA reaffirming me Administration's position mat E.0.12291 applies to •Vl agency policy guidance
mat affects tbe public. Snefa policy guidanrg inriiirfei nr* m\y rpgnlatipnt '*••* ire p'KlifhM for notice and
comment, but also strategy statements, guidelines, policy numuals. gram aixl k» procedures. Advaoce Notice*
of Proposed Rule Mating, press releases and other doctimenm announcing or hnptem<«ti'«g regulatory policy that
affects me public." EX). 12498 directs the agencies to consider benefit analysis in setting regulatory priorities.
Further, the 1991/92 regulatory moratorium diieai federal agenrie* *yi gftipMie te likely fotts and benefits of
kgislative proposals underactive consideration by Congress or lobe proposed by me agency."
11
-------
of benefit analysis in the OMB review process, try to allocate their scarce benefit estimate
resources to issues where the benefit information is most likely to make a difference.
The need to prioritize the level of effort has led to an informal set of "acceptable" cost-
effectiveness (e.g., cost per unit risk reduction or cost per unit effluent reduction) cutoffs in some
broad categories.2 Policy options with costs clearly below the "going rate" are good candidates
for minimal benefits analysts. Options with costs in excess of the cutoff warrant additional
benefits analysis or other justification. Cost-effectiveness cutoffs are really one type of transfer,
applying the same criteria of implicit benefits from one setting to another. Cost-effectiveness
cutoffs are only a benefits transfer to the extent that benefits information is considered when
establishing the cutoff levels. However, cutoffs are typically applied as a coefficient transfer not
a benefits function transfer, which limits the ability to custom fit the cutoffs to the specifics of
each situation, thereby increasing the uncertainty.
One recent regulatory action did not have significant harriers to considering benefit
analysis as a central pan of the regulatory process. The Clean Air Act fi 169A protects visibility
at national parks and wildemess,arcas, for example. If EPA determines that a visibility
impairment exists, then EPA must determine the appropriate response. In selecting the BART
level of abatement effort, the Clean Air Act 5169A states that the decision "shall take into
consideration the costs of compliance, the energy and nonair quality environmental impacts of
compliance, any existing pollution control technology in use at the source, the remaining useful
life of the source, and the degree of improvement in visibility." Although this legislative
language does not require using economic benefit analysis, it clearly opens the door.
Although §169A was added to the CAA in 1977, EPA has required emission abatement
to protect visibility only once. In 1990 EPA proposed a determination that the Navajo
Generating Station (NGS), a large coal-fired electric generating facility in Page. Arizona, caused
significant visibility impairment at the Grand Canyon National Park. When EPA proposed the
emission reduction in February 1991 EPA said that it "was not required as a part of the BART
analysis to estimate monetary benefits associated with improving visibility in the Grand Canyon.
However, as a check of reasonableness for its approach, EPA evaluated and considered the
benefit analysis developed as a part of the RIA" (Federal Register, 1991. p. 5,182). EPA used a
benefits transfer based on an existing contingent valuation study to ettiroate monetary benefits.
The draft RIA concluded that benefits may exceed costs with a fairly wide uncertainty range.
2For instmce, in 1985 EPA ettablitheJ "policy-derived" caa-eflectiveiKu guidelines for air poDutioo New Source
PHftnnance Standards of S3,000/megajnun for panJculWemaoer and Jl^SO/megagnm for both sulfur diodde
and volatile organic compounds (Elktns and Ruuell, 1985).
12
-------
I
Prior to proposal, the NGS commissioned a pilot contingent valuation study directly concerned
with the benefits of reducing sulfur dioxide emissions from NGS. The NGS study concluded
that costs exceeded benefits. EPA invited comments during proposal on both benefit studies.
However, in the final rale, EPA stated, "[bjecause the benefits analysis forms no part of [the]
legal basis for today's action, EPA is not responding to those comments" (Federal Register,
1992, p. 50,184). The final rule requires NGS to reduce its sulfur emissions by 90 percent It
seems that each side used sufficient benefit analysis to counter the benefit-cost conclusions
presented by the other, perhaps causing the benefits analyses to be ride-stepped in the official
decision-making process.
State and local agencies are also becoming more interested in benefits analysis. For
example, the South Coast Air Quality Management District (SCAQMD) is the local air pollution
control authority in the Los Angeles area. Under the California Clean Air Act, in 1989 the
SCAQMD approved a massive plan to reduce air pollution in the South Coast. As pan of
preparing the plan, the SCAQMD asked the California State University Fullerton Foundation to
prepare an economic evaluation of the potential health benefits of improving air quality in the
South Coast The report examined the benefits from a number of health and welfare endpoints
associated with various pollutants (Hall et al., 1989). Part of the motivation for this study was to
help build public support for the pollution control measures set forth under the plan by
demonstrating that substantial benefits would accrue as a result of the control costs incurred.
Another example of state interest in benefits is the New York State Energy Research and
Development Authority (NYSERDA). New York has a policy of considering the full social
costs in electric utility planning. NYSERDA asked the Pace University Center for
Environmental Legal Studies Energy Project to prepare a study of the environmental externality
costs of electric utility operations (Ottinger et aL, 1990). The study examines the social costs of
available methods of generating electricity as well as the social costs of demand-side
management programs.
Legislative Branch
Legislative development is the third broad government arena for benefit analysis. The
U.S. Congress or state legislatures make many fundamental choices long before the specific
regulations are promulgated or damage suits litigated. Congress is increasingly interested in
benefit analysis and has recently either prepared or required several major benefit studies. Three
examples of Congress's recent interest in benefits are the inclusion of benefits assessment in the
change to the National Acid Precipitation Assessment Program (N APAP), the Office of
13
-------
\
Technology Assessment (a congressional entity) report Catching Our Breath: Next Steps far }
Reducing Urban Ozone, and a retrospective and prospective report on benefits and costs required A
by §812 of the Clean Air Act J
*
The KAPAP State of Science and Technology (SOS/T) reports include a review of the •
stale of knowledge about physical and economic benefits for environmental effects categories I
associated with acid rain (Brown et aL. 1990). The NAPAP J 990 Integrated Assessment Report
develops a "quality of information** ranking system for all information in the SOS/T, including J
monetized benefits information. In general, NAPAP used fairly rigorous criteria for assessing
the quality benefits information. The Integrated Assessment includes monetized benefit |
information on only four environmental endpoints: national agriculture, forests in the southeast,
recreational fishing in the Adirondack Mountain region, and urban visibility in the east Eight ~l
other endpoints3 of concern are qualitatively discussed but monetized benefit estimates are not •
developed because of NAPAP's assessment of inadequate information on physical damages, ~&
valuation, or bom. The benefits estimation techniques used in the Integrated Assessment include I
supply/demand analysis for commercial crops and forest products, travel-cost for recreational
fishing, and a blend of meta-analysis, expert judgment and benefits transfer for visibility. j
The congressional committees working on reauthorizing the Clean Air Act requested that
the Office of Technology Assessment prepare the analysis in Coft&irv our Breath. Ozone is the
United States' most widespread and persistent air pollution problem. In spite of considerable
effort and much progress, 45 percent of the U.S. population lived in metropolitan areas that did J
not meet the ozone air quality standards in 1988 (EPA. 1990). Because of the abatement
activities already in place, further progress on ozone wfll be increasingly expensive. OTA's I
analysis focused on two environmental endpoints of ozone. The benefits analysis used expert
judgment based on existing literature for the value of reducing health effects and supply/demand ' 1
analysis for commercial crop effects. *
Section 812 of the Clean Air Act Amendments of 1990 expands the scope of the existing J
§312 report on the cost of federal air pollution programs. The goal of the expanded Cost of .
Clean Air report is to include monetary benefit analysis in a comprehensive examination of the ~1
full social and private costs and benefits of the Clean Air Act. The first report must estimate the
costs and benefits of all air programs prior to the 1990 Amendments. After the "retrospective** "I
report is issued. EPA must periodically update the retrospective report and prepare a •*
"prospective** report, with projections of the costs and benefits of further progress in reducing air <•
^Wildlife, oilier terrestrial ecosystems, water-based ncraukm. commercial fiihiqg, other aquatic ecosystems,
building material, cultural mmn^aHi ""I human hftlih ^§tf
-------
*
t
•I
pollution. The Amendments create an Advisory Council on Clean Air Compliance Analysis to
peer review the data, methodology, and findings in the report and to make recommendations to
EPA.*
Nongovernment-Sponsored Benefit Analysis
Recognizing that benefit analysis plays a role in the public decision process, various
groups outside die government also produce benefit analyses. These efforts range from
publicizing existing work to undertaking substantive new efforts. Affected parties in legal or
regulatory proceedings have various legal and procedural opportunities to provide benefits
information. But outside groups also provide benefits information in other settings as well The
motives for providing such information likely range across the spectrum, from pro bono
provision of information to narrow strategic advocacy. Sometimes the analysis is obviously tied
to a particular action pending in Congress or an issue emerging in the national environmental or
political landscape. Two recent examples are the series of articles written by Portney and
Krupnick (Portney, 1990; Krupnick and Portney, 1991), and the American Lung Association's
latest survey of studies on the health costs of air pollution (Cannon, 1990).
STRATEGIC CONSIDERATIONS: BENEFIT ANALYSIS IN THE REAL WORLD
The acceptable level of "good enough" benefit analysis is not determined solely by the
particular legal situation. A number of strategic or tactical issues face all the interested parties
who have the option of producing benefit analysis. Economic researchers are seldom the
ultimate public policy decision makers, and economic efficiency is not necessarily the primary
concern of all parties. Benefit analysis is typically prepared at the request of, and for the
purposes of, someone else. The "client" must decide to accept a given level of benefits effort
(perhaps a level already provided by someone else), or to undertake more extensive analysis.
That decision is basically driven by the question, "Are further efforts likely to make a difference
that I will like?" The researcher can provide useful opinions about what additional efforts will
likely produce and an evaluation of the influence on uncertainty from more information, but the
decision to go forward or not rests with the client's evaluation of whether the analysis will
further their interests. Ideally, of course, the client's interests include making rational and
socially beneficial decisions based on objective information. However, this is not always the
case, and the researcher may be more vulnerable to manipulation and/or misinterpretation if
unaware of all the client's motives for requesting benefit analysis.
4TbeiiulialmembmoftheGxindlareR.aunmi^
Gates, P. Forney, R. Schmalrmice. T. Tietenberg. and K. ViscusL
15
-------
I
Benefits researchers arc doing a great disservice to themselves, their client, and the . J
profession in general if they allow strategic or tactical considerations to influence the content of a ^.
benefit analysis. Current standard practices of careful reporting of data or results taken from • I
other sources, open disclosure of new data, survey instruments and methods, detailed
descriptions of assumptions, biases and omissions, careful attention to economic theory and - |
statistical procedure, and adequate quality control procedures are important to maintain, no
matter how the client intends to use the results. Bad analysis is never good enough, despite the "i
client's interests, J
However, even if benefit analysis can be totally inoculated from deliberate strategic mis- ~|
preparation, the client may still face various strategic considerations. For instance, an argument
that a new analysis must be prepared because the level of uncertainty in the existing benefits *1
transfer is unacceptable raay be a pretext, where the real motivation is a stalling tactic. Anew *
benefit study that costs less than the present value of a delayed decision can be an economically -i
rational move by a client with adequate resources to invest I
J
Another strategic consideration that a client must carefully evaluate is the amount of 1
scrutiny that the benefit analysis will undergo. If the client could be assured of an impartial
review by knowledgeable people, the decision could focus on issues of reducing uncertainty and
providing better information, for example. However, in the real world benefit analysis is often
reviewed in an adversarial setting where the audience (e.g., decision makers, juries) may not _
have a great deal of technical expertise. A greater level of effort may be required, not because J
more information is really needed, but to protect the analysis from being discredited in the eyes
of the nonexpert audience by voluminous criticisms. . I
Setting a precedent can also be an important issue that influences the level of analysis one •
side or the other is willing to support The total return to an investment in additional research J
may be much greater than the expected return from the current situation. If one side establishes a
precedent in a small case on whether benefits can be measured, determines the appropriate way J
to estimate benefits, or a benefits function, the larger payoff may come later in a different and
larger case. |
Our final point on strategic issues is the different ability of parties to afford additional *
efforts. Consider a David and Goliath situation, where a government or a court is considering J
whether to require a solution to a particular pollution problem. The affected parties may not
have equal ability to provide additional analysts. If the side with the largest icsources perceives J
the outcome may be more favorable if they provide additional information that reduces the
j
-------
1
benefits uncertainty, they can provide the further analysis. The other side's income or wealth
limits can prevent them from exercising a similar option. This issue does not only exist in "big
corporation versus the little guy" settings: the resources of a well-funded national advocacy
group can far surpass the resources of a property owner or small business. The "decision maker"
must keep these tactical issues in mind when reviewing additional information that has been
submitted. Silence from one side may be more reflective of current wealth than of the magnitude
of the actual benefits. Newly provided analysis may reflect an much information on the
submitter's analysis of the likely outcome as it does of the real issues in the case.
FUTURE DIRECTIONS
Benefit estimation is both an art and a science, combining theory from the social
sciences, techniques from statistics, and sound judgment on the part of the practitioner. Progress
will be made as we improve our art, our techniques, and our science. One often noted weakness
in the current state of economic science in general is the relative infirequency with which results
are tested. One cornerstone of the scientific method is the replicability of results, but the
economics profession does not usually emphasize repeating analysis. The aversion to repeating
analyses is not due to malicious intent but to scarce resources, ever expanding research agendas,
and a pressing need to try to provide answers to the crucial problems confronting society.
However, it does result in greater uncertainty in our results, frequently conflicting conclusions,
and diminished acceptability of our results. This problem is endemic to most of economics but is
particularly relevant to the issue of benefits transfer. Much of the uncertainty associated with
benefits transfer comes from the limited knowledge we have about how different specifics about
the assessment situation in question will influence the estimates. As we gain a better
understanding of the effects that variations in our techniques have on benefit estimates for a
single situation and on the differences identical techniques produce when used in different
situations, we will improve our ability to use benefits transfer techniques and understand the
associated uncertainties. Consequently, our ability to meet the need of the decision maker at the
lowest possible cost will also improve.
Atkinson, S.E., T.D. Crocker, and J.F. Shogren. 1992. "Bayesian Exchangeability, Benefits
Transfer, and Research Efficiency." Water Resources Journal 23(March):715-722.
Brown etal. 1990. "State of Science and Technology. Report #27." Methods for Valuing Acidic
Deposition and Air Pollution Effects. National Acid Precipitation Assessment Program.
Cannon, J.S. 1990. The Health Costs of Air Pollution, Third Edition: 1984-1989. American
Lung Association.
17
-------
EUdns, C.L., and M. Russell. September 11,1985. "Guidelines for Cost-Effectiveness of New
Source Perfonnance Standards." Memorandum to AJ. Barnes.
Federal Register. 1980. "Final Rule for Natural Resource Damage Assessments under the
Comprehensive Environmental Response, Compensation and Liability Act of 1980
(CERCLA)." 51(148)37,647-27,753.
Federal Register. Februarys. 1991. 56(27), page 5,182.
Federal Register. October 3.1992. 56X192). page 50,184. .
Fisher, A.C., J.V. Krutilla, and CJ. Cicchettl 1972. The Economics of Environmental
Preservation: A Theoretical and Empirical Analysis." American Economic Review.
Freeman, A.M. m. 1984. "On the Tactics of Benefit Estimation Under Executive Order 12291." J
In Environmental Policy under Reagan's Executive Order: The Rote of Benefit-Cost
Analysis, V.K. Smith, ed., pp. 167-186. Chapel Hill: University of North Carolina Press. ~i
Hall, J., et aL June 1989. Economic Assessment of At Health Benefits from Improvement in Air
Quality in me South Coast Air Basin. Final Report to the South Coast Air Quality ^
Management District • ]
Knipnick, A., and P. Portney. 1991. "Controlling Urban Air Pollution: A Benefit-Cost
Assessment" Science. 1
Ottinger,RX.,etal. 1990. Environmental Externality Costs From Electric Utility Operations.
Draft Final Report for NYSERDA. ,
Portney, P. 1990. "Policy Watch: Economics and the Clean Air Act** Journal of Economic V '
Perspectives.
"Regulatory Impact Analysis Guidance." 1989. Regulatory Program of the United States J
Government: April 1.1938 - March 31,19S9. Appendixr
Smith, V.K, 1992. "On Separating Defensible Benefits Transfers From 'Smoke and Mirrors'." I
Water Resources Journal 28(March):685-694.
U.S. Environmental Protection Agency. 1990. National Air Quality and Emissions Trends 1
Report, 1988 (EPA-450/4-90-002). J
J
J
18
-------
WHAT IS CONSUMER'S SURPLUS FOR A DAY OF USE? AND
WHAT DOES IT TELL US ABOUT CONSUMER'S SURPLUS?
Edward R. Morey*
ABSTRACT
Compensating variation Jbr a day of use is a well-defined concept for a change in die
price of a recreational site but is not, in general, a well-defined concept for a change in die
characteristics of a site. Sufficient conditions f or when it is well-defined f or characteristics
changes are identified. These sufficient conditions are assumed in most discrete-choke
models of recreational participation and site choke. When well-defined, compensating
variation for a day of use multiplied by the number of days in die original state (proposed
state) is a Laspeyres index (Paasche index) that bounds die compensating variation (CV)
from below (above). The first approximation is a linear approximation to die CV, and die
second approximation is a linear approximation to die equivalent variation. The average of
these two approximations is an almost second-order approximation to die CV and is akin to
die Harberger triangle. Simulation results indicate die bias in *^v>- B***^ approximations can
be small or large, and die bias in the average of diese two linear approximations while often
1'te small can be large if die proposed changed will result in a large percentage change in
predicted number of days.
-I
Consumer's surplus for a day of use is a common way to express the benefits a
representative individual derives from a recreational site. The U.S. Forest Service uses
consumer's surplus for a day of use as die bask measure of a site's recreational value. Walsh,
Johnson, and McKean (1991) surveyed twenty yean of empirical research on die recreational
value of our national forests. They note, The standard unit of measurement is an activity day,
defined as one person on-rite for any part of a calendar da/* (p. 176). Derivation of day of we
measures is common in both the travel-cost and contingent valuation literature and is particularly
common in the discrete-choice variants of these methodologies. A ***" fMtnp|?ff iff Bwkytapl.
Hancmann. and Strand (1984); Carson, Hanmiann, and Wegge (1987); Cameron (1988); and
Cameron and James (1987).
•
Why the attraction to consumer's surplus for a day of use when the desired welfare
measure for policy analysis is not consumer's surplus for a unit consumed but «*«**•*!
consumer's surplus? For a given time period such as a year, tiw policy maker wants to know
how each individual values a change in prices or site characteristics rather than his or her value
•Univi^ty of Colorado, Department of Ecooomka. I want ID tfamkTiykrBtagban, Robert Rowe,V. Kerry
Smith. and the many participants of (Us conference go benefit Busier who fcvced me 10 vigorously defend me
arguments in this peper. Aoy remaning errors venafonniiaely my own.
-------
per day of use for that change.1 However, policy makers and economists alike are attracted to "
for a day of use measures for a number of reasons, one being consumer's surplus day of use Wi
lends itself to use in benefit transfers. The notion is that once a representative individual's |
consumer's surplus for a day of use has been estimated for X-ing at one site where "X" is a
recreational activity such as fishing or hiking, the analyst can obtain that individual's consumer's ~ j
surplus for the site or any similar site by multiplying consumer's surplus for a day of use at the
first site by the number of days spent X-ing at the site to be valued. 1
This paper examines the concept of consumer's surplus for a unit of use and identifies its ..
relationship to consumer's surplus per unit of time. Does consumer's surplus for a unit of use J
stand alone as a well-defined concept, and if so, should it be the standard-bearer for transferring
benefit measures from one site to another? I
1
We begin our examination of these issues with a thought experiment Consider the
maximum you would pay to have the price you pay for the next Coke you drink reduced by
$0.50. Your answer is $0.50. Further note that this is how much you would pay each and every
time you purchase a Coke to have the price of that Coke reduced by $0.50. Fifty cents is your J
consumer's surplus for a unit of use for having the price of Coke reduced by $0.50 (Le., it's your
per-Coke consumer's surplus for the price reduction).
Consider now a similar thought experiment for a reduction in the cost of a day at a .
recreational site.2 For simplicity, assume a world of three commodities: two types of activities, J
days at a recreational site and days at home, and a numeraire good mat can be consumed
anywhere. What is the maximum amount an individual would pay each time he or she spent a J
day at the site to have the cost of that day reduced from Pj to P ] where Pt is the cost for the day?
The answer is (P^ - P J), which is the individual's day of use compensating variation (CV) for the 1
price change, denoted CVDU.3
CVDU is represented graphically in Figure 1 as the vertical distance ab, whereas the
individual's CV associated with the change is the area P° acP,1. Obviously, CVDU * CV. The
issue is, therefore, how CV can be derived, or approximated, from the CVDU.
1 Consumer's siaptosfc defined here as eiUief the compens^
the change. It is deTuied for a specific tilne period soch as a year or season. 1
2Later. I consider toe more oomplicaied issue of itaeonsuinerU surplus ami the cxmsumerUsan>lus for a tmit of use *
for a change in the characteristics of a recreational site.
3Noe thai CV for a day of u«. CVDU. is not the same as the CV for a day.
-------
Hfcksian Demand Function evaluated at the utility level u °
ManhaKan
Demand Function
T • Number of days at the site
Figure 1. Per Day of Use Compensating Variation
Consider multiplying CVDU by the number of days at the site.4 The figure obtained depends on
whether CVDU is multiplied by the number of days at the site in the original state, the number of
days in the proposed state, or some average of the two. Define CV° »(T° x CVDU), where T° is
the number of days when Pt = Pj. Graphically, CV° is the area PjabPj. Define CVJ •
-------
Figure 1 suggests, CV^ is a Laspeyres index that bounds the CV from below, and CV J is a
Paasche index that bounds the CV from above.
Theorem 1:
CV?«[T°(P°-P})]SCV (i) I
and J
CV;.[Tl(p?-PJ)]iCV. (2) I
Proof mat [l? (p°- PJ)] * CV ,
Define me indirect utility function for the season as V « V(Y, PI,, PO where Y is income,
Pbisthecostofeachdayathome.Vo.VCY.PjP^.andVl.VCY.PjpJ). Dual to this indirect |
utility function is the expenditure function E«E(V,Pb, PI). Define T as me number of days at
the site, H as the number of days at home, and let N denote the quantity of the numeraire i
consumed (Le,,N»Y-PiT-PjiH). .1
By definition, the CV for a change from p.P to
s
By definition of the expenditure function
, Pj. P " ) « PfP * P JH° * NO (4) '
and
E(vl,Pj,Pj)=P|Tl + PjH» + Nl (5) J
Substitute Eq. (4) into Eq. (3) to obtain 1
(6) ,
)
I
-------
Now note that
because T°. H°, and N° are by definition capable of producing V°. Therefore, P}T° + PjH<» * N«
are sufficient expenditures to produce V° given pj and Pj. However, E(V0, PJj, P|) is by
definition the minimum expenditures required to produce V° given PJ and Pj.
GivenEq.(6)andEq.(7),
(8)
SCV.
The proof that [H x (P°-p} )] ^CV is analogous to the proof that [T»x(Pj-pJ )]
Farther note that from Eq. (6), Eq. (7), and the definition of CV^, it follows that
As Eq. (9) indicates, the bias in CV^ is how much the expenditures to produce V° would decline
at the proposed prices if the individual is allowed to adjust his allocation from {TO, H°, N°) to
CV° also a linear approximation to the CV for a change in Pt, that is
CV = E(VI, Pj, Pj) - E(VO, Pj. P{)
(10)
by Taylor's Theorem
»l-pj) sinceE(VI.PJ.P|)=E(V°,PJX)*Y
by Shepard* s lemma
By an analogous argument, CV a linear approximation to the equivalent variation.
-------
I
Note that CV° *s tf"B io * linear ^ppmrimaticm to the CV that it essen*"lly diif- to HicVs
(1942 and 1946). The Hicksian approximation to the CV, which is in terms of quantity changes
rather than price changes, is
*
CV-Pl(Xl~X°)(D5eweiU987) (11) |
where X»(H,T,N) and P»OV PI, I).5 ThetdbreC^imgta DC labeled fficks^ i
approximation to the CV. •
Summarizing to here, [T°(Pj - P{)] and |Tl(Pj- P})] are respectively lower and upper 1
boimds on the (^, and [T^-PJ)] is, madditiori, a Imear approximation to m^ These
results make consumer's surplus for a day of use. (P0-?1). useful . 1
Unfortunately, neidierCV^ or Cvj win always closely approximate the CV. Put simply,
the actual degree of bias in these linear approximations.depends on the individual's preferences |
and the magnitude of me price change. The bias can be small or large. For example, in Figure 1
the bias is significant visually. Intuitively, the bias inCV^andCvJ results because neither 1
measure considers the substitutability between days at home and days at the site. The degree of
bias in each of these measures is an increasing function of the marginal rate of substitution
between days at home and days at me site and of the magnitude of the price change; the greater
the change inT that will result from the proposed price change, the greater me bias. .
In contrast to these linear approximations, the average of CV^ andCVJiso&muta
second^rd^ar^noxiniationtomeCVforachangeinPt. Denote this average CV^* I
L(TUT°) (12)
t always better approximate die CV man either CVj or CV|. J
In contrast to CV^, an exaasecondKirderaprroximation tome CV for a changed 1
(13) .
Pl (X1 - X°)« P[ (T1-T0)+ij (H1 - H8)+(H1 - N°)
-------
I
by Taylor's Theorem
because
by Shepard's lemma
Comparing Eq. (12) and (13), the difference between CV^* and an exact second-order
approximation to the CV is the difference between (T1 - T°) and the change in Pt, (Pt -P^),
multiplied by the slope of the Hicksian demand function for T evaluated at the initial utility level
and prices,
Note that a different almost second-order approximation to the CV is the well-known
Harberger triangle,* (Po(X» -X°) +(Pl - F>) (Xl - X<>)). The (ti
nce between CVj*6 and
the Harberger triangle is CV*1* is an almost second-order approximation to the CV in terms of
the price change, and the Harberger triangle is an almost second-order approximation to the CV
in terms of die quantity changes. In this sense CV**6 might be labeled the price-change
equivalent to the Harberger triangle
Summarizing the last few paragraphs, CV for a day of use can be used to obtain an
almost second-order approximation to the CV by multiplying CV for a day of use by average
Wdtznun(1988).
af the Harhgflg irimgk tet Hifffaereef (1971). DiewCTt (1976 and 1987). md
-------
number of days at the site in the initial and proposed states. In general, this approximation is
better than the approximation obtained by multiplying CV for a day of use by the number of days
at the site in one of the states.
To get a feel for how large biases in CVj, Cvj, and CV^** can be, I ran 100 simulations.
Simulations tenus nothing about how small or large the bias will be in any particular real world
example. They are by definition assumption-specific; a particular preference ordering is
assumed, and then the bias is determined for different price changes for mat preference ordering. '
The simulations reported here are based on a simple repeated discrete-choice random-utility
model that explains the probability of visiting the site on any given day. No claim is made that
mis discrete-choice model reflects tram. The largest bias I generated is a case where the price
reduction causes the probability of visiting the site each day to increase from effectively zero to 2
percent For this case, CV = $18.25, CV? = $0.0025, and CVJ « $203.40.
For comparison, a case where the price reduction causes the probability to increase from
4 to 9 percent resulted in a CV of $58.95. aCVj of $35.34, and a Cv} of $90.56, and a case
where a price increase causes the probability to decrease from 4 to 1 percent results in a CV of
-$2239, a CV? of -$35.34, and aCV} of -$1330. For the 100 simulations, neither CV? nor CV}
closely approximate the CV unless the price change caused the probability to change by less thwi
10 percent, and men CV.CV^. and CvJ are all effectively zero. For example, a price decrease
mat caused the probability to increase from 2 to 2.04 percent (a 2 percent change) resulted in a
CV of $0.3233, a CVj of $03205, and a CVJ of $03262, but a price decrease that caused the
probability to decrease from 34 to 26 percent (a 30 percent change) resulted in a CV of -$10130,
a CV? of -$113.60, and a CVJ of -$89JO. These simulation results are just an example, but they
Ao indicate the potential for a Hat and «ne tint tnrmMM pf rtw» fignifi^ncy pf thf prfcy rhangf
1
•I
much more closely approximates me CV. For the eight simulation results noted
above, the CV's and their corresponding CV^**s are ($18.25 and $101.70), {$58.95 and
$62.95), {-$2239 and -$2432), {$03233 and $03234), and {-$10130 and .$10130). Except
f or the first set, CV and CV^* are all similar. In the first case, a five-fold difference exists
between the CV and the CV^*6. Again, these simulation results should not be taken too
seriously, but they do suggest that the CV^* closely approximates the CV except in cases in
which the price change will cause a great change in the number of days at the site. However,
bias is significant because any policy mat increases demand from effectively zero to a small
number of days win involve a large multiplicative change in total demand.
8
-------
should the researcher spend much time guilding the valuation lily, knowing that the final benefit
estimate is only as good as its weakest component? When an of the original valuation studies
have significant problems, either in their own right or for benefit transfer, does the researcher
press ahead or refuse to play? While refusal to come up with an estimate may not be an option
for a benefit analysis on a single pathway (assuming the decision to begin the study embodies
some judgment mat some type of estimate wfll result), it is a real option for social costing, where
many pathways will clearly be left blank. Tberefoce,addmg one more to the list is unlikely to
raise serious objections.
Protocols are perhaps most needed to guide the use of multiple studies on a given effect.
each study with significant flaws, to establish a range of uncertainty. Existing practices vary
.widely. Take the use of symptom-day values in a benefit transfer. Three contingent valuation
studies provide such values, each with significant problems, each giving values mat are in a
range of a priori plausibility. But because the values themselves are small ($2 to 2C/day), small
absolute differences between them can translate into large percentage differences and significant
dependence of the benefit estimates on the- values ghofen Some researchers average the
midpoint values and obtain a range by averaging 95 percent values. Others use only midpoint
values from the mice studies to represent low, mid, and high estimate of unit values. Others
give up and use judgment Others go with one study judged to be the "best"
Although die above areas could benefit from analysis and codification, one particular area
suggested for codification may not yield many benefits: establishing detailed criteria for
evaluating original studies. Beyond stating the obvious—mat studies are "good" if they are
based on acceptable theory, the theory links to wen done empirics, and essential results are
reported—what more can we do to evaluate studies? The weighting of these criteria is the
crucial element; yet weights depend on the use to which the studies wffl be put, the policy
setting, and the skills of the researcher in getting around problems or supplementing a smdy with
other data, for example. A premium should be placed on flexibility for the researcher to include
studies felt to be most appropriate for me problem at hand; the major responsibility in return for
this freedom being to document choices.
The NUSAP system (based on work by Funtowicz and Ravetz) being used for the DOE
Fuel Cycle study may be a useful tool for documenting choices of studies «mit in particular, die
uncertainties felt by the researcher in making benefit transfers. NUSAP is an acronym for the
evaluative categories in mis quality and uncertainty message system (Numerical entry. Units,
Spread of values, Assessment of values, and Pjedigree). A separate set of entries would be used
to document choices about emissions, concentrations, impacts, and monetization. Each of these
-------
dements contains subelements, ratings are given for some of the subelements, and the researcher
is encouraged to provide comments explaining the ratings and any other information provided by
the entries (see Table 1). The system as we use it does not involve weighting the various entries
to come up with a score associated with each choice. Rather, it is used to qualify the choice for
thereat or ultimate user of the benefit transfer analysis. This tool would work equally well for
documenting the quality and uncertainties of a single original study as for documenting choices
in the benefit transfer exercise.
RESEARCH AGENDA
To meet the demand for reliable frfnffi! analyses based on secondary sources, major
research efforts are needed. The research agenda spans the following options:
• Develop methods to make better use of existing studies in the benefit transfer process.
• Improve the quality of original studies so mat the results of secondary studies will be
more credible.
• Routinely include in the original study design elements to aid in benefit transfers.
i original research with the sole purpose of obtaining results to be used in benefit
rs.
• Develop incentives for researchers to engage in research supporting benefit transfers.
Making Better Use of Origin*) State ,'i
To use original valuation studies, researchers must know about them. Many literature
reviews of the benefits of environmental improvements exist, but focus varies and is generally
limited to one category or subcategory. Major efforts are beginning to develop bibliographies
covering the benefits analysis literature. The Environmental Protection Agency's (EPA's)
bibliography is available on diskette, but it is stffl by no means comprehensive. Bibliographies
Oat cut across an benefit categories are being developed in the above cited efforts associated
with estimating the social costs of electricity. Efforts to ttanrtantiTr these databases and perhaps
merge mem are needed. In addition protocols for mdicalmg where reports and other unpublished
materials can be obtained are sorely needed. Once the studies are obtained, protocols for their
use in a benefit transfer are Ftr-ded but currc^y do not **"f*, as noted above.
Original studies can also be more efficiently used to the extent mat their results can be
combined into either a meta-analysis or, if the original data can be obtained, into new analyses on
the combined samples. Such analyses could, in meory, estimate values or functions that
eliminate (or at least reduce) the need for ad toe consideration of multiple studies for
J
•I
10
-------
I
From such a random-utility model we can derive an expected CV per day for any change in the
price or characteristics of the site. This constant CV per day can, for example, be multiplied by
the number of days in me year to get the CV for the year. From this discrete-choice model we
can also derive a CV for a day of use. Note CV for a day of use is not the same thing as CV per
day. CV for a day of use is what CV per day would be if the individual were constrained to
spend the day at the site. The individual is not constrained in mis way.
Theorem 1 and die approximation results imply the following for mis simple discrete-
choice model of recreational demand: CV for a day of use multiplied by the number of days
each year to the site in the original state is both a lower bound and a linear approximation to the
yearly CV associated with the change; CV for a day of use multiplied by the number of days
each year to the site in the proposed state is both an upper bound on the yearly CV and a linear
approximation to the yearly equivalent variation, EV, associated with the change; and CV for a
day of use multiplied by the average of the number of days at the site in the two states is almost a
second-order approximation to die CV associated with the change. The simulation results
discussed earlier were all derived from a discrete-choice rando"! utility model. Therefore, even
though the original discussion of simulation results described the CVs as those for price changes,
they could for this model also be described as CVs resulting from changes in the characteristics
of the site. This assertion is true became any change in the characteristics of the site HE* a
quality-equivalent price change if we assume that the utility the individual receives on a day is
only a function of whether he or she spends that day at a site, the amount of the numeraire
consumed that day, and the characteristics of the site.
CONCLUSIONS
Care is required when using consumer's surplus for a day of use. Consumer's surplus for
a day of use exists for any change in die price of a day at a recreational site and is equal to the
price change. However, if the change mvnlvef ^ change fa die f hiirartrristics of the rite, a
constant CV per day of use does not, in general, exist In addition, even when a constant
consumer's surplus for a day of use does exist, multiplying it by the number of days at the site in
the original state provides only a lower bound on the consumer's surplus, and multiplying it by
the number of days at the site in the proposed state provides only an upper bound on the
consumer's surplus. SmiuUtions show u^ DUU m Aese approximations (»n be small or large.
The average of the two bounds often closely approximates the consumer's surplus, but even this
average can be significantly biased for numerous proposed policies.
11
-------
Bockstael. Nancy E^ Michael W. Hanemann, and Ivar E. Strand Jr. 1984. "Measuring the
Benefits of Water Quality Improvements Using Recreation Demand Models: VoLD."
Prepared for the Office of Policy Analysis, U.S. Environmental Protection Agency,
Washington, DC
I
Referendum Data: Maximum Likelihood Banmimon py \~cnawou *.
Journal of Environmental Economics and Management 15:355-379.
^CWIMM vj •— ,-.. — _
Cameron, Trady A., and Michelle D.James. 1987. "Efficient Estimation Methods for Use With I
"Closed-Ended" Contingem Valuation Survey Date." The Review of Economics and
Statistics 69:269-276. I
Quson, Richard T.,MichadW.Haiieniaim,aiidTliomasWegge. 1987. "Southcentral Alaska
Sport Fishmg Economic Study." Pnqpared for Alaskaltepariinem of Fish and Game, -,
Anchorage, AK. Sacramento, CA: Jones and Stokes Associates. I
DieweruW.Erwin. 1976. "Harberger'i Welfare Indicator and Revealed Preference Theory ."
American Economic Review 66(1):143-152, I .
Diewert, W. Erwin. 1987. "Cost Functions." In The New Pelgrave Dictionary of Economics,
Volume lt John Eatwell, Murray Milgate, and Peter Newman, eds., New York: Stockton •
Press Limited. I
Harberger, Arnold C 1971. Three Basic Postulates of Applied Welfare Economics: An
Interpretive Essay.** Journal of Economic Literature 9(3):785-797. L
Hicks, JJL 1942. "Consumer's Sarpms and Index Numbers.** Review of Economic Studies
9:126-137. \
Hicks. JJL 1946. Value and Capital 2nd Ed. Oxford: daredon Press.
Walsh, Richard B.,Donn Johnson, and John McKean. 1990. "Review of Outdoor Recreation I
Economic Demand Studies Wim Non*Market Benefit F*^1^^*^ 1968-1988.** In
Advances in Allied Microeconomics: VolS: Recent Developments in i Modeling of
Technical Change; and Modeling Demand/or and Valuation of Recreation Resources. \
AJ^. Link and V.!LSniim, eds., Greenwich. CT: JAI Press. I
Weitzman, Martin L. 1988. "Qmsomer's Surphts as an Exact Approximation When Prkes are \
Appropriately Deflated.** The Quarterly Journal of Economics 103(3):543-SS3. J
1
-------
ISSUES IN BENEFITS TRANSFER
Trudy Ann Cameron*
ABSTRACT
These comments cover four separate issues in benefits transfer. The first is an idea for
using weighted maximum likelihood estimation to recalibrate study sample models to reflect
policy population relative frequencies of different sociodemographic groups and
environmental attributes. These recalibrated models are then transferred to the study context
The second issue highlights the substantial value for benefits transfer of an estimation
methodology proposed in the international development literature by Edward Learner. The
third issue is a description of a recent survey and evaluation prepared for the National
Research Council concerning the "combination of information" (CI) in a wide array of
different disciplines. This report very closely parallels the insights drawn by many of the
participants in the 1992 AERE workshop. Finally I make a recommendation concerning
competitive funding for the incremental effort necessary for documenting and preparing data
associated with primary studies that have substantial promise for benefits transfer
applications.
Environmental benefits assessments are now mandated for many benefit-cost analyses of
public projects, and these assessments also form an essential component of much environmental
litigation. Original studies, unique to the particular valuation problem in question, are typically
very expensive and highly time-consuming because household surveys must usually be
conducted to gamer the appropriate data. As a consequence, researchers are pressured to look for
"good enough numbers" provided by some existing, sufficiently similar assessment
The demand for benefits estimates that can be selected "off the shelf' from an inventory
of estimates is overwhelming. For example, if an oil spill lolls 200 sea birds, researchers would
find simply averaging the dollar values attached to dead sea birds in half a dozen existing studies
convenient to estimate a satisfactory dollar value of each of these particular birds, in this
particular area.
Of course, the advisability of this strategy of borrowing estimates for the new valuation
problem will depend on the similarity of the two contexts. In a few cases, finding a similar study
may be relatively easy. In other cases, arguing that the values from the "study" case are
transferable to the "policy" case may be less valid. In still other cases, no existing values may be
available for any similar scenario (i.e., species, type of damage or enhancement, or locale).
Given that benefits transfer is widely practiced, assessing suitable protocols for making such
transfers is important
'University of California, Department of Economics.
-------
Benefits transfer practices were the subject of a recent special section of the journal
Water Resources Research. This collection of papers maps out many important issues in this
area. It also showcases work on the overall practice of benefits transfer, rather than specific
examples.1
This paper addresses four distinct issues relevant to benefits transfer. I describe an idea
for using weighted maximum likelihood estimation to recalibrate study sample models to reflect
policy population relative frequencies of different sociodemographic groups and environmental
attributes.2 These recalibrated models are then transferred to the study context I review and
highlight the substantial value for benefits transfer of an estimation methodology proposed in the
international development literature by Edward Learner. I then describe a recent survey and
evaluation prepared for the National Research Council concerning the "combination of
information*' (CI) in a wide array of different disciplines. This report very closely parallels the
insights drawn by many of the participants in the 1992 AERE workshop. Finally, I advocate
competitive funding for the incremental effort necessary for documenting and preparing data
associated with primary studies having substantial promise for benefits transfer applications.
REWEIGHTTNG STUDY SAMPLE TO REFLECT POLICY POPULATION
Li ordinary least squares estimation (OLS), a sample that is nonrepresentative only in
terms of the distribution of an exogenous variable presents no problem for estimation. In
contrast, if the sample is nonrepresentative in terms of an endogenous variable, potential exists
for sampling bias in the estimation results. In general, in any estimation algorithm, if an
observation's presence or absence in the estimating sample is in any way related to the
magnitude of the outcome researchers are trying to explain, potential exists for bias in the
estimates.
The case study in which I participated emphasized random utility modeling (RUM) of
recreational site choices. These models are estimated by maximum likelihood (ML) methods. A
long tradition in models like this is employing weighted exogenous sample maximum likelihood
(WESML) estimation when the estimating sample is not representative of the desired study
population, but the approximate distribution of respondent attributes in the study population is
known.
r
lT1iMe papen include Atkinson. Crocker and Sbogreo (1992). BoytewdBergttrom (1992), Brookshire and Ndll
(1992), Desvousges. Naughton, tad Parsons (1992), Loomis (1992), Luktn, Johnson, and Kibter (1992),
McCooneU (1992), Smith (1992), and Walsh, Johnson, and MeKen (1992).
211us tenninolagy—"study" versus "policy** samples Mvl/brpO|Milatioo&—was atkipted during the Wciriubop
wffl be adhered to throughout this paper.
-------
I
GROUNDWATER VALUATION: DOUGHERTY COUNTY, GEORGIA
John C. Bergstrom and Kevin J. Boyle*
ABSTRACT
The benefit transfer problem addressed here involves using existing valuation data to
transfer estimates of groundwater quality benefits to Dougherty County, Georgia.
Groundwater provides the sole source of almost all drinking water supplies in the county. In
addition, the availability of abundant groundwater supplies, combined with good sandy soil
and a ff»lrt climate, make this coy"ty • major agricultural production region. In the
Dougherty County region, a high potential exists for chemical fertilizers and pesticides used
in agricultural production to leach through the soil and contaminate groundwater supplies.
We evaluated groundwater valuation estimates from several previous studies as
potential candidates for transfer to Dougherty County. Because of a number of limitations,
the valuation estimates reported in previous studies provide, at best, "ball park" estimates of
groundwater protection benefits in Dougherty County and therefore are suitable for only a
"scoping" type analysis. The "transferability of existing valuation estimates to Dougherty
County might be improved by reestimating valuation models from existing data, obtaining
additional secondary data from each existing study site, and conducting a small and
inexpensive survey at the policy site (Dougherty County) to collect primary data on a limited
ly and demand uncertainty). Benefit
groundwater protection.
numer of key valuation
variables (e.g., subjective
transfer holds promise as a potential alternative for
However, much more research is needed to establish acceptable protocols for transferring
benefit estimates from one site to another.
In many regions of the U.S. groundwater provides the major source of water for
municipal, industrial, and agricultural activities. The continued use of groundwater to support
these economic activities can be threatened by the activities themselves. For example, toxic
chemicals from municipal and industrial waste dumps may leach through the soil and
contaminate groundwater supplies. Chemical fertilizers and pesticides applied on agricultural
land may also result in toxic chemicals leaching through the soil and contaminating groundwater
supplies. One question of general interest is, "What are the benefits to the general public in a
specific area of 'safe* groundwater quality (where safe implies that chemical concentrations in
*UnivenityorGeaftfa|gBdUiiiveni^orKUoe,ieipeciivdy. We would like tottamk Steven Edwvds and Bnoe
Undiayfw proves irfcrautka from the* Member* of the cue
•tody group included David Brooksfaire (Unlventty of New Me*ifflXL»C«iic»( Arson* NMiooai
Labontny), Steve Cratchfield (USDA, Economic Research Service), Martin David (Univenity of Wisconsin),
Ricfcrd Duboorg (Univenity College London). Stephen Fnber (Univmity of PHuborgb). Job Hoehn
(MirMpm Sfflfl.
Research Service). Marc Ribaodo (USDA. Ecoi«i^Rc«an±Senrice).indR£xtoeyWekher(NOAA). AB
encn. omitsfani, sad opinions ire solely attributable to me tumon.
-------
the water are within EPA health advisory levels)? Benefit transfer provides a potential means of
addressing this question.
This paper proposes a protocol for transferring existing ground water quality benefits
using a case study approach. We present background information on the valuation problem for
the case study "policy site" and discuss individual and aggregate values. We present a proposed
benefit transfer protocol for die case study and assess die applicability of odsting ground water
valuation data at "study rites." Finally we conduct a validity check of the proposed protocol and
discuss implications for future benefit transfer research.
VALUATION PROBLEM BACKGROUND
Dougherty County, located in southwest Georgia on the southern Atlantic Coastal Plain,
is underlain by a deep succession of sand, clay, and carbonate rocks that form a large aquifer
system (Rouhani and Hall, 1988). Oroundwater provides the source of almost all drinking water
supplies in Dougherty Comity, which includes the City of Albany (Pierce, Barbar, and Stiles,
1982). The geographic and physical features of Dougherty County are illustrated in die maps
provided in Appendix A.
The availability of abundant gtoundwater supplies, combined with good sandy soil and a
mild climate, makes agriculture the largest industry in the county. Major agricultural products in
the county include peanuts, soybeans, wheat, and com. This crop production in the county
involves heavy use of chemical fertilizers and pesticides. Some of these chemicals may be
persistent and eventually leach through the soil and contaminate groundwater supplies. Because
of the way groundwater moves underground, surface contamination in one area can spread to
groundwater supplies many miles away (Cohen, Creeger, and Enfield, 1984; Kundell, 1980; Sun,
1990).
Contamination of groundwater by agricultural chemicals was first discovered in the late
1970s. By 1986, EPA groundwater testing studies had detected 19 pesticides in groundwater
supplies in 24 states where the source of contamination was most likely agricultural application
(U.S. EPA, 1987). Farms in Georgia and across the U.S. commonly apply large amounts of
nitrogen fertilizer to crops. Nitrogen in fertilizer, after it leaches through the soil, may show up
as nitrate in groundwater supplies. In 1986, an EPA study found that 2.7 percent of rural wells in
the U.S. had nitrate concentrations preceding the EPA health advisory level of 10 ppm (parts per
million). Nitrate has been found in groundwater samples tested in Georgia, Florida, North
Carolina, South Carolina, and Virginia (Hayes, Maslia, and Meeks, 1983; McConnel et aL,
1984; Williams etaL, 1988).
1
1
J
J
3
-------
1
The empirical evidence from groundwater testing studies which is faMy sparse suggests
jtijft concentrations of agricultural chffnic^l contaminants (pesticides and nitrates) in the
Dougherty County area are within EPA standards for safe drinking water (Georgia DNR, 1989;
Nielson and Lee, 1987; Sun, 1990; Williams ct at, 1988). Nielson and Lee (1987), however,
identify the Dougherty County area as a region with potential for groundwater contamination by
agricultural chemicals. Because groundwater is the major source of drinking water in Dougherty
Comity (including both municipal and private wells), groondwater contamination by agricultural
chemicals is a potential public health threat Potential negative health effects associated with
ingesting chemical contaminants are siimniarii«d by the U.S. EPA (1989).
Using the potential Pareto-improvement criteria as a decision rale, a groundwater
protection program would be justified if the benefits of the program exceed the costs. The
overall objective of this case study is to estimate the benefits of groundwater protection in
Dougherty County via benefits transfer. The major challenge as to develop a protocol for using
existing groundwater valuation data at identified study sites to address the specific valuation
problem in Dougherty County (the policy site).
VALUE MEASURE CONCEPTS
A theoretically appropriate individual value measure requires a clear definition of the
commodity or service to be valued. Figure 1 illustrates how we can define the commodity or
service of interest in our case study. The initial concern in the case study is with the uses of
chemicals by the agricultural industry in Dougherty County. These uses involve human
activities such as mixing chemicals at wholesale and retail farm stores, mixing and applying
chemicals on farms, and disposing of used chemical containers.
Chemical uses combine with physical pathways to create potential groundwater
contamination situations. For example, improperly mixing highly concentrated chemical
solutions near unprotected wellheads create a situation in which groundwater contamination may
easily OCCUr. Qmnndwater cftntaipiMtjnn nmy m\*n
-------
Chemical Uses
Physical Pathways to Contamination |
Monitoring
(Test Results)
i
Current Water QuaKiy
AQ-(Q1-QO)
Potential for Future
Contamination
Future Water Quality
(Q1)
I
1
Figure 1. Definition of Commody or Service to be Valued
groundwater quality (Q°). The results of cuneat groundwater monitoring, combined with an
of fat\"T r****flKfl uses and potential pathways to contamination, provide information
on probable future water quality (Q1). The probable change in water quality (AQ) is then defined
As discussed earlier, monitoring data suggests that groundwater quality in Dougherty
County is currently "safe." However, because of existing chemical uses and physical features, a
relatively high potential exists for groundwater q^iafry to become *Hnisafe" m u^ future itam
increased agricultural chemical connrniinatfan Hence, Q^Q0, which implies mat AQ
represents an uncertain decrease in water quality. Uncertainty enters the policy analysis in terms
of whether Q° win be maintained or Q1 will occur. In addition, if Q1 occurs, the timing of this
water quality degradation may be a random event
Suppose next that a groundwater protection policy, denoted as Z. is proposed for
Dougherty County to prevent a degradation m water quality from Q° to Q1. Thus, in this case
stodytheobjxtiveistovaliKthegroiindwato We
J
J
1
J
-------
1
I
I
t
I
\
I
I
I
Fovir marinas arc located ather on the Km or within a mile of the Arthur Kill or Kill van
KulL One of the marinas offers a public boat ramp while the others predominantly provide slip
storage for moored boats. IB addition, six city and county parks are within the area. Park
officials indicated that these parks were used for picnicking, bird watching, and other activities
such as softbalL
In addition to use services, the Arthur Kill area contains wetlands oat may provide nonnse
services. The wetlands system in the Arthur Kffl area covets approximately 400 acres of
freshwater and saltwater tidal marshes and creeks. The areas where potential effects may be found
are along the Kffl between Bridge Creek (norm of Ooethals Bridge) and me Isle of Meadows (at
the mourn of Fresh Kills). This area covers approximately 127 acres of wetlands (B-Laing, 1990)
and supports a variety of wading and seabird species as weB as several hundred invertebrate
species. The freshwater marshes support an additional 20 to 30 species of invertebrates and
vertebrates suitable as food for the birds in the area. The area directly contributing to habitat
functions covers approximately 25 to 40 acres (the sum of the acreage used for feeding and
nesting).
Biologists afsyss wetlands in terms of their functions using a qualitative method of
evaluation called WET, which stands for Wetlands Evaluation Techniques. WET analyzes the
wetlands area in terms of social signifirancr, effectiveness, and opportunity. The Exxon
technical team conducted a WET analysis on the Arthur Kill region. Its results are cumulative
for the many oil spills that occurred in the region in a short time period, implying that the effects
are likely to be greater than just those from the Exxon spflL Condnsions about the functions of
the Arthur Kill wetlands include the following:
-«---»-! -~.KMfei
•uC %^^HU41MtH**BU- wuw * • • • •! —.—— .
• These wetlands serve as a filtering system by trapping andimfatt, pathogens, and toxfc
g^ifrfffufiys fad removing |fren> frpm wHBf ttipspot. Again, me extent of mis function
is uncertain because of the water traffic.
• These wetlands provide an important educational and research function. In particular*
the Harbor Herons Project has served to educate the pubBc. The fact that these
%
-------
iyoO» VOluniCCIa CHKagCU in muii |»ujov.ta uiviuviug tu* uiuuuug ui UBIIMUJ&* ouu
presenting infonnatton about tf»e area and its species. This project generated a great
deal of media coverage (Parsons, 1986).
* The last function the Arthur Kill wetlands provide is erosion control protection for the
area. The region provides moderate erosion control, particularly along the shorelines of
the wetlands where peat sediment is stabilized by the intertidal marshes, which
contributes to a stable shoreline and deters erosion of die mainland (Winfield. 1990). •
In summary, tins natural resource setting provides te backdrop for a case study using the
benefits transfer methodology in an NRDA context The setting enables researchers to evaluate
bom use and nonuse natural resource services. Use values are the values associated with natural
resource services where physical and/or visual cxtttart between people and tite natural resource
occur. Nonuse values do not require contact; rather these services are the resdt of a resource
providing well-being to people or other resources simply by existing.
Bom National Oceanic and Atmospheric Administration (NOAA) and Exxon prepared
damage estirnatesiismg the transfaniethodology. The parties were able to reach a negotiated
settlement based on these estimates. Because te estimates prepared by ttieNOAAhave not
been made public, this paper relies on the estimates prepared by Exxon's experts, which have
been made public (Desvousges and Mfflflten, 1991).
TRANSFER STUDIES IN AN NRDA CONTEXT
The Arthur Kill oil spill is typical of many NRDA cases. The size and/or location of such
spills often make a full-blown damage assessment inefficient because the assessment itself could
cost more than the damage. In these mstances, using the transfer metto
damage is more efficient Benefits transfer methodology also can provide a useful screening
devise for targeting assessments that will require more detailed Type B assessments.1
During the AERE workshop, participants discussed using transfermetiiodologyin
NRDAs. T^levd of omifort among partidpants musing the ttansferm
I on the status of the assessment and the
amount of probable scrutiny it wfll receive. As pan of this disctisrior*. the participants discussed
a continuum of NRDAs: on the left side are imtialscreenmgasseasments and on the right ride is
1TaeTypeAawdeI is a rioaffe proem thtt MI a
IBQI& QBtfOE DQsT ^sVUUJL BOQrt^OOHBlOO Utt^tBG •DC
-------
1
1
a full-blown study to support litigation (see figure 2). The continuum depicts the role of the •
— .».__.».* An *!M WDT1 A nmnvMCM frran an initial •MMtxmMit tn neffotiatftd SCttlCfnCIlt tO •
litimkm. scrutinv increases. Thus, the imprecision associated with using the transfer 1
methodology may be more of an issue when litigation is pending. •
Screening N*Bf*f*?? .„._.. 1
Aiionmnont SrtUemsnt ubgalion H
L_ L__ 1 1
Lttto Scrutiny Much Scrutiny •
Figure 2. Continuum of Valuation Scrutiny In an NRDA Context 1
When the level of scrutiny is relatively low, the willingness to use the transfer 1
methodology is nign. none ot our group memueis expressea any nesiiauon aooui using uranucis ••
for an initial NRDA or for a negotiated settlement However, most members were reluctant to 1
adopt the transfer methodology when litigation is involved- Because the level of scrutiny is much 1
higher in a litigation context, most of our group members thought oat the margin of error inherent 1
ma transfer study was not defensible. 1
Finally ftp transfer methftdology can also \t mmf hi «rtaWifhinE HP^ implementing 1
NRDA policy Fbrnumplf Af ^Typf A2"f»e«F|twi!tflfp^*«yNOAAtoe«tim«tefliedf|irnagp I
caused by certain types of oil spills is a transfer model The budget and tune constraints for 1
NRDA policy making are similar to the types of constraints that make transfer methodology •
transfer studies ma policy-related context •
• I
DATA AND METHODOLOGY ISSUES 1
Like any transfer study, data and methodological issues need to be resolved for NRDA •
fnmefiMC tn lv» fffrvtiv* Tn mir Hicrn*
-------
of damaged resources or services). valuation of the interrupted or rliminatrd services, and the
valuation methodology.
The development of the quantity data to be used in an NRDA transfer study often proves
Recall that, in a transfer study, the value or price of the affected resource or
to frf cMlfngg eca , ,
service is transferred from other studies. However, the researcher must determine the quantity of
affected resources or services before applying the transferred values. Determining the quantities
for in NRDA typically requires judgment because the quantities of it«ouices or services are not
observable, or if observable, ate not readily available. If historical data are available, they may
provide some useful guidance, but judgment may still be necessary.
In the Arthur Kill oil spill, die Exxon estimate addressed three types of interrupted or lost
services: fishing and boating access, near-water iecieatira(paik use), and wetlands services.
The access data used in Exxon's estimate were based on interviews wim local marina operators.
The number of marina slips and an estimate of typical c«cnpancy during the winter months were
combined to estimate the number of affected boats in marina slips.
The park-use data were also baaed on interviews with key informants (park officials in
this case). They estimated park use in terms off the number of viators during the off-season.
Finally , biologists estimated the number of acres of affected wetlands based on their field
Relying on key informants to develop the quantity estimates is not unusual in an NRDA.
In many cases, no better source of data is available. However* using key informants may
introduce moral hazard into the picture beouise they niay have a vested intn^stm the outcome of
ifa. damage yirtirfntes. Key fafflnnimy may realize th*t tfar interviewer is somehow associated
wim the recent spill, and the informant may provide biased estimates.3 When relying on key
informants for the quantity data, researchers should use their best judgment and be aware of the
possibility of moral hazard.
Biologists or other types of scientists often provide oner types of datfl. such as quantity
estimates, In many instances, rstimatre by scientists are the best source of the necessary data.
Our gram expressed sxnnecoaceins abort Scientists often
approach issues differently from economists, thus prodncmg data that are not useful to
economists. Our group discussicfi indicated that eccwNnito
their future efforts better than they have historically.
'Ahhoogh
povide
nuy be booed K>
btto|«NRDArab, they nay B«ke<
Hie data they
-------
I
I
I
I
I
r
The second type of data Usue we discussed was valuation issues. In a transfer study, the
quality of the estimate depends partly on the availabilily and quality of the original studies and
their suitability to transfer. These concerns lie not unique to NRDA transfer studies. Adopting
die transfer methodology means that the researcher adopts the values from the original study and
any inherent weaknesses in them.
In addition to methodological issues, we confront issues of "sameness" as well The
NRD A estimate based on a transferred value has mote credibility when the affected service is
very similar to me service on which thetransfer value is based. For example, an NRD A estimate
for cold-water fishing in the Northeast may not be well represented by a value for warm-water
fishing in Calif omia. SeasonaUty is an important consideration for niany recreation estimates.
Most recreation studies are based on the "high season," die season when feat particular recreation
activity is at its peak.. Ignoring the seasflnality feme fr «. ppngfer fftydy can reynlt in grmr in the
NRDA estimate.
As part of the discussion, we informally polled our group members on their i
it of
the adequacy of existing studies for transferring use values. We asked our group members to rate
the existing studies on a scale of 1 to 5, wim 1 bemg inadequate and 5 very adequate. Ibis
assessment included the number of studies, the quality of those studies (inclusion of substitute
sites, assumptions, parameters), and the *tansferability" of the stwties. Table 1 shows the
general adequacy ratings, although me adequacy of available studies win vary m particular cases
(e.g., locations, season, activity). For many types of use services, the group consensus was that
existing studies are generally not adequate for transfer. We concluded tfftt existing studies on
big-time sport fishing and big-game hunting are more adequate for transfer purposes nan die
other V9t categories considered. Existing studies on other uses such as swimming and wildlife
viewing did not receive a favorable rating in terms of adequacy.
Data issues for nonuse values are paro^culariy controversial. Even in a full-blown
analysis, nonuse values are extremely difficult to estimate, Economists have used contingent
valuation (CV) to estimate nonuse values, and disagreeinemexisu) about its vaUdity for this use.
The difficulty of the situation is amplified in a transfer study.
-------
TABLE 1. GROUP ASSESSMENT OF THE ADEQUACY OF EXISTING STUDIES
FOR TRANSFERRING USE VALUES
Use Category
Fishing
Big time
Boating
Motorized
Nonmotorized
Swimming
Beach Use
Shoreline Use
Wildlife Viewing
Hunting
Big game
Waterfowl
AdcQUBcy*
4
3
1
2
1
1
1
1
4
3
J
We specifically evaluated die adequacy of studies f or wetland values. In the Arthur Kill
study, the biologists determined that the wetlands in the Kfll area were only serving some of their
intended functions. A study that focused on this particular subset of finv^cms in the
geographic area did not exist The studies thai do exist do an incomplete job of valuing
wetlands, even in general tenns. Table 2 smimarizea the available wetlands studies.
The methodological issnes we discussed focused on me unit of valuation. Use services
studies havefcwr possible choices for the unit of valuation. The first is the unit-day value,
where, Ior example, the valve of a fishing day or * boating day is transferred from a study to the
NRDAsite. Although this approach has the advantage of simplicity, Ac differences between tiie
sites that may influence the demand for services ate essentially ignored.
The second approach uses a valuation equation. In mistype of transfer, th
from an existing study are applied to the means (or tepreseiitativevahjes) of the same variables
for me NRDAsite. This approach offers an improvement over the noil-day value, but it is
10
-------
1
I
1
£*8J
jit
e
liiSI!
I
I
11!
t
i ill ii
! PPI sil i
imiiiiiifill* ti i
if
I,,
I
1
v
!!
t
I 11111*
•B'B fi w*
^*.
iis
11
i
ii
•8
11
8
liil
-------
* *!**!
Mum
ii«
u w •« «>
!*•«.!».
I
f
II
lid
i«||i|=iq
• B B>V «
-------
,41
i
I
I
I
t
I
r
i
•5*
o,
•g!
tPPH s
iig?fifk
s.
•§
f lSv3-g«Si
i §£H
8 si*
8 «*e-8
~ • M E ^
£3 g
^ •»"
* -1!
II !|1
2$ 111
I 8
l£ If
: w
i!«
I I
!l
a
i.
fgi«
lee
1^
11
*B
i
j»5
'IIs,
Oaifrf
ii
i
-------
1
•5
i?
I*
I
rl
1
i£
l ip
•}F
lit I
If! if
i —^
!*
I
*
a
* in
*
IB
!S
11
I
tl
12
-------
[
frequently difficult to find an appropriate equation to transfer and the comparable data for the
NRDA site.
The third approach is a generalized model from which values can be transferred. Such a
model requires much more information that the previous two approaches, but it offers the
advantage of better estimating the site-specific value. The group members discussed the
possibility of adapting the Random Utility Model (RUM) for transfer.
A final option for valuation is the meta-analysis approach. This approach compiles all
available values and their influences and produces a value that accounts for the many possible
influences. Like the generalized model above, the data requirements are extensive. (For nonuse
values, whether such an analysis can be performed given the currently available studies is
unclear.)
The methodology adopted in an NRDA transfer study depends in part on the timing, the
funding, and the available data. Our group discussion indicated mat we would like to see a
movement toward using the generalized model.
RESEARCH AGENDA
Our group discussion revealed that much research still needs to be done on use and
nonuse values for NRDA transfer purposes. We focused on three primary research items: the
design and undertaking of a "grand" study, more and better original studies, and a technique to
generalize RUMs.
The first research agenda item (deemed most important by the group) was the design of
the grand study. Such a study would encompass all types of services and the influences on the
demand for these services. The study would be suitable for transfer purposes and would be
linked to ecological models.
The second research item is the need for more and better quality original studies. Our
group thought more studies on use values would be helpful, particularly on those types of values
for which few studies exist, such as swimming and boating. Bat more important are studies on
nonuse values. Such research should address fundamental issues associated with credible
valuation procedures. Consensus on transferring nonuse values depends on consensus on
estimating credible nonuse values. The group concluded mat good studies on wetlands and
seabirds would go far in filling our needs for nonuse estimates. New studies undertaken should be
designed with transfer in mind.
13
-------
RUM moaeu» un
for using RUMs in a transfer setting.
a study.
infonnative.
B-Laing Associates.
"*1
Research Triangle Instittitt.
14
-------
f AJAAAJ-«\J » rka-vjr* »»v/.
I
4
I
I
I
I
I
t
I
I
I
APPLICATION OF THE TYPE A MODEL
Carol Adairc Jones*
ABSTRACT
The Type A mode) is the single largest benefits transfer model for natural
resource damage assessment and the only one that has regulatory status for litigation .
under CERCLA and the Clean Water ACL In this case study, we focus on the Type A
model procedures for valuing losses in recreational services due to fish lolls and
fishery closures resulting from an oil or chemical spill. In addition we discuss how to
value recreational fishery injuries.
The natural resource damage assessment model for coastal, and marine
environments, the 'Type A model." is the single largest benefits transfer model for
natural resource damage assessment and the only one that has regulatory status for
litigation under CERCLA and the Clean Water ACL The model provides a simplified
assessment procedure for short-term releases of oil and hazardous substances. It
represents a low-cost alternative to Type B damage assessments, which may require
detailed field observations and extensive collection and analysis of chemical, biological,
and behavioral data.
The first-generation Type A model, under review here, was promulgated under
rule-making by the U.S. Department of Interior (DOI) in 1987. It covers the coastal and
marine environment of the U.S. DOI is required to revise the model every two years; this
year, the agency intends to propose a new Great Lakes version, as well as a substantially
revised coastal and marine version of the model.
In this case study, we focus on the Type A model procedures for valuing losses in
recreational services due to fish kills and fishery closures resulting from an oil or
chemical spill. The Type A model incorporates the data and algorithms to calculate
fishery injuries, measured as the reduction in (fish stocks and) recreational fishery catch
"NOAA Damage Assessment Center. Members of the case study group included Mark Downing (Texas
AftM). Rick Duoford (Research Triangle Institute). Michael Hanemaan (University of California-
Berkeley), Christopher Hansen (U.S. Forest Service), Robert Leewonby (NOAA). Edward Morey
(University of Colorado-Boulder), Jim Opalucb (University of Rhode Island), Richard Ready
(University of Kentucky), Dan Schruefer (NOAA), Thomas Wegge (Jones and Slakes AnociatcsX and
Peter WUcy (NOAA).
-------
AGROUND- TVPEAMODELTORCOASTALANDMARINE
ENVIRONMENTS
(sec Figure 1).
ts it disperses,
aiculfltions.
tions.
The biological effects module calculates losses to biological populations through
time. The calculations include the following: die direct mortality to adult, juvenile, and
larval biota due to toxic concentrations; recruitment losses due to stock effects; and the.
indirect mortality and weight loss to adult, juvenile, and larval biota doe to the loss of
foodstuff in the food web.
The economic damages module calculates the dollar values for injuries to biota
based on use values. It also calculates me tosses due to dc«ares of fishing, waterfowl
hunting, or beach areas.
The caVnilatiffls rely on geographic data bases mat contain avenge resource
distributions for multiple habitat types within ten geographic regions throughout the
coastal US. based on the classification scheme developed in Cowirdin et al (1979).
Marine and estuarine systems are subdivided into snbtidal and intertidal subsystems men
broken down into additional habitat classes (based on shoreline type or bottom type).
After the authors factored in the Hkefihood of each province-system«subsystem-class
combination and the feasibility of collecting data for each likely grouping, they created a
dftrtvw with 36 mtertidal and 55 snbtidal ecosystem types with seasonal variations.
Figure 2 provides a map of the ten regions, and Table 1 hsts the habitat classifications.
The species in the database are classified into 13 categories) including ^THC fish
categories. The nine fish categories represent 141 species, including both finfish and
.1
-------
I
F
i
i
i
r
i
i
User Input:
Spill Type, Location, Date,
Habitat Classification,
Beach/Hunting/Fishing Closures
Physical Fates
Submodel
Chemical
Data Base
•PHYS_BIO.LNK" FILE
Surface Water Column
Bottom Concentrations
Biological Effects
Submodel
Biological
DataBase
"BIO_ECON.LNK- FILE
Btomass Reductions
..
Economic
Economic Damages
Submodel
MONETARY DAMAGES
Rgure 1. Model System Overview (NRDAM/CME)
-------
I
*
*
1.
1
I
Boundaries of 10 Marine and Estuarlne Provinces
Source: Type A Documentation
-------
TABLE 1. HABITAT CLASSIFICATIONS
I
f
I
I
\
I
Ecosystem Types
A. 10 Marine and Estuarine Provinces
1. Atlantic and Gulf
PI. Acadian (Northeast: north of Cape Cod)
P2. Virginian (Mid-Atlantic: Cape Cod to Cape Hatteras)
P3. Carolinian (South-Atlantic: Cape Hatteras to Cape Canaveral)
P4. Louisianian (Gulf Coast: Cedar Key, Florida to Port Aransas, Texas)
P5. West Indian (South Florida, South Texas, West Indian Islands)
2. Pacific
P6. Califomian (California: south of Cape Mendocino)
P7. Columbian (Pacific Northwest: Cape Mendocino to Vancouver Island)
PS. Fjord (Gulf of Alaska: south of Aleutian chain)
P9. Arctic (Alaska: North of Aleutian Chain)
P10. Pacific Insular (Hawaii and other Pacific islands)
a. Subtidal Bottom Types
S-B1. Rock bottom
Cobble (unconsolidated)
Sand (unconsolidated)
Mud (unconsolidated)
Rooted vascular aquatic bed (grasses)
Macroalgal aquatic bed (e.g., kelp)
S-B2.
S-B3.
S-B4.
S-B5.
S-B6.
S-B7. Coral reef
S-B8. Molluskreef
S-B9 Worm reef
b. Intel-tidal Bottom Types
I-B1. Rocky shore
I-B2. Cobbled beach
I-B3. Sandy beach
I-B4. Muddy shore
I-BS. Saltmarsh (cordgrass)
I-B6. Trees (coastal wetlands)
I-B7. Coral reef
I-B8. Molluskreef
-------
invertebrates (see Table 2). Four categories of species information are included: aauii
biomass, by species; larval numbers, by species category; mortality and growth
parameters by species category; and primary and secondary productivity values.
The model is not intended to represent any specific localized populations of
estuarine or marine situations: the databases represent average values for representative
types of ecosystems. Consequently, to capture die necessary breadth of geographic
coverage, the Type A Model has sacrificed geographic specificity.
CASE STUDY PROBLEM: VALUING RECREATIONAL FISH-KILLS AND
FISHERY CLOSURES
Injury Quantification
Short-term (acute toxicity) losses are calculated separately for adults and larvae
based on the toxicity information in the chemical database and die species distribution
data. The model also calculates long-term losses due to the acute mortality to adult,
juvenile, and larval biota due to toxic concentrations; the reduced recruitment into the
adult fishery due to acute toxicity kills of larvae, juveniles, and adults; and the indirect
mortality to adult, juvenile, and larval biota due to loss of foodstuff in the food web.
The fishery population dynamics in the model are based on the assumptions that
die instantaneous catch rate (or catchabUity coefficient), the instantaneous natural
mortality, and the growth function for individuals remain constant, and that egg
production and larval numbers return to pre-spUl levels immediately following
dissipation of the spill. The architects of the model justify these assumptions on the
grounds that die model is designed for spills of short duration.
Lost catch due to closure of an area to fishing is also calculated based on the
biomass in die closed area. Because some of die lost catch in die closure area is due to
mortality from acute toxicity, only die lost catch due to die closure in excess of die acute
toxicity losses is added to die long-term losses to calculate total catch loss.
-------
I
t
I
I
I
I
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET
f
I
I
f
I
Specks
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
15
Amf
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
J-fc
33
34
35
36
37
38
39
40
42
43
•/*_*__*_•_•
T«*w§ory
Category"
1
1
2
2
2
2
2
3
3
3
3
4
4
4
^
4
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
1
2
A
2
6
6
6
6
6
6
6
2
2
BT«_.
MJ
________ «_«*»
F'lnniif DID
2 Planktivorouc ficb
3 Piscivorous fish
Common Name
A^trimfl iShffrt
Alcwifc (and Blueback Herring)
Menhaden Atlantic and Gulf
Atlantic Herring
Butterfish
Pollock
Atlantic Mackerel
Blocfish
StripedBass
Monkfish (Goosefisb)
Weakfish (Grey Sea Trout)
Tuna
Swordfish
Sharks
T""ll I\IT
Docfish
__^V£ftB0_ta
YeUowtail Flounder
Summer Flounder (Fluke)
American Plaice
Witch Flounder
Winter Flounder (Blackback)
Atlantic Cod
Haddock
Redfisb (Ocean Perch)
Stiver Hake (Whiting)
Red Hake
While Hake
Scup
Tilefisb
Black Sea Bass
Atlantic Wolffish
Hickory Shad
-hTtt-it _Lf ft_*_tw-
opQmftD MftCKBrei
Harvestfish
Atlantic Croaker
Drums
Spot
YeUow Perch
Carp
Eels
Atlantic Thread Herring
Anchovy, Atlantic
S D-Bcn-l-ufa S
6 Scmi-demen-l fi-h 9
7 Mottusks 10
Scientific Nun?
Alosa sapidissima
Alosa puudoharengtu. A. aestivatis
Brevoonia ryrannus, B. patronus
Cbipea harengus harengus
Pcprilus triacanihus
FoUachuuvirau
Scomber Kombrus
Pomtaomtu saltatrii
Moroni jaxatitis
Lophuis aneriamus
Cynotdon ngalis
Thumuutpp.
Xphias gladius
Squalu j Qranthias
Lunmdafemginea
Paralictahysdeauma
Hippoglossmdes plaessoides
Gfyptocephaius cynoglossus
Pseudoplaavnecus americmus
Gtofba morhua
Metanog numw aeglefinia
Sebastesjuciatus
MtrlNCciiu hiltntorix
Urophycis duos
Urophycis tenuis
Sunotomtu chrysops
Lopholatiba chamaeleonticeps,
Caulolatilus microps
Cattrapriftu ttriata
Anarehichas bipus
AJotamediocris
Scomberomona macuiaau
PepriUualepidotiu
Micropogonias mdulatus
gf^fff^Sffaf
LriffstffwtHif mHtfitfmf
Percajlavacau
Cypruuu carpio
Anoni1Hfr.--kM
Optahonemtogliiuan
Anchoaspp-
(continued)
DtC'ipodi 1 1 Wiierf owl
Squid 12 Shonbirds
Munmak 13 Scabudt
4 Top carnivonif
-------
-------
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET (CONTINUED)
Species
Number Category*
44 2
45 6
46 6
47 6
48 6
49 3
50 3
51 3
52 3
53 3
54 3
55 3
56 5
57 5
58 6
59 6
60 6
61 6
62 6
63 6
64 6
65 6
66 6
67 6
68 6
69 6
70 6
71 6
72 2
73 6
74 6
75 3
76 6
77 6
78 1
79 1
80 1
81 1
82 1
83 2
84 2
85 2
86 2
•Category Key
1 AnadnuDoot fuh
2 PUnktrmoBifiih
3 Pifavorou* fii b
Common Name
Striped Mullet
Sbeepsbead
Spotted Sea Trout
Scientific Name
Mugilcephalus
Archotargus probatocephalus
Cynosdon nebulosus
Sand Sea Trout (White Sea Trout) Cynosdon arenariiu
Sea Catfish
Atlantic Halibut
Bonito (Tunny)
CrevaUeJack
Greater Amberjack
Jacks, Other
Blue Runner
Dolphins
Flounder, Southern
Flounder, Gulf
Drum, Red
Drum, Black
Porgies
Florida Pompano
Grunts
Pinfish
Kingfish
Sheepshead
Cock
Tautog
Groupers
Snapper, Red
Snapper, Other
Whiting (Southern Hakes)
Spanish Sardine
Silver Jenny
Bonefish
Barracuda
Sea Bass
Trifiterfish
• ••oc1— • ••••
Salmon. Sockeye (« Red)
Salmon, Chum (*Keta)
Salmon, Pink
Catinrvt Phhtnnt (• 7hie\
Salmon, Coho(« Silver)
Mackerel, Pacific
Mackerel Jack
Anchovy. Pacific
Herring, Sea (Pacific)
5 Dementlfiih
6 Semi-dementi fUh
7 Moilwks
Anus/tits
Hippoglossus hippoglossus
Euthynnus alletteratus
Caranx hippos
Seriola dumerUi
Carangidae
Caranx crysos
Coryphaenidae
Paralichthys lethostigma
Paralichthys albiquna
SdaenopsoceUaus
Pogonias cromis
Sparidae
Trachinotus carolinus
Hf^iflMr*
Lagodonrhombodies
Uenttcirrhuj spp.
Archosargiu probatocephalus
Brosmebrosme
Tautogaonitis
Epinepheliu spp., Mycteroperca spp.
Lutjanus campechantu
f nffmf^f
Urophycafloridanus
SardaeOaaitrua
Eucinostomits gula
Albuiavulpes
^ — »^ — — » -*— ~
iiprjia iiniar
Tfimi'iiiuf
flflWtuW
Oncorhynchus nerka
Oncorhynchus keta
Oncorhynchus tshawytscha
Oncorhynchus kuiach
Scomber japonicus
Trachurus symmetricus
EngraiiUs mordax
Oupea harengus paUasi
(continued)
t Decapod* 11 Waterfowl
9 Sqoid 12 ShORbMi
10 Manmth 13 Sednidt
V
1
-------
1
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET (CONTINUED)
1
I
I
f
I
I
I
\
I
r
i
i
r
i
Species
Number Category*
87 5
88 5
89 6
90 6
91 6
92 6
93 6
94 6
95 6
96 6
97 2
98 2
99 5
100 5
101 5
102 5
103 7
104 6
105 6
106 6
107 6
108 6
199 6
Invertebrates
201 7
202 7
203 7
204 8
205 8
206 8
207 9
208 7
209 8
210 8
211 7
212 7
213 7
214 7
215 8
216 7
217 8
218 8
1 Audnmoas fidi
2 nAOICu^WDttB OMB
3 Piiti varans fith
Common Name .
Bounder, Pacific
Halibut, Pacific
Perch, Pacific Ocean
Rockfish, Other
Perch, Other
Sablefish (Black Cod)
Cod, Tree (Pacific)
Lmgcod
Hake. Pacific (Whiting)
Sea Bass
Pollock, Walkye
Mackerel, Atka
SofeYeUowfin
Flounder, Anowtcoth
Turbot, Greenland
Plaice, Alaska
Smelt
Bounder, Starry
Sole, Butter
Sole, Dover
Sole, English
Sole, Rock
Other Fish
Snrfdam
Ocean Quahog
Atlanta Sea Scallop
Northern Shrimp
Red Crab
Squid, Atlantic
Blue Mussel
Bloc Crab (Hard Shell)
BhrCraMSoft Shell)
Soft Gam
Oyster, Atlantic
Hard Qam (Qoanog)
Conch
Shrimp (Brown, Pink, White)
Calico Scallop
Crabs (general)
Stone dab
S Daneralfiifa I
7 Molhuki 10
Scientific Name
PlmnrnMtiitap
Hippogloaanu sttnolepis
Sebostts oluttts
Stbastestpp.
Hyperprosopm spp.
Anoplopomafimbna
Optuodonelongatus
tferlitcctiu pfoductus
Theragradulcogramma
Piauvgnanmus motiopterygius
Ltmandaasptra
AOureahes aomias
KtManitiushippoglossouies
Osmeridae
Paraliclahysitellatus
ttopiettaisolepis
Mienaomuspacifiais
PGtophfysvttubts
deoeric)
.Cnivuln tnli/linim/l
vfnmmw m-irrm-fmtnmf
Artico fffflnrfffB
Hotnanuaneriaouu
Pandabuborealis
Gayonqumqtudau
LoKgoptolei.nieiaUcebrosus
Jafvfibtf tdlllis
CdBmeaessapithu
Ityaortiiaria
Cfosuuttto virginica
Men«nariamercenaria
Pnanutfp.
ArgopeaaigiVbus
deneric)
Mauppemenetuaia
(cooUmwd)
Dacapodt 11 Watefowl
Squid 12 SbonbMt
•NAftmiiiAlft 19 Saigtlntflc
4 Topcaraivonu
-------
TABLE 2. SPECIES LIST AND CATEGOR12AT1UIN f UK
DATA SET (CONTINUED)
Specks
Number Category*
219 8
220 7
221 8
222 8
223 9
224 8
225 8
226 7
227 7
228 7
229 7
230 7
231 7
232 7
233 7
299 7
Birds
301 11
302 11
303 11
304 11
305 11
306 11
307 11
308 11
311 12
312 12
313 12
314 12
315 12
316 12
321 13
«5 14
j£+ ij
323 13
324 13
325 13
326 13
327 13
328 13
•CrtcforjKcy
1 AiMdromouf fiib
2 Ptanktjvorow fiib
31UMj4uMW«Ulft flBh
mdwuimu IKQ
4 Top cttnivanii
Common Name
Lobster, Spiny
Abalone
Crab, Dungencss
Shrimp, Pacific
Squid. Pacific
Crab, Snow (Tanner)
Crab, King
Clam, Butler
Clam, Horse
Clam, Geoduc
Clam, Manila
Oyster, Pacific
Oyster. Olympic
Atlantic Bay Scallop
Pacific Sea Scallop
Other In vertibreates
Marsh Ducks
Diving Ducks
Mergansers
Whistling Ducks
Stiff-Tailed Ducks
Coots
Geese
Swans
Sandpipers
Plovers
•¥•
TOTDStOOCS
Oyster Catchers
Phalaropes
Avocetes, Stilts
Gulls. Terns
fVwnwontc
VAJiuiOfMHa
Auks
Shearwaters
Storm Petrels
Pelicans
Frigaiebiids
Gannets, Boobies
5 Demersal fith
6 Soni-^leatcnal fiih
7 MoUwki
SclenttflcName
Panuliris spp.
Haiiotistpp.
Cancer magaicr
Pandalus boreatis
Loligo opakxxia, Berrytetahis magister.
Onychoteuihisbonalijaponicus
Qiionoccctes
Paralithodes catiasdiettica, P. platypus
Saxidomus nuttalli
Tresuscapax
Paaopea generosa
Tapes ptulippinarvm
Crassoarea gigas
Ostrea Ittrida
Argopccten irradiems
Pecten caurinus
(generic)
Anatinae
Aymyinae
Merginae
1 JCnmt*vj jUitiMC
Oxyofinae
Rallidae
Anserinae
Cygnmae
Smtopacktac
dandriidae
AnhrizJdae
HaematDpodidae
Phalaropodidae
fka(_MKIlMn_grtj,lnM
Ka^tm Tifii»ii»lnr
Laridae
phmlMYnrmmrwl*^
Alcidae
PjoceOarUdae
Mydrobaiidae
Prtrranidae
Fregatidae
Snlidae
8 Decapod* tl Waterfowl
9 Squid 12 Sbonbink
10 Munmalt 13 SMbtidt
10
-------
Valuation of Damages
I
I
I
I
I
I
I
I
Translation from Change in Stock to Change in Trip Catch and Number of Affected
Trips
In the biological submodel, the fish stock is allocated to recreational catch
mortality, commercial catch mortality, and natural mortality based on share parameters
for each species in the database. The predicted reduction in stock due to a spill is also .
allocated to those categories, assuming constant proportions. Jim Opaluch, one of the
authors of the economic module (and a participant in the case study group), indicated that
an assumption implicit in the valuation procedure was that all species are highly mobile;
with this assumption, the change in fish stock will be spread over a wide geographical
area and generally will produce a small change in catch rate (trip quality) over a large
aber of trips.
The value per fish, catchability coefficient, level of fishing effort, and cost per
unit effort parameters are assumed lo be unaffected by the spill. Consequently, the
decline in recreational fishing catch due to a spill is calculated as the recreational fishing
share of the stock (a parameter in the database) times the change in the fishery stock
calculated in the biological module.
Valuation of Ike Change in Catch Rates
The valuation procedure then assigns the reduction in recreational stock size at a
rate of one fewer fish per angler. In the calculation, die number of anglers affected just
equals the change in the recreational stock size; there is no independent calculation of
total trips affected. This procedure is a creative way to avoid explicitly characterizing the
levels of fishing participation affected by die spill (which is likely to be larger than the
spill area because of fish mobility).
To generate the recreational fishing values for the Type A model, the authors
relied on two studies providing an e«H*Mfr- of the change in the value of recreational
fishing trips with a unit change in catch rate. Rowe et al. (1985) provide consumer
surplus estimates for trips to California, Oregon, and Washington marine fisheries from
separate random utility models for each state. For selected species, me scenario valued
was the increase in the catch rate of one species by one fish/trip at all site/mode
combinations where the species is caught Norton, Smith, and Strand (1983) provide
estimates of the changes in consumer surplus with changes in catch rates for several
I
11
-------
sinpco
model.
Because these two studies valued only a few species, the modelers needed a
procedure to provide values for other species. They calculated the change in consumer
surplus on a weight basis for the available species. Judging that the variation in the value
per pound did not appear to vary greatly across the species valued in the studies, they
employed the simple mean of the estimates ($1.84/lb) in the model to value losses of all
species.
QUESTIONS DISCUSSED IN THE CASE STUDY SESSION
We discussed whether the current procedures for valuing recreational fishing
injuries in the Type A model can be improved. We considered the adjustments that
would contribute the most to improving the estimates and the adjustments that are
currently feasible.
The group proposed separate discussions of the injury from fish lolls, which we
believed was appropriately valued as a change in quality of the recreational fishery, and
the injury from fishery closures, which we thought might better be modeled as a change
in the quantity of resources available. We consider each modeling context separately
below. For most possible extensions, we concluded that data are insufficient to determine
whether such changes would represent substantial refinements to the model calculations.
The discussion produced a series of recommendations for further research. In the final
section, we discuss criteria to be used in selecting studies for inclusion in the model
database.
MODELING ISSUES
Population Effects Due To Fish Kills And Their Impact On Fish Population
Dynamics
Currently, the effect of fish kills is modeled as a change in the quality of
recreational fishing trips that affects the trip value but does not affect total participation in
the fishery. A single value per gram of fish killed appears in the model: the variation
across species in damages per fish killed is completely driven by variation in average
weight across species. In addition, the value does not vary with the size of the spill (and
the effect on stock and catch rates) or the extent to which available substitutes are
J
J
12
-------
I
I
I
I
f
I
I
similarly affected. We discussed several possible extensions to the modeling, as reported
below.
• Expand the single recreational fish value included in the database to a matrix of
values, including variations in the value of lost fish by
—fish species,
—geographical area of spill, and
—user types.
Most members of the group thought incorporating species and geographical
variations could be an important contribution to the model and believed that some
additional values have appeared in the literature since the model was first developed. We
did not think that incorporating variations in consumer surplus values by user types
would make an important contribution.
• Adapt the modeling and expand the value database to incorporate variations in
die change in consumer surplus per unit change in catch depending on
—the level of the change in catch per trip (Le., avoiding the assumption that the
change in consumer surplus is linear in catch); and
—the extent to which substitutes are affected (which will vary substantially
depending on whether the affected species have localized populations or are
highly mobile over a wide area).
To implement either, it would be necessary to change the modeling to identify the
geographic zone of impact (taking into account the mobility of the species) and the
number of trips taken to that zone. With tins information, an estimated change in catch
per affected trip could be calculated (rather than implicitly assigning a reduction of one
fish per trip.). In addition, the Type A model would need a matrix of values in the
database, capturing the nonlinearities and substitution possibilities in the values.
Are the size of the change in catch per trip and the extent of tile substitutes
affected important sources of variation in value? The group discussion was inconclusive:
we concluded that research is needed to explore these issues. To the extent mat spills
valued with the model are relatively small and the species are mobile, nonlinearities in
the change in consumer surplus with a change in catch rates are not likely to have a large
effect on values. For spills heavily injuring highly localized species, the variation in the
change in catch rate may be much greater, for this context, exploring the possibility of
substantial nonlinearities is more important. Impacts on localized groupings of species
also raise questions regarding the treatment of variations in substitution possibilities.
13
-------
Some preliminary analysis by Graham-Tomasi ana sung wuu
fishing model (Jones and Sung, 1991) suggests that variation in substitution possibilities
has a far greater effect on the value per lost fish than variation in the quantity of fish lost
per trip.
Are these changes feasible? Unfortunately, we had serious questions about the
availability of necessary data. The NMFS marine recreational surveys were cited as a
possible source of data on trips. In addition, we discussed how to implement the
variations in value with nonlinearities and substitution possibilities. Because of the
difficulty of establishing a formula, some individuals in the group suggested creating
categories of "small/medium/large effects" and assigning spills to suitable categories.
However the distinctions are to be implemented, additional research needs to be done to
generate the necessary values for making such distinctions.
• Incorporate changes in fishing participation as a result of spill-induced quality
changes in the fisheries.
Currently, the Type A model treats fishing participation levels as constant when
fishing quality changes based on the assumption of mobile fish species. With this
assumption, the population changes generally being modeled would yield small changes
over a wide geographic area. We concluded that further research would be useful to
identify how elastic trip participation is to quality changes (at die level of quality changes
involved) and the extent to which damages are underestimated by excluding this category
of effects.
Some recent preliminary analysis of the Michigan recreational fishery model
performed by Graham-Tomasi and Sung indicates that, though the participation elasticity
is not large, the share of damages contributed by that behavioral response may be
substantial.
Incorporating this extension in the model would require developing a generic
participation equation. Before this equation could be added, we would need to include
the modeling and database adjustments required to implement it Those adjustments
would build into the model the capacity to identify the zone of impact on the fisheries
(taking into account fish mobility) then determining the impact on trip catch in the
affected zone and the total number of trips in the affected zone.
14
-------
I
I
I
I
I
I
I
I
t
I
.An additional requirement would be to ensure that the modeling in the fishery
dynamics and the valuation portions of the model are consistent regarding trip
participation. We believed ensuring this consistency would not be difficult
Fishery Closures
Fish not caught because of a closure are valued using the same procedures as for
fish kills, that is, the total number of trips is assumed constant, but the value of each
affected trip is reduced because of the lower catch rate. This procedure implicitly
assumes a small closure area and the existence of (perfect) substitute sites sufficiently
nearby so that additional travel costs are essentially zero.
We believed considering modeling closures as a change in quantity of fishing
resources would be appropriate. In this case, the correct calculation of damages for a
change in quantity of recreational fishing services would be the change in trips times the
consumer surplus per trip. Ideally, in the studies providing die basis for tile consumer
surplus of a lost fishing trip, the species and site characteristics are similar to the closure
area, and the substitution possibilities are similar in both study and spill contexts.
This extension would seem to be more important in cases in which most close
substitution opportunities are not available. The current procedures appear adequate in
cases of a small area of closure.
Incorporating this extension would require trip participation rates and additional
consumer surplus values on a per-trip basis. More studies are likely to be available for
valuing fishing trips (as needed in mis extension, modeling a change in quantity) man for
valuing changes in the catch rate on trips (as needed for a change in quality).
RECOMMENDATIONS FOR FURTHER RESEARCH
We generally felt mat additional work is needed to explore whether substantial
variations exist in consumer surplus for a change in catch per trip by species, geographic
area, size of the effect, and the extent of substitution possibilities mat are affected. The
group agreed mat the current set of random utility models that have been estimated
provides a good basis for such analysis. The participation question also needs to be
explored; mis research can be done with the participation models linked to random utility
models or with the earlier generation travel cost models, employing equations estimating
total trips.
IS
-------
SELECTION OF STUDIES FOR INCLUSION IN THE DATABASE
The selection of studies and specific consumer surplus value calculations from the
studies is critical to the model database. We addressed the following issue: What criteria
should be applied to exercise quality control in the choice of studies used to estimate
consumer values? We identified three sets of criteria that may be relevant to the selection
of studies:
• relevance of the consumer surplus measure to the context (change in quality, •
loss of access)
• quality of study (meets minimum standards)
• comparability of context between the study site and the spill site
However, we did not agree on how to apply the criteria. We did not believe that
the current literature provides enough basis to decide what factors are operationally
important in determining "comparability." And we concluded mat the quality judgment
needs to be made within the context of the study's objective and its use in the transfer.
Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of
Wetlands and Deepwater Habitats of the United States. Office of Biolog
Services, Fish and Wildlife Service, U.S. Department of the Interior, FWS/OBS-
79/31.
Jones, Carol Adaire, and Yusen Sung. July 31,1991. Valuation of Environmental
Quality at Michigan Recreational Fishing Sites: Methodological Issues and
Policy Applications. EPA Contract No. CR-816247-01-2.
Measuring Damages to Coastal and Marine Natural Resources: Concepts and Data
Relevant to CERCLA Type A Damage Assessments. Volumes I and H, and
Appendices A through H (Volume fi). NTIS, PB87-142485.
Norton, Virgil, Terry Smith, and Ivar Strand. 1983. Stripers: The Economic Value of
the Atlantic Coast Commercial and Recreational Striped Bass Fisheries.
University of Maryland Sea Grant Publication No. UM-SG-TS-83-12.
Rowe, Robert W. 1985. Valuing Marine Recreational Fishing on the Pacific Coast.
National Marine Fisheries Service Administration Report No. LJ-8-18C. June.
Type A Recreational Fishing Case Study.
Type A Regulations: 52 FR 9042, March 20,1987 and Technical Corrections, Type A:
S3 FR 9769, and Availability of Corrected Type A Model: S3 FR 9819, both
March 25,1988.
16
-------
I
1
I
I
I
LONG-TERM HEALTH RISKS*
PIGEON RIVER, NORTH CAROLINA.
Susan B. Kask*
ABSTRACT
Executive Order 12291 requires benefit-cost analysis for all government legislation.
Does this mean that for each piece of environmental legislation we must provide new health
benefits estimates for each illness and each toxin to value benefits? Estimating the benefits
of a reduction in health risks is a difficult task for the policy researcher. In this paper we
present a protocol for transferring health benefits from a study site to a different policy site
and provide an example of its application.
I
Protection of public health is a primary goal of much of U.S. environmental legislation
because environmental pollution can have a variety of negative effects on public health. For
example air pollution can cause itchy eyes, chronic respiratory disease, and even death for those
most sensitive. These effects, however, occur with some probability. Environmental pollution
increases the risk of exposure to a contaminant, which in turn increases the risk of adverse health
effects (see Figure 1). A benefit from reduced pollution is the reduction in the risk of these
health effects. To evaluate the benefits from environmental pollution control legislation, we
must account for these health benefits.
I
Pollutant
Releases
to the
'Environment
— >
Pollutant
Concentration
>
Risk
of
Exposure
Risk of
Adverse
Health
Effects
\
I
I
I
Figure 1. The Link Between Pollution and Health
Estimating the benefits of risk reduction is difficult for the policy researcher. How much
individuals value a reduction in their future risk of contracting cancer or chronic illness from a
reduction in pollution is a challenge to estimate. Furthermore, estimating the value of reduced
* Western Carolina University, Economics and Finance Department Members of the case study group included
Sergio Ardila (the Inter-American Development Bank). Robert Benens (Oregon State University), Alan
Krupnick (Resources for die Future), Spencer Pearce (Consultant). Eirik Romstad (Agricultural University of
Norway), Richard Ruppert, and John StoU (University of Wisconsin-Green Bay).
I
-------
risk of acute illness or discomfort from a variety of symptoms is equally problematic. Must we
provide a new estimate for each illness, for each toxin, to value benefits? Studies exist that value
accidental death, death at some future date, and reductions in illness days, for example. Can we
use these studies as proxy estimates across illnesses and toxins? Can they be transferred
spatially? This paper explores the potential to transfer health benefits.
We present a basic model underlying health benefit estimates. We also present the
primary issues and a proposed protocol for benefits transfer. To demonstrate the protocol and
illustrate the pitfalls of transfer, we consider a case study. Finally we present our conclusions
and recommendations for future research.
J
CONVENTIONAL THEORY OF HEALTH BENEFITS MEASUREMENT
The typical model for measuring health benefits usually begins with a damage or
production function that links self-insurance activities (e.g., medical treatment, purchase of air
conditioners, diet, and exercise) to health. We denote this function as
H * H(Z)
where Z is a vector of self-insurance activities and H is a state of health. In some cases H is also
a function of the level of pollutant (Shogren and Crocker, 1991). The production function may
be represented with a two-state model with state 0 representing good health and state 1
representing death (Smith and Desvousges, 1987), or alternatively, H may represent an index or a
continuum of health outcomes (Dickie and Gerking, 1991; Shogren and Crocker, 1991). Here
we assume a two-state world for illustrative purposes.
»
As shown in Figure 1, pollution affects health through the risk of exposure and the risk of
adverse health effects given exposure. We can include pollution into a probability density
function Q, representing the probability of having good health. This probability depends on the
level of pollution in the environment, which in turn affects the level of exposure of an individual,
and the individual's level of private self-protection. This probability function is
Q * (KX,Q)
where X is the level of private self-protection and Q is the level of some pollutant in the
Venvironment. An alternative approach is found in Smith and Desvousges (1987) where they
-------
I
I
I
I
I
I
I
separate the risk of exposure and the risk of illness, and the level ot poianani auccu. u»w »««. ~-
exposure. . .
Each individual has an indirect utility function
V * V[M,H(Z)]
where M is their income and His their level of health. In a two-stale world where HQ is good
health and HI is poor health, consumers maximize expected utility given some level of pollution
(X, Qo)V[M, Ho(Z)] + [l-Q(ic, Qo)]V[M,
Their willingness to pay (WTP) for a small change in Q given self-protection is the difference
between die level of utility in each state divided by the expected marginal utility of income.
V(M,Ho)-(M.Hi)
WTP
Alternatively, a discrete decrease in Q from QQ to Qi is represented as
ic(X, Qo)V(M, Ho) + [1 - *(X, Qo)]V(M. Hi) *
«(X, Qi)V(M - P, Ho) + U - *(X, Qi)]V(M - P, Hi)
where P represents the WTP1 to maintain the initial level of utility at the new level of pollution
(a Hicksian compensating measure of welfare change). Using a variety of benefits estimation
techniques, we can estimate the value of P given self-protection expenditures.
A PROTOCOL FOR HEALTH BENEFITS TRANSFER
The overriding concern for public health behind much of U. S. environmental legislation,
and Executive Order 12291 suggests a significant demand exists, and will continue to exist, for
benefit estimates of reduced risk to health. Evaluation of these benefits will require expensive
and time-consuming projects for each substance and health effect. Benefits transfer may provide
a solution to satisfying the need for benefits analysis for the variety of environmental legislation
and regulation in the U.S. However, the transfer approach poses potential rislcs: poor quality
1 Smith nd Devoosgcs (1987) refer to tins value as an option price.
-------
benefits transfers may lead to incorrect policy choices (Dcsvousges, Naughton, and Parsons,
1992). A sound approach to transfer is necessary.
Benefits transfers apply existing benefit estimates from a study site to a policy site.
Researchers must transfer the issue or commodity from a particular policy site into something
that can be interpreted using existing information (Smith, 1992). What criteria should we use to
transfer health benefits from a study site to a policy site? Table 1 lists our general recommended
approach for a transfer analysis. We focus on Stage 2, Transfer Criteria, in more detail below.
We identify three areas as the primary focus for a transfer protocol: commodity specification,
market and exchange mechanism, and site and sample characteristics. We discuss each below.
TABLE 1. GENERAL APPROACH FOR TRANSFER ANALYSIS
• Define the purpose of the estimates and the level of precision needed.
* Use proposed transfer criteria (commodity, sample, market, site) to describe study site.
• Select an existing benefit study or studies that satisfy the transfer criteria, keeping in
mind estimates' purpose and precision.
• Determine the appropriate transfer method (e.g., point estimate or confidence interval,
function transfer, Bayesian approach, or meta-analysis).
The Transfer Protocol: Commodity Specification
One of the most important steps in a benefits estimation and benefits transfer is careful
specification of the commodity to be valued. How should we define our commodity when
valuing health benefits? Table 2 identifies six areas for clarification in commodity specification.
Response/Causal Agent: Should we define our commodity based on the substance or
the end result (morbidity/mortality or both)? We «tcoT"mg"^ that the commodity in health
transfer studies be defined by die end result, die risk of illness or death. We posit tiiat ultimately
die consumer cares about die health effect (Le., die itchy eyes, coughing, birth defects) and not so
much the source or pollutant that causes die healtii effect. If diis position proves defensible, tiien
benefits transfer exercises become significantly less complicated because we can consider
reductions in cancer risk from exposure to benzene in die air, for example, die same as a
reduction of cancer risk from dioxin exposure in die water. This position, however, may not hold
true for pollution sources mat have variations in avoidance opportunities and, as discussed in
-------
I
I
I
I
I
I
I
I
I
I
I
1
I
TXBLE2. ^rOMMENDEDCOMMODrrVSPEC^CATIONCRITERIA
, t Should we oenne ou* Commodity based on the
Response/Causal agent substance or the health effect?
<»
Are we changing risk through changes in
Riskdefmition ySSSSS^Sit^^^^tO^v^^t
Is there a latency period between exposure and
Temporal dimensions occurrence of health effect?
Exposure pathway
Exposure level
Does exposure occur through water, air, and food,
for example?
Is exposure cumulative or acute?
more detail below, morbidity effects. Thus, the role of me causal agent in risk valuation
responses is an important research issue.
If we base our commodity specification on the end result, the illness, we then should
consider the potential to transfer values across illnesses. For example, can we transfer the health
benefit estimates for a reduction in the risk of death from lung cancer to liver cancer? To best
answer this question let us consider die three general categories for valuation in health benefit
studies: death, illness with no death, and illness followed by death. In me first case, individuals
value mortality alone. A pure morbidity value is provided in the second case and a combined
value in the third. Returning to our question above, an individual may not value death from lung
cancer the same as death from liver cancer, because this is actually a combined value and the
morbidity characteristics may vary across disease. Variation in morbidity across diseases may
include differences in severity or timing for example.
This potential for variation in morbidity characteristics may also cause problems for
trailer across pollutant sources for the same disease. For example, consumers may value
reduced risk of lung cancer from dioxin exposure the same as reduced risk of lung cancer from
asbestos, only if the morbidity characteristics and avoidance opportunities are the same between
causal agents.
-------
Symptoms and the potential for death should be the primary factors used to define the
commodity in a health benefits transfer study. However, the pollutant source may be more
important if avoidance opportunities, or morbidity effects, vary across sources. The cause of the
symptoms, or death (e.g., lung cancer versus liver cancer) may also be important to value
estimates because morbidity characteristics may vary.
Although we have three general categories for valuing health benefits, no studies have yet
valued combined mortality and morbidity impacts. We recommend researchers use mortality
estimates as lower bounds in the absence of combined studies. Because morbidity is already an
element in these measures, adding morbidity and mortality values may result in double counting.
Finally, the units of measurement for the commodity defined are important If health risks are
portrayed as unit days of a symptom, the researcher must consider the problems of over or under
estimation surrounding unit day measures (Morey, 1992).
Risk Definition: Environmentally related health effects can range from acute illness and
discomfort, which may occur with a high probability, to sudden death that may occur with a low
probability. The components of risk include both the probability of a health effect occurring as
well as the severity of that health effect Ehrlich and Becker (1972) recognize that risk can be
reduced by decreasing either element In a laboratory environment, Shogren (1990) found
reductions in probability were preferred to severity reduction. Whether policy changes the
severity of the event or the probability of its occurrence can inf"^?? how consumers value a
change in the overall risk Therefore, when evaluating study and policy sites, researchers must
clarify the component of risk that the proposed policy is changing — probability or severity.
Secondly, considering the direction and magnitude of the risk change is important Does the
probability or severity of the policy under consideration increase or decrease? In the absence of
information on symmetry, researchers should be cautious in transferring the health benefit
estimates from an increase in probability at a study site to a policy site where a decrease in
probability occurs.
Temporal Dimensions: Health effects from environmental hazards range from acute
effects to chronic latent health effects. The temporal dimension of health effects
includes the length of time the illness occurs and the time period between exposure and
occurrence of the illness or death. We cannot assui"ff that consumers will value latent health
effects the same as immediate effects nor assume they would value chronic and acute effects in
the same fashion. Therefore, looking for similarities in the temporal dimensions of the health
effects between the policy site and the study site is important Presumably, temporal dimensions
are similar when the health effect is constant across sites.
-------
i
I
I
I
I
I
1
I
I
I
I
I
I
Voluntary and Involuntary Dimension: Although we have stated that the pollutant or
source of a disease may be unimportant when transferring health benefit estimates, in one case
characteristics of the source become important: the voluntary/involuntary nature of exposure to a
health hazard. Environmental health risks are typically involuntary (a person is unknowingly
exposed) as compared to health risks from smoking, drinking, and driving, for example (a person
chooses to incur the risk). Valuation of voluntary risks may be quite different from involuntary
(Stair, 1969; Starr, 1979); thus they should not be used interchangeably. The distinction occurs
because voluntary risks imply some form of control over die risk, and perceived control can
influence me value of risk reduction.
Exposure Pathway: Although we have ruled out the importance of the pollutant's
source in value estimates, we may find that die exposure pathway affects consumer values. This
effect would become relevant if exposure pathways influence our ability to avoid a hazard or the
voluntary nature of exposure. For example, individuals may perceive greater control over the
quality of their water and food than over air quality.
Exposure Level: Exposure to environmental pollutants can range from short time
periods with high doses to long time periods with low doses. How consumers value a change in
health risk will be influenced by these exposure levels* because they inflwuyr consumer
probability perceptions and time preferences. Therefore, researchers must choose study sites
with similar exposure levels as policy sites for benefits transfer.
Transfer Protocol: Sample and Site Characteristics
Researchers classify sample and site characteristics in two general areas: the
socioeconomic characteristics of the sample and the location and temporal characteristics of the
site. Characteristics that should be highlighted in a health benefits transfer study are discussed
below.
Sodoecomomte Characteristics Sample characteristics such as income, education, age,
awareness of risk, frtisffli"?- health, M^ bip*K"^ risk may affect benefit ftstimatffs Because the
sample in a study site is probably not identical to die policy site, researchers must find study site
value fgtif«««tes ft»"! have well-developed valuation models. These models should include the
socioeconomic factors that influence estimates and thus provide more insight into the
relationship between demographic rh^n>rtCTistics of the sample and values estMMted, Good
understanding and documentation of study site demographics will allow researchers to identify
the sample characteristics that vary across study and policy sites.
I
-------
Location and Temporal Characteristics: lust as socioeconomic characteristics affect
benefit estimates, the researcher must also be aware of certain site characteristics that influence
values. For example, location characteristics possibly important to health benefits estimation
include the presence of insurance programs, access to medical care, potential for avoidance
opportunities, climate, time period of exposure, and baseline exposure levels. The analyst should
establish a relationship between these location and temporal characteristics and the values given
at the study site. As above, reporting of these characteristics for die study site is important
Finally, as with an original benefits estimation study, analysts must consider the size of the
population affected to calculate total benefits.
Transfer Protocol: Market and Exchange Mechanisms
Psychologists discovered that alternative means of framing a problem can systematically
influence choice and values (e.g., Tversky and Kahneman, 1981). Three important factors
regarding framing effects of a risk valuation problem are the risk reduction technology, the
exchange medium, and the type of question (WTP/willingness to accept fWTA]). Finally, an
additional market issue is the presence of nonuse values in the market The importance of these
issues for benefits transfer is discussed below.
Risk Reduction Technology: Evidence suggests that alternative risk reduction strategies
influence valuation. Individuals can produce a given reduction privately or collectively.
Individual preference for private or collective reduction depends on the payment's perceived
productivity. Collective reduction may prove more efficient given scale economies, because
many private actions are too expensive or complicated to be economically feasible (Shogren,
1990). However if excessive free-riding is perceived, private reduction may be valued more
highly. Thus, determining the risk reduction strategies most appropriate for the policy site is
important Figure 2 illustrates the individual's choice of risk reduction actions.
Exchange Medium: One of the most important factors in designing a valuation study is
the exchange medium (or "payment vehicle"). Consumers can pay to reduce the risk of adverse
health effects through wages, taxes, or prices. The medium can influence values given; thus
using a realistic medium for the policy site is important for both benefits transfer, as well as
original benefits studies.
Nonuse Values and WTP/WTA: Analysts must determine whether nonuse values are
relevant and what welfare change measure is appropriate for the policy site. Nonuse values
include the health effects of children, other relatives, neighbors, and friends. Consumers may
value the health of others as well as their own health. However, the extent to which these nonuse
8
-------
e «
>
UJ
-------
values may be embedded within current value statements given by individuals is unclear.
Although not readily available, some measure of nonuse values might be appropriate in health
transfer studies.
Selecting between Hicksian compensating and equivalent measures and using WTP or
WTA depends on the property rights allocation and the direction of the policy change for the
particular policy site. Therefore, well-defined property rights and risk reduction should be
consistent across the sites. Otherwise, extrapolating one value measure for another is
questionable given the theoretically predicted and empirically observed divergence in WTP and
WTA for improved health quality.
Study Selection
Following the transfer protocol suggested above, an analyst can select the study sites
most appropriate for valuation at the new policy site. We recommend that existing contingent
valuation method (CVM) studies be given priority because the alternative approaches have an
array of problems. CVM studies are preferred because of their potential to capture morbidity and
the diversity of possible samples (i.e., general population versus white mate workers).
If CVM studies are unavailable, we recommend the few averting behavior studies and
experimental laboratory studies. Hedonic wage models are given a lower priority because of the
narrow sample group and the focus on risk of accidental death. Cost of illness is given the
lowest priority because of its weak theoretical underpinning.
Additional selection criteria may include the theoretical soundness of the study, level of
information reported, and purpose of estimates and level of precision required. Of course the
study site should match policy site specifications to a level the researcher considers acceptable.
A CASE STUDY: LONG-TERM HEALTH RISKS FROM SURFACE WATER
POLLUTION
A classic case of exposure to a long-term health risk is found in Western North Carolina.
Champion Paper currently discharges approximately 43 million gallons of coffee-colored
wastewater into the Pigeon River daily. In addition to the discoloration, a potentially more
serious problem is the risk to public health from the dioxin and other toxins present in the
discharge. The state of North Carolina is considering a weakening of the maximum allowable
dioxin limit of 14 parts per trillion (ppt). What are the benefits of maintaining the limit or the
costs of raising the limit? This case study provides a working example of the need to transfer
benefit estimates and the many potential problems for the valuation of changes in long-term
health risks from surface water contamination.
J
]
10
-------
I
I
I
I
I
I
I
I
I
1
I
The Site: The Pigeon River originates in Haywood County, North Carolina, as a pristine
stream in the Pisgah National Forest. The river flows north, 10 miles, to Canton, where
Champion paper discharges their effluent The river continues northwest, 16 miles, crossing the
Tennessee state border past seven small communities in both states until it reaches Newport, in
Cocke County, Tennessee. Thirty-six miles from the mill, the river empties into Douglas Lake.
The 1990 mean flow rates, north of Canton, vary from a low fall flow of 88 cfs to a high of
10,900 cfs in the spring. The river is regulated by Lake Logan and Walters Lake.
The Pigeon flows through mountainous terrain between the Great Smokey and Bald
Mountains. The river above Canton is used both as a municipal drinking water source, rated
WS3, and for recreational activities such as swimming, boating, and fishing. Downstream from
Canton, the river has been rated as Class C water for boating and fishing only; immersion is not
recommended. A posted advisory recommends against eating fish caught in the river north of
Canton. The 10-mile stretch from Walters Lake to the state line is considered a good "brown"
water rafting run and is sometimes used by recreationists in the area. In Tennessee, the river is
classified and protected for industrial water use, fish and aquatic life, recreational activities
including swimming, irrigation, and livestock and wildlife watering. But, because of the present
level of discharge the river does not meet state requirements for aquatic life or recreational uses.
Tennessee has posted a warning against eating fish from the river. In addition, the present high
color level prohibits any additional waste discharge; thus the river is not used for any other
industrial discharge in Tennessee.
Water Contamination: In 1989, industrial water use accounted for 8S.6 percent of
water used in Haywood County. Fifty-one percent of industrial water is used by Champion
Paper in a pulp mill?, paper naXi? and their utilities and filter plants.4 They produce food board
and fine paper using an integrated bleached kraft pulp and paper manufacturing process.
Pollutants pTMfnt in
argf hi fithfr ni
n*1 **$*&***& **y EPA
given in Table 3. In addition to the pouutants in Table 3, die discharge also affects the stream's
temperature and acidity. The average winter effluent temperature is 29.8°C and die summer
temperature is 37.9°C. Acidity levels range from pH 6.4 to &2. EPA temperature limits for
effluent are between 29°to 32°C, with a 13°C maximum increase in stream temperature. The
acidity limits are pH 6 to 9.
2Indndes drip coating,
3 Produces fine paper, food bond, and dried pulp.
*EPA Fora* 1 and 2C submitted by J. R. Kuptthcfc to EPA Region IV Office. Admta, GA.
11
-------
TABLE 3. DISCHARGE POLLUTANTS FOR CHAMPION FA»>t,K
CANTON, NORTH CAROLINA (1989)
ii>
1989 Sample Values
Effluent Characteristic
Biochemical Ox Demand (5 Day)
Total Suspended Solids
Fecal Coliforro
True Color
2,4,6 Trichlorophenol
Pentachlorophenol
Zinc (one sample)
Chloroform (w/ plant modification)
23,7,8 TCDD (dioxin)
23,7,8 TCDF(furans)
Daily
Average
12.5 mg/1
1 1,331 Ibs/day
50/100 ml
1,043 std. units
Daily
Max
44.4 mg/1
38,449 Ibs/d
650/100 ml
2,035 std. units
< 10 US/I
<50|ig/l
80»ig/l
238 mg/1
6.61 pg/1
5.62 pg/1
Daily Average
Standard Limits
30mg/l
42,012 Ibs/d
200/lOOml
50 std unit
3.3mg/l
0.014 pg/1
1
Commodity Specification: Long-Term Health Risks from Dioxin
Response/Causal Agent: Dioxin exposure causes a range of health risks from life-
threatening cancers of the soft tissues to nonlife-threatening skin problems, fertility problems,
and birth defects.5 In addition, evidence suggests dioxin can cause immune system suppression
in mice at low dose levels, and it is a known promoter of other carcinogens.6 Dioxin can
contaminate the air, water, and soil, and exposure occurs through three possible pathways:
inhalation, absorption, or ingestion. Dioxin is more easily absorbed in small doses.
Increasing the exposure levels of dioxin may increase the risk of immunosuppressant
health effects,7 and if accumulated exposure levels increase,8 the population may have a risk of
cancer. Therefore, we may specify our commodity as a particular set of symptoms such as
increased disease days from failure of the immune system to fight colds, flu, and other common
'See Schmidt (1992).
«See Schmidt (1992).
7See Schmidt (1992).
'Dioxin hy a long half-life, causing potential •Ttf!i?*at*nT> in the body.
12
-------
I
I
I
I
I
ailments, and as an increase in the risk of chronic illness. We may also specify the commodity as
an increased risk of cancer mortality.
Elevated cancer mortality risk is evident in the health statistics for the area. Both
Haywood and Cocke Counties have cancer rates greater than the national average (see Table 4).
Cancer mortality rates for the two counties range from 7 percent to 35 percent greater than the
national average,9 Chemical workers exposed to dioxin in the U.S. and Germany have been
found to have cancer mortality rates 15 percent to 24 percent greater than their national averages
for all cancers. In the U.S. those with long-term exposures to dioxin at chemical plants had rates *
87 percent above normal in one study and nine times higher man the general population in
another.10
I
TABLE 4. AGE-ADJUSTED CANCER MORTALITY RATES (PER 100,000 PERSONS)
Yew
1979 -1981
1982-1984
1985 -1987
1988 -1990
Haywood*
135.44
167.89
179.24
NA
Cocke
14U
158.4
153.7
151.9
132.0
133.0
132.7
133.7
I
I
I
•Ibeie data ire quoted for yon 1979 through 1981.1981 through 1985.1984 through 1988.
*US. data are for yean 1979.1981.1984. and 1989. respectively.
*
Risk Definition! The policy BiHfpr co*MiK*ffration (increasing the irn>Tiirr"Ttn exposure
limits) affects the probability of exposure and thus the probability of immune suppression health
effects, as well as the r^obabflity of cancer mortality.
Temporal Dimension: Although the immune system effects occur soon after exposure,
cancer has a latency period. The immune system problems persist as long as a potent level of the
\
''Voett figures do not coucci for otter
exposure
be attributed toldy to dioxio
13
-------
chemical remains in the body and thus cause chronic problems given the long half-life of the
chemical11 The cancers arc also chronic.
Voluntary or Involuntary: Exposure to the hazard in our case study is both voluntary
and involuntary. Paper mill workers and those who live in the communities surrounding the mill
voluntarily expose themselves to the hazard, assuming they are aware of the chemical's
presence.12 Although we recognize their relative ability to relocate, downstream residents are
involuntarily exposed.
Exposure Levels: The policy site population has been exposed to low dose levels for
long time periods. Present exposure levels for the communities surrounding the mill and the
downstream communities are considered low. Mill workers, however, may have higher exposure
levels. A July 1989 EPA Fact Sheet (EPA, 1988) on the Pigeon River in North Carolina reported
dioxin levels in fish fillet samples of 2.3 to 80 ppt and wholefish levels of 36 to 91 ppt In
Tennessee they found 0.17 to 29.3 ppt in fillets.13 The NC state limit for dioxin is 0.014 pg/1 or
14 ppt
Both states have given advisories against eating fish from the Pigeon River, and neither
state has classified the river for use as domestic water supply. Residents along the river or users
of the river have had a lifetime of exposure if they have any regular contact with the river, for
example, through recreational activities such as fishing and boating or through drinking from
contaminated wells. Tests performed in 1987 by the Tennessee Health Department found toxins,
such as furans, contaminating wells of Hartford residents.
Policy Site and Sample Characteristics
Sodoeconomic: Both Cocke and Haywood Counties are rural areas. Table 5
summarizes tile 1990 demographic data for these two counties.
Location and Temporal: Both government and private insurance programs are
available to consumers in both counties; medical care is similar to that available in rural areas in
theU.S. Exposure has occurred over a period of 80 years, the time frame in which the
UAB references to humane system problems are presently hypo&etkaltacaose evidence of thuhealm effect h«
only teen found in nice.
l*T^tt*aa^tiieie^cmaiiA^cm^aAtowawmaAootcvAlD, Ite coos of moving could be seen ss
conservative estimates for benefits of berth risk reductions. See averting behavteHienure(Abdina. Ranch,
and Epp. 1992).
13Tbe higher levels were obtained from the whole body of a booon-fceding white tucker. Tests of surface-feeding
nnfish yielded a dioxin level of 12 ppt llievah^aiinn^kvelsoouUbeiwrtiaUyeiujdainedbyfoodsoinxe.
14
-------
I
I
I
I
I
I
I
I
I
I
i
I
I
I
I
I
I
I
TABLE 5. 1990 DEMOGRAPHIC INFORMATION
AND COCKE COUNTY, TN
Haywood County,
North Carolina
Population
* *'* »
Mean Household Income
MffOT Education '
Male/Female Distribution
Racial distribution (W/B)
Age Distribution
>65
Median Age
Household Size (mean)
poundaries for me maikefs
M.rkrt and Exchange Medi
water quality.
46,942
$22,698
12.1
47/53
98/1.4%
184%
20.8%
39.9
2.4
15
Codec County,
29,141
$17,1624 -
12 (median)
48/52
97/2.1%
12.9%
24.0%
35.2
Z58
-lhe
-------
the high incineration necessary for cleaning up toxic
Collective action appears to be the most likely cleanup strategy for source reduction. Individuals
can, however, pursue private averting behaviors such as purchasing bottled water or avoiding the
river for recreational activities such as swimming and fishing. When transferring values we may
consider either collective action or private action values, but the latter may not reflect reduction
in the substance from all pathways (i.e., air. water, and soil).
Exchange Medium: The policy site medium would likely be a city water price or taxes,
both of which can be applied to a collective reduction strategy.
Nonuse Values and WTPAVTA: Nonuse values arc likely present for children,
relatives, and possibly others for both the morbidity and mortality impacts. At the policy site,
communities have the property right to clean water, bui the Mule is responsible for enforcement
of that right Citizens must convince their government of their preferences; thus we would
measure a consumer's WTP to avoid an increase in the dioxin limit (a Hicksian equivalent
measure of welfare change).
Benefits Transfer: Valuing the Benefits of Maintaining 14 ppt Limit on Dioxin
In this case study we want to estimate the ex ante economic value to avoid an increase in
dioxin limits. Because we have defined our commodity as the probability of morbidity and
mortality effects from long-term low dose levels of exposure, we are estimating the value of
avoiding an increase in the probability of chronic morbidity or cancer mortality, or both.
A significant amount of research estimates economic values for a reduction in the risk of
morbidity or mortality (Gegax, Gerking, and Schulze, 1991; Gerking and Stanley. 1986; Smith
and Desvousges, 1987; Viscussi, Magat, and Huber, 1991). Other studies, such as Berger et al.
(1987), provide economic values for symptom-free days. Many of these studies have focused on
short-term risks where the time between the cause and effect is immediate (accidental death) and
on acute health effects such as bums and coughs. Few studies have looked at the chronic and/or
latent health effects characteristic of our policy site. Using the criteria suggested earlier, we
selected four studies as potential study sites: Viscusi, Magat, and Huber (1991); Gegax,
Gerking, and Schulze (1991); Smith and Desvousges (1987); and Berger et al. (1987)." Table 6
summarizes the characteristics of these studies.
J
15Given the shortage erf morbidity studies available, UK Berger et al. (1987) study was selected although it focuses
oo short-term morbidity effects.
16
-------
1
I
I
I
I
1
,8
r
«*
II
ill*
-R
£
a
•s
io
I
3
I
i
I
zzzzzzzz
* 5 S N 13 ^ fi
*"* & E? ^ _ ts <**
-------
£
>•
•a
o!
[00 «•* 9
8
•5
f
I
3
18
-------
I
I
I
I
I
I
I
I
I
1
Both Viscusi, Magat, and Huber
(1987) are CVM morbidity studies, while the Gegax, Getting, and Schulze (1991) and Smith ana
Desvousges (1987) are mortality studies. Note Gegax, Gerktng, and Schulze is a hedonic wage
study and Smith and Desvousges is a CVM study. The Gegax, Coking, and Schulze study
measures WTA for an increase in perceived risk of accidental death. The other studies measure
WTP for decreases in the health risks (Viscusi, Magat, and Huber and Smith and Desvousges)
and WTP to get an increase in symptom-free days (Berger, Blomquist, Kenkel, and Tolley). .
Which study should we use for benefits transfer?
Study Selection: Given the specification of our commodity we choose Viscusi, Magat,
and Huber (1991) and Smith and Desvousges (1987) as our possible studies. Both studies value
chronic or latent health effects, which are similar to the same effects from dioxin exposure.16
Smith and Desvousges (a mortality study) and Viscusi, Magat, and Huber (a morbidity study)
provide demographic information and a sensitivity analysis of their results. Both also value a
change in probability not severity. Table 7 compares the Viscusi, Magat, and Huber and Smith
and Desvousges study sites with our policy site.
Although several characteristics of the study sites make mem appealing for a benefits
transfer, the sites also have several important problems. First, a critical problem is the difference
in the direction of change for the study sites and our policy site. Viscusi, Magat, and Huber
(1991) looks at risk decreases; Smith and Desvousges (1987) look at bom increase and decreases.
Smith and Desvousges find that consumer values are higher for WTP to decrease risk than WTP
to avoid an increase.
If we agree with their findings, we can consider the study sites as upper bound estimates.
Second, the policy site includes both chronic morbidity and latent mortality effects, while the
study sites include only one or the other. As recommended above, mortality figures may be
considered lower bounds. Therefore, we might consider the economic values from both study
sites as upper bounds but also consider the Smith and Desvousges (1987) study values as lower
bounds. The transfer is imprecise because no benefit estimate applies perfectly. The analyst
must now recall the purpose for the estimate and determine the need for accuracy. Finally, must
we adjust for the demographic differences in education and income levels at the policy site?
sges (1987) study potability of death tram expontt to hi
Vfecnii, Mactt,
dta^exponre, any erf the syiqMcra my testate
Although chrtnk bronchitis nuynca be iipe^
19
-------
TABLE 7. COMPARISON OF POLICY SITE TO STUDY SITE
Morbidity/Mortality
Rlskdef.
Temporal dimensions
Voluntary/involuntary
Exposure pathway
Exposure level
SocioecoDomlc
Household income
(mean)
Years education
Male/female
distribution
Racial distribution
W/B
%>65
Household size
% Households with
children < 18
Exchange Mech.
ReducLTech.
Nonuse
WTP/WTA
Viscusi et al., 1991
Morbidity
• Probability of chronic
bronchitis (
-------
^to « «tnnsfening •» «t»uon Ul"
V^T^er- ^^r££^-*-^'«"'i*<'T».
—-
! route. -«tt for the Smith M»d Desvousges (1987)
i
i ^S^-^r-tr^—-
I W^" ,M,,«edfiC«6»«-«»*e'-ed-OB"aidyVdlieS
1 ^^^^sssssss^
KU-*-**-** ^^...^.mowxx^^
»^-»^r^r«»*-s2?--.
r^^zs&-^rj^rssr-r-
The Smith and Desvou
death risk increase range from $17.71 to $47.47. The Viscusi, Magat, an
observation values range from $1.50 to $80.00 per 1/100,000 decrease in probability of chronic
bronchitis, with a mean of $8.83. The Smith and Desvousges and Viscusi, Magat, and Huber
probability levels are significantly different with Smith and Desvousges levels ranging from
conditional probability of death of 1/10 to 1/300. Given mat thcsr two studies use different
approaches and our concerns for double counting raised earlier, these values should be neither
or added together.
d
\
i
• approacocs •»» ~— -—
, compared nor added together.
» .. .1 _. u^»hi ctiidie!
I
I
I
I
I
•
I
Recall that both studies* estimates may be considered upper bouns.
Desvousges (1987) study uses probabilities higher man nose we might expect for dioxin, me
nd Huber (1991) study is valuing acute morbidity effects that may be more
d Viscusi. Magat, and Huber
Viscusi, Magat, and Huber
severe man the acute effects expected from dioxin exposure, and Viscus. a,
measures values for risk reduction. In bom studies the demographics may also suggest higher
values for the study sites due to higher levels of income and education.
.
Iteiidcarexparareisab
Iteiidcarexparareisabo
"See Smith ttdDesvooige* , .
tKfcQdtiortii&fetecanaiiiaMlpriba^
•cataUjyljOOO.
l^tecosi Mas*, ^ Huber (1991) do »i«*hivtty«aJyttt^
terignffitm Atr^erfu*ak»i»prob«yiw»a»bkdirectfy
-------
The actual choice of a WTP figure must depend on the researcher's policy needs. If only
rough estimates are required, the above studies may provide adequate guesses. However if more
precise measures are needed, researchers may wish to conduct an original benefit estimation
study.
CONCLUSIONS, LIMITATIONS, AND FUTURE RESEARCH NEEDS
Benefits transfer is significantly more difficult to apply man to discuss in theory. The
most important limitation is the difficulty in finding reasonably similar commodity specification
between the new policy and old study sites. The variation across studies in commodity
specification makes transfers difficult To ease this problem we suggested assuming the causal
agent does not matter. However, in our study the variation in direction and magnitude of
probability change, the severity of health effects, and the appropriate welfare measure posed
significant challenges for transfer. Exacerbating this problem is the singular focus of studies on
either morbidity or mortality. Although most long-term health risks from environmental
substances include both categories of health risks, the relationship between them has not been
examined in the literature. Aggregation through the independent valuation and summation of
mortality and morbidity impacts may introduce a systematic bias in estimates (Hoehn and
Randall, 1989). This topic is important for future research.
After the above limitations have been adequately addressed, we can then turn our
research focus to the relationships between the demographic, location, and temporal variables to
value estimates. Further research might also include more studies in developing nations to
enhance our understanding of demographic and cultural variables on economic values and our
potential for international transfers. In addition, the role of prior information on values and
Baysian exchangeability should be studied in more detail (Atkinson, Crocker, and Shogren,
1992). The importance for benefits transfer of documentation and presentation of demand
equations cannot be overstated. A collective effort to organize existing studies and databases is
needed to enhance researchers' ability to conduct transfers.
Further study of disease attributes, causes, and source as they relate to values is
warranted. Can we use hedonic methods to evaluate the relationship between disease attributes
and values? Finally, researchers' have not exhausted the various questions surrounding valuation
methodology as applied to health risk values nor the potential for nonuse values.
22
-------
I
\ BIBLIOGRAPHY
BIBLIOGRAPHY
!Abdatta,C.W.,B.A. Roach, and D.J. Epp. 1992. Valuing Environmental Quality Changes
Using Averting Expenditures: An Application to Groundwater Contamination." La
Economics 68(2):163-169.
\ Champion. April 27, 1989. Letter to U.S. EPA and attached U.S. EPA Forms 1 and 2C.
J.Shogren. 1992. "Bayesian Exchangeability, Benefit Tra
3:7 15-722.
J
and Research cenc.
Berger,M.,G.Blomquiit,D. Kenkcl, andG.ToUcy. 1987. "Valuing Changes in Health Riss:
1 A Comparison of Alternative Measures." Southern Economic Journal 53567-984,
Center for Health and Environmental Statistics. 1988. "Leading Causes of Mortality: North
I Carolina Vital Statistics.*' 2.
Cropper, M., and F. Sussman. 1990. Valuing Future Risks to Life." Journal of Environmental
. Economics and Management 19:160-174.
' Cropper, M.L., and W£.Oates. 1992. "Environmental Economics: A Survey." Journal of
Economic Literature 30(2):675-740.
I Desvonsges. W.H., M.C. Naughton, and G JL Parsons. 1992. "Benefit Transfer. Conceptual
Pmhtemit fo tempting ^fwif Quality P^TVgfits tfcmg Existing SfrMHcs ** Wafer
• Resources Research 28(3):675-683.
* Dickie, M., and S. Gerking. 1991. "Willingness to Pay for Ozone Control: Inferences from the
Demand for Medical Care.** Journal of Environment Economics and Management
I 21:1-16.
"Dioxiji Re-examined " 1991. Tne Economist 87.
I "Dioxin Via Skin: A Hazard at Low DosesT 1989. Science News 141.
Ehrlich, L, and G. Becker. 1972. "Market Insurance, Self-Protection, and Self-Insurance."
\ Journal of Political Economy 80:623-648.
TDA Finds Dioxin in Milk." 1989. Science News 165.
1 Fisher, A., L.G. Chestnut, and D.M, Violette. 1989. "Ihe Value of Reducing Risks from Death:
A Note on New Evidence." Journal of Policy Analysis and Management 8188-100.
1 Gegax, D., S. Gerking, and W. Schnlze. 1991. -Perceived Risk and the Marginal Value of
Safety.** Review of Economics and Statistics 73389-5%.
I Gerking, S., and L.R. Stanley. 1986. "An Economic Analysis of Air Pollution and Health: The
Case of St Louis.** The Review of Economics and Statistics 68(1).
1 Gould, M. 1988. "The Most Potent CarcinogenT* Resources ¥12-5.
Hayes, D.,etaL 1992. Valuing Fcvd Safety in Experimental Auction Markets. Mimeo.lowa:
I Iowa State University.
I
-------
Health Planning Commission. 1992. Picture of the Present. Nashville, TN: Tennessee's Health
Services, Parts 1 and 2.
"
Hoehn, JJP., and A. Randall. 1989. 'Too Many Proposals Pass the Benefit Cost Test
American Economic Review 789(3):544-551.
Jones-Lee, M. 1974. 'The Value of Changes in Probability of Death or Injury." Journal of
Political Economics 82:835-849.
Kamrin, M., and P.W. Rodgers. 1985. Dioxins in the Environment. Washington: Hemisphere
Publishing Corp.
Kask, S., and S. Maani. 1992. "Uncertainty, Information and Hedonic Pricing." Land
Economics 68:170-184.
Krupnick, A., and M. Cropper. 1990. 'The Effect of Information on Health Risk Valuation."
Working paper. Washington, DC: Resources for the Future.
Magat, W., W.K. Viscusi, and J. Huber. 1988. "Paired Comparison and Contingent Valuation
Approaches to Morbidity Risk Valuation." Journal of Environment Economics and-
Management 15:395-411.
Morey, E. 1992. "What is Consumer's Surplus Per day of Use! And What Does It Tell Us
About Consumer's Surplus?" Paper presented at the 1992 AERE Benefits Transfer
Procedures, Problems, and Research Needs Workshop, Snowbird, UT, June 3-5.
Morrison, C. 1989. "Pollutants No Threat, Champion Official Says." Asheville Citizen Times,
Ashevilk NC, March 8,1989, No. 61.
Raloff, J. 1989. "Dioxin: Paper's Trace." Science News 135:104-106.
"Roast Those Dioxins Away." 1988. Science News 134:158.
Schmidt, K.F. 1992. "Dioxin's Other Face: Portrait of an Environmental 'Hormone.'" Science
News 141:24-27.
Shogren.J. 1990. 'Impact of Self-Protection and Self-Insurance on Individual Response to
Risk." Journal of Risk and Uncertainty 3:191-204.
Shogren, J., and TD. Crocker. 1991. "Risk. Self-Protection, and Ex Ante Economic Value,"
Journal of Environment Economics and Management 20:1-15.
Smith, V.K.. and WJi Desvousges. 1987. "An Empirical Analysis of the Economic Value of
Risk Change." Journal of Political Economy 95:89-115.
Smith, V.K. 1992. "On Separating Defensible Benefit Transfers From 'Smoke and Mirrors *"
Water Resources Research 28(3):685-694.
Starr, C. 1969. "Social Benefit Versus Technological Risk." Science 165:1232.
Starr, C. 1979. Current Issues in Energy: A Selection of Papers by Chauncey Starr. Oxford-
Perganon Press.
J
J
I
I
I
I
I
I
I
I
J
1
24
-------
-------
[
State Crater for Health Statistics. 1985. Leading Causes of Mortality: Nor A Carolina Vital
Statistics. Vol. 2,1985 and 1981.
State of Tennessee Department of Health. November, 1991. "Annual Bulletin of Vital Statistics
for Year 1989." Nashville, Tennessee.
Strum, Carol Van, and Paul E. Merrell. 1988. Reproductive Risks from Consumption of
Dioxin- and Furan-Contaminated Milk Packaged in Paper Containers. Tidewater, OR:
Alder Hill Associates.
"Survey of Buying Power Demographics, USA" 1991. Sales and Marketing Management
Magazine.
This'Nontoxic* Dioxin Isn't" 1988. Science News 133.
Tversky, A., and D. Kahneman. 1981. "The Framing of Decisions and the Psychology of
Choice" Science211:453-458.
Two E.P.A. Studies Confirm Threat to Fish of Dioxin from Paper Plants." The New York
Times, The Environment, March 14, p. 23.
U.S. Environmental Protection Agency. 1988. Assessment of Dioxin Contamination of Water,
Sediment, and Fish in &e Pigeon River System (A Synoptic Study). EPA Report No. 001.
U.S. Environmental Protection Agency (EPA). 1989a. "Authorization to Discharge Under the
National Pollutant Discharge Elimination System, Permit #NC0000272." September.
U.S. Environmental Protection Agency (EPA). 1989b. "Fact Sheet Amendment Based on
Comments Received My 12,1985 to August 25,1989." September.
U.S. Environmental Protection Agency (EPA). 1989c. "Fact Sheet Application for National
Pollutant Discharge Elimination System. Permit to Discharge Treated Waste Water to
U.S. Waters " Appl. 4NC0000272, July.
U.S. Environmental Protection Agency (EPA). 1989d. "Public Notice #89NC003." September.
U.S. Geological Survey. 1991. "Water Resources Data, N.C, Water Year 1990." Water-Data
Report, NC-90-1.
Viscusi,W.K.,W.Magat,andJ.Huber. 1991. "Pricing Environmental Health Risks: Survey
Assessment of Risk-Risk and Risk-Dollar Tradeoffs for Chronic Bronchitis." Journal of
Environmental Economics and Management 21:32-51.
25
-------
I
-------
RECREATIONAL FISHING VALUATION: ACID RAIN
PROVISIONS OF THE CLEAN AIR ACT AMENDMENTS
Mary Jo Kealy, Susan Herrod, George Parsons, and Mark Montgomery*
ABSTRACT
Our work group developed a research protocol to assess the likely magnitude of the
economic benefits of improved or nondegraded recreational fishing mat are expected to result
from implementing the Clean Air Act Amendments of 1990. We used data for the study site
from the 1990 NAPAP Integrated Assessment, which includes Maine, New Hampshire,
Vermont, and New York. The policy site includes Pennsylvania, Virginia, West Virginia,
Maryland, New Jersey, and Delaware.
\
Congress mandated in §812 of the Clean Air Act Amendments of 1990 (CAAA) that
EPA conduct a comprehensive analysis of the impact of the CAAA on the U.S. economy, public
health, and the environment This analysis is to include costs, benefits, and other effects
associated with compliance with each standard issued for emissions of sulfur dioxide (SO2) and
nitrogen oxides. Title IV of the CAAA mandates a reduction in SOj emissions of 10 million
tons per year, with a national cap on SO2 taking effect in the year 2000.
With the reduction in these precursors to acidic deposition, water quality improvements
are expected. A potentially significant source of economic benefits from improved water quality
is enhanced recreational fishing. This case study involves developing a research protocol to
assess the likely magnitude of the economic benefits of improved or nondegraded recreational
fishing that are expected to result from the implementation of the CAAA to control precursors of
acidic deposition.
Although substantial improvements (nondegradations) in water chemistry and fish
populations may be attributed to the CAAA for three regions of die country (Le., Adirondack
region in New York, Mid-Atlantic Highlands, and Mid-Atlantic Coastal Plains), a preliminary
economic assessment has been completed for the Adirondacks only. This area together with
*U.S. Environmental Protection Agency. VS. Environmental Protection Agency. University of Defame, GrianeB
College, respectively. Members of tfae ere sbify group aKtoattTrady&nertt
Angeles), JenOd Fletcher (West Virginia University). Myrick Freon* (Bowdoto CoDe|e). Don Gmw
(Environmental Law mainrteX Reed Johnson (Research Tringle Insure), Dona Ltwson (NOAA Damage
Assessment Center), GregMidads (Abt Associates, Inc.), Andrew Mute (McMoaer University), State Navmd
(Noragric, Agricultural Univenity of Norway), and Robert Unsworn (Industrial Economics, lac.). The views
ejo?rcs^ by teuton of jhis paper to not iiecessa
Agency. Responsibility for errors nd "*"*****¥!£ renuins with the authors.
-------
three other northeastern states (Le., Maine, New Hampshire, and Vermont) was studied as part of
the National Acid Precipitation Assessment Program (NAPAP) and preliminary results were
included in the 1990 Integrated Assessment .
At the time of the Assessment, the Adirondacks (and the rest of the Northeast) was the
only affected region of the country for which all of the linkages from emissions to fish
population declines were established. Therefore, the limited resources for the economic analysis
were devoted to assessing damages to the recreational fishery in this region. Finally, the analysis
was limited to losses to anglers, and researchers made no attempt to assess any potential nonuse
values associated with die changed water chemistry and biota. The contingent valuation method
of assessing nonuse values was considered too controversial to survive die NAPAP peer-review
process.
The Assessment's National Surface Water Survey (NSWS) encompassed all of the
regions of the country thought to suffer advene water chemistry conditions from acidic
deposition. However, the Assessment ascertained that only the Adirondack and Mid-Atlantic
regions have potentially high losses in waters suitable for die survival of certain fish populations.
The economic losses to recreational fishermen in the Mid-Atlantic regions still need to be
assessed. Unfortunately, the linkages from emissions to fish populations are less definitive for
the Mid-Atlantic regions, particularly die Coastal Plains, man for die Adirondacks. Moreover,
relative to the costs of controlling emissions, die benefits of improved fish habitat and
populations are likely to be quite small so that a full-scale original study may not be warranted.
However, two arguments can be made for a less ambitious analysis. First, on a regional
scale, die damaged conditions of die fishery may represent a significant loss and a
disproportionate burden. Second, recreational fishing damages from fish population losses are
but one effect of acidic deposition to be considered along with other damages such as, health
effects, impaired visibility, and materials damages. Note tiiat wim die probable exception of
health effects, each of these effect categories inffo^ff OTCT and nonuse values diat are affected by
acidic deposition. Therefore, although a full-scale original study of recreational fishing in die
Mid-Atlantic region may not be warranted by definitive science or die relative costs and benefits
of §812 of dttCAAA, a less ambitious HMcspftfiit of die likely extent of damages is appropriate
in tiiis policy context One of die goals of tiiis benefit transfer research protocol exercise is to
describe die extent of analysis required by die policy context
Consistent wim die Assessment, die research protocol described here does not address
nonuse values. That topic warrants separate treatment and is beyond our scope.
•
f
i
i
-------
THE BENEFIT TRANSFER RESEARCH PROTOCOL
i
t
i
i
A benefit transfer can involve a fairly simple practice such as applying estimates of
benefits from one study to an entirely new situation. If multiple, related studies are available,
researchers may construct weighted averages of benefit estimates. The original functions that
generated the benefit estimates can themselves be transferred, and available data from the policy
site can be used in place of the means from the study sites to simulate the models. Ever
increasing levels of effort can be directed toward methods of assembling, analyzing, evaluating,
combining, and interpreting existing information on how people are affected by a change in
conditions, and these methods all qualify as benefit transfers.
In this paper, we develop a benefit transfer protocol for exploiting existing data collected
in an original study, rather than the values or functions estimated from these data. By having
access to the data, researchers are not restricted by the modeling assumptions of the original
study. Furthermore, we can consider methods of combining the existing data with data from the
policy site.
The four types of data needed in an assessment of recreational fishing benefits are
• behavioral data (e.g.. where do anglers fish and how often?);
• population and angler characteristics (e.g., income, age, tastes, and attitudes);
• site characteristics (e.g., fishing quality, size of the water body, cost of access,
geographic distribution of waterbodies by type and in relation to the angling
population); and
• policy variables (e.g., fish catch rates, presence of fish species, Acidic Stress Indexes).
Our original data for the study site are from the 1990 NAPAP Integrated Assessment, which
includes Maine, New Hampshire, Vermont, and New York (Shankk et aL, 1990). The policy
site includes the Mid- Atlantic states of Pennsylvania, Virginia, West Virginia, Maryland, New
Jersey, and Delaware. The data from die Northeast on recreation behavior, site characteristics,
population and angler characteristics, and policy variables, may be used alone or in combination
with policy site data on these parameters. Presently, population characteristics are readily
available for the Mid- Atlantic regions, and we anticipate die future availability of some policy
site data on angler characteristics and recreation behavior (e.g.. National Recreation Survey).
Site characteristic data exist for die policy site, but accessing diese data and linking diem with
die recreation behavior model is a labor-intensive task. Finally, aggregate data on die range of
changes in die policy relevant variables are available in die policy region, but tiiese data may not
import well into die recreation behavior models dial rely on "rite"-specific data.
-------
We develop a benefit transfer research protocol that breaks the analysis down into stages.
The progression from one stage to the next is based on a value of information analysis similar to
the one presented in Deck and Chestnut (1992) and based on Freeman (1984). The titles for
some of the stages of the research protocol have been generalized, however, to accommodate our
more encompassing interpretation of the types of analyses that qualify as "transfers." At each
stage of the analysis, we attempt to evaluate the benefits and costs of proceeding to the
subsequent stage. We based the decision on the cost of obtaining increments in the quality of
benefit information relative to an assessment of how important the quality increment is to the
policy context Finally, in our conclusions we suggest some changes in the way we do empirical
research to make benefit transfer practical as well as defensible.
Stage 1 begins with the Qualitative Assessment of the economic significance of the
damaged recreational fishery. Assuming significant damages have occurred and the policy will
result in a reduction in damages, the Transfer Scoping Analysis is designed. The purpose of
this second stage of the exercise is to assess the availability and relevance of existing information
(e.g., studies, reports, databases). The third, or Benefit Transfer Computation/Estimation,
stage is to determine how best to synthesize, analyze, and otherwise interpret the relevant
information to quantify the economic benefits associated with the policy. Here, we attempt to
specify and estimate recreational fishing demand models using study site data (i.e., the
Northeastern states) alone or in combination with other available data sources. If these data
sources are inadequate for providing credible estimates of the recreational fishing benefits of
reductions in acidic deposition in the Mid-Atlantic Highlands and Mid-Atlantic Coastal Plains,
then, moving to the fourth stage may be necessary. The Update/Validate stage involves at least
some primary data collection (e.g., a pilot study) and model estimation, most likely using
procedures for combining data from different sources. The forthcoming National Recreation
Survey is described briefly because it may provide relevant, but thin, site-specific data that can
be combined with other data to update or validate an existing model For completeness, the fifth
step in the Deck and Chestnut (1992) proposed protocol is an original study. We omit this step
because it does not involve a "transfer" at all.
I
.1
1
1
Stage 1: The Qualitative
nt
The objective of the qualitative assessment is to determine the likely economic
significance of die changed condition due to the policy. Two important factors influence any
conclusions mat can be drawn at this preliminary stage of the analysis. The first relates to the
magnitude of the change in the condition of the environment that results from the policy and
whether an economically relevant endpoint can be measured. The second involves the sensitivity
-------
than just with the specific visibility applications in this benefits transfer. The remaining
comments focused on technical issues such as rehabilitation of existing studies, weighting
of results, and sensitivity analyses, for example.
GENERAL BENEFIT TRANSFER ISSUES
In the process of discussing this case study, group members raised several general
benefit transfer issues. Although we chose to focus on the specifics of our case study, we
list these more general issues to provide a more complete picture of the concerns/thoughts
about benefit transfer raised by the group.
• Values through time: Changes in values, changes in income, and discounting
questions must all be addressed when projecting benefits over some extended
time period.
* Peer review: Questions about whether to use study results that have not been
fully peer reviewed or published in peer-reviewed journals are frequently
encountered. Questions were also raised about what sort of peer-review process
is appropriate for benefit transfer. Some review is always desirable, although
peer-review publication is not always practical.
• Statistics: We generally agreed mat more information man only mean results of
available studies should be used when conducting transfers. Some quantitative
characterization of uncertainty or distributions of study results should be carried
into the transfer.
• Economic theory: Concerns were raised about the consistency of implicit
assumptions in benefits transfer with economic theory.
• Costs of being wrong: Costs of being wrong should be considered in
evaluating the efficacy of a benefit transfer.
• Underlying study issues: A benefits transfer cannot ignore and is at risk of
amplifying uncertainties in the results of underlying studies. This uncertainty
includes limitations of each study method, such as CVM, travel cost, or hedonic
property value. Questions of aggregation and total values versus component
values may also be important. Before we transfer estimates we need to evaluate
thoroughly what the available estimates tell us about the original study scenario.
Role of expert opinion: Most transfer exercises involve some ji
part of the researcher. Expert opinion should be acknowledged i
assumptions identified.
ton the
REFERENCES
Brookshire, D.S., R. d* Arge, W.D. Schulze, and M. Thayer. 1979. Methods Development
for Assessing Air Pollution Control Benefits. VoL 2: Experiments in Valuing
Non-Market Goods: A Case Study of Alternative Benefit Measures of Air
Pollution in the South Coast Air Basin of Southern California. Prepared for the
U.S. Environmental Protection Agency, Washington, DC.
15
-------
Chestnut, L.G., and RJ>. Rowe. 1990a. "Economic Valuation of Changes in Visibility:
A State of the Science Assessment for NAPAP." In Methods for Valuing Acidic
Deposition and Air Pollution Effects. Section BS. National Acid Precipitation
Assessment Program, Washington, DC.
Chestnut, L.G. and R.D. Rowe. 1990b. Preservation Values for Visibility Protection at
National Parks. Draft Final Report prepared for U.S. Environmental Protection
Agency, Research Triangle Park, NC, and National Park Service, Denver, CO,
February.
Loehman, E, D. Boldt, and K. Chaikin. 1981. Measuring the Benefits of Air Quality
Improvements in the San Francisco Bay Area. Prepared for the U.S.
Environmental Protection Agency. Menlo Park, CA: SRI International.
McClelland, G., W. Schulze, D. Waldman, I. Irwin, D. Schenk, T. Stewart. L. Deck, and
M. Thayer. June 1991. Valuing Eastern Visibility: A Field Test of the Contingent
Valuation Method. Draft Report prepared under U.S. Environmental Protection
Agency's Cooperative Agreement #CR-815183-01-3, Washington, DC.
McClelland, G., W. Schulze, I. Irwin, D. Schenk. D. Waldman, T. Stewart, L. Deck. P.
Slovic, S. Lictenstein, and M. Thayer. March 1990. Valuing Visibility: A Field
Test of the Contingent Valuation Method. Draft Report prepared under U.S.
Environmental Protection Agency's Cooperative Agreement #CR-812054,
Washington, DC.
National Acid
Assessment
nation Assessment Program (NAPAP). 1991. 1990 Integrated
Washington, DC.
Rae,D-A. 1984. Benefits of Visual Air QuaUty in Cincinnati~-Results of a Contingent
Ranking Survey. Prepared for the Electric Research Power Institute by Charles
River Associates. 4RP-1742.
ToUey, G A., A. Randall, G. Blomquist, R. Fabian, G. Fisbelson, A. Frankel. J. Hoehn, R.
Krumm, E. Mensah, and T. Smith. 1986. Establishing and Valuing the Effects of
Improved Visibility in Eastern United States. Prepared for the U.S.
Environmental Protection Agency, Washington, DC.
Trijonis, J., M. Thayer, J. Murdoch, and R. Hagement 1984. Air Qualify Benefits
Analysis far Los Angeles and San Francisco Based on Housing Values and
Visibility. Prepared for die California Air Resources Board, Sacramento, CA.
Trijonis, J., M. Pitchford, W. Malm, W. White, and R.Husar. 1990. Causes and Effects.
of Visibility Reduction: Existing Conditions and Historical Trends— National
Acid Precipitation Assessment Program (NAPAP), SOOT 24.
1»
r
j
j
16
-------
.
ISSUES IN BENEFITS TRANSFER
Trudy Ann Cameron*
ABSTRACT
These comments cover four separate issues in benefits transfer. The first is an idea for
using weighted maximum likelihood estimation to recalibrate study sample models to reflect
policy population relative frequencies of different sociodemograpbic groups and
environmental attributes. These recalibrated models are then transferred to the study context
The second issue highlights the substantial value for benefits transfer of an estimation
methodology proposed in the international development literature by Edward Learner. The
third issue is a description of a recent survey and evaluation prepared for the National
Research Council concerning the "combination of information" (CQ in a wide array of
different disciplines. This report very closely parallels the insights drawn by many of the
participants in the 1992 AERE workshop. Finally I make a recommendation concerning
competitive funding for the incremental effort necessary for documenting and preparing data
associated with primary studies that have substantial promise for benefits transfer
applications.
Environmental benefits assessments are now mandated for many benefit-cost analyses of
public projects, and these assessments also form an essential component of much environmental
litigation. Original studies, unique to the particular valuation problem in question, are typically
very expensive and highly time-consuming because household surveys must usually be
conducted to gamer the appropriate data. As a consequence, researchers are pressured to look for
"good enough numbers" provided by some existing, sufficiently similar assessment
The demand for benefits estimates that can be selected "off the shelf from an inventory
of estimates is overwhelming. For example, if an oil spill kills 200 sea birds, researchers would
find simply averaging the dollar values attached to dead sea birds in half a dozen existing studies
convenient to estimate a satisfactory dollar value of each of these particular birds, in this
particular area.
Of course, the advisability of this strategy of borrowing estimates for the new valuation
problem will depend on the similarity of the two contexts. In a few cases, finding a similar study
may be relatively easy. In other cases, arguing that the values from the "study** case are
transferable to the "policy" case may be less valid. In still other cases, no existing values may be
available for any similar scenario (i.e., species, type of damage or enhancement, or locale).
Given that benefits transfer is widely practiced, assessing suitable protocols for making such
transfers is important
'University of California, Department of E
-------
Benefits transfer practices were the subject of a recent special section of the journal
Water Resources Research. This collection of papers maps out many important issues in this
area. It also showcases work on the overall practice of benefits transfer, rather than specific
examples.1
This paper addresses four distinct issues relevant to benefits transfer. I describe an idea
for using weighted maximum likelihood estimation to recalibrate study sample models to reflect
policy population relative frequencies of different sociodemographic groups and environmental
attributes.2 These recalibrated models are then transferred to the study context I review and
highlight the substantial value for benefits transfer of an estimation methodology proposed in the
international development literature by Edward Learner. 1 then describe a recent survey and
evaluation prepared for the National Research Council concerning.the "combination of
information" (CI) in a wide array of different disciplines. This report very closely parallels the
insights drawn by many of the participants in the 1992 AERE workshop. Finally, I advocate
competitive funding for the incremental effort necessary for documenting and preparing data
associated with primary studies having substantial promise for benefits transfer applications.
REWEIGHTING STUDY SAMPLE TO REFLECT POLICY POPULATION
In ordinary least squares estimation (OLS), a sample that is nonrepresentative only in
terms of the distribution of an exogenous variable presents no problem for estimation. In
contrast, if the sample is nonrepresentative in terms of an endogenous variable, potential exists
for sampling bias in the estimation results. In general, in any estimation algorithm, if an
observation's presence or absence in the estimating sample is in any way related to the
magnitude of the outcome researchers are trying to explain, potential exists for bias in the
The case study in which I participated emphasized random utility modeling (RUM) of
recreational site choices. These models are estimated by maximum likelihood (ML) methods. A
long tradition in models like this is employing weighted exogenous sample maximum likelihood
(WESML) estimation when the estimating sample is not representative of the desired study
population, but the approximate distribution of respondent attributes in the study population is
known.
l1haiB papas include Atkinson, Crocker and Sbogren (1992), B<>yte and B>licy"un^ct»rt/drpcipnhtinnt «nt adopted daring the Worictbop and
win be adhered to throughout this paper.
i
1
1
4
J
-------
Suppose that the study population distribution is defined over attributes X and choices
j € C. This is a joint distribution, which can be decomposed as a conditional distribution times a
marginal distribution:
f(j,X)*P(jlX)p(X)
(1)
Now, if the study sample happened to be truly representative of the study population, the
likelihood function for the individual choices observed in the sample would be given as follows
(where yy = 1 if individual i chooses alternative j and yij = 0 otherwise):
p(Xi)
This calculation results in a formula for the log-likelihood given by
(3)
By exploiting the decomposition of the joint distribution into a conditional distribution times a
marginal distribution, the log-likelihood function in Eq. (3), to be maximized over the unknown
parameters p, consists of a sum of two components. The second component does not depend on
the parameters P, so it can be ignored, and the optimization of log £ can proceed simply by
maximizing the first term in Eq. (3). Weights are unnecessary.
However, most benefits assessments require voluntary participation of members of the
affected study population in the survey necessary to gather the data. In RUM models,
researchers now generally acknowledge that nonparticipation should be included as a relevant
choice along with specific site choices conditional on participation. Whether contacted
individuals opt to comply by completing their questionnaire or interview will determine their
presence in the final estimating sample for the study.3 Nonparticipants in the associated
recreational activity are typically less likely to be interested in the survey and hence less likely to
appear in the final sample. Because of this tendency, most modem RUM applications involve
fiindamftH^y choice-based samples.
Ben-Akiva and Lerman (1985) demonstrate that unweighted MLE is still feasible for the
standard multinomial logit specifications typically used to estimate RUM models, providing the
*ri*** «m »nH up tying tmnitteA ft
complete nonresponse.
mpl» harmm. nf item mwiMpmii* nr
-------
choice model has a full set of J-1 alternative-specific constants (Le.. site-specific dummy
variables plus a nonpartitipation dummy variable). Exogenous information concerning the true
study population distribution of attributes X is still required for the process of adjusting the
estimated probabilities after the estimation process. Manski and Lerman (1977) call this
approach "exogenous sample maximum likelihood*' (ESML).
However, in benefits transfer applications, the last thing a researcher wants in the model
for the original study sample is a set of site-specific dummy variables, for the following reason.
Using these dummy variables is akin to estimating entity-specific fixed effects in a panel data
model for pooled time-series and cross-section data. Providing no new entities appear in the data
set for which a policy forecast is desired, these fixed effects are fine. But if new entities will
appear, the researcher will have no fixed effects to use for them. Random-effects models for the
study sample are preferred under these conditions.
Benefits transfer exercises require, by definition, that models calibrated for one set of site
choices be applied to different sites (or at different time periods). This feature precludes using
ESML estimation for RUM models destined for transfer exercises. A formal choice-based
sample maximum likelihood estimator is dearly indicated in this context Unfortunately, mis
estimator is somewhat intractable. A consistent estimator for p* that represents a tractable
alternative is the WESML estimator.
The WESML estimator is typically implemented by partitioning the estimating sample
into G groups (or "cells**) defined over intervals of the values of some subset of the exogenous
variables. The group-specific weights, w(, are given by f*?V£. where the numerator is the
population relative frequency of individuals in group g, and the denominator is the sample
relative frequency of individuals in group g. With Nf designating the number of sample
observations in group g, the WESML tog likelihood function is given by
(4)
Proving that this estimator for p is consistent under very general conditions is daunting.
Furthermore, the WESML estimator is not fully efficient even asymptotically, so its variance-
covariance matrix is more complex than that of a true maximum likelihood Estimator (see
Manski and Lerman, 1977). Even its corrected vaiiance-covariance matrix (outlined in Ben-
Altiva and Lerman (1985, p. 239) does not attain the Cramer-Rao tower bound. Thus these are
compromise estimators; computational tractability is gained at the expense of fun statistical
efficiency. They are nevertheless highly practical.
-------
To illustrate how WESML estimators might apply in benefits transfer situations, a simple
numerical example may be helpful. Consider a RUM model where only two variables affect
choice: respondent income and catch rates. Suppose that the study population is one million
people with joint frequencies for income and catch rates as given in Table 1 A. (Note that the
groups in this example are extremely coarse and that frequencies are measured in 10,000's.)
Suppose that a study sample of 50 respondents yields the joint sample frequencies shown in
Table IB. To inflate or deflate the influence of each sample observation so that the weighted
study sample mimics the study population distribution of attributes, the weights will be as given
in Table 1C.
WESML estimation will produce a set of utility parameters. (J, that can be argued to
represent the best parameterization of a "typical" or "average" set of preferences for the study
population. For benefits transfer, however, we would prefer to have a set of parameters, p*, that
represent the typical preferences of the "policy" population. If the researcher has access to the
full set of data used to calibrate the original study sample model and obtaining an approximate
joint distribution of the exogenous variables for the policy population is possible, the following
modified weighting scheme seems appropriate. Intuitively, researchers would simply construct a
set of weights for use in the WESML algorithm that serve to make the study sample
representative of the policy population, rather than the study population.
To continue the simple illustration, suppose that the policy population (also one million
people) has the joint distribution of exogenous variables given in Table 2A. The set of weights
necessary to make the sample with frequencies as in Table IB representative of this alternative
population appears in Table 2B. WESML estimation of the RUM specification using these
weights will produce a different set of estimates for the jJ vector of preference function
parameters—one that better approximates the typical preferences of mis new population.
Reviewing the data requirements necessary to make this rewtighting scheme work is
useful First imagine the ideal case. With unlimited data on a vector of individual-specific
sociodemographic variables, X, and a vector of individual-specific environmental amenities, Z,
researchers might imagine calibrating a full parametric continuous joint density function fP(X, Z)
based on exogenous sample data for the policy population. Researchers would analogously
calibrate a full parametric continuous joint density f*(X, Z) for the study sample.4 With these
*Ja oar earlier numerical example, fundamentally contiiiDOUs distributions for fa^qfUf gnd catch
aggregated into four cells to that a simple discrete distributions could be used to form ibe weights.
were
-------
TABLE 1A. STUDY POPULATION FREQUENCIES (1
-------
TABLE 2A. POLICY POPULATION FREQUENCIES (104)
Income Catch
Low
High
Total
Low
High
25
25
15
35
40
60
Total
50
50
100
TABLE 2B. WEIGHTS TO MAKE STUDY SAMPLE ESTIMATES REFLECT
POLICY POPULATION FREQUENCIES
Income Catch
Low
High
Low
High
1.25
1.25
1.5
0.7
two continuous joint densities, researchers could then calculate (unique) individual-specific
weights based on the ratio
~
for each individual's own vector of values for X and Z.5
This level of detail is highly improbable for current real applications. Multivariate joint
densities are simply too difficult to calibrate unless normality is invoked and even this
assumption may often be questionable. Furthermore, the raw data necessary to calibrate the full
joint density function fP(X, Z) are not typically available, at least with current information
technologies. For sociodemographic variables, official Census descriptive statistics will
sometimes provide two- or even three-way cross-tabulations of variables such as age, income,
and ethnicity, but these cross-tabulations are rarely available for specific subpopulations. Much
of the raw data exist; the infrastructure for extracting arbitrarily designated subsets of the
population is simply not yet as readily accessible as researchers might like. Data on the
environmental attributes are even more scarce, and when they are available, researchers must
frequently assume statistical independence between the X and die Z variables because these are
typically drawn from different sources. Because full vectors of both X and Z values are not
^Recall that tfaf weights in oar numerical i
four groups were defined
nple were only group-specific, not mdmdaal-specific, mnA that only
-------
extracted from the same individuals, the joint density cannot be estimated. Information
technology promises great strides in this area in the future, however.
In the meantime, researchers will have to make do with nonparametric frequency
information over matching "cells" in die policy population and the study sample. This method
requires comparable domains for f*(X, Z) and fP(X, Z). If the domains did not overlap, weights
could not be constructed. The number of partitions along each dimension of (X, Z) will be
dictated by the study sample's size. If some cells are empty, they can frequently be merged with
adjoining nonempty cells for both the study sample and the policy population. However, if too
many cells that are well-represented in the policy population are empty in the study sample,
researchers will have problems. In general, the more refined the cells, the better, but a tradeoff
exists between resolution (the fineness of the cell partitions) and cell frequency deficiencies.
Cell designations are entirely subjective.
Researchers have argued that simply transferring point estimates of benefits from a study
area to a policy area is generally not wise (Loomis, 1992). Point estimates depend on a vector of
estimated parameters as well as a matrix of exogenous variables. Thus, mis argument
recommends (correctly) that transferring the point estimate of mean value from the study to the
policy area is unwise because fundamentally different values of the exogenous variables may
apply in die policy area. Instead, transferring the entire model is preferable, applying it to new
(mean) values of the exogenous variables for the policy population. The reweighting scheme
described here goes one step further than ''model transfer." It avoids not only the assumption
that the exogenous variables are identical in the two regions but also the assumption that typical
preferences for die study region and die policy region are identical.
Preferences may indeed be systematically different if die study involves endogenous
location choice or if fundamental preferences are not uniformly distributed across die entire
country (we usually assume dial diey are). The disadvantage is dutrecalibration of die study
model wirn different weights requires tiiat die full study data set be available. The full data set
will not always be available, aUhongh pressure is mounting in die economics discipline to
preserve estimating samples and documentation as a condition for publication.
LEAMER'S BAYESIAN DATA-POOLING MODEL
Edward Learner (1991) has recently proposed a Bayesian econometric medwdology tiiat
appears to have much to offer benefits transfer practitioners in terms of focusing our agenda for
improving quantitative procedures. The current framework for Learner's model is OLS
regression, and die application he uses to illustrate die approach is a convenience sample of data
f
J
1
8
-------
pertaining to GNP growth in developed and developing countries. His application tests the so-
called "convergence hypothesis" (that higher initial GNP implies slower growth rates across
countries). His two samples are developed countries (assumed to provide good quality data) and
developing countries (assumed to provide poorer quality data). Although Learner's application is
not benefits transfer, he injects valuable rigor into the explicit modeling of many judgments
similar to those made in every application of benefits transfer.
The problem is one of combining information about some economic quantity from two
data sets of differing quality. Data pooling appears in benefits transfer exercises when
alternative study samples are combined either to provide transferable benefits estimates or
transferable models. It also takes place when study samples are pooled with small-scale policy
samples to "update" the study information with policy area information.
Learner's method is Bayesian and uses prior information about regression coefficients.
Estimates from pooled data depend on three types of parameters:
8 = the investigator's lack of confidence about the prior,
p - the subjective degree of similarity between the "study*'and the''policy"
relationships \
Xi = the amount of contamination of (for example) the "study** (i=l) and the "policy"
(i=2) data caused by such things as measurement errors, left-out variables, and
simultaneity, for example.
Learner's basic specification for the pooling of contaminated data across data sets i=l,2 is
as follows:
(5)
where the Pi are the true parameters and 6j is a bias vector due to the statistical pathologies of the
data. From this specification, extreme muMcollmearity clearly exists. Nevertheless, Learner
shows mat the informational deficiencies of the underidentified model can be overcome with
prior information. He assumes that 0i - N(0, Vi) and resorts to the random coefficients model
given by
ftl-Cl-M)
(6)
-------
where p is the most likely common structural parameter vector and U measures departures from
this vector. Learner notes that this parameterization conveniently allows a relative lack of
information about P but confidence that the difference between Pi and Pi is small (i.e., for large
U and p near unity).
The prior covariance matrix for the model in Eq. (5) is men given by
Var
Pi
P2
61
.82.
ru pu o
Pu u o
0
LO
0
0
0
0
0
C7)
Still, depending on the number of variables in the vector X, this can represent a daunting number
of unknown parameters about which prior values must be asserted. The number of prior
parameters can be reduced substantially by adopting the constraint Vi = XjU where Xi measures
the relative importance of experimental contamination (Le.t a high value of Xj means that the
investigator wishes to discount the information in mat sample).
The number of prior parameters can be further reduced by making U « SZUo, where UQ is
the prior on the amount of noise in each of the pi vectors. 5 is then interpreted as the "discount
rate" on the prior variances. With these simplifications (for greater tractability), the researcher
now needs to specify priors only for the vector p and the matrix U, as well as the scalars 5, p, Xi,
and X2 (in the two-sample case).
The innovations in Learner's approach (despite the current estimator being demonstrated
only for the OLS context) include the following:
• specifying a generalized random coefficients model for combining information;
• incorporating errors-in-variables concerns and other pathologies, which allow
assumptions about the extent of these pathologies to differ across samples; and
• adhering to the desirable Bayesian econometric paradigm.
Conceptually, this approach has much to offer benefits transfer research. It formalizes explicitly
what we all do while searching for "relevant!" studies to be used for benefits transfer. Consider
the Xt (unreliability) parameters. The larger Xi is, the less weight is put on sample i's results in
averaging its information with the prior. By discarding studies, we implicitly assume mat Xi goes
to infinity; by using a study, we implicitly assume that Xi goes to zero. A better strategy would
be to use expert judgment about the qualities of different studies (and their relevance) to assign
I
1
10
-------
I
4
0 < Xi < •» appropriately for each study. Learner's conceptualization forces us to reveal our
assumptions explicitly and allows for intermediate values of the Xj parameters, rather than
limiting them to the extremes of zero or infinity.
It will be some time yet before Learner's OLS procedures are adapted to MLE contexts
and then to RUM parameter estimation tasks. The computer algorithms are complicated even in
the context of OLS. However, Learner offers benefits transfer theorists and practitioners
something to strive for. His insights could lead to some very useful dissertation work in the
hands of an environmental econoraetrician. The benefits transfer literature directly needs
statistical methodologies that force practitioners to be specific about their priors overall (as on p
and U) and their priors as they embark on the blending of multiple sources of information
(namely 5. p, and the Xj's).
NATIONAL RESEARCH COUNCIL REPORT ON CI
A subcommittee of the National Research Council recently convened a panel to study and
report on "Statistical Issues and Opportunities for Research in the Combination of Information"
(Gaver et al., 1992). This report has just recently been completed, and almost all of its findings
are relevant to the discussion at the AERE benefits transfer workshop.6 The practice of
combining information apparently takes place in almost every quantitative discipline with
important lessons being learned at different rates by different groups. The terminology varies
across fields. For example, it is called "data fusion" in the defense industry and "meta-analysis"
in several social sciences. The report provides a wealth of information and insight into research
opportunities by examining a broad range of case studies in different disciplines.'
Because the report will be readily available, thi* P&pcr merely summarizes and
paraphrases its main conclusions, many of which echo the sentiments of the different teams
working on case studies at the AERE workshop. (The quotes in the folio wing points are drawn
from the conclusions section of Gaver et aL, 1992).
• "Authors and journal editors should attempt to raise the level of quantitative
explicitness in the reporting of research findings." Documenting data and models is a
clear necessity for improved benefits transfer exercises. Ideally, all study sample data
would be freely available, allowing the widest variety of transfer tt**M"qw;. including
re-weighting.
• "CI based only on P-values should be avoided in favor of estimates of quantities of
direct decision-making relevance, together with uncertainty estimates." The crudest
methods of CI across studies will ascertain whether a particular explanatory variable is
6I am grateful to David Draper (of RAND art UCLA) for providing a prelinunvy copy of this report
11
-------
a significant determinant of the outcome variable and allow these results to be "ballots"
in a vote on whether the variable explains the outcome. Slightly more sophisticated
methods use the unit-free prob-value (or P-value) associated with the coefficient on the
crucial variable in different studies, averaging these continuous quantities, possibly
with sample-size weights, to ascertain the overall judgment of whether the variable
explains the outcome. This recommendation.advocates that significance or
nonsignificance is not the important issue; rather, the magnitude of the effect of the
variable on outcome ought to receive the attention in CI exercises.
• It is worth investigating the costs and benefits associated with going beyond numerical
summaries to data registries or archives (for both published and unpublished studies)."
This issue is addressed by David's (1992) paper on data accessibility.
• "Increase the explicitness in the formulation of models that express judgments about
how information sources to be combined (subjects, variables, research studies, bodies
of expert opinion) are similar (exchangeable) and how they differ." This point
corresponds directly to the advances offered in the paper by Learner (1991) outlined in
the previous section.
• "The practice of CI could benefit from increased use of sensitivity analysis and
predictive validation."
1 "Hierarchical statistical models are a useful framework for CL Use in fields where they
are not yet routinely employed is to be encouraged, as is an increase in the coverage of
such models in intermediate and advanced mflistify courses." Econometricians do not
routinely leach or use these methods, bat these methods merit close scrutiny for
application to benefits transfer.
"Q modeling could be improved by increased use of random effects models in
preference to the current default of fixed effects.** This terminology is somewhat
confusing to econometricians.7 Translated, this recommendation advocates random
coefficients models, rather than the more familiar nonrandom coefficients models. "At
a minimum, we believe that researchers will often find it useful to perform a sensitivity
analysis in which both kinds of models are fit, and the substantive conclusions from the
two approaches are compared"
Researchers need a "gene
perform interactive Baye
Learner has advocated in „
to be enhanced and disiifiniiffjff4 more broadly.
t • • tr-
i andwu m hie
ctive Bayesiani
algorithms clearly need
More meta-analysis should be undertaken. Researchers need to do "more work on the
design of meta-analyses and related CI exercises" and pay "increased attention to
alternative analytic approaches."
"Workers using CI procedures.. ."in benefits transfer would profit from a study of Q
methods used in other fields, and funding agencies should give a higherpriority to
"cross-disciplinary conferences on methods for combining information.
]
.]
"T
7It is used differently in the econometric analysis of paWl data.
12
-------
•
i
f
THE PUBLIC GOODS NATURE OF WELL-DOCUMENTED DATA SETS
Well-documented data sets in general machine-readable form are a valuable public good.
They are rarely available because the private costs to researchers of providing the data almost
always outweigh private benefits. Journals are now making an effort to internalize some of these
costs by requiring either that the data be available or that they be supplied on diskette when the
paper is submitted for review.
In addition, establishing competitively allocated resources to support post-study data
documentation and archiving for future benefits transfer exercises would be very useful to these
exercises. This program would have to be on-going, selecting only those data sets each year that
clearly have promise for future use in transfer exercises by other researchers. The incremental
cost of cleaning up and annotating a data set for public consumption rises quickly with the time
elapsed since completion of the original study. But in many cases, the incremental cost to the
research team of retaining a research assistant for an additional month after completing the main
project is relatively small (at least compared to the cost of going back to the data after several
years have passed or of collecting new data).
In many cases, the research team responsible for collecting and processing the data set
will have a proprietary interest in using that data for a set of studies before they become widely
available to everyone. We must acknowledge that the compensation for much contract work is
often taken (by academics) in the form of future publications employing the data made available
by the original survey study. In these cases, proprietary interest might be a negotiable item in a
proposal for incremental data documentation funding. The research team could include a time
limit within which the delivered cleaned-up data would not be disseminated to other users. This
time limit would allow the documentation phase to proceed in a timely fashion without the
possible cost to the original research team of lost proprietary rights to the data conferred
implicitly by unintelligible or nonexistent documentation.
CONCLUSIONS
Benefits transfer, a widespread practice that has been ongoing, will continue to take
place. In the face of tightening budgets, the need for "off the shelf estimates of economic-
environmental benefits for policy and litigation will continue to increase. Therefore formulating
and promulgating a set of guidelines for these exercises are valuable endeavors. These
guidelines could be similar to the accepted standards for antitrust litigation. Without such
protocols, highly varying standards of accuracy might implicitly be applied in different cases.
13
-------
Workshop participants did not expect to produce a completed set of such protocols, and
they did not However, die participants seemed to experience a collective "consciousness-
raising" concerning the problems of benefits transfer. The opportunity for each group to conduct
an intensive post-mortem on a particular benefits transfer case emphasized the common
problems; the summary presentations allowed each group to articulate its own unique findings
for the benefit of members of other groups. At a minimum, all participants left the workshop
with a greater appreciation for the enormity of the challenge.
This area is ripe for productive applied research in this area. The subject of benefits
transfer protocols may be less glamorous than alternative theoretical topics in the area of
environmental economics. "Publication bias" favors new research on new topics, rather than
pragmatic issues such as benefits transfer. However, the workshop highlighted the scope of
applicability of research on the problem. In terms of influencing potentially huge reallocations
of society's resources through policy making or litigation, the benefits transfer research has
profound relevance.
J
1
1992. "Bayesian
Water Resources
Atkinson, Scott E., Thomas D. Crocker, and Jason F. :
Exchangeability, Benefits Transfer, and Research \
Research 28:715-722.
Ben-Akiva, Moshe, and Steven R. Lerman. 1985. Discrete Choice Analysis: Theory and
Application to Travel Demand, Cambridge: MIT Press.
Boyle, Kevin J., and John C. Bergstrom. 1991 "Benefits Transfer Studies: Myths, Pragmatism,
and Idealism." Water Resources Research 28:6*57-663.
Brookshire, David S., and Helen ILNeilL 1992. "Benefit Transfers: Conceptual and Empirical
Issues." Water Resources Research 28:651-655.
David, Martin. 1992. "Benefiting Benefits Transfer Information Systems for Complex
Scientific Data," Paper presented at the 1992 AERE Benefits Transfer Procedur
Problems, and Research Needs Workshop, Snowbird, TJT, lime 3-5.
Des
,
, William R, Michael CNaughum, and George R. Parsons. 1992. "Benefit
r Conceptual Problems in Estimating Water Quality Benefits Using Existing
Studies." Water Resources Research 28:675-683.
Oaver, DJ>, Jr., David Draper. P.K. Goel, J.B. Greenhouse, L.V. Hedges, ON. Morris, and C.
Watemaux. 1992. On Combining Information: Statistical Issues and Opportunities far
Research, Draft report of the Panel of Statistical Issira and Opporbimties for Research in
die Combination of Information, Committee on Applied and Theoretical Statistics, Board
lAppt
«iSl!
on Mathematical Sciences, Commission on Physical Sciences, Mathematics t
Applications, National Research Council.
14
-------
Learner, Edward E. October 1991. "Eastern Data and Western Attitudes." Technical Working
Paper #114. National Bureau of Economic Research.
Loomis,John. 1992. 'The Evolution of a More Rigorous Approach to Benefit Transfer: Benefit
Function Transfer" Water Resources Research 28:701-705.
Luken Ralph A., F. Reed Johnson, and Virginia Kibler. 1992. "Benefits and Costs of Pulp and
Paper Effluent Controls Under the Clean Water Act" Water Resources Research 28:665-
674.
Manski, C, and S. Lerman. 1977. "The Estimation of Choice Probabilities from Choice-Based
Samples." Econometrica 45:1977-1988.
McConnell, K.E 1992. "Model Building and Judgment Implications for Benefit Transfers
with Travel Cost Models." Water Resources Research 28:695-700.
Smith, V. Kerry. 1992. "On Separating Defensible Benefit Transfers From 'Smoke and
Mirrors'." Water Resources Research 28:685-694.
Walsh. Richard G., Donn M. Johnson, and John R. McKean. 1992. "Benefit Transfer of
Outdoor Recreation Demand Studies, 1968-1988." Water Resources Research
28:707-713.
15
-------
J
f
-------
I
4
i
BENEFIT TRANSFER AND SOCIAL COSTING
Alan I. Krupnick*
ABSTRACT
Increasing demand for benefit analyses that are too comprehensive for original research
lo be feasible and static or falling research budgets put a high value on the wise use of
existing benefit studies to estimate benefits associated with new policies and problems. In
mis paper I define the sources of the increased demand for benefit analyses, identify the types
of benefits most useful to benefit transfer now, examine the protocols for conducting benefit
transfers, and suggest a future research agenda.
Interest in developing and applying techniques for benefit transfer is growing rapidly.
Benefit transfer is the application of original damage or benefit studies made in a given policy
context and location (what Desvousges, Naughton, and Parsons [1992] refer to as A policy site) to
another context and/or location (what these authors refer to as a study site). Burgeoning demand
for benefit analyses that are too comprehensive for original research to be feasible together with
static or falling research budgets put a premium on the wise use of existing benefits studies to
estimate benefits associated with new policies, problems, or simply new locations. The idea of
designing future original research to enhance the reliability of benefit transfers presents
particularly interesting challenges.
This paper has three purposes: to delineate the sources of this burgeoning demand, with
particularly attention to the movement led by Public Utility Commissions (PUCs) to incorporate
all of the externalities of electricity generation into utility decision making; to identify the types
of benefits that are most amenable to benefit transfer now; and to examine protocols for
conducting benefit transfers and suggest a future research agenda to make benefit transfers
easier, reliable, and more consistent with welfare economics.
SOURCES OF DEMAND FOR BENEFIT TRANSFER
Since environmental and natural resource economics began as a discipline in die early
1970s, the primary demand for analyzing the benefits of environmental improvement came from
U.S. government agencies interested in establishing "unit-day" recreation values for evaluating
projects and policies affecting water resources and from agencies needing to comply with E.O.
12291, which mandates benefit -cost analyses for all "major" regulations. These needs translated
into research budgets for original research in estimating policy-related environmental benefits,
while also giving rise to using the original research results in what we would now label "benefit
'Senior Fellow. Resources for the Future.
-------
transfer" exercises to comply with the Executive Order. One of the most visible and successful
of these secondary studies was EPA's benefit-cost analysis of the lead-pnasedown regulations
(U.S. Environmental Protection Agency, 1985), which used original benefit studies to provide
estimates of the value of statistical lives and values of avoiding a variety of acute health effects
to argue that the phasedown made economic sense.
More recently, passage of the CERCLA (Superfund) law has propelled interest in benefit
transfer and resulted in the embodiment of this concept in the Type A natural resource damage
assessment model (now being updated), which estimates damages to recreational and commercial
fishing from a given type and size of oil spill in a given location using existing literature (see
Jones, 1992).
But each of these needs is relatively narrow, involving damage to, at most, a few
nonmarket commodities and usually by only one cause (e.g., lead or an oil spill). The limited
scope of these demands sets them apart from the newest demand for benefit transfers—mat of
stale PUCs who wish to formally introduce «tjfnM«t of the external costs of alternative means
for generating electricity into utility decision making. All externalities associated with the fuel
cycle supporting each generation technology need to be addressed. For the coal cycle, this
means addressing externalities from acid mine drainage to environmental effects of air emissions
at the generation stage. Some 29 states are considering or requiring that the planning f or new
investments accounts for residual environmental damages from alternative generation
technologies (Cohen et aL, 1990).
Unfortunately, but not surprisingly, no original studies provide comprehensive estimates
of these damages;1 even imagining how an original study would be conducted, assuming mat the
money to pay for it could be found, is difficult Even if some studies of mis type were
conducted, the location specificity of environmental damages (Le*. their sensitivity to the
location of the new power plant, irrespective of the technology creating these damages) would
still necessitate using techniques for transferring die comprehensive results of these studies to the
study fife. Thus, assuming fry* states are prepared to "qplfiPCTt social costing, researchers must
devise and codify methods for consistently using benefit transfer tpchniflwff to estimate
1Onmger et al. (1990). like other wort: in ibis met, use credo benefit tansfertedrnkines to estimate damages snd
ignore the location-specificity of impacts. Otter cunpTehe&sive estimates of the extend costs of electricity tu
abatement costs as a proxy for damage (Bonow. Bfewaid. and Matron, 1991).
I
J
3
J
J(
-------
incremental damages in each state as well as across different potential power plant sites within a
state.
Major on-going studies are already codifying benefit transfer techniques but without
carefully considering the models they are using. The U.S. Department of Energy (DOE) is
funding a study conducted by Oak Ridge National Lab and Resources for the Future that is
designed to develop methods and estimate the externalities from alternative fuel cycles used in
generating electricity at two "reference environments." No original research is in the work plan;
rather benefit transfer (as well as health, biological, and meteorological science transfer) is to be
used to the fullest extent possible in the context of a damage function approach.2 Economists,
engineers, and natural scientists in Europe, with funding from the European Community, are
following the identical work plan and methods while sharing some of the research effort to
estimate comparable externalities for potential power plant sites in Europe.
New York State is funding Hagler, Bailly to do a more ambitious external costing study
that builds on the DOE research to develop a computer model for utilities to use in estimating the
external costs associated with any proposed new capacity expansion. In addition, smaller studies
with similar objectives are on-going in Wisconsin (Research Triangle institute) and California
(National Economic Research Associates and Regional Economic Research, Inc.). For the most
part, each of these studies, facing the enormity of their tasks, which take in virtually an the
benefit estimation literature, is primarily assembling and evaluating literature to provide any
estimates of damage, without paying much attention to theoretical prerogatives and constraints
discussed at the workshop.
A final, potentially major source of demand for benefit transfers comes from international
aid organizations such as the World Bank and the U.S. Agency for International Development
These groups are responsible for capturing the environmental effects of their lending in
developing countries, but with very few exceptions (Whittington et aL, 1989), no original studies
of the benefits of environmental improvements in frw countries exist Here, protocols for
benefit transfer that take into account different personal and market characteristics are
^Al the valuation i
involves)
pactt^ problematic for a benefits transfer for
became the absolute and relative nagnbodesofeoviroiimentilctan^esassociatBdwiiha
nower plau or an eatiie bid cyck wffl, m gcoenl. be qinte
p^yMty v^tnft findi^f (>tr pn^flMit pflp{yptratin»if) yw< tfif it^ n gf changes is moch broader. Ibe damage
fbnaioB approach is not without Us problems, however, because n^anxoachoainot capon WTPctfinose who
•void impacts. We nuy say that the dunagefinctioo approach fc good for id^
for effects caused by air pollution and other effects when avoidance behavkr ray not be pervasive) but not in
me long-run, where avoidance opportunities, such as residential tocatioo derisions, an mote viabk and a^
-------
particularly important, as differences in incomes, institutions, cultures, climate, and resources,
for example, are surely far larger between a developed and developing country than among states
in the U.S. (in the case of social costing of electricity). The existence of widespread subsidies on
energy and other commodities greatly distorts relative prices, adding the identification of shadow
prices to the long list of challenges to benefit transfer.
Researchers even debate whether benefit transfer is legitimate for certain types of
nonmarket commodities affected by programs in developing countries. The basic tenet of
individual sovereignty underlying benefit estimation may not be applicable in societies that
emphasize group welfare. And the profound influence of poverty in developing countries on
willingness to pay raises questions about whether any benefit transfer technique involving U.S.
income elasticities of demand can be justified.
WHICH BENEFITS CAN BE TRANSFERRED NOW?
Benefits can be characterized into four groups by their effects on the following: health,
output, economic assets, and environmental assets— with my subjective ratings on the ease with
which benefit transfers can be conducted, given the existing state of the original research
literature, the characteristics of the commodity being valued (e.g., its dependence on personal
characteristics, site and regional characteristics, and extent of me market questions), and die
degree of codification of the literature for benefit transfer. The perspective in making these
judgments is that of the PUC evaluating the methods used to provide estimates of social costs. It
is recognized that the scope of the task requires some degree of "quick and dirty" analysis, rather
than the courtroom-proof reliability of natural resource damage assessment estimates.
Two of the four categories can be pretty much ignored: damage to output and to
economic assets. Damages to output, for example crop damage from air pollution or damages to
commercial fishing from a spill, are easy enough to estimate using original research and
garnering market price and supply and demand elasticities, for example, for the products, as
warranted. On the other hand, damages to economic assets cannot reliably be estimated in
original studies, let alone in a benefit transfer. Materials inventories are still lacking, and no
major modeling efforts for valuing the complex behavioral linkages necessary toe a defensible
riaiy benefit estimate have been i|*«d*a*A*n jn many years.
Probably the health effects category is the easiest for making credible benefit transfers.
Once atmospheric or other natural processes are taken into account (e.gM when estimating the
effect of reduced emissions on ambient air quality), the researcher can presume to a first
approximation, that the health effects and the values people place on avoiding these effects are
I
I
1
I
1
I
1
I
4
1
J
J
j
J
-------
I
4
i
reasonably similar across locations. The extent of the maiket is clean people living in the air
basin in which the postulated air quality change occurs.
Codification has proceeded for many years. Estimates of the value of a statistical life
taken from summary reviews and specific studies are widely used, multiplied by expected deaths
"delayed" to obtain the mortality benefits from a particular program, investment, or other
exogenous change in baseline conditions. A similar protocol is followed in using the literature
on the values of avoiding acute health effects to estimate the benefits of baseline pollution
reductions (see Hall et al., 1989; Krupnick and Portney. 1991; and National Economic Research
Associates, 1990) for benefit transfer studies for improving air quality in Los Angeles that
include estimates of mortality and morbidity benefits). Indeed, "spreadsheet" models axe
available that first match estimates of changes in air pollution concentrations to dose-response
functions for a wide variety of health effects and then match these to unit values for avoiding
these effects to obtain health benefit estimates for environmental improvements.
Yet, the benefit transfers are of the crudest type: they use unit values and unaided
judgment to combine the different values obtained from the literature. Few of the spreadsheets
use valuation functions in the benefit transfer, for example, of the kind arising from regression
analysis explaining variation in willingness-to-pay (WTP) responses. Hie methods for
establishing error bounds and best estimates are ad hoc and heterogeneous across benefit transfer
studies.
The original studies do not always lend themselves to transfers. Virtually the entire
mortality risk valuation literature addresses accidental deaths in prime-age adults, a setting
inappropriate for all environmental mortality except perhaps accidental toxic waste releases and
similar catastrophes. One study (Mitchell and Carson. 1986) addresses the latency issue so
important to valuing deaths due to cancer but is silent on die effect of prior health status and age
on valuation. These issues are important in environmentally related deaths to those with heart
disease and chronic lung disease. Further, researchers trying to use mis study to value noncancer
related deaths may find that it postulated risk changes outside the risk changes associated with
power plant emissions. Also no reliable studies are available to value life-years saved (except in
occupational accidents) even though this health endpoint can be estimated by health scientists.
The most problematic area for benefit transfer is damage to environmental \
although there is some differentiation among these subcategories. Benefit transfer of recreation
values or demand functions presents one of the greatest challenges. Accounting for regional
factors (such as the range and quality of substitute sites) and site-specific factors (such as
-------
congestion) is likely to be difficult Furthermore no acceptable procedures exist for determining
the "spatial extent of the market" That is, debate is still lively on methods for determining the
size of the population that would be or is affected by a recreation quality or quantity change.
Codification of the chain of effects from concentration change to valuation is absent, with
the exception of the Type A model noted above. Because benefit transfers have generally
followed the procedure of using unit-day values, these values exist in great profusion for all types
of uses and environments (Walsh, Johnson, and McKean, 1988). But applying these values to
specific sites is problematic, more so than applying unit values to health because of the
presumption that WTP to avoid health effects is less influenced by region and site variables than
WTP for recreation. Codification of recreational fishing damages from oil spills in the Type A
model represents a useful prototype for the future development of portable, PC-based models for
use in benefit transfer. However, this particular model uses a unit-day value approach for the
valuation step.
Likewise, the recreation literature is of somewhat limited usefulness in estimating social
costs because the majority of the literature focuses on changes in the availability of resources not
on changes in their quality. Few studies incorporate explanatory variables that map back into
readily measured physical quantities, such as water turbidity, nutrient concentrations, and the
like. Most of the literature values catch rate changes.
Benefit transfer for valuing visibility also presents formidable challenges because of the
sensitivity of values to region, site, and personal characteristics. Characterizing the policy and
study site is particularly difficult for visibility tufncfit transfers. Although visual range can be
characterized in a relatively straightforward way. the vista being affected is particularly difficult
to characterize, beyond "urban," "rural," and "lecieatkmal area," whkft is unlikely to be
sufficient In addition, the extent of the market problem is even more difficult man that for
recreation because "use" as a function of distance to the site can be observed for recreation, but
not for some visibility problems (e.g., urban visibility).
The literature on visibility benefits is fairly conducive to benefit transfer (see Chestnut
and Rowe, 1992). Studies of visibility values in multiple cities (Tollcy et aL, 1988) are available,
which then permit examination of city-specific factors affecting values and derivation of
functional relationships to predict WTP, given t^***- bflpr^'Ttf visual range and the ftyy of the
change (National Acid Precipitation Assessment Program, 1989). A number of examples of
benefit transfers involving visibility (Rowe, Chestnut, and Sknrnanich, 1990; Chestnut and
Rowe, 1988) are available. The Electric Power Research Institute (EPRI) (1991), which
J
1
-------
I
i
examines benefits from improved visibility in the eastern U.S. from reductions in SO2 emissions,
is a particularly good example of a benefit transfer where all the steps of the damage function
approach were linked together (i.e., emissions to concentrations, concentrations to optics., optics
to perceptions, and perceptions to value).
The major problem with benefit transfer in this category is the original studies.
Significant debate surrounds protocols for eliciting values in contingent valuation studies. For
example, the size of photographs shown to respondents appears to influence WTP. Concerns
about joint valuation of visibility and health (i.e., that visibility is used as a proxy for health
effects) and about embedding are also important From the perspective of the social costs of
electricity issues, research efforts have concentrated too much on national parks in the southwest
and not enough on valuing visibility effects at more mundane locations, both rural and urban.3
The literature on nonuse values for environmental assets clearly cannot yet support
benefit transfers associated with social costing of electricity, because most of the studies are for
non-marginal changes in unique environments (species extinction, loss of an ecosystem) while
the effects of a single power plant on any species or ecosystem is likely to be small and on
unique areas or species (after compliance with the Endangered Species Act and other federal
legislation) negligible. An exception might be nonuse values for visibility at national parks, such
as the Grand Canyon, associated with power plant emissions (Decisions Focus, Inc., 1990).
Admitting nonuse values into the benefit transfer exercise has the potential for
complicating matters enormously. For instance, in the presence of altruism about people's
health, the "extent of the market" issue, which is so easy to dismiss when researchers are
considering only "use" values, must be addressed anew.
For social costing of electricity, the bottom line is mat environmental benefit transfers are
most feasible and reasonable for the health benefit category (although some serious problems
remain) and are not needed for crop damage estimation. Recreation damage estimation
associated with a new power plant is, generally, beyond our abilities, not because the economics
isn't up to it but because of gaps in the science and the lack of baseline recreation participation
information specific to reference environments of interest. Visibility damages fail for similar
reasons—scientific linkages between emissions and changes in visual range are absent Nonuse
value estimation studies for marginal changes in resource quality or quantity are virtually
nonexistent Given these problems, researchers must conclude that estimates of damages
resulting from benefit transfers are not sufficient or reliable enough to support more than a rank
'California dries and Denver have also been the subject of multiple benefit studies.
»
7
-------
ordering of new generation technology options on the basis of social costs. That is, reliance on
benefit transfers to support social cost dispatch or social cost pricing of electricity is probably
pushing benefit transfer (and original study) techniques beyond their capabilities.
PROTOCOLS
Researchers confronting the need to estimate the benefits of environmental improvements
but who, for one reason or another, cannot conduct original research to estimate such benefits,
currently either must rely on simplistic protocols for conducting their benefit transfer study or
find no guidance, except from what they can glean from other examples of benefit transfers. For
instance, the U.S. Forest Service sanctions the use of "unit-day values** for estimating recreation
benefits. But such values are averages over a wide range of site characteristics and policy
scenarios (most examining the value of recreation at a site rather the change in value associated
with a change in site quality) that may be inappropriate for the study site.
Reliance on existing benefit transfer studies is also risky because such studies are not
designed for educating the practitioner on how a reasonable benefit transfer should be (or was)
done, making communication about such protocols dependent on the often haphazard and
incomplete reporting of such procedures. Further, as different benefit transfer studies use
different protocols, the researcher is left with the task of sorting them out This task should be a
subject of a generalizable research effort not reinvented every time by each researcher.
The papers published in Water Resources Research as well as the participants in the
workshop are in close agreement on general protocols for using existing studies, so I do not need
to recount mem in detail here. The care and effort used in conducting a benefit transfer—indeed,
whether researchers should attempt it at all—depend on the commodity being valued; differences
in regional, site, and personal characteristics; and the nature of die original literature being relied
on for the benefit transfer. Given that a benefit transfer is called for, much emphasis is placed on
using demand or value functions where possible, as opposed to using average unit values—be
they for a day of recreation or a day of coughing avoided. Using the function approach puts
some additional burden on the researcher (data must be gathered on the variables at the sooty site
found by the original study to affect WTP, for instance); indeed* without careful reporting of
results in the original study, this approach may be impossible.
f
Nevertheless, in the practical application of these broad guidelines, many choices are
available with few guidelines to follow. What does the researcher do when the valuation
literature is based on changes in physical effects (e.g., catch rates) but no link exists from catch
rates to fish populations or changes in water quality? When the underlying science is poor.
t
j
J
8
-------
should the researcher spend much time guilding the valuation lily, knowing that the final benefit
estimate is only as good as its weakest component? When all of the original valuation studies
have significant problems, either in their own right or for benefit transfer, does the researcher
press ahead or refuse to play? While refusal to come up with an estimate may not be an option
for a benefit analysis on a single pathway (assuming the decision to begin the study embodies
some judgment that some type of estimate will result), it is a real option for social costing, where
many pathways will clearly be left blank. Therefore, adding one more to the list is unlikely to
raise serious objections.
Protocols are perhaps most needed to guide the use of multiple studies on a given effect,
each study with significant flaws, to establish a range of uncertainty. Existing practices vary
widely. Take the use of symptom-day values in a benefit transfer. Three contingent valuation
studies provide such values, each with significant problems, each giving values that are in a
range of a priori plausibility. But because the values themselves are small ($2 to 20/day), small
absolute differences between them can translate into large percentage differences and significant
dependence of the benefit estimates on the values chosen. Some researchers average the
midpoint values and obtain a range by averaging 95 percent values. Others use only midpoint
values from the three studies to represent low, mid, and high estimates of unit values. Others
give up and use judgment Others go with one study judged to be the "best"
Although the above areas could benefit from analysis and codification, one particular
suggested for codification may not yield many benefits: establishing detailed criteria for
evaluating original studies. Beyond stating the obvious—that studies are "good" if they are
based on acceptable theory, the theory links to well done empirics, and essential results are
reported—what more can we do to evaluate studies? The weighting of these criteria is the
crucial element; yet weights depend on the use to which the studies will be put, the policy
setting, and the skills of the researcher in getting around problems or supplementing a study with
other data, for example. A premium should be placed on flexibility for the researcher to include
studies felt to be most appropriate for the problem at hand; the major responsibility in return for
this freedom being to document choices.
The NUSAP system (based on work by Funtowicz and Ravetz) being used for the DOE
Fuel Cycle study may be a useful tool for documenting choices of studies and, in particular, the
BfKrnaiiitifs felt by the researcher in making benefit transfers. NUSAP is an acronym for the
evaluative categories in mis quality and uncertainty message system (Numerical entry, Units,
Spread of values, Assessment of values, and Pjedigree). A separate set of entries would be used
to document choices about emissions, concentrations, impacts, and monetization. Each of these
-------
dements contains subelements, ratings are given for some of the subelements, and the researcher
is encouraged to provide comments explaining the ratings and any other information provided by
die entries (see Table 1). The system as we use it does not involve weighting the various entries
to come up with a score associated with each choice. Rather, it is used to qualify the choice for
the leader or ultimate user of the benefit transfer analysis. This tool would work equally well for
documenting the quality and uncertainties of a single original study as for documenting choices
in the benefit transfer exercise.
RESEARCH AGENDA
To meet the demand for reliable benefit analyses based on secondary sources, major
research efforts are needed. The research agenda spans the following options:
• Develop methods to make better use of existing studies in the benefit transfer process.
• Improve the quality of original studies so that the results of secondary studies will be
more credible.
• Routinely include in the original study design elements to aid in benefit transfers.
i original research with the sole purpose of obtaining results to be used in benefit
transfers.
• Develop incentives for researchers to engage in research supporting benefit transfers.
Making Better Use of Original Studies
To use original valuation studies, researchers must know about them. Many literature
reviews of die benefits of environmental improvements exist, but focus varies and is generally
limited to one category or subcategory. Major efforts are beginning to develop bibliographies
covering die benefits analysis literature. The Environmental Protection Agency's (EPA*s)
bibliography is available on diskette, but it is still by no means comprehensive. Bibliographies
that cut across all benefit categories are being developed in die above cited efforts associated
with estimating die social costs of electricity. Efforts to standardize diese databases and perhaps
merge diem are needed. In addition protocols for indicating where reports and odier unpublished
materials can be obtained are sorely needed. Once the studies are obtained, protocols for dieir
use in a benefit transfer are needed but currently do not exist, as noted above.
Original studies can also be more efficiently used to die extern dial dieir results can be
combined into either a meta-analysis or, if the original data can be obtained, into new analyses on
die combined samples. Such analyses could, hi dieory, estimate values or functions that
eliminate (or at least reduce) die need for ad hoc consideration of multiple studies for
.
j
T
10
-------
1
I
(
•l
I
I
I
Si
II
iii
|m i
.1
6i
i
•e«
_ . Bfl
I If
§
I
e«
§
£
5
*
ill
s| S »
v> I ^a
i,«
"s^
,,l?
OKI
ib
Sod
e»*o g^ t
Illlils
il
VM
1
IB
Lli!!1
N!
is
I
i
i«l«*aii
hlilil
c
JS
Sslaffw-l*
Iflli!|^!8
*«
!
' fit
!*4
!>'
!JB
;s
*'
I-BI?
1
fl
>i
-
S
I
tll!S
-------
s
I
f4
!
"8
'»«•
«•
i
i
•a •
5$
*»
s
!HHI
1
i
1:
y
,!
A
1'
;l«-
III
Si
f
S*i
?1«
I
fi! I
i
•&
*5
»I
S
65
lilllllli ISii
s
1
:£
«J K
If
!i*
ilW
« I *-a
v !
ifl »
I iili
il
rss
i
s
f
if
• •
fe-s
SI
-«
S5
1
!i
8W
nii«!]
itliiil
I'l
1
•M O
?l
w
A
12
-------
P
I
t
I
1
I
I
frequently difficult to find an appropriate equation to transfer and the comparable data for the
NRD A site.
The third approach is a generalized model from which values can be transferred. Such a
model requires much more information that the previous two approaches, but it offers the
advantage of better estimating the site-specific value. The group members discussed the
possibility of adapting the Random Utility Model (RUM) for transfer.
A final option for valuation is the meta-analysis approach. This approach compiles all
available values and their influences and produces a value that accounts for the many possible
influences. Like the generalized model above, the data requirements are extensive. (For nonuse
values, whether such an analysis can be performed given the currently available studies is
unclear.)
The methodology adopted in an NRD A transfer study depends in pan on the timing, the
funding, and the available data. Our group discussion indicated that we would like to see a
movement toward using the generalized model.
RESEARCH AGENDA
Our group discussion revealed that much research still needs to be done on use and
nonuse values for NRDA transfer purposes. We focused on three primary research items: the
design and undertaking of a "grand" study, more and better original studies, and a technique to
generalize RUMs.
The first research agenda item (deemed most important by the group) was the design of
the grand study. Such a study would encompass all types of services and the influences on the
demand for these services. The study would be suitable for transfer purposes and would be
linked to ecological models.
The second research item is the need for more and better quality original studies. Our
group thought more studies on use values would be helpful, particularly on those types of values
for which few studies exist, such as swimming and boating. But more important are studies on
nonuse values. Such research should address fundamental issues associated with credible
valuation procedures. Consensus on transferring nonuse values depends on consensus on
estimating credible nonuse values. The group concluded that good studies on wetlands and
aeabirds would go far in filling our needs for nonuse estimates. New studies undertaken should be
designed with transfer in mind.
13
-------
Finally, we decided that our discipline should take steps to generalize RUM models tor
use in transfer. The goal of this research would be to evaluate how a RUM could be used in the
transfer process. For example, would it be possible to design a large-scale data collection, such
as a multistate region, that would support a general RUM model? Alternatively, another strategy
might be to divide the collected data into subsets that could be used to estimate a RUM for a
specific set of sites relevant for the transfer problem. Finally, the group agreed that better data
lit essential for using RUMs in a transfer setting.
This agenda is ambitious and requires funding. Sponsors of new studies have their own
specific needs, and those needs may not correspond with transfer study needs. This last point
may be particularly true of NRDA litigation situations. Sponsors of any such study have their
own timetable and agenda and may not be willing to subsidize the purely research components of
a study.
Finally, our group concluded that, as a discipline, we need to change our attitudes about
replication. Such studies would be extremely helpful for transfer purposes, but traditionally such
studies are not publishable. Consequently, researchers do not undertake replicative studies, or if
they do, they are not published and generally not readily available to other researchers. However,
we discussed the need to consider using experimental designs to evaluate the validity and
reliability of the previous study. Research progress from simple replications would be far less
informative.
ing Associates. August 1990. Arthur Kill, Kill van Kull, and Tributaries: General Wetland
unwary. Final draft report prepared for Exxon Company, USA.
*
Desvousgcs, William H., and A. Jeanne MiHiken. 1991. An Economic Assessment of Natural
Resource Damages from the Arthur Kill Oil Spill Draft final report prepared for Exxon
Company, USA, Research Triangle Park, NC: Research Triangle Institute.
Exxon Company, USA, Internal company documents.
Parsons, K-C. 1986. The Harbor Herons Project: 1986. New York: New York Audubon
Society.
The Trust for Public Land and New York City Audubon Society. February 1990. The Harbor
Herons Report: A Strategy for Preserving a Unique Urban Wildlife Habitat and Wetland
Resource in Northwestern Stolen Island. New York, NY.
Urbont, D. 1990. The Harbor Herons and the Arthur Kill. The Ethical Culture School: Andrew
Lazes Publisher.
Winfield, Ted, ENTRDC, Inc. August 29,1990. Personal communication with Kristy Mathews,
Research Triangle Institute.
14
B-Laing
Summary,
i
-------
than just with the specific visibility applications in this benefits transfer. The remaining
comments focused on technical issues such as rehabilitation of existing studies, weighting
of results, and sensitivity analyses, for example.
GENERAL BENEFIT TRANSFER ISSUES
In the process of discussing this case study, group members raised several general
benefit transfer issues. Although we chose to focus on the specifics of our case study, we
list these more general issues to provide a more complete picture of the concerns/thoughts
about benefit transfer raised by the group.
• Values through time: Changes in values, changes in income, and discounting
questions roust all be addressed when projecting benefits over some extended
time period.
• Peer review: Questions about whether to use study results that have not been
fully peer reviewed or published in peer-reviewed journals are frequently
encountered. Questions were also raised about what sort of peer-review process
is appropriate for benefit transfer. Some review is always desirable, although
peer-review publication is not always practical.
• Statistics: We generally agreed mat more information than only mean results of
available studies should be used when conducting transfers. Some quantitative
characterization of uncertainty or distributions of study results should be carried
into the transfer.
• Economic theory: Concerns were raised about the consistency of implicit
assumptions in benefits transfer with economic theory.
• Costs of bring wrong: Costs of being wrong should be considered in
evaluating the efficacy of a benefit transfer.
• Underlying study issues: A benefits transfer cannot ignore and is at risk of
amplifying uncertainties in die results of underlying studies. This uncertainty
includes limitations of each study method, such as CVM, travel cost, or hedotuc
property value. Questions of aggregation and total values versus component
values may also be important Before we transfer estimates we need to evaluate
thoroughly what the available estimates tell us about the original study scenario.
• Role of expert opinion: Most transfer exercises involve some judgment on the
part of the researcher. Expert opinion should be acknowledged and key
assumptions identified.
REFERENCES
Brookshire, D.S., R. cT Arge, W.D. Schulze, and M Thayer. 1979. Methods Development
far Assessing Air Pollution Control Benefits. VoL 2: Experiments in Valuing
Non-Market Goods: A Case Study of Alternative Benefit Measures of Air
Pollution in the South Coast Air Basin of Southern California. Prepared for the
U.S. Environmental Protection Agency. Washington, DC.
15
-------
Chestnut, L.G., and R.D, Rowe. 1990a. "Economic Valuation of Changes in Visibility:
A State of the Science Assessment for NAPAP." In Methods for Valuing Acidic
Deposition and Air Pollution Effects. Section B5. National Acid Precipitation
Assessment Program, Washington, DC.
Chestnut, L.G. and R.D. Rowe. 1990b, Preservation Values for Visibility Protection at
National Parks. Draft Final Report prepared for U.S. Environmental Protection
Agency, Research Triangle Park, NC, and National Park Service, Denver, CO,
February.
Loehraan, £., D. Boldt, and K. Chaikin. 1981. Measuring the Benefits of Air Quality
Improvements in the San Francisco Bay Area. Prepared for the U.S.
Environmental Protection Agency. Menlo Park, CA: SRI International.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck, and
M. Thayer. June 1991. Valuing Eastern Visibility: A Field Test of the Contingent
Valuation Method. Draft Report prepared under U.S. Environmental Protection
Agency's Cooperative Agreement 4CR-815183-01-3, Washington, DC.
McClelland, G., W. Schulze, J. Irwin, D. Schenk, D. Waldman, T. Stewart, L. Deck, P.
Slovic, S. Uctenstein, and M. Thayer. March 1990. Valuing Visibility: A Field
Test of the Contingent Valuation Method. Draft Report prepared under U.S.
Environmental Protection Agency's Cooperative Agreement #CR-812054,
Washington, DC.
National Acid
Assessment Pi
Assessment Report. Washington,
(NAPAP). 1991. 1990 Integrated
Rae,DA. 1984. Benefits of Visual Air QuaUty in Cincinnati--Results of a Contingent
Ranking Survey. Prepared for the Electric Research Power Institute by Charles
River Associates. 4RP-1742.
Tolley, GA., A. Randall, G. Blomquist, R. Fabian, G. Fishelson, A. Frankel. I. Hoehn, R.
Krumm, E. Mensah, and T. Smith. 1986. Establishing and Valuing the Effects of
Improved Visibility in Eastern United States. Prepared for die U.S.
Environmental Protection Agency, Washington, DC.
Trijoms, J., M. Thaver, J. Murdoch, and R. Hagement. 1984. Air Quality Benefits
Analysis for Los Angeles and San Francisco Based on Housing Values and
Prepared for the California Air Resources Board, Sacramento, CA.
Trijonis. J., M. Pitehford, W. Malm, W. White, and R.Husar. 1990. Causes and Effects.
of Visibility Reduction: Existing Conditions and Historical Trends— National
Acid Precipitation Assessment Program (NAPAP), SOS/T 24.
J
J
J
I-
16
-------
RECREATIONAL
APPLICATION OF THE TYPE A MODEL
Carol Adairc Jones*
ABSTRACT
The Type A model is the single largest benefits transfer model for natural
resource damage assessment and the only one that has regulatory status for litigation .
under CERCLA and the Clean Water Act In this case study, we focus on the Type A
model procedures for valuing losses in recreational services due to fish kills and
fishery closures resulting from an oil or chemical spill. In addition we discuss how to
value recreational fishery injuries.
I
I
t
I
I
I
I
I
The natural resource damage assessment model for coastal, and marine
environments, the 'Type A model." is the single largest benefits transfer model for
natural resource damage assessment and the only one that has regulatory status for
litigation under CERCLA and the Clean Water ACL The model provides a simplified
assessment procedure for short-term releases of oil and hazardous substances. It
represents a low-cost alternative to Type B damage assessments, which may require
detailed field observations and extensive collection and analysis of chemical, biological,
and behavioral data.
The first-generation Type A model, under review here, was promulgated under
rule-making by the U.S. Department of Interior (DOI) in 1987. It covers the coastal and
marine environment of the U.S. DOI is required to revise the model every two years; this
year, the agency intends to propose a new Great Lakes version, as well as a substantially
*
revised coastal and marine version of the model.
In this case study, we focus on the Type A model procedures for valuing losses in
recreational services due to fish kills and fishery closures resulting from an oil or
chemical spill. The Type A model incorporates the data and algorithms to calculate
fishery injuries, measured as the reduction in (fish stocks and) recreational fishery catch
*NOAA Damage Assessment Center. Members of the case study group included Mark Downing (Texas
A&M), Rick Dunford (Research Triangle Institute), Michael Hanemann (University of California-
Berkeley), Christopher Hansen (U.S. Forest Service), Robert Leeworthy (NOAA), Edward Morey
(University of Colorado-Boulder), Jim Opaluch (University of Rhode Island), Richard Ready
(University of Kentucky). Dan Schniefer (NOAA), Thomas Wegge (Jones and Stokes Associates), and
Peter Wiley (NOAA).
1
-------
by weight due to direct kills, recruitment losses, food web effects, and closures. The
problem posed in this case study is how to value such recreational fishery injuries.
BACKGROUND: TYPE A MODEL FOR COASTAL AND MARINE
ENVIRONMENTS
The model relies on computer modeling to predict the fates and effects of spills
and value the injuries. The essence of the operation is contained in three modules:
physical fates submodel, biological effects submodel, and economic damages submodel
(see Figure 1).
The physical fates module models the path of contamination as it disperses,
determining the concentration of the spilled substance over time and by location within
the study area. This module incorporates a chemical database with the physical and
chemical properties of 469 substances, used in tile species-by-species mortality
calculations.
The biological effects module calculates losses to biological populations through
time. The calculations include the following: the direct mortality to adult, juvenile, and
larval biota due to toxic concentrations; recruitment losses due to stock effects; and the
indirect mortality and weight loss to adult, juvenile, and larval biota due to the loss of
foodstuff in the food web.
The economic damages module calculates the dollar values for injuries to biota
based on use values. It also calculates the losses due to closures of fishing, waterfowl
hunting, or beach areas.
The calculations rely on geographic data bases mat contain avenge resource
distributions for multiple habitat types within ten geographic regions throughout the
coastal US, based on the classification scheme developed in Cowardin et aL (1979).
Marine and estuarine systems are subdivided into subtidal and intertidal subsystems then
broken down into additional habitat classes (based on shoreline type or bottom type).
After the authors factored in the likelihood of each pfovmce-system-subsystem-class
combination and the feasibility of collecting data for each likely grouping, they created a
database with 36 intemdal and 55 subtidal ecosystem types with seasonal variations.
Figure 2 provides a map of the ten regions, and Table 1 lists the habitat classifications
The species in the database are classified into 13 categories, including nine fish
categories. The nine fish categories represent 141 species, including bom finfish and
-------
r
I
I
I
I
I
I
I
•i
User Input:
Spill Type, Location, Date,
Habitat Classification.
Beach/Hunting/Fishing Closures
Physical Fates
Submodel
Chemical
Data Base
"PHYS_BIO.LNK" FILE
Surface Water Column
Bottom Concentrations
•MMMMMW
Biological Effects
Submodel
Biological
DataBase
•BICLECON.LNK" FILE
Biomass Reductions
--.
Economic
(DataBase
Economic Damages
Submodel
MONETARY DAMAGES
Figure 1. Model System Overview (NRDAM/CME)
-------
Source:
4
-------
TABLE 1. HABITAT CLASSIFICATIONS
I
1
r
i
Ecosystem Types
A. 10 Marine and Estuarine Provinces
1. Atlantic and Gulf
PL Acadian (Northeast: north of Cape Cod)
P2. Virginian (Mid-Atlantic: Cape Cod to Cape Hatteras)
P3. Carolinian (South-Atlantic: Cape Hatteras to Cape Canaveral)
P4. Louisianian (Gulf Coast: Cedar Key, Florida to Port Aransas, Texas)
P5. West Indian (South Florida, South Texas, West Indian Islands)
2. Pacific
P6. Califomian (California: south of Cape Mendocino)
P7. Columbian (Pacific Northwest: Cape Mendocino to Vancouver Island)
P8. Fjord (Gulf of Alaska: south of Aleutian chain)
P9. Arctic (Alaska: North of Aleutian Chain)
P10. Pacific Insular (Hawaii and other Pacific islands)
a. Subtidal Bottom Types
S-B1. Rock bottom
S-B2. Cobble (unconsotidated)
S-B3. Sand (unconsolidated)
S-B4. Mud (unconsolidated)
S-B5. Rooted vascular aquatic bed (grasses)
S-B6. Macroalgal aquatic bed (e.g., kelp)
S-B7. Coral reef
S-B8. Molluskrcef
S-B9 Worm reef
b. Intertidal Bottom Types
I-B1. Rocky shore
I-B2. Cobbled beach
I-B3. Sandy beach
1-B4. Muddy shore
I-BS. Saltmarsh (cordgrass)
I-B6. Trees (coastal wetlands)
I-B7. Coral reef
I-B8. MoUuskreef
-------
invertebrates (see Table 2). Four categories of species information are included: adult
biomass, by species; larval numbers, by species category; mortality and growth
parameters by species category; and primary and secondary productivity values.
The model is not intended to represent any specific localized populations of
estuarine or marine situations: the databases represent average values for representative
types of ecosystems. Consequently, to capture the necessary breadth of geographic
coverage, the Type A Model has sacrificed geographic specificity.
CASE STUDY PROBLEM: VALUING RECREATIONAL FISH-KILLS AND
FISHERY CLOSURES
Injury Quantification
Short-term (acute toxicity) losses are calculated separately for adults and larvae
based on the toxicity information in the chemical database and the species distribution
data. The model also calculates long-term losses due to the acute mortality to adult,
juvenile, and larval biota due to toxic concentrations; the reduced recruitment into the
adult fishery due to acute toxicity kills of larvae, juveniles, and adults; and the indirect
mortality to adult, juvenile, and larval biota due to loss of foodstuff in the food web.
The fishery population dynamics in the model are based on the assumptions that
the instantaneous catch rate (or catchability coefficient), the instantaneous natural
mortality, and the growth function for individuals remain constant, and that egg
production and larval numbers return to pre-spill levels immediately following
dissipation of the spill. The architects of the model justify these assumptions on the
grounds that the model is designed for spills of short duration.
Lost catch due to closure of an area to fishing is also calculated based on the
biomass in the closed area. Because some of the lost catch in the closure area is due to
mortality from acute toxicity, only the lost catch due to the closure in excess of the acute
toxicity losses is added to the long-term losses to calculate total catch loss.
-------
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET
i
r
r
i
Species
Number Category*
1 1
2 1
3 2
4 2
5 2
6 2
7 2
8 3
9 3
10 3
11 3
12 4
13 4
14 4
15 4
16 5
17 5
18 5
19 5
20 5
21 6
22 6
23 6
24 6
25 6
26 6
27 6
28 6
29 6
30 6
31 1
32 2
33 2
34 6
35 6
36 6
37 6
38 6
39 6
40 6
42 2
43 2
1 Anadromous fiih
2 PUnktivorous fiih
3 Piscivorous fob
4 Top ctraivonu
Common Name
American Shad
Alewife (and Blueback Herring)
Menhaden Atlantic and Gulf
Atlantic Herring
Butterfish
Pollock
Atlantic Mackerel
Biuefish
Striped Bass
Monkfisb (Goosefish)
Weakfisb (Grey Sea Trout)
Tuna
Swordfish
Sharks
Dogfish
Yellowtail Flounder
Summer Flounder (Fluke)
American Plaice
Witch Flounder
Winter Bounder (Blackback)
Atlantic Cod
Haddock
Redfish (Ocean Perch)
Silver Hake (Whiting)
Red Hake
White Hake
Scup
Tilefisb
Black Sea Bass
Atlantic Wolfiish
Hickory Shad
King Mackerel
Spanish Mackerel
Harvestfish
Atlantic Croaker
Drums
Spot
Yellow Perch
Carp
Eels
Atlantic Thread Herring
Anchovy, Atlantic
5 Dementi fuh S
6 Semi-dementi fisb 9
7 Mollusks 10
Scientific Name
Alosa sapidissima
Alosa pscudoharengus, A. aestivalis
Brevoonia tyrannus, B. patronus
Clupea harengusharengus
PeprUus triacanthus
Poltachiusvirens
Scomber tcombnts
Pomatomtu saltatrix
MOTOM saxatilis
Lophius americanus
Cynoscion regalis
Thiuuuts spp.
Xipkias gladius
Odontaspididae, Carcbarbinidae, etc.
Squalus acanthias
Limandaferruginea
Paralichthys dentatus
Hippoglossoides platessoides
Gtypiocephalus cynoglossus
Pseudopleuronectes americanus
Gladus morhua
Melanogrammus aeglefinus
Sebastes fasciatus
Merluccius bilinearis
Urophycis chuss
Vrophycis tenuis
Stenotomus chrysops
LoptolatUiu chamaeleonticeps.
Caulolatilus microps
Centropristis striala
Anarchichas lupus
Alosa mediocris
Scomberomonu cavaQa
Scomberomorus macuUuus
Ptprilus obutidotus
• itgrw itir *fu*jn*mf»ma
Mifrfttfft tfftti/f T utuiul/ituf
Sfi/ifituinf
Leiottomus xantHurus
Percaflavescens
Cyprinus carpio
AngniUifonnes
Opisthonemaoglinum
Anchoaspp.
(continued)
Decapod* 11 Waterfowl
Squid 12 Sboiebtrds
Mammals 13 Seabiids
-------
-------
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET (CONTINUED)
Species
Number
Category*
Common N»rn*_
Scientific Name
44
45
46
47
4g
49
5Q
51
52
53
54
55
56
57
53
59
60
6i
62
63
64
65
66
67
gg
69
70
71
72
73
74
75
76
77
7g
79
82
g3
g4
g5
2
6
6
6
6
3
3
3
3
3
3
3
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
2
6
6
3
6
6
1
1
1
1
!
2
2
2
Striped Mullet
Sbeepsbead
Spotted Sea Troul
Sand Sea Trout (White Sea Trout)
SeaCatfish
Atlantic Halibut
Bonito (Tunny)
CrcvalleJack
Greater Amberjack
Jacks, Other
Blue Runner
Dolphins
Rounder. Southern
Flounder, Gulf
Drum, Red
Drum. Black
Porgies
Florida Pompano
Kingfish
Sheepshead
Cock
Taut°8
Groupers
Snapper.***
Snapper. Other
Whiting (Southern Hakes)
Spanish Sardine
Silver Jenny
Bonefish
Barracuda
SeaBass
Triggerfish
Salmon. Sockeye(- Red)
Salmon. Cham («Keta)
Salmon, Pink
Salmon. Chinook (-King)
Salmon, Coho(- Silver)
Mackerel. Pacific
MackereUack
Anchovy, Pacific
Mugil cephalus
Archosargus probatocephalus
Cynoscion nebulosus
Cynoscion arenarius
Ariusfetis
Hippogtossus hippoglossus
Euthynnus alletteratus
Caranx hippos
Seriola dumerili
Carangidae
Caranx crysos
Coryphaenidae
Paralichihys lethoaigma
Paralichthys aUnauaa
Sciaenops ocellatus
Pogonias cromis
Sparidae
Trachinotus carolinus
ididae
Lagodon rhombodies
Mauicirrhus spp.
Archosargus probatocephalus
Brosme brotme
Tautogaonitis
Epinephelus spp., Mycteroperca spp.
Lujanus campechaiMS
Urophycisfloridanus
Sardinella aurita
Eucinostoma gula
Albula wipes
Sphyraenidac
Oncorhynchusntrka
Oncorhynchus keta
Oncorhynchus gorbuscha
Oncorhynchus tshawytscha
Oncorhynchus kistach
Scomber japonicus
Trachurus symmetricus
Engraulis mordax
Cbipea harengus pallasi
(continued)
"Category Key
1 Antdromout fish
2 PlMktivoeou* fish
3 Pi»avorout fitb
4 Top cmivonis
5 Danemlfiib
6 Sam-dementi fuh
7 MoUutks
8 Decapod*
9 Squid
10 Mammali
11 Waterfowl
12 ShoiebW*
13 Scabnds
-------
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR BIOLOGICAL
DATA SET (CONTINUED)
I
I
1
I
1
I
1
\
1
I
I
I
I
I
Species
Number Category*
87 5
88 5
89 6
90 6
91 6
92 6
93 6
94 6
95 6
96 6
97 2
98 2
99 5
100 5
101 5
102 5
103 7
104 6
105 6
106 6
107 6
108 6
199 6
Invertebrates
201 7
202 7
203 7
204 8
205 8
206 8
207 9
208 7
209 8
M*V7 0
210 8
211 7
212 7
213 7
214 7
215 8
216 7
217 8
218 8
•Category Key
1 Anadromous fish
2 FlBBkii vorous fish
3 Piscivorous fish
4 Top caraivonis
Common Name •
Flounder, Pacific
Halibut, Pacific
Perch, Pacific Ocean
Rockfish, Other
Perch, Other
Sablefish (Black Cod)
Cod, True (Pacific)
Lingcod
Hake, Pacific (Whiting)
Sea Bass
Pollock, Walleye
Mackerel Atka
Sole, YeUowfm
Flounder, Arrowtooth
Turbot, Greenland
Plaice, Alaska
Smelt
Flounder. Starry
Sole, Butter
Sole, Dover
Sole, English
Sole. Rock
Other Fish
Surf Clam
Ocean Quabog
Atlanta Sea Scallop
American Lobster
Northern Shrimp
Red Crab
Squid, Atlantic
Blue Mussel
Bin* CtA fHanl Stem
** Msw ^•AtnV yiaiaW OUwUy
Blue Crab (Soft Shell)
SoftCLvn
Oyster, Atlantic
Hard Chun (Quanog)
Conch
Shrimp (Brown, Pink. White)
Calico Scallop
Crabs (general)
Stone Crab
5 Demersal fish S
6 Semi-demersal fish 9
7 Mollukc 10
Scientific Name
Pleuronectidae
Hippoglossus stmolepis
Sebastes alutus
Sebastes spp.
Embiotoca spp.. Amphistichus spp.,
Hyperprosopon spp.
Anoplopomafimbria
Gaits macroccphatus
Ophiodon elongatus
Merluccius productus
Serranidae
Theragra chalcograauna
Pleurogramnus mmopterygaa
Lunaniaaspera
Athereahes aomias
Reinhardtius hippoglossoides
PlcuToneaes quadrituberculatus
Osmeridae
Pamlichthys stellatus
Isopsetta isoUpis
Microttomui pacfficus
Panphrys veitulus
Lepidopsetta biltneata
(generic)
Spisulasolidissima
Aftico islondicQ
Placopectcn magcllanicus
Homana americanus
Pmdalus borealis
Gttyon auinouedens
Loligo pealei, Ufa iOtctbrosus
HytOusedtdis
Collintcttf HirHrfid*
Uyaarenaha
Crassottrea virgmica
Mercaiaria nercenaria
Strombustpp.
Paaeusspp.
Argopectengibbus
(generic)
• m . •. . »
mtiuppe mtntnano
(cootinaed)
Decapods 11 Waterfowl
Squid 12 Sborebirdi
Mammals 13 Seated*
-------
TABLE 2. SPECIES LIST AND CATEGORIZATION FOR
DATA SET (CONTINUED)
Specks
Number Category"
219 8
220 7
221 8
222 8
223 9
224 8
225 8
226 7
227 7
228 7
229 7
230 7
231 7
232 7
233 7
299 7
Birds
301 11
302 11
303 11
304 11
305 11
306 11
307 11
308 11
311 12
312 12
313 12
^14 19
31** !«•
315 12
««jr «•%
310 12
321 13
VY) 1^
3*J- 13
323 13
324 13
325 13
326 13
327 13
328 13
•Category Key
1 Anadranoui fob
2ttl__Jrt'uj -m. A*k
rUBKuvuvuui rtin
3 Picdvorous fit b
4 TopCK&ivoius
Common Name
Lobster, Spiny
Abalone
Crab. Dungeness
Shrimp, Pacific
Squid, Pacific
Crab, Snow (Tanner)
Crab, King
Clam, Butter
Clam. Horse
Clam, Geoduc
Clam, Manila
Oyster, Pacific
Oyster, Olympic
Atlantic Bay Scallop
Pacific Sea Scallop
Other Invertibreates
Marsh Ducks
Diving Ducks
Mergansers
Whistling Ducks
Stiff-Tailed Ducks
Coots
Geese
Swans
Sandpipers
Plovers
Turnstones
Ctvtipr PatrhMX
\jjov&i \ »an JHJ a
Phalaropes
Avocetes, Stilts
Gulls. Terns
fVwnvwofitc
vAJUDuniDis
Auks
Shearwaters
Storm Petrels
Pelican1*
Frigatebirds
Gannets, Boobies
S Dementi fiib
6 Semi-demcrMl fob
7 MolliMks
Scientific Name
Panuliris spp.
Haliotis spp.
Cancer magister
Pandalus borealis
LoligoopaltaxHs. Benyteuthis magister,
Onychoteuthis boreali japonicus
Chionoecetes
Paralilhodes camtschatica, P. platypus
Saxubmiu nunalli
Tresuscapax
Panopea generosa
Tapes philippinarum
Crassoarea gigas
Ostrea lurida
Argopecten irradians
Pecten cautious
(generic)
Anatinac
Aymyinac
Merginae
Dendrocygninae
Oxynrinae
Rallidae
Anserinac
Cygninae
ScotopacMflf
Charadriidae
AphrizMar
Tf umfmliLtwummAlAn^
nxxuaiopouiuac
Phalaropodidae
Kccuvinxcnaae
Lvidae
PhfllfVfYVYW^vli^
Alcidae
Pf^w>lfV
Hydrobatidae
prVrmi^»'
Fiegatidae
Sulidae
8 Decapod* 11 Waterfowl
9 Squid 12 SboieMrdi
10 Mimoult 13 Seated*
10
-------
Valuation of Damages
1
I
I
I
I
I
I
Translation from Change in Stock to Change in Trip Catch and Number of Affected
Trips
In the biological submodel, the fish stock is allocated to recreational catch
mortality, commercial catch mortality, and natural mortality based on share parameters
for each species in the database. The predicted reduction in stock due to a spill is also .
allocated to those categories, assuming constant proportions, Jim Opaluch, one of the
authors of the economic module (and a participant in the case study group), indicated that
an assumption implicit in the valuation procedure was that all species are highly mobile;
with this assumption, the change in fish stock will be spread over a wide geographical
area and generally will produce a small change in catch rate (trip quality) over a large
number of trips.
The value per fish, catchability coefficient, level of fishing effort, and cost per
unit effort parameters are assumed to be unaffected by the spill. Consequently, the
decline in recreational fishing catch due to a spill is calculated as the recreational fishing
share of the stock (a parameter in the database) times the change in the fishery stock
calculated in die biological module.
Valuation of the Change in Catch Rales
The valuation procedure then assigns the reduction in recreational stock size at a
rate of one fewer fish per angler. In the calculation, the number of anglers affected just
equals the change in the recreational stock size; there is no independent calculation of
total trips affected. This procedure is a creative way to avoid explicitly characterizing the
levels of fishing participation affected by the spill (which is likely to be larger than die
spill area because of fish mobility).
To generate the recreational fishing values for the Type A model, the authors
relied on two studies providing an estimate of the change in the value of recreational
fishing trips with a unit change in catch rate. Rowe ct al. (1985) provide consumer
surplus estimates for trips to California, Oregon, and Washington marine fisheries from
separate random utility models for each state. For selected species, the scenario valued
was the increase in the catch rate of one species by one fish/trip at all site/mode
combinations where the species is caught Norton, Smith, and Strand (1983) provide
estimates of the changes in consumer surplus with changes in catch rates for several
I
11
-------
striped bass nsnenes on ine c,asi V^OUM. mey c
model.
Because these two studies valued only a few species, the modelers needed a
procedure to provide values for other species. They calculated the change in consumer
surplus on a weight basis for the available species. Judging that the variation in the value
per pound did not appear to vary greatly across the species valued in the studies, they
employed the simple mean of the estimates ($1.84/lb) in the model to value losses of all
species.
QUESTIONS DISCUSSED IN THE CASE STUDY SESSION
We discussed whether the current procedures for valuing recreational fishing
injuries in the Type A model can be improved. We considered the adjustments that
would contribute the most to improving the estimates and the adjustments that are
currently feasible.
The group proposed separate discussions of the injury from fish kills, which we
believed was appropriately valued as a change in quality of the recreational fishery, and
the injury from fishery closures, which we thought might better be modeled as a change
in the quantity of resources available. We consider each modeling context separately
below. For most possible extensions, we concluded that data are insufficient to determine
whether such changes would represent substantial refinements to the model calculations.
The discussion produced a series of recommendations for further research. In the final
section, we discuss criteria to be used in selecting studies for inclusion in the model
database.
MODELING ISSUES
Population Effects Due To Fish Kills And Their Impact On Fish Population
Dynamics
Currently, the effect of fish kills is modeled as a change in the quality of
recreational fishing trips that affects the trip value but does not affect total participation in
the fishery. A single value per gram of fish killed appears in me model: the variation
across species in damages per fish killed is completely driven by variation in average
weight across species. In addition, the value does not vary with the size of the spill (and
the effect on stock and catch rates) or the extent to which available substitutes are
j
12
-------
I
I
I
I
t
I
r
i
i
similarly affected. We discussed several possible extensions to the modeling, as reported
below.
• Expand the single recreational fish value included in the database to a matrix of
values, including variations in the value of lost fish by
—fish species,
—geographical area of spill, and
—user types.
Most members of the group thought incorporating species and geographical
variations could be an important contribution to the model and believed that some
additional values have appeared in the literature since the model was first developed. We
did not think that incorporating variations in consumer surplus values by user types
would make an important contribution.
• Adapt the modeling and expand the value database to incorporate variations in
the change in consumer surplus per unit change in catch depending on
—the level of the change in catch per trip (i.e., avoiding the assumption that the
change in consumer surplus is linear in catch); and
—the extent to which substitutes are affected (which will vary substantially
depending on whether the affected species have localized populations or are
highly mobile over a wide area).
To implement either, it would be necessary to change the modeling to identify the
geographic zone of impact (taking into account the mobility of the species) and the
number of trips taken to that zone. With this information, an estimated change in catch
per affected trip could be calculated (rather man implicitly assigning a reduction of one
fish per trip.). In addition, the Type A model would need a matrix of values in the
database, capturing the nonlinearities and substitution possibilities in the values.
Are the size of the change in catch per trip and the extent of die substitutes
affected important sources of variation in value? The group discussion was inconclusive:
we concluded that research is needed to explore these issues. To the extent that spills
valued with the model are relatively small and the species are mobile, nonlinearities in
the change in consumer surplus with a change in catch rates are not likely to have a large
effect on values. For spills heavily injuring highly localized species, the variation in the
change in catch rate may be much greater, for this context, exploring the possibility of
substantial nonlinearities is more important. Impacts on localized groupings of species
also raise questions regarding the treatment of variations in substitution possibilities.
13
-------
Some preliminary analysis by Graham-Tomasi and Sung with the Michigan recreational
fishing model (Jones and Sung, 1991) suggests that variation in substitution possibilities
has a far greater effect on the value per lost fish than variation in the quantity of fish lost
per trip.
Are these changes feasible? Unfortunately, we had serious questions about the
availability of necessary data. The NMFS marine recreational surveys were cited as a
possible source of data on trips. In addition, we discussed how to implement the
variations in value with nonlinearities and substitution possibilities. Because of the
difficulty of establishing a formula, some individuals in the group suggested creating
categories of "small/medium/large effects" and assigning spills to suitable categories.
However the distinctions are to be implemented, additional research needs to be done to
generate the necessary values for making such distinctions.
• Incorporate changes in fishing participation as a result of spill-induced quality
changes in the fisheries.
Currently, the Type A model treats fishing participation levels as constant when
fishing quality changes based on the assumption of mobile fish species. With this
assumption, the population changes generally being modeled would yield small changes
over a wide geographic area. We concluded that further research would be useful to
identify how elastic trip participation is to quality changes (at the level of quality changes
involved) and the extent to which damages are underestimated by excluding this category
of effects.
Some recent preliminary analysis of the Michigan recreational fishery model
performed by Graham-Tomasi and Sung indicates that, though the participation elasticity
is not large, the share of damages contributed by that behavioral response may be
substantial.
Incorporating this extension in the model would require developing a generic
participation equation. Before this equation could be added, we would need to include
the modeling and database adjustments required to implement it Those adjustments
would build into the model the capacity to identify the zone of impact on the fisheries
(taking into account fish mobility) men determining the impact on trip catch in the
affected zone and the total number of trips in the affected zone.
14
-------
1
I
I
I
I
I
1
I
An additional requirement would be to ensure that the modeling in the fishery
dynamics and the valuation portions of the model are consistent regarding trip
participation. We believed ensuring this consistency would not be difficult.
Fishery Closures
Fish not caught because of a closure are valued using the same procedures as for
fish kills, that is, the total number of trips is assumed constant, but the value of each
affected trip is reduced because of the lower catch rate. This procedure implicitly
assumes a small closure area and the existence of (perfect) substitute sites sufficiently
nearby so that additional travel costs are essentially zero.
We believed considering modeling closures as a change in quantity of fishing
resources would be appropriate. In this case, the correct calculation of damages for a
change in quantity of recreational fishing services would be the change in trips times the
consumer surplus per trip. Ideally, in the studies providing the basis for the consumer
surplus of a lost fishing trip, the species and site characteristics are similar to the closure
area, and the substitution possibilities are similar in both study and spill contexts.
This extension would seem to be more important in cases in which most close
substitution opportunities are not available. The current procedures appear adequate in
cases of a small area of closure.
Incorporating this extension would require trip participation rates and additional
consumer surplus values on a per-trip basis. More studies are likely to be available for
. valuing fishing trips (as needed in mis extension, modeling a change in quantity) than for
valuing changes in the catch rale on trips (as needed for a change in quality).
RECOMMENDATIONS FOR FURTHER RESEARCH
We generally felt mat additional work is needed to explore whether substantial
variations exist in consumer surplus for a change in catch per trip by species, geographic
area, size of the effect, and the extent of substitution possibilities mat are affected. The
group agreed that the current set of random utility models that have been estimated
provides a good basis for such analysis. The participation question also needs to be
explored; this research can be done with die participation models linked to random utility
models or with the earlier generation travel cost models, employing equations estimating
total trips.
15
-------
SELECTION OF STUDIES FOR INCLUSION IN THE DATABASE
The selection of studies and specific consumer surplus value calculations from the
studies is critical to the model database. We addressed the following issue: What criteria
should be applied to exercise quality control in the choice of studies used to estimate
consumer values? We identified three sets of criteria that may be relevant to the selection
of studies:
• relevance of the consumer surplus measure to the context (change in quality,
loss of access)
• quality of study (meets minimum standards)
• comparability of context between the study site and the spill site
However, we did not agree on how to apply the criteria. We did not believe that
the current literature provides enough basis to decide what factors are operationally
important in determining "comparability." And we concluded that the quality judgment
needs to be made within the context of the study's objective and its use in the transfer.
REFERENCES
Cowardin,L.M.,V. Carter, F.C.Golet, and ET.LaRoe. 1979. Classification of
Wetlands and Deepwater Habitats of the United States. Office of Biological
Services, Fish and Wildlife Service, U.S. Department of the Interior, FWS/OBS-
79/31.
Jones, Carol Adaire, and Yusen Sung. July 31,1991. Valuation of Environmental
Quality at Michigan Recreational Fishing Sites: Methodological Issues and
Policy Applications. EPA Contract No. CR-816247-01-2.
Measuring Damages to Coastal and Marine Natural Resources: Concepts and Data
Relevant to CERCLA Type A Damage Assessments. Volumes I and n, and
Appendices A through H (Volume fi). NTIS, PB87-142485.
Norton, Virgil, Terry Smith, and Ivar Strand. 1983. Stripers: The Economic Value of
the Atlantic Coast Commercial and Recreational Striped Bass Fisheries.
University of Maryland Sea Grant Publication No. UM-SG-TS-83-12.
Rowe, Robert W. 1985. Valuing Marine Recreational Fishing on the Pacific Coast.
Nadonal Marine Fisheries Service Administration Report No. LJ-8-18C. June.
Type A Recreational Fishing Case Study.
Type A Regulations: 52 FR 9042, March 20,1987 and Technical Corrections, Type A:
53 FR 9769, and Availability of Corrected Type A Model: 53 FR 9819, both
March 25,1988.
16
-------
I
I
LONG-TERM HEALTH RISKS VALUATION:
PIGEON RIVER, NORTH CAROLINA
Susan B. Kask*
ABSTRACT
Executive Order 12291 requires benefit-cost analysis for all government legislation.
Does this mean that for each piece of environmental legislation we must provide new health
benefits estimates for each illness and each toxin to value benefits? Estimating the benefits
of a reduction in health risks is a difficult task for the policy researcher. In this paper we
present a protocol for transferring health benefits from a study site to a different policy site
and provide an example of its application.
Protection of public health is a primary goal of much of U.S. environmental legislation
because environmental pollution can have a variety of negative effects on public health. For
example air pollution can cause itchy eyes, chronic respiratory disease, and even death for those
most sensitive. These effects, however, occur with some probability. Environmental pollution
increases the risk of exposure to a contaminant, which in turn increases the risk of adverse health
effects (see Figure 1). A benefit from reduced pollution is the reduction in the risk of these
health effects. To evaluate the benefits from environmental pollution control legislation, we
must account for these health benefits.
1
r
i
i
i
i
Pollutant
Releases
to the
'Environment
>.
Pollutant
Concentration
>,
Risk
of
Exposure
>,
Risk of
Adverse
Health
Effects
Figure 1. The Link Between Pollution and Health
Estimating the benefits of risk reduction is difficult for the policy researcher. How much
individuals value a reduction in their future risk of contracting cancer or chronic illness from a
reduction in pollution is a challenge to estimate. Furthermore, estimating the value of reduced
Members of the case study group included
Sergio Ardila (the Inter-American Development Bank), Robert Benens (Oregon State University), Alan
Knipokk (Resources for the Future), Spencer Pence (Consultant), Eirik Romstad (Agricultural University of
Norway). Richard Ruppert, and John Stoll (University of Wisconsin-Green Bay).
-------
risk of acute illness or discomfort from a variety of symptoms is equally problematic. Must we
provide a new estimate for each illness, for each toxin, to value benefits? Studies exist that value
accidental death, death at some future date, and reductions in illness days, for example. Can we
use these studies as proxy estimates across illnesses and toxins? Can they be transferred
spatially? This paper explores the potential to transfer health benefits.
We present a basic model underlying health benefit estimates. We also present the
primary issues and a proposed protocol for benefits transfer. To demonstrate the protocol and
illustrate the pitfalls of transfer, we consider a case study. Finally we present our conclusions
and recommendations for future research.
CONVENTIONAL THEORY OF HEALTH BENEFITS MEASUREMENT
The typical model for measuring health benefits usually begins with a damage or
production function that links self-insurance activities (e.g., medical treatment, purchase of air
conditioners, diet, and exercise) to health. We denote this function as
H = H(Z)
where Z is a vector of self-insurance activities and H is a state of health. In some cases H is also
a function of the level of pollutant (Shogren and Crocker, 1991). The production function may
be represented with a two-state model with state 0 representing good health and state 1
representing death (Smith and Desvousges, 1987), or alternatively, H may represent an index or a
continuum of health outcomes (Dickie and Gerking, 1991; Shogren and Crocker, 1991). Here
we assume a two-state world for illustrative purposes.
*
As shown in Figure 1, pollution affects health through the risk of exposure and the risk of
adverse health effects given exposure. We can include pollution into a probability density
function Q, representing the probability of having good health. This probability depends on the
level of pollution in the environment, which in turn affects the level of exposure of an individual,
and the individual's level of private self-protection. This probability function is
Q * Q(X,Q)
where X is die level of private self-protection and Q is the level of some pollutant in the
\environmenL An alternative approach is found in Smith and Desvousges (1987) where they
-------
I
I
I
I
I
1
f
separate the risk of exposure and the risk of illness, and the level of pollutant aliecis ine
exposure. . .
Each individual has an indirect utility function
*
V = V[M,H(Z)]
where M is their income and H is their level of health. In a two-state world where HO is good
health and HI is poor health, consumers maximize expected utility given some level of pollution
n + (i-tOVj,
m avi
Alternatively, a discrete decrease in Q from QQ to Qi is represented as
n(X, Qo)V(M, Ho) •»• [1 - n(X, Qo)]V(M, HI)
where P represents the WTP1 to maintain the initial level of utility at the new level of pollution
(a Hicksian compensating measure of welfare change). Using a variety of benefits estimation
techniques, we can estimate the value of P given self-protection expenditures.
A PROTOCOL FOR HEALTH BENEFITS TRANSFER
The overriding concern for public health behind much of U. S. environmental legislation,
aad Executive Order 12291 suggests a significant demand exists, and will continue to exist, for
benefit estimates of reduced risk to health. Evaluation of these benefits will require expensive
and time-consuming projects for each substance and health effect Benefits transfer may provide
a solution to satisfying the need for benefits analysis for the variety of environmental legislation
and regulation in the U. S. However, the transfer approach poses potential risks: poor quality
* Smith and Devousges (1987) refer to this value as an option price.
-------
benefits transfers may lead to incorrect policy choices (Desvousges, Naughton, and Parsons,
1992). A sound approach to transfer is necessary.
Benefits transfers apply existing benefit estimates from a study site to a policy site.
Researchers must transfer the issue or commodity from a particular policy site into something
that can be interpreted using existing information (Smith, 1992). What criteria should we use to
transfer health benefits from a study site to a policy site? Table 1 lists bur general recommended
approach for a transfer analysis. We focus on Stage 2, Transfer Criteria, in more detail below.
We identify three areas as the primary focus for a transfer protocol: commodity specification,
market and exchange mechanism, and site and sample characteristics. We discuss each below.
TABLE 1. GENERAL APPROACH FOR TRANSFER ANALYSIS
• Define the purpose of the estimates and the level of precision needed.
• Use proposed transfer criteria (commodity, sample, market, site) to describe study site.
• Select an existing benefit study or studies that satisfy the transfer criteria, keeping in
mind estimates' purpose and precision.
• Determine the appropriate transfer method (e.g., point estimate or confidence interval,
function transfer, Bayesian approach, or meta-analysis).
The Transfer Protocol: Commodity Specification
One of the most important steps in a benefits estimation and benefits transfer is careful
specification of the commodity to be valued. How should we define our commodity when
valuing health benefits? Table 2 identifies six areas for clarification in commodity specification.
Response/Causal Agent: Should we define our commodity based on the substance or
the end result (morbidity/mortality or both)? We recommend mat the commodity in health
transfer studies be defined by the end result, the risk of illness or death. We posit that ultimately
the consumer cares about the health effect (ie., the itchy eyes, coughing, birth defects) and not so
much the source or pollutant that causes the health effect If mis position proves defensible, then
benefits transfer exercises become significantly less complicated because we can consider
reductions in cancer risk from exposure to benzene in the air, for example, the same as a
reduction of cancer risk from dioxin exposure in the water. This position, however, may not hold
true for pollution sources mat have variations in avoidance opportunities and, as discussed in
-------
I
I
I
I
I
TABLE 2. RECOMMENDED COMMODITY SPECIFICATION CRITERIA
I
I
\
i
i
Response/Causal agent
Risk definition
Temporal dimensions
Should we define our commodity based,on the
substance or the health effect?
Are we changing risk through changes in
probability, in severity of a health effect, or both?
Is there a latency
occurrence of health effect?
Voluntary and involuntary dimension Is exposure voluntary or involuntary?
Exposure pathway
Exposure level
Does exposure occur through water, air, and food,
for example?
Is exposure cumulative or acute?
more detail below, morbidity effects. Thus, the role of the causal agent in risk valuation
responses is an important research issue.
If we base our commodity specification on the end result, the illness, we then should
consider the potential to transfer values across illnesses. For example, can we transfer the health
benefit estimates for a reduction hi the risk of death from lung cancer to liver cancer? To best
answer this question let us consider the three general categories for valuation in health benefit
studies: death, illness with no death, and illness followed by death. In die first case, individuals
value mortality alone. A pure morbidity value is provided in the second case and a combined
value in the third. Returning to our question above, an individual may not value death from lung
cancer the same as death from liver cancer, because this is actually a combined value and the
morbidity characteristics may vary across disease. Variation in morbidity across diseases may
include differences in severity or timing for example.
This potential for variation in morbidity characteristics may also causf problems for
transfer across pollutant sources for the same disease. For example, consumers may value
reduced risk of lung cancer from dioxin exposure me same as reduced risk of lung cancer from
asbestos, only if the morbidity characteristics and avoidance opportunities are the same between
causal agents.
-------
Symptoms and the potential for death should be the primary factors used to define the
commodity in a health benefits transfer study. However, the pollutant source may be more
important if avoidance opportunities, or morbidity effects, vary across sources. The cause of the
symptoms, or death (e.g., king cancer versus liver cancer) may also be important to-value
estimates because morbidity characteristics may vary.
Although we have three general categories for valuing health benefits, no studies have yet
valued combined mortality and morbidity impacts. We recommend researchers use mortality
estimates as lower bounds in the absence of combined studies. Because morbidity is already an
element in these measures, adding morbidity and mortality values may result in double counting.
Finally, the units of measurement for the commodity defined are important If health risks are
portrayed as unit days of a symptom, the researcher must consider the problems of over or under
estimation surrounding unit day measures (Morey, 1992).
Risk Definition: Environmentally related health effects can range from acute illness and
discomfort, which may occur with a high probability, to sudden death that may occur with a low
probability. The components of risk include both the probability of a health effect occurring as
well as the severity of that health effect Ehrlich and Becker (1972) recognize that risk can be
reduced by decreasing either element In a laboratory environment, Shogren (1990) found
reductions in probability were preferred to severity reduction. Whether policy changes the
severity of the event or the probability of its occurrence can influence how consumers value a
change in the overall risk. Therefore, when evaluating study and policy sites, researchers must
clarify the component of risk that the proposed policy is changing—probability or severity.
Secondly, considering the direction and magnitude of the risk change is important Does the
probability or severity of the policy under consideration increase or decrease? In the absence of
information on symmetry, researchers should be cautious in transferring the health benefit
estimates from an increase in probability at a study site to a policy site where a decrease in
probability occurs.
Temporal Dimensions: Health effects from environmental hazards range from acute
immediate effects to chronic latent health effects. The temporal dimension of health effects
includes the length of time the illness occurs and the time period between exposure and
occurrence of the illness or death. We cannot assume that consumers will value latent health
effects the same as immediate effects nor assume they would value chronic and acute effects in
the same fashion. Therefore, looking for similarities in the temporal dimensions of the health
effects between the policy site and the study site is important Presumably, temporal dimensions
are similar when the health effect is constant across sites.
-------
I
I
I
I
I
I
I
I
I
I
I
Voluntary and Involuntary Dimension: Although we have stated that the pollutant or
source of a disease may be unimportant when transferring health benefit estimates, in one case
characteristics of the source become important: the voluntary/involuntary nature of exposure to a
health hazard. Environmental health risks are typically involuntary (a person is unknowingly
exposed) as compared to health risks from smoking, drinking, and driving, for example (a person
chooses to incur the risk). Valuation of voluntary risks may be quite different from involuntary
(Starr, 1969; Starr, 1979); thus they should not be used interchangeably. The distinction occurs
because voluntary risks imply some form of control over the risk, and perceived control can
influence the value of risk reduction.
Exposure Pathway: Although we have ruled out the importance of the pollutant's
source in value estimates, we may find that the exposure pathway affects consumer values. This
effect would become relevant if exposure pathways influence our ability to avoid a hazard or the
voluntary nature of exposure. For example, individuals may perceive greater control over the
quality of their water and food than over air quality.
Exposure Level: Exposure to environmental pollutants can range from short time
periods with high doses to long time periods with low doses. How consumers value a change in
health risk will be influenced by these exposure levels, because they influence consumer
probability perceptions and time preferences. Therefore, researchers must choose study sites
with similar exposure levels as policy sites for benefits transfer.
Transfer Protocol: Sample and Site Characteristics
Researchers classify sample and site characteristics in two general areas: the
soctoeconomic characteristics of the sample and the location and temporal characteristics of the
site. Characteristics that should be highlighted in a health benefits transfer study are discussed
below.
Sodoeoonomic Characteristics: Sample characteristics such as income, education, age,
awareness of risk, baseline health, and baseline risk may affect benefit estimates. Because the
sample in a study site is probably not identical to the policy site, researchers must find study site
value estimates that have well-developed valuation models. These models should include the
socioeconomic factors that influence estimates and thus provide more insight into the
relationship between demographic characteristics of the sample and values estimated. Good
understanding and documentation of study site demographics will allow researchers to identify
the sample characteristics that vary across study and policy sites.
I
-------
Location and Temporal Characteristics: lust as socioeconomic characteristics affect
benefit estimates, the researcher must also be aware of certain site characteristics that influence
values. For example, location characteristics possibly important to health benefits estimation
include the presence of insurance programs, access to medical care, potential for avoidance
opportunities, climate, time period of exposure, and baseline exposure levels. The analyst should
establish a relationship between these location and temporal characteristics and the values given
at the study site. As above, reporting of these characteristics for the study site is important.
Finally, as with an original benefits estimation study, analysts must consider the size of the
population affected to calculate total benefits.
Transfer Protocol: Market and Exchange Mechanisms
Psychologists discovered that alternative means of framing a problem can systematically
influence choice and values (e.g., Tversky and Kahneman, 1981). Three important factors
regarding framing effects of a risk valuation problem are the risk reduction technology, the
exchange medium, and the type of question (WTP/willingness to accept [WTA]). Finally, an
additional market issue is the presence of nonuse values in the market. The importance of these
issues for benefits transfer is discussed below.
Risk Reduction Technology: Evidence suggests that alternative risk reduction strategies
influence valuation. Individuals can produce a given reduction privately or collectively.
Individual preference for private or collective reduction depends on the payment's perceived
productivity. Collective reduction may prove more efficient given scale economies, because
many private actions are too expensive or complicated to be economically feasible (Shogren,
1990). However if excessive free-riding is perceived, private reduction may be valued more
highly. Thus, determining the risk reduction strategies most appropriate for the policy site is
important. Figure 2 illustrates the individual's choice of risk reduction actions.
Exchange Medium: One of the most important factors in designing a valuation study is
the exchange medium (or "payment vehicle**). Consumers can pay to reduce the risk of adverse
health effects through wages, taxes, or prices. The medium can influence values given; thus
using a realistic medium for the policy site is important for both benefits transfer, as well as
original benefits studies.
Nonuse Values and WTP/WTA: Analysts must determine whether nonuse values are
relevant and what welfare change measure is appropriate for the policy site. Nonuse values
include the health effects of children, other relatives, neighbors, and friends. Consumers may
value the health of others as well as their own health. However, the extent to which these nonuse
-------
I
I
1
I
DC
,K
0)
8
•5
5
I
-------
values may be embedded within current value statements given by individuals is unclear.
Although not readily available, some measure of nonuse values might be appropriate in health
transfer studies.
Selecting between Hicksian compensating and equivalent measures and using WTP or
WTA depends on the property rights allocation and the direction of the policy change for the
particular policy site. Therefore, well-defined property rights and risk reduction should be
consistent across the sites. Otherwise, extrapolating one value measure for another is
questionable given the theoretically predicted and empirically observed divergence in WTP and
WTA for improved health quality.
Study Selection
Following the transfer protocol suggested above, an analyst can select the study sites
most appropriate for valuation at the new policy site. We recommend that existing contingent
valuation method (CVM) studies be given priority because the alternative approaches have an
array of problems. CVM studies are preferred because of their potential to capture morbidity and
the diversity of possible samples (i.e., general population versus white male workers).
If CVM studies are unavailable, we recommend the few averting behavior studies and
experimental laboratory studies. Hedonic wage models are given a lower priority because of the
narrow sample group and the focus on risk of accidental death. Cost of illness is given the
lowest priority because of its weak theoretical underpinning.
Additional selection criteria may include the theoretical soundness of the study, level of
information reported, and purpose of estimates and level of precision required. Of course the
study site should match policy site specifications to a level the researcher considers acceptable.
A CASE STUDY: LONG-TERM HEALTH RISKS FROM SURFACE WATER
POLLUTION
A classic case of exposure to a long-term health risk is found in Western North Carolina.
Champion Paper currently discharges approximately 43 million gallons of coffee-colored
wastewater into the Pigeon River daily. In addition to the discoloration, a potentially more
serious problem is the risk to public health from the dioxin and other toxins present in the
discharge. The state of North Carolina is considering a weakening of the tnaTimnm allowable
dioxin limit of 14 parts per trillion (ppt). What are the benefits of maintaining the limit or the
costs of raising the limit? This case study provides a working example of the need to transfer
benefit estimates and the many potential problems for the valuation of changes in long-term
health risks from surface water contamination.
10
-------
I
I
I
I
I
I
I
I
I
Hie Site: The Pigeon River originates in Haywood County, North Carolina, as a pristine
stream in the Pisgah National Forest The river flows north, 10 miles, to Canton, where
Champion paper discharges their effluent The river continues northwest, 16 miles, crossing the
Tennessee state border past seven small communities in both states until it reaches Newport, in
Cocke County, Tennessee. Thirty-six miles from the mill, the river empties into Douglas Lake.
The 1990 mean flow rates, north of Canton, vary from a low fall flow of 88 cf s to a high of
10,900 cfs in the spring. The river is regulated by Lake Logan and Walters Lake.
The Pigeon flows through mountainous terrain between the Great Smokey and Bald
Mountains. The river above Canton is used both as a municipal drinking water source, rated
WS3, and for recreational activities such as swimming, boating, and fishing. Downstream from
Canton, the river has been rated as Class C water for boating and fishing only; immersion is not
recommended. A posted advisory recommends against eating fish caught in the river north of
Canton. The 10-mile stretch from Walters Lake to the state line is considered a good "brown"
water rafting run and is sometimes used by recreationists in the area. In Tennessee, the river is
classified and protected for industrial water use, fish and aquatic life, recreational activities
including swimming, irrigation, and livestock and wildlife watering. But, because of the present
level of discharge the river does not meet state requirements for aquatic life or recreational uses.
Tennessee has posted a warning against eating fish from the river, la addition, the present high
color level prohibits any additional waste discharge; thus the river is not used for any other
industrial discharge in Tennessee.
Water Contamination: In 1989, industrial water use accounted for 85.6 percent of
water used in Haywood County. Fifty-one percent of industrial water is used by Champion
Paper in a pulp mill2, paper mill,3 and their utilities and filter plants.4 They produce food board
and fine paper using an integrated bleached kraft pulp and paper manufacturing process.
Pollutants present in the Hicfhaty. in either ««
quantities or regulated by EPA are
given in Table 3. In addition to the pollutants in Table 3, die discharge *!«* affects the stream's
temperature and acidity. The average winter effluent temperature is 29.8°C and the summer
temperature is 37.9°C. Acidity levels range from pH 6.4 to 8.2. EPA temperature limits for
effluent are between 29°to 32°C, with a 13°C if»^y"*um increase in stream temperature. The
acidity limits arc pH 6 to 9.
2lDdndes chip cooking, pulp waiting, icreeoing and teaching, recovery adgcnen
production tfchkxinediaj&tefcrteKttng.
^Produces fine paper, food bond, and dried pulp.
4EPA Anns 1 and 2C snhniaed by J. R. Kupatrick to EPA Reikn TV Office. Aunta,GA.
of cooking chemicals. «J
11
-------
TABLE 3. DISCHARGE POLLUTANTS FOR CHAMPION PAPER MILL IN
CANTON, NORTH CAROLINA (1989)
1989 Sample Values
Effluent Characteristic
Daily
Average
Daily Daily Average
Max Standard Limits
Biochemical Ox Demand (5 Day)
Total Suspended Solids
Fecal Coliform
True Color
2,4,6 Trichlorophenol
Pentachlorophenol
Zinc (one sample)
Chloroform (w/ plant modification)
23,7,8 TCDD (dioxin)
23,7,8 TCDF(furans)
12.5 mg/1 44.4 rag/1
1 1,331 Ibs/day 38,449 Ibs/d
50/100 ml 650/100 ml
1,043 std. units 2,035 std. units
<50ng/l
80ng/l
238 mg/1
6.61 pg/1
5.62 pg/1
30mg/l
42,012 Ibs/d
200/lOOml
50 std unit
3.3mg/l
0.014 pg/1
Commodity Specification: Long-Term Health Risks from Dioxin
Response/Causal Agent: Dioxin exposure causes a range of health risks from life-
threatening cancers of the soft tissues to nonlife-thrcatening skin problems, fertility problems,
and birth defects.5 In addition, evidence suggests dioxin can cause immune system suppression
in mice at low dose levels, and it is a known promoter of other carcinogens.6 Dioxin can
contaminate the air, water, and soil, and exposure occurs through three possible pathways:
inhalation, absorption, or ingestion. Dioxin is more easily absorbed in small doses.
Increasing the exposure levels of dioxin may increase the risk of immunosuppressant
health effects,7 and if accumulated exposure levels increase,8 the population may have a risk of
cancer. Therefore, we may specify our commodity as a particular set of symptoms such as
increased disease days from failure of the immune system to fight colds, flu, and other common
5See Schmidt (1992).
*See Schmidt (1992).
7See Schmidt (1992).
'Dioxin has a long half-life, causing potential accumulation in the body.
12
-------
I
I
I
I
ailments, and as an increase in the risk of chronic illness. We may also specify the commodity as
an increased risk of cancer mortality.
Elevated cancer mortality risk is evident in the health statistics for the area. Both
Haywood and Cocke Counties have cancer rates greater than the national average (see Table 4).
Cancer mortality rates for the two counties range from 7 percent to 35 percent greater than the
national average.9 Chemical workers exposed to dioxin in the U.S. and Germany have been
found to have cancer mortality rates IS percent to 24 percent greater than their national averages
for all cancers. In the U.S. those with long-term exposures to dioxin at chemical plants had rates
87 percent above normal in one study and nine times higher than the general population in
another.™
TABLE 4. AGE-ADJUSTED CANCER MORTALITY RATES (PER 100,000 PERSONS)
I
1
I
I
I
Year
1979 - 1981
1982 • 1984
1985 - 1987
1988 - 1990
Havwood"
135.44
167.89
179.24
NA
Cocke
141.2
158.4
153.7
151.9
U.S>
132.0
133.0
132.7
133.7
"These data are quoted for yean, 1979 through 1981,1981 through 1985,1984 through 1988.
HIS. data are for yean 1979,1981,1984, and 1989, respectively.
Risk Definition: The policy under consideration (increasing the maximum exposure
limits) affects the probability of exposure and thus the probability of immune suppression health
effects, as well as the probability of cancer mortality.
Temporal Dimension: Although the immune system effects occur soon after exposure,
cancer has a latency period. Tne immune system problems persist as long as a potent level of the
*These figures do not comet for other
exposure.
"See Schmidt (1992).
sing behaviors and thus cannot be attributed solely to dioxin
13
-------
chemical remains in the body and thus cause chronic problems given the long half-life of the
chemical.11 The cancers are also chronic.
Voluntary or Involuntary: Exposure to the hazard in our case study is both voluntary
and involuntary. Paper mill workers and those who live in the communities surrounding the mill
voluntarily expose themselves to the hazard, assuming they are aware of the chemical's
presence.12 Although we recognize their relative ability to relocate, downstream residents are
involuntarily exposed.
Exposure Levels: The policy site population has been exposed to low dose levels for
long time periods. Present exposure levels for the communities surrounding the mill and the
downstream communities are considered low. Mill workers, however, may have higher exposure
levels. A July 1989 EPA Fact Sheet (EPA, 1988) on the Pigeon River in North Carolina reported
dioxin levels in fish fillet samples of 2.3 to 80 ppt and wholefish levels of 36 to 91 ppt. In
Tennessee they found 0.17 to 29.3 ppt in fillets.13 The NC state limit for dioxin is 0.014 pg/1 or
14 ppt
Both states have given advisories against eating fish from the Pigeon River, and neither
state has classified the river for use as domestic water supply. Residents along the river or users
of the river have had a lifetime of exposure if they have any regular contact with the river, for
example, through recreational activities such as fishing and boating or through drinking from
contaminated wells. Tests performed in 1987 by the Tennessee Health Department found toxins,
such as furans, contaminating wells of Hartford residents.
Policy Site and Sample Characteristics
Sodoeconomic: Both Cocke and Haywood Counties are rural areas. Table 5
summarizes the 1990 demographic data for these two counties.
Location and Temporal: Both government and private insurance programs are
available to consumers in both counties; medical care is similar to mat available in rural areas in
the U.S. Exposure has occurred over a period of 80 years, the time frame in which the
1 1 AH references ID immune system problems are presently hypothetical because evidence of this health effect has
only been found in mice.
1*ntis assumes these households em afford to move and choose not to. Hie costs of moving ondd be seen as
tivf yjtinuu«e «nr hMvfitt nf health tide n*hirtin«c See averting behtvior HtefSfi&e (AbdaQa, Roach,
and Epp, 1992).
13The higher levels were obtained from the whole body of a boncm-feeding white sucker. Tests of surface-feeding
sonfish yielded a dioxin level of 12 ppt. Hie variation in die levels could be partially explained by food source.
14
-------
I
I
I
I
I
I
I
I
TABLE 5. 1990 DEMOGRAPHIC INFORMATION FOR HAYWOO1) Court A *
AND COCKE COUNTY, TN
Population
Mean Household Income
Mean Education
Male/Female Distribution
Racial distribution (W/B)
Age Distribution
>65
Median Age
Household Size (mean)
Haywood County,,
North Carolina
46,942
$22,698
12.1
47/53
98/1.4%
20.8%
39.9
2.4
Cocke County,
Tennessee
29,141
$17,624
12 (median)
48/52
97/2.1%
12.9%
24.0%
35.2
2.58
I
I
I
I
I
I
I
mill has been operating. Avoidance opportunities are limited but include staying away from the
river, not eating die fish, not working at the mill, and moving. Although all of these would
reduce exposure, airborne and soil contamination are unavoidable to area residents. Finally, the
geographic extent of the market would include those who live in the vicinity of the contaminated
rjortion of the river and in the vicinity of the mill Given the central location of the river and/or
mill in each county and the location of the mill in the two-county area, we can use the county
boundaries for the market's geographic definition.14
Market and Exchange Mechanisms
Bisk Reduction Technology: Dtoxin has a half-life of 7 years, giving a long
detoxification time frame. Thus, some type of reduction strategy is necessary. Source reduction
must occur from either voluntary reduction by industry or government enforcement Because of
the limited number of highly contaminated sites the private sector has little incentive to provide
t4We include Cocke County to ibis cmdy because ifcey «e(Brecuy iffeciBdljyNorti Cwflina legislation relating to
r quality.
I
15
-------
the high incineration necessary for cleaning up toxic sons or s>iuug& nuu... .^
Collective action appears to be the most likely cleanup strategy for source reduction. Individuals
can, however, pursue private averting behaviors such as purchasing bottled water or avoiding the
river for recreational activities such as swimming and fishing. When transferring values we may
consider either collective action or private action values, but the latter may not reflect reduction
in the substance from all pathways (i.e., air. water, and soil).
Exchange Medium: The policy site medium would likely be a city water price or taxes,
both of which can be applied to a collective reduction strategy.
Nonuse Values and WTP/WTA: Nonuse values are likely present for children,
relatives, and possibly others for both the morbidity and mortality impacts. At the policy site,
communities have the property right to clean water, but the stale is responsible for enforcement
of that right. Citizens must convince their government of their preferences; thus we would
measure a consumer's WTP to avoid an increase in the dioxin limit (a Hicksian equivalent
measure of welfare change).
Benefits Transfer: Valuing the Benefits of Maintaining 14 ppt Limit on Dioxin
In this case study we want to estimate the ex ante economic value to avoid an increase in
dioxin limits. Because we have defined our commodity as the probability of morbidity and
mortality effects from long-term low dose levels of exposure, we are estimating the value of
avoiding an increase in the probability of chronic morbidity or cancer mortality, or both.
A significant amount of research estimates economic values for a reduction in the risk of
morbidity or mortality (Gegax, Gerking, and Schulze, 1991; Gerking and Stanley, 1986; Smith
and Desvousges, 1987; Viscussi, Magat, and Huber, 1991). Other studies, such as Berger et al.
(1987), provide economic values for symptom-free days. Many of these studies have focused on
short-term risks where the time between the cause and effect is immediate (accidental death) and
on acute health effects such as burns and coughs. Few studies have looked at the chronic and/or
latent health effects characteristic of our policy site. Using the criteria suggested earlier, we
selected four studies as potential study sites: Viscusi, Magat, and Huber (1991); Gegax,
Gerking, and Schulze (1991); Smith and Desvousges (1987); and Berger et al. (1987)." Table 6
summarizes the characteristics of these studies.
1
15Givcn tiie shortage of morbidity studies available, the Berger et al. (1987) study was selected although ii focuses
on short-term morbidity effects.
16
-------
I
I
1
i
r
£
tn
&
If
l°l£
im
•a
e
i
*
fi
I S E
,1 1
ZZZZZZZ'Z
s
I a I
•5
€
-» •»
•P i
£ *v
I
17
-------
2
6
oo SSr-i
&
fe?
!
OS
g
*
28
2;d
«»*?
18
8 glS
t 28
-------
I
I
I
I
I
I
I
I
I
Both Viscusi, Magat, and Huber (1991) arm
(1987) are CVM morbidity studies, while the Gegax, Gerking, and Schulze (1991) and Smith and
Desvousges (1987) are mortality studies. Note Gegax, Gerking, and Schulze is a hedonic wage
study and Smith and Desvousges is a CVM study. The Gegax, Gerking, and Schulze study
measures WTA for an increase in perceived risk of accidental death. The other studies measure
WTP for decreases in the health risks (Viscusi, Magat, and Huber and Smith and Desvousges)
and WTP to get an increase in symptom-free days (Berger, Blomquist, Kenkel, and Tolley).
Which study should we use for benefits transfer?
Study Selection: Given the specification of our commodity we choose Viscusi, Magat,
and Huber (1991) and Smith and Desvousges (1987) as our possible studies. Both studies value
chronic or latent health effects, which are similar to the same effects from dioxin exposure.16
Smith and Desvousges (a mortality study) and Viscusi, Magat, and Huber (a morbidity study)
provide demographic information and a sensitivity analysis of their results. Both also value a
change in probability not severity. Table 7 compares the Viscusi, Magat, and Huber and Smith
and Desvousges study sites with our policy site.
Although several characteristics of the study sites make them appealing for a benefits
transfer, the sites also have several important problems. First, a critical problem is the difference
in the direction of change for the study sites and our policy site. Viscusi, Magat, and Huber
(1991) looks at risk decreases; Smith and Desvousges (1987) look at both increase and decreases.
Smith and Desvousges find that consumer values are higher for WTP to decrease risk than WTP
to avoid an increase.
If we agree with their findings, we can consider the study sites as upper bound estimates.
Second, the policy site includes both chronic morbidity and latent mortality effects, while the
study sites include only one or the other. As recommcndfid above, mortality figures may be
considered lower bounds. Therefore, we might consider the economic values from both study
sites as upper bounds but also consider the Smith and Desvousges (1987) study values as lower
bounds. The transfer is imprecise because no benefit estimate applies perfectly. The analyst
must now recall the purpose for the estimate and determine the need for accuracy. Finally, must
we adjust for the demographic differences in education and income levels at the policy site?
16Smith andDesvomges (1987) study potability of death from exposure to hazarious wastes, art Viscusi, Magat.
and Huber (1991) «adyieveie(±rancbionchiti*. AWwighdironic bronchitis inay not be a spedftt effect from
dioxin exposure, many of the symptoms nay be similar tnt leu severe.
19
-------
TABLE 7. COMPARISON OF POLICY SITE TO STUDY SITE
Morbidity/Mortality
Riskdef.
Temporal dimensions
Voluntary/involuntary
Exposure pathway
Exposure level
Socioeoonomic
Household income
(mean)
Years education
Male/female
distribution
Racial distribution
W/B
%>65
Household size
% Households with
children < 18
Exchange Mech.
Redact Tech.
Nonuse
WIP/WTA
Vlscusl et al. 1991
Morbidity
• Probability of chronic
bronchitis (1)
(-1/100,000)
Chronic fitness (serious)
Involuntary
Air not specified
Low
$35,000 -$37,000
14«
50/50
2.71
Paired Comparisons
Private
No
WTP
Smith and Desvousges,
1987
Mortality
* Probability of
exposure (4)
(-0.05/50 ... -5/50)
Latent Effect (serious)
Involuntary
Ingestion air
Variable
$32.500
14b
39/61
97/3
17.2
2.7
36
Taxes Prices
Collective
No
WTP
Policy Site
Morbidity and Morality
• Probability (t)
Chronic and Latent
(mild-serious) (serious)
Involuntary/Voluntary
Air, Soil, Ingestion
Low
A &
$22,700 $17300
12" 12*
47/53 48/52
98/1.4 97/2.1
18.2 12.9
2.4 238
30 38
Taxes or Utility Prices
Collective
Yes
WTP
•Mean
^Median
20
-------
I
I
I
Value Transfer: We must determine whether we are transferring an equation or a
specific estimate from the study sites. Whether we use an equation or a specific estimate
depends primarily on the information available from the study sites. If an equation and the
relevant data are available from our policy site,.transfer of an equation would be the preferred
route.
In our particular case, a transfer equation exists for the Smith and Desvousges (1987)
study for the risk increase case.17 Given specific exposure and conditional risk levels,18 age,
income, the number of children in a household, and attitudes to hazardous wastes for the policy
site population, a researcher can calculate an estimate for policy site WTP to avoid a probability
increase. A transfer function is not readily available in the Viscusi, Magat, and Huber (1991)
paper.19
If a transfer equation is not available, a specific estimate can be used. Our study values
would depend on the dose response for dioxin, which establishes the relationship between
proposed policy and probability of the health effect. This relationship can be used to determine
the appropriate risk change for analysis.
The Smith and Desvousges (1987) mean values for WTP to avoid a 1/100,000 end point
death risk increase range from $17.71 to $47.47. The Viscusi, Magat, and Huber (1991)
observation values range from $1.50 to $80.00 per 1/100,000 decrease in probability of chronic
bronchitis, with a mean of $8.83. The Smith and Desvousges and Viscusi, Magat, and Huber
probability levels are significantly different with Smith and Desvousges levels ranging from
conditional probability of death of 1/10 to 1/300. Given mat these two studies use different
approaches and our concerns for double counting raised earlier, these values should be neither
compared nor added together.
Recall that both studies* estimates may be considered upper bounds. The Smith and
Desvousges (1987) study uses probabilities higher man those we might expect for dioxin, the
Viscusi, Magat, and Huber (1991) study is valuing acute morbidity effects that may be more
severe than the acute effects expected from dioxin exposure, and Viscusi, Magat, and Huber
measures values for risk reduction. In both studies the demographics may also suggest higher
values for the study sites due to higher levels of income and education.
"See Smith and Detvonsges (1987), page 103.
ring seated by 1JOOO. Ite tide of exposure is abo
•catod by 1JOOO.
l*Visatti, Magat, nd Huber (1991) do a Beasiuviry natysu but find the various demogrqihic variables ate
tnrigirifirmt
21
-------
The actual choice of a WTP figure must depend on the researcher's policy needs. If only >
rough estimates are required, the above studies may provide adequate guesses. However if more
precise measures are needed, researchers may wish to conduct an original benefit estimation
study.
CONCLUSIONS, LIMITATIONS, AND FUTURE RESEARCH NEEDS
Benefits transfer is significantly more difficult to apply than to discuss in theory. The
most important limitation is the difficulty in finding reasonably similar commodity specification
between the new policy and old study sites. The variation across studies in commodity
specification makes transfers difficult. To ease this problem we suggested assuming the causal
agent does not matter. However, in our study the variation in direction and magnitude of
probability change, the severity of health effects, and the appropriate welfare measure posed
significant challenges for transfer. Exacerbating this problem is the singular focus of studies on
either morbidity or mortality. Although most long-term health risks from environmental
substances include both categories of health risks, the relationship between them has not been
examined in the literature. Aggregation through the independent valuation and summation of
mortality and morbidity impacts may introduce a systematic bias in estimates (Hoehn and
Randall, 1989). This topic is important for future research.
After the above limitations have been adequately addressed, we can then turn our
research focus to the relationships between the demographic, location, and temporal variables ID
value estimates. Further research might also include more studies in developing nations to
enhance our understanding of demographic and cultural variables on economic values and our
potential for international transfers. In addition, the role of prior information on values and
Baysian exchangeability should be studied in more detail (Atkinson, Crocker, and Shogren,
1992). The importance for benefits transfer of documentation and presentation of demand
equations cannot be overstated. A collective effort to organize existing studies and databases is
needed to enhance researchers* ability to conduct transfers.
Further study of disease attributes, causes, and source as they relate to values is
warranted. Can we use hedonic methods to evaluate the relationship between disease attributes
and values? Finally, researchers' have not exhausted the various questions surrounding valuation
methodology as applied to health risk values nor the potential for nonuse values.
22
-------
,gc,a
Land
BIBLIOGRAPHY
Abdalla, C.W., B.A. Roach, and D.J. Epp. 1992. "Valuing Environmental Quality Changes
Using Averting Expenditures: An Application to Groundwater Contamination." Lou
Economics 68(2): 163-169.
Champion. April 27,1989. Letter to U.S. EPA and attached U.S. EPA Forms 1 and 2C.
Atkinson, S.E., T. Crocker, and I. Shogren. 1992. "Bayesian Exchangeability, Benefit Transfer,
and Research Efficiency." Water Resources Research 28(3):715-722.
Berger, M., G. Blomquist, D. Kenkel, and G. Tolley. 1987. "Valuing Changes in Health Risks:
A Comparison of Alternative Measures." Southern Economic Journal 53:967-984.
Center for Health and Environmental Statistics. 1988. "Leading Causes of Mortality: North
i Carolina Vital Statistics." 2.
Cropper, M., and F. Sussman. 1990. "Valuing Future Risks to Life." Journal of Environmental
Economics and Management 19:160-174.
' Cropper, M.L., and W.E. Gates. 1992. "Environmental Economics: A Survey." Journal of
Economic Literature 30(2):675-740.
I Desvousges,W.H.,M.C. Naughton, and G.R. Parsons. 1992. "Benefit Transfer: Conceptual
Problems in Estimating Water Quality Benefits Using Existing Studies." Water
• Resources Research 28(3):675-683.
* Dickie, M., and S. Gerking. 1991. "Willingness to Pay for Ozone Control: Inferences from the
Demand for Medical Care." Journal of Environmental Economics and Management
I 21:1-16.
"Dioxin Re-examined." 1991. Hie Economist 87.
I "Dioxin Via Skin: A Hazard at Low Dosesr 1989. Science News 141.
. Ehrlich.L, and G.Becker. 1972. "Market Insurance, Self-Protection, and Self-Insurance"
l Journal of Political Economy 80:623-648.
"FDA Finds Dioxin in Milk," 1989. Science News 165.
I Fisher, A., L.G. Chestnut, and D.M.Violeue. 1989. The Value of Reducing Risks from Death:
A Note on New Evidence." Journal of Policy Analysis and Management 8:88-100.
\ Gegax, D., S. Getting, and W. Schulze. 1991. -Perceived Risk and the Marginal Value of
Safety." Review of Economics and Statistics 73389-596.
I Gerking, S., and L.R. Stanley. 1986. "An Economic Analysis of Air Pollution and Health: The
Case of St Louis." The Review of Economics and Statistics 68(1).
I Gould, M. 1988. The Most Potent Carcinogen?* Resources 92:2-5.
Hayes, D.,etal. 1992. Valuing Food Safety in Experimental Auction Markets. Mimeo.lowa:
Iowa State University.
23
\
-------
Health Planning Commission. 1992 .Picture of the Present Nashville, TN: Tennessee's Health '
Services, Parts 1 and 2.
Hoehn, J.P., and A. Randall. 1989. 'Too Many Proposals Pass the Benefit Cost Test" j
American Economic Review 789(3):544-551.
Jones-Lee, M. 1974. "The Value of Changes in Probability of Death or Injury." Journal of I
Political Economics 82:835-849. *
Kamrin, M., and P.W. Rodgers. 1985. Diorins in the Environment. Washington: Hemisphere I
Publishing Corp. J
Kask, S., and S. Maani. 1992. "Uncertainty, Information and Hedonic Pricing." Land *
Economics 68:170-184. J
Krupnick, A., and M. Cropper. 1990. "The Effect of Information on Health Risk Valuation."
Working paper. Washington, DC: Resources for the Future. I
Magat, W., W.K. Viscusi, and J. Huber. 1988. "Paired Comparison and Contingent Valuation
Approaches to Morbidity Risk Valuation." Journal of Environment Economics and- I
Management 15:395-411. |
Morcy, E. 1992. "What is Consumer's Surplus Per day of Use! And What Does It Tell Us .
About Consumer's Surplus?" Paper presented at the 1992 AERE Benefits Transfer. I
Procedures, Problems, and Research Needs Workshop, Snowbird, UT, June 3-5. "
Morrison, C. 1989. "Pollutants No Threat, Champion Official Says." Ashevitte Citizen Times, I
Asheville NC, March 8, 1989, No. 61. I
Raloff, J. 1989. "Dioxin: Paper's Trace." Science News 135:104-106. •
"Roast Those Dioxins Away." 1988. Science News 134:158.
Schmidt, KJ7. 1992. "Dioxin's Other Face: Portrait of an Environmental 'Hormone.'" Science I
News 141:24-27. I
Shogren, J. 1990. "Impact of Self-Protection and Setf-Insurance on Individual Response to I
Risk." Journal of Risk and Uncertainty 3: 191-204. |
Shogren, J., and T.D. Crocker. 1991. "Risk, Self-Protection, and Ex Ante Economic Value." .
Journal of Environment Economics and Management 20:1-15. I
Smith, V.K., and W.H. Desvousges. 1987. "An Empirical Analysis of the Economic Value of
Risk Change." Journal of Political Economy 95:89-115. I
Smith, V.K. 1992. "On Separating Defensible Benefit Transfers From 'Smoke and Mirrors.'"
Water Resources Research 28(3):685-694. . i
Starr, C. 1969. "Social Benefit Versus Technological Risk." Science 165:1232.
Starr, C. 1979. Current Issues in Energy: A Selection of Papers by Chauncey Starr. Oxford: I
Perganon Press. ' '
-------
-------
State Center for Health Statistics. 1985. Leading Causes of Mortality: North Carolina Vita!
Statistics. Vol. 2.1985 and 1981.
State of Tennessee Department of Health. November, 1991. "Annual Bulletin of Vital Statistics
for Year 1989." Nashville, Tennessee.
Strum, Carol Van, and Paul E. Merrell. 1988. Reproductive Risks from Consumption of
Dioxin- and Furan-Contaminated Milk Packaged in Paper Containers. Tidewater, OR:
Alder Hill Associates.
"Survey of Buying Power Demographics, USA." 1991. Sales and Marketing Management
Magazine.
"This 'Nontoxic' Dioxin Isn't." 1988. Science News 133.
Tversky, A., and D. Kahneman. 1981. "The Framing of Decisions and the Psychology of
Choice." Science 211:453-458.
'Two E.P.A. Studies Confirm Threat to Fish of Dioxin from Paper Plants." The New York
Times, The Environment, March 14, p. 23.
U.S. Environmental Protection Agency. 1988. Assessment of Dioxin Contamination of Water,
Sediment, and Fish in the Pigeon River System (A Synoptic Study). EPA Report No. 001.
U.S. Environmental Protection Agency (EPA). 1989a. "Authorization to Discharge Under the
National Pollutant Discharge Elimination System, Permit #NC0000272." September.
U.S. Environmental Protection Agency (EPA). 1989b. Tact Sheet Amendment Based on
Comments Received July 12,1985 to August 25,1989." September.
U.S. Environmental Protection Agency (EPA). 1989c. Tact Sheet: Application for National
Pollutant Discharge Elimination System, Permit to Discharge Treated Waste Water to
U.S. Waters." Appl. #NC0000272f July.
U.S. Environmental Protection Agency (EPA). 1989d. "Public Notice #89NC003." September.
U.S. Geological Survey. 1991. "Water Resources Data, N.C., Water Year 1990." Water-Data
Report, NC-90-1.
Viscusi,W.IC,W.Magat,andJ. Huber. 1991. "Pricing Environmental Health Risks: Survey
Assessment of Risk-Risk and Risk-Dollar Tradeoffs for Chronic Bronchitis." Journal of
Environmental Economics and Management 21:32-51.
25
-------
r
-------
RECREATIONAL FISHING VALUATION: ACID RAIN
PROVISIONS OF THE CLEAN AIR ACT AMENDMENTS
*
Mary Jo Kealy, Susan Herrod, George Parsons, and Mark Montgomery*
ABSTRACT
Our work group developed a research protocol to assess the likely magnitude of the
economic benefits of improved or nondegraded recreational fishing that are expected to result
from implementing the Clean Air Act Amendments of 1990. We used data for the study site
from the 1990 NAPAP Integrated Assessment, which includes Maine, New Hampshire,
Vermont, and New York. The policy site includes Pennsylvania, Virginia, West Virginia,
Maryland, New Jersey, and Delaware.
\
Congress mandated in §812 of the Clean Air Act Amendments of 1990 (CAAA) that
EPA conduct a comprehensive analysis of the impact of the CAAA on the U.S. economy, public
health, and die environment This analysis is to include costs, benefits, and other effects
associated with compliance with each standard issued for emissions of sulfur dioxide (SO^ and
nitrogen oxides. Title IV of the CAAA mandates a reduction in SOj emissions of 10 million
tons per year, with a national cap on SO2 taking effect in the year 2000.
With the reduction in these precursors to acidic deposition, water quality improvements
are expected. A potentially significant source of economic benefits from improved water quality
is enhanced recreational fishing. This case study involves developing a research protocol to
assess the likely magnitude of the economic benefits of improved or nondegraded recreational
fishing that are expected to result from the implementation of die CAAA to control precursors of
acidic deposition.
Although substantial improvements (nondegradations) in water chemistry and fish
populations may be attributed to die CAAA for duee regions of die country (Le., Adirondack
region in New York, Mid-Atlantic Highlands, and Mid-Atlantic Coastal Plains), a preliminary
economic assessment has been completed for die Adirondacks only. This area together with
'U-S.Envir
ltd Protection Agency, U.S. Environmental Protection Agency, Univenity of Delaware, GrinneD
College, respectively. Members of die case study group included Trudy Cameron (Uni verity of Quifanna-Lot
Angdes),Je^Ftetcber(WestVirgmuilJniversity).Myrtt
(Environmental Law Institute), Reed Johnson (Research Triangle Institute), Doom Lawson (NOAA Damage
Assessment Center), Greg Michaels (Abe Amoriaiei, Inc.), Andrew MuOer (McMnster University), State Navrud
(Noragric, Agricultural University of Norway), and Robert Unswonb (Industrial Economics, be.). Hie views
expressed by the authors of Jnis paper do not necessarily represent tiwie of uieU.S. Environmental Protection
Agency. Responsibility for errors ""^ nfnjyffapy TMimtnt with thf authors.
-------
three other northeastern states (i.e., Maine, New Hampshire, and Vermont) was studied as pan of
die National Acid Precipitation Assessment Program (NAPAP) and preliminary results were
included in the 1990 Integrated Assessment.
At die time of the Assessment, the Adirondacks (and the rest of the Northeast) was the
only affected region of the country for which all of die linkages from emissions to fish
population declines were established. Therefore, die limited resources for die economic analysis
were devoted to assessing damages to die recreational fishery in this region. Finally, die analysis
was limited to losses to anglers, and researchers made no attempt to assess any potential nonuse
values associated with die changed water chemistry and biota. The contingent valuation method
of assessing nonuse values was considered too controversial to survive die NAPAP peer-review
process.
The Assessment's National Surface Water Survey (NSWS) encompassed all of die
regions of the country thought to suffer adverse water chemistry conditions from acidic
deposition. However, die Assessment ascertained that only die Adirondack and Mid-Atlantic
regions have potentially high losses in waters suitable for die survival of certain fish populations.
The economic losses to recreational fishermen in die Mid-Atlantic regions still need to be
assessed. Unfortunately, die linkages from emissions to fish populations are less definitive for
die Mid-Atlantic regions, particularly die Coastal Plains, tiian for die Adirondacks. Moreover,
relative to die costs of controlling emissions, die benefits of improved fish habitat and
populations are likely to be quite small so dial a full-scale original study may not be warranted.
However, two arguments can be made for a less ambitious analysis. First, on a regional
scale, die damaged conditions of die fishery may represent a significant loss and a
disproportionate burden. Second, recreational fishing damages from fish population losses are
but one effect of acidic deposition to be considered along wim odier damages such as, health
effects, impaired visibility, and materials damages. Note diat widi die probable exception of
heahh effects, each of tiiese effect categories includes uses and nonuse values dtat are affected by
acidic deposition. Therefore, aUhough a full-scale original study of recreational fishing in die
Mid-Atlantic region may not be warranted by definitive science or die relative costs and benefits
of §812 of die CAAA, a less ambitious assessment of die likely extent of damages is appropriate
in tiiis policy context One of die goals of tiiis benefit transfer research protocol exercise is to
describe die extent of analysis required by die policy context
Consistent witii die Assessment, die research protocol described here does not address
nonuse values. That topic warrants separate treatment and is beyond our scope.
F
i
-------
THE BENEFIT TRANSFER RESEARCH PROTOCOL
f
I
I
A benefit transfer can involve a fairly simple practice such as applying estimates of
benefits from one study to an entirely new situation. If multiple, related studies are available,
researchers may construct weighted averages of benefit estimates. The original functions that
generated the benefit estimates can themselves be transferred, and available data from the policy
site can be used in place of the means from the study sites to simulate the models. Ever
increasing levels of effort can be directed toward methods of assembling, analyzing, evaluating,
combining, and interpreting existing information on how people are affected by a change in
conditions, and these methods all qualify as benefit transfers.
In this paper, we develop a benefit transfer protocol for exploiting existing data collected
in an original study, rather than the values or functions estimated from these data. By having
access to the data, researchers are not restricted by the modeling assumptions of the original
study. Furthermore, we can consider methods of combining the existing data with data from the
policy site.
The four types of data needed in an assessment of recreational fishing benefits are
• behavioral data (e.g., where do anglers fish and how often?);
• population and angler characteristics (e.g., income, age, tastes, and attitudes);
• site characteristics (e.g., fishing quality, size of the water body, cost of access,
geographic distribution of waterbodies by type and in relation to the angling
population); and
• policy variables (e.g., fish catch rates, presence of fish species, Acidic Stress Indexes).
Our original data for the study site are from the 1990 NAPAP Integrated Assessment, which
includes Maine, New Hampshire, Vermont, and New York (Shankle et al., 1990). The policy
ate includes the Mid- Atlantic states of Pennsylvania, Virginia, West Virginia, Maryland, New
Jersey, and Delaware. The data from the Northeast on recreation behavior, site characteristics,
population and angler characteristics, and policy variables, may be used alone or in combination
with policy site data on these parameters. Presently, population characteristics are readily
available for the Mid- Atlantic regions, and we anticipate the future availability of some policy
site data on angler characteristics and recreation behavior (e.g., National Recreation Survey).
Site characteristic data exist for the policy site, but accessing these data and linking them with
the recreation behavior model is a labor-intensive task. Finally, aggregate data on the range of
changes in the policy relevant variables are available in the policy region, but these data may not
import well into the recreation behavior models that rely on "rite"-specific data.
-------
We develop a benefit transfer research protocol that breaks the analysis down into stages.
The progression from one stage to the next is based on a value of information analysis similar to
the one presented in Deck and Chestnut (1992) and based on Freeman (1984). The titles for
some of the stages of the research protocol have been generalized, however, to accommodate our
more encompassing interpretation of the types of analyses that qualify as "transfers." At each
stage of the analysis, we attempt to evaluate the benefits and costs of proceeding to the
subsequent stage. We based the decision on the cost of obtaining increments in the quality of
benefit information relative to an assessment of how important the quality increment is to the
policy context Finally, in our conclusions we suggest some changes in the way we do empirical
research to make benefit transfer practical as well as defensible.
Stage 1 begins with the Qualitative Assessment of the economic significance of the
damaged recreational fishery. Assuming significant damages have occurred and the policy will
result in a reduction in damages, the Transfer Scoping Analysis is designed. The purpose of
this second stage of the exercise is to assess the availability and relevance of existing information
(e.g., studies, reports, databases). The third, or Benefit Transfer Computation/Estimation,
stage is to determine how best to synthesize, analyze, and otherwise interpret the relevant
information to quantify the economic benefits associated with the policy. Here, we attempt to
specify and estimate recreational fishing demand models using study site data (i.e., the
Northeastern states) alone or in combination with other available data sources. If these data
sources are inadequate for providing credible estimates of the recreational fishing benefits of
reductions in acidic deposition in the Mid-Atlantic Highlands and Mid-Atlantic Coastal Plains,
then, moving to the fourth stage may be necessary. The Update/Validate stage involves at least
some primary data collection (e.g., a pilot study) and model estimation, most likely using
procedures for combining data from different sources. The forthcoming National Recreation
Survey is described briefly because it may provide relevant, but thin, site-specific data that can
be combined with other data to update or validate an existing model. For completeness, the fifth
step in the Deck and Chestnut (1992) proposed protocol is an original study. We omit this step
because it does not involve a "transfer" at all.
i
Stage 1: The Qualitative Assessment
The objective of the qualitative assessment is to determine the likely economic
significance of the changed condition due to the policy. Two important factors influence any
conclusions that can be drawn at this preliminary stage of the analysis. The first relates to the
magnitude of the change in the condition of the environment that results from the policy and
whether an economically relevant endpoint can be measured. The second involves the sensitivity
-------
To illustrate how WESML estimators might apply in benefits transfer situations, a simple
numerical example may be helpful. Consider a RUM model where only two variables affect
choice: respondent income and catch rates. Suppose that the study population is one million
people with joint frequencies for income and catch rates as given in Table 1 A. (Note that the
groups in this example are extremely coarse and that frequencies are measured in 10,000's.)
Suppose that a study sample of 50 respondents yields the joint sample frequencies shown in
Table IB. To inflate or deflate the influence of each sample observation so that the weighted
study sample mimics the study population distribution of attributes, the weights will be as given
in Table 1C.
WESML estimation will produce a set of utility parameters, p, that can be argued to
represent the best parameterization of a "typical" or "average" set of preferences for the study
population. For benefits transfer, however, we would prefer to have a set of parameters, p, that
represent the typical preferences of the "policy" population. If the researcher has access to the
full set of data used to calibrate the original study sample model and obtaining an approximate
joint distribution of the exogenous variables for the policy population is possible, the following
modified weighting scheme seems appropriate. Intuitively, researchers would simply construct a
set of weights for use in the WESML algorithm that serve to make the study sample
representative of the policy population, rather than the study population.
To continue the simple illustration, suppose that the policy population (also one million
people) has the joint distribution of exogenous variables given in Table 2A. The set of weights
necessary to make the sample with frequencies as in Table IB representative of this alternative
population appears in Table 2B. WESML estimation of the RUM specification using these
weights will produce a different set of estimates for the P vector of preference function
parameters—one that better approximates the typical preferences of this new population.
Reviewing the data requirements necessary to make mis reweighting scheme work is
useful, first imagine the ideal case. With unlimited data on a vector of individual-specific
sociodemographic variables, X, and a vector of individual-specific environmental amenities, Z,
researchers might imagine calibrating a full parametric continuous joint density function fP(X, Z)
based on exogenous sample data for the policy population. Researchers would analogously
calibrate a full parametric continuous joint density f*(X, Z) for the study sample.4 With these
4In our earlier numerical example, fundamentally continuous distributions for income and catch tales were
aggregated into four cells so that a simple discrete distributions could be used to form the weights.
-------
TABLE 1A. STUDY POPULATION FREQUENCIES (104) ,
Income Catch
Low
High
Total
Low
40
20
60
High
10
30
40
Total
50
50
100
TABLE IB. STUDY SAMPLE FREQUENCIES
Income Catch
Low
High
Total
Low
High
10
10
5
25
15
35
Total
20
30
50
TABLE 1C WEIGHTS TO MAKE STUDY SAMPLE ESTIMATES REFLECT STUDY
POPULATION FREQUENCIES
Income Catch
Low
High
Low
2
1
High
1
0.6
-------
TABLE 2A. POLICY POPULATION FREQUENCIES (10«) -
Income Catch
Low
High
Total
Low
High
25
25
15
35
40
60
Total
50
50
100
TABLE 2B. WEIGHTS TO MAKE STUDY SAMPLE ESTIMATES REFLECT
POLICY POPULATION FREQUENCIES
Income Catch
Low
High
Low
High
1.25
1.25
1.5
0.7
two continuous joint densities, researchers could then calculate (unique) individual-specific
fl)fY 7^
weights based on the ratio f»Vv' Zl *°r cac'1 individual's own vector of values for X and Z.5
This level of detail is highly improbable for current real applications. Multivariate joint
densities are simply too difficult to calibrate unless normality is invoked and even this
assumption may often be questionable. Furthermore, the raw data necessary to calibrate the full
joint density function fP(X, Z) are not typically available, at least with current information
technologies. For sociodemographic variables, official Census descriptive statistics will
sometimes provide two- or even three-way cross-tabulations of variables such as age, income,
and ethnicity, but these cross-tabulations are rarely available for specific subpopulations. Much
of the raw data exist; the infrastructure for extracting arbitrarily designated subsets of the
population is simply not yet as readily accessible as researchers might like. Data on the
environmental attributes are even more scarce, and when they are available, researchers must
frequently assume statistical independence between the X and the Z variables because these are
typically drawn from different sources. Because full vectors of both X and Z values are not
^Recall that the weights in our numerical example were only group-specific, not individual-specific, and that only
four groups wen defined.
-------
extracted from the same individuals, the joint density cannot be estimated. Information
technology promises great strides in this area in the future, however.
In the meantime, researchers will have to make do with nonparametric frequency
information over matching "cells" in the policy population and the study sample. This method
requires comparable domains for f*(X, Z) and fP(X, Z). If the domains did not overlap, weights
could not be constructed. The number of partitions along each dimension of (X, Z) will be
dictated by the study sample's size. If some cells are empty, they can frequently be merged with
adjoining nonempty cells for both the study sample and the policy population. However, if too
many cells that are well-represented in the policy population are empty in the study sample,
researchers will have problems. In general, the more refined the cells, the better, but a tradeoff
exists between resolution (the fineness of the cell partitions) and cell frequency deficiencies.
Cell designations are entirely subjective.
Researchers have argued that simply transferring point estimates of benefits from a study
area to a policy area is generally not wise (Loomis, 1992). Point estimates depend on a vector of
estimated parameters as well as a matrix of exogenous variables. Thus, this argument
recommends (correctly) that transferring the point estimate of mean value from the study to the
policy area is unwise because fundamentally different values of the exogenous variables may
apply in the policy area. Instead, transferring the entire model is preferable, applying it to new
(mean) values of the exogenous variables for the policy population. The reweighting scheme
described here goes one step further than "model transfer.*' It avoids not only the assumption
that the exogenous variables are identical in the two regions but also the assumption that typical
preferences for the study region and the policy region are identical.
Preferences may indeed be systematically different if the study involves endogenous
location choice or if fundamental preferences are not uniformly distributed across the entire
country (we usually assume that they are). The disadvantage is that recalibration of the study
model with different weights requires that the full study data set be available. The full data set
will not always be available, although pressure is mounting in the economics discipline to
preserve estimating samples and documentation as a condition for publication.
LEAMER'S BAYESIAN DATA-POOLING MODEL
Edward Learner (1991) has recently proposed a Bayesian econometric methodology that
appears to have much to offer benefits transfer practitioners in terms of focusing our agenda for
improving quantitative procedures. The current framework for Learner's model is OLS
regression, and the application he uses to illustrate the approach is a convenience sample of data
-------
pertaining to GNP growth in developed and developing countries. His application tests the so-
called "convergence hypothesis" (that higher initial GNP implies slower growth rates across
countries). His two samples are developed countries (assumed to provide good quality data) and
developing countries (assumed to provide poorer quality data). Although Learner's application is
not benefits transfer, he injects valuable rigor into the explicit modeling of many judgments
similar to those made in every application of benefits transfer.
The problem is one of combining information about some economic quantity from two
data sets of differing quality. Data pooling appears in benefits transfer exercises when
alternative study samples are combined either to provide transferable benefits estimates or
transferable models. It also takes place when study samples are pooled with small-scale policy
samples to "update" the study information with policy area information.
Learner's method is Bayesian and uses prior information about regression coefficients.
Estimates from pooled data depend on three types of parameters:
5 = the investigator's lack of confidence about the prior,
p ss the subjective degree of similarity between the "study" and the "policy"
relationships;
Xi = the amount of contamination of (for example) the "study" (i = 1) and the "policy"
(i-2) data caused by such things as measurement errors, left-out variables, and
simultaneity, for example.
Learner's basic specification for the pooling of contaminated data across data sets i=l,2 is
as follows:
Yi =
(5)
where the p*i are the true parameters and 81 is a bias vector due to the statistical pathologies of the
data. From this specification, extreme multicollinearity clearly exists. Nevertheless, Learner
shows that the informational deficiencies of the underidentified model can be overcome with
prior information. He assumes that 6j - N(0, VO and resorts to the random coefficients model
given by
ft]- •€]«)
(6)
-------
where p is the most likely common structural parameter vector and U measures departures from
this vector. Learner notes that this parameterization conveniently allows a relative lack of
information about P but confidence that the difference between Pi and fc is small (i.e., for large
U and p near unity).
The prior covariance matrix for the model in Eq. (5) is then given by
V > Var
Pi
P2
61
L62.
U pu 0 0
pU U 0 °
0 0 V! 0
LO 0 0
Still, depending on the number of variables in the vector X, this can represent a daunting number
of unknown parameters about which prior values must be asserted. The number of prior
parameters can be reduced substantially by adopting the constraint Vj = XjjU where X* measures
the relative importance of experimental contamination (Le., a high value of Xj means that the
investigator wishes to discount the information in that sample).
The number of prior parameters can be further reduced by making U = o^Uo* where Uo is
the prior on the amount of noise in each of the Pi vectors. 6 is then interpreted as the "discount
rate" on the prior variances. With these simplifications (for greater tractability), the researcher
now needs to specify priors only for the vector p and the matrix U, as well as the scalars 5, p, Xi,
and X2 (in the two-sample case).
The innovations in Learner's approach (despite the current estimator being demonstrated
only for the OLS context) include the following:
• specifying a generalized random coefficients model for combining information;
• incorporating errors-in-variables concerns and other pathologies, which allow
assumptions about the extent of these pathologies to differ across samples; and
• adhering to the desirable Bayesian econometric paradigm.
Conceptually, this approach has much ID offer benefits transfer research. It formalizes explicitly
what we all do while searching for "relevanr studies to be used for benefits transfer. Consider
the Xi (unreliability) parameters. The larger X* is, the less weight is put on sample i's results in
averaging its information with the prior. By discarding studies, we implicitly assume that Xi goes
to infinity; by using a study, we implicitly assume that Xi goes to zero. A better strategy would
be to use expert judgment about the qualities of different studies (and their relevance) to assign
l»
i
10
-------
f
0 < Xi < «*» appropriately for each study. Learner's conceptualization forces us to reveal our
assumptions explicitly and allows for intermediate values of the Xi parameters, rather than
limiting them to the extremes of zero or infinity.
It will be some time yet before Learner's OLS procedures are adapted to MLE contexts
and then to RUM parameter estimation tasks. Hie computer algorithms are complicated even in
the context of OLS. However, Learner offers benefits transfer theorists and practitioners
something to strive for. His insights could lead to some very useful dissertation work in the
hands of an environmental econometrician. The benefits transfer literature directly needs
statistical methodologies that force practitioners to be specific about their priors overall (as on p
and U) and their priors as they embark on the blending of multiple sources of information
(namely 8, p, and the V s).
NATIONAL RESEARCH COUNCIL REPORT ON CI
A subcommittee of the National Research Council recently convened a panel to study and
report on "Statistical Issues and Opportunities for Research in the Combination of Information"
(Gaver et al., 1992). This report has just recently been completed, and almost all of its findings
are relevant to the discussion at the AERE benefits transfer workshop.6 The practice of
combining information apparently takes place in almost every quantitative discipline with
important lessons being learned at different rates by different groups. The terminology varies
across fields. For example, it is called "data fusion" in the defense industry and "meta-analysis"
in several social sciences. The report provides a wealth of information and insight into research
opportunities by examining a broad range of case studies in different disciplines.'
Because the report will be readily available, this paper merely summarizes and
paraphrases its main conclusions, many of which echo the sentiments of the different teams
working on case studies at the AERE workshop. (The quotes in die following points are drawn
from the conclusions section of Gaver et at, 1992).
• "Authors and journal editors should attempt to raise the level of quantitative
explicitness hi the reporting of research findings." Documenting data and models is a
clear necessity for improved benefits transfer exercises. Ideally, all study sample data
would be freely available, allowing the widest variety of transfer techniques, including
re-weighting.
• "CI based only on P-values should be avoided in favor of estimates of quantities of
direct decision-making relevance, together with uncertainty estimates." The crudest
methods of CI across studies will ascertain whether a particular explanatory variable is
6I am grateful to David Draper (of RAND and UCLA) for providing a preliminary copy of this report
11
-------
a significant determinant of the outcome variable and allow these results to be "ballots"
in a vote on whether the variable explains the outcome. Slightly more sophisticated
methods use the unit-free prob-value (or P-value) associated with the coefficient on the
crucial variable in different studies, averaging these continuous quantities, possibly
with sample-size weights, to ascertain the overall judgment of whether the variable
explains the outcome. This recommendation.advocates that significance or
nonsignificance is not the important issue; rather, the magnitude of the effect of the
variable on outcome ought to receive the attention in CI exercises.
* "It is worth investigating the costs and benefits associated with going beyond numerical
summaries to data registries or archives (for both published and unpublished studies)."
This issue is addressed by David's (1992) paper on data accessibility.
• "Increase the explicitness in the formulation of models that express judgments about
how information sources to be combined (subjects, variables, research studies, bodies
of expert opinion) are similar (exchangeable) and how they differ." This point
corresponds directly to the advances offered in the paper by Learner (1991) outlined in
the previous section.
• "The practice of CI could benefit from increased use of sensitivity analysis and
predictive validation."
* "Hierarchical statistical models are a useful framework for CL Use in fields where they
are not yet routinely employed is to be encouraged, as is an increase in the coverage of
such models in intermediate and advanced statistics courses." Econometricians do not
routinely teach or use these methods, but these methods merit close scrutiny for
application to benefits transfer.
• "CI modeling could be improved by increased use of random effects models in
preference to the current default of fixed effects." This terminology is somewhat
confusing to econometricians.7 Translated, this recommendation advocates random
coefficients models, rather than the more familiar nonrandom coefficients models. "At
a minimum, we believe that researchers will often find it useful to perform a sensitivity
analysis in which both lands of models are fit, and the substantive conclusions from the
two approaches are compared."
• Researchers need a "general-purpose computing package allowing researchers to
perform interactive Bayesian "
Learner J>»» advocated intent
to be enhanced and 4tiw"inTi*»d more broadly.
• More meta-analysis should be undertaken. Researchers need to do "more work on the
design of meta-analyses and related CI exercises" and pay "increased attention to
alternative analytic approaches."
• "Workers using CI procedures..." in benefits transfer would profit from a study of CI
methods used in other fields, and funding agencies should give a higher priority to
"cross-disciplinary conferences on methods for combining information."
analysis in hierarchical models in a routine manner."
tive Bayesian software, but these algorithms clearly need
V
j
'it is used differently in the econometric nalysis of panel
12
-------
THE PUBLIC GOODS NATURE OF WELL-DOCUMENTED DATA SETS
Well-documented data sets in general machine-readable form are a valuable-public good.
They are rarely available because the private costs to researchers of providing the data almost
always outweigh private benefits. Journals are now making an effort to internalize some of these
costs by requiring either that the data be available or that they be supplied on diskette when the
paper is submitted for review.
In addition, establishing competitively allocated resources to support post-study data
documentation and archiving for future benefits transfer exercises would be very useful to these
exercises. This program would have to be on-going, selecting only those data sets each year that
clearly have promise for future use in transfer exercises by other researchers. The incremental
cost of cleaning up and annotating a data set for public consumption rises quickly with the time
elapsed since completion of the original study. But in many cases, the incremental cost to the
research team of retaining a research assistant for an additional month after completing the main
project is relatively small (at least compared to the cost of going back to the data after several
years have passed or of collecting new data).
In many cases, the research team responsible for collecting and processing the data set
will have a proprietary interest in using that data for a set of studies before they become widely
available to everyone. We must acknowledge that the compensation for much contract work is
often taken (by academics) in the form of future publications employing the data made available
by die original survey study. In these cases, proprietary interest might be a negotiable item in a
proposal for incremental data documentation funding. The research team could include a time
limit within which the delivered cleaned-up data would not be disseminated to other users. This
time limit would allow the documentation phase to proceed in a timely fashion without the
possible cost to the original research team of lost proprietary rights to the data conferred
implicitly by unintelligible or nonexistent documentation.
CONCLUSIONS
Benefits transfer, a widespread practice that has been ongoing, will continue to take
place. In the face of tightening budgets, the need for "off the shelf* estimates of economic-
environmental benefits for policy and litigation will continue to increase. Therefore formulating
and promulgating a set of guidelines for these exercises are valuable endeavors. These
guidelines could be similar to the accepted standards for antitrust litigation. Without such
protocols, highly varying standards of accuracy might implicitly be applied in different cases.
13
-------
Workshop participants did not expect to produce a completed set of such protocols, and
they did not However, the participants seemed to experience a collective "consciousness-
raising" concerning the problems of benefits transfer. The opportunity for each group to conduct
an intensive post-mortem on a particular benefits transfer case emphasized the common
problems; the summary presentations allowed each group to articulate its own unique findings
for the benefit of members of other groups. At a minimum, all participants left the workshop
with a greater appreciation for the enormity of the challenge.
This area is ripe for productive applied research in this area. The subject of benefits
transfer protocols may be less glamorous than alternative theoretical topics in the area of
environmental economics. "Publication bias" favors new research on new topics, rather man
pragmatic issues such as benefits transfer. However, the workshop highlighted die scope of
applicability of research on the problem. In terms of influencing potentially huge realisations
of society's resources through policy making or litigation, the benefits transfer research has
profound relevance.
V
1
REFERENCES
Atkinson, Scott E., Thomas D. Crocker, and Jason F. Shogren. 1992. "Bayesian
Exchangeability, Benefits Transfer, and Research Efficiency." Water Resources
Research 28:715-722.
Ben- Akiva, Moshe, and Steven R. Lerman. 1985. Discrete Choice Analysis: Theory and
Application to Travel Demand. Cambridge: MTT Press.
Boyle, Kevin J., and John C. Bergstrom. 1992. "Benefits Transfer Studies: Myths, Pragmatism,
and Idealism." Water Resources Research 28:657-663.
Brookshire, David S., and Helen R. Neill. 1992. "Benefit Transfers: Conceptual and Empirical
" Water Resources Research 28:651-655.
David, Martin. 1992. "Benefiting Benefits Transfer. Information Systems for Complex
Scientific Data." Paper presented at the 1992 AERE Benefits Transfer Procedures,
Problems, and Research Needs Workshop, Snowbird, UT, June 3-5.
Desvousges, William R, Michael C. Naughton. and George R. Parsons, 1992. "Benefit
Transfer Conceptual Problems in Estimating Water Quality Benefits Using Existing
Studies." Water Resources Research 28:675-683.
Gaver, D.P, Jr., David Draper, P.K. God, JS. Greenhouse. L.V. Hedges, C.N. Morris, and C.
Watemaux. 1992. On Combining Information: Statistical Issues and Opportunities for
Research. Draft report of the Panel of Statistical Issues and Opportunities for Research in
die Combination of Information, Committee on Applied and Theoretical Statistics, Board
on Mathematical Sciences, Commission on Physical Sciences, Mathematics, and
Applications, National Research Council
14
-------
Learner, Edward E. October 1991. "Eastern Data and Western Attitudes." Technical Working
Paper #114. National Bureau of Economic Research.
Loomis, John. 1992. "The Evolution of a More Rigorous Approach to Benefit Transfer: Benefit
Function Transfer" Water Resources Research 28:701-705.
Luken, Ralph A., F. Reed Johnson, and Virginia Kibler. 1992. "Benefits and Costs of Pulp and
Paper Effluent Controls Under the Clean Water Act" Water Resources Research 28:665-
674.
Manski, C, and S. Lerman. 1977. "The Estimation of Choice Probabilities from Choice-Based
Samples." Econometrica 45:1977-1988.
McConnell, K.E. 1992. "Model Building and Judgment: Implications for Benefit Transfers
with Travel Cost Models." Water Resources Research 28:695-700.
Smith, V. Kerry. 1992. "On Separating Defensible Benefit Transfers From 'Smoke and
Mirrors'." Water Resources Research 28:685-694.
Walsh, Richard G., Dorm M. Johnson, and John R. McKean. 1992. "Benefit Transfer of
Outdoor Recreation Demand Studies, 1968-1988." Water Resources Research
28:707-713.
15
-------
]
J
-------
BENEFIT TRANSFER AND SOCIAL COSTING
Alan J. Krupnick*
ABSTRACT
Increasing demand for benefit analyses that are too comprehensive for original research
to be feasible and static or falling research budgets put a high value on the wise use of
existing benefit studies to estimate benefits associated with new policies and problems. In
this paper I define the sources of the increased demand for benefit analyses, identify the types
of benefits most useful to benefit transfer now, examine the protocols for conducting benefit
transfers, and suggest a future research agenda.
Interest in developing and applying techniques for benefit transfer is growing rapidly.
Benefit transfer is the application of original damage or benefit studies made in a given policy
context and location (what Desvousges, Naughton, and Parsons [1992] refer to as a policy site) to
another context and/or location (what these authors refer to as a study site). Burgeoning demand
for benefit analyses that are too comprehensive for original research to be feasible together with
static or falling research budgets put a premium on the wise use of existing benefits studies to
estimate benefits associated with new policies, problems, or simply new locations. The idea of
designing future original research to enhance the reliability of benefit transfers presents
particularly interesting challenges.
This paper has three purposes: to delineate the sources of this burgeoning demand, with
particularly attention to the movement led by Public Utility Commissions (PUCs) to incorporate
all of the externalities of electricity generation into utility decision making; to identify the types
of benefits that are most amenable to benefit transfer now; and to examine protocols for
conducting benefit transfers and suggest a future research agenda to make benefit transfers
easier, reliable, and more consistent with welfare economics.
SOURCES OF DEMAND FOR BENEFIT TRANSFER
Since environmental and natural resource economics began as a discipline in the early
1970s, the primary demand for analyzing the benefits of environmental improvement came from
U.S. government agencies interested in establishing "unit-day" recreation values for evaluating
projects and policies affecting water resources and from agencies needing to comply with E.O.
12291, which mandates benefit -cost analyses for all "major" regulations. These needs translated
into research budgets for original research in estimating policy-related environmental benefits,
while also giving rise to using the original research results in what we would now label "benefit
"Senior Fellow, Resources for the Future.
-------
transfer" exercises to comply with the Executive Order. One of the most visible and successful
of these secondary studies was EPA's benefit-cost analysis of the lead-phasedown regulations
(U.S. Environmental Protection Agency, 1985), which used original benefit studies to provide
estimates of the value of statistical lives and values of avoiding a variety of acute health effects
to argue that the phasedown made economic sense.
More recently, passage of the CERCLA (Superfund) law has propelled interest in benefit
transfer and resulted in the embodiment of this concept in the Type A natural resource damage
assessment model (now being updated), which estimates damages to recreational and commercial
fishing from a given type and size of oil spill in a given location using existing literature (see
Jones, 1992).
But each of these needs is relatively narrow, involving damage to, at most, a few
nonmarket commodities and usually by only one cause (e.g., lead or an oil spill). The limited
scope of these demands sets them apart from the newest demand for benefit transfers—that of
state PUCs who wish to formally introduce estimates of the external costs of alternative means
for generating electricity into utility decision making. All externalities associated with the fuel
cycle supporting each generation technology need to be addressed. For the coal cycle, this
means addressing externalities from acid mine drainage to environmental effects of air emissions
at the generation stage. Some 29 states are considering or requiring that the planning for new
investments accounts for residual environmental damages from alternative generation
technologies (Cohen et al., 1990).
Unfortunately, but not surprisingly, no original studies provide comprehensive estimates
of these damages;1 even imagining how an original study would be conducted, assuming that the
money to pay for it could be found, is difficult Even if some studies of mis type were
conducted, the location specificity of environmental damages (Le., their sensitivity to the
location of the new power plant, irrespective of the technology creating these damages) would
still necessitate using techniques for transferring the comprehensive results of these studies to the
study site. Thus, assuming that states are prepared to implement social costing, researchers must
devise and codify methods for consistently using benefit transfer techniques to estimate
*Otttf gT ft til
we W
flt *«an*fff
yrimntc damage afrf
ignore the location-specificity of impacts. Otter comprehensive estimates of the external costs of electricity use
qnfm costs as a proxy for damage (Beroow, Biewald, and Matron, 1991).
T
f
-------
incremental damages in each state as well as across different potential power plant sites within a
state.
Major on-going studies are already codifying benefit transfer techniques but without
carefully considering the models they are using. The U.S. Department of Energy (DOE) is
funding a study conducted by Oak Ridge National Lab and Resources for the Future that is
designed to develop methods and estimate the externalities from alternative fuel cycles used in
generating electricity at two "reference environments." No original research is in the work plan;
rather benefit transfer (as well as health, biological, and meteorological science transfer) is to be
used to the fullest extent possible in the context of a damage function approach.2 Economists,
engineers, and natural scientists in Europe, with funding from the European Community, are
following the identical work plan and methods while sharing some of the research effort to
estimate comparable externalities for potential power plant sites in Europe.
New York State is funding Hagler, Bailly to do a more ambitious external costing study
that builds on the DOE research to develop a computer model for utilities to use in estimating the
external costs associated with any proposed new capacity expansion. In addition, smaller studies
with similar objectives are on-going in Wisconsin (Research Triangle Institute) and California
(National Economic Research Associates and Regional Economic Research, Inc.). For the most
part, each of these studies, facing die enormity of their tasks, which take in virtually all the
benefit estimation literature, is primarily assembling and evaluating literature to provide any
estimates of damage, without paying much attention to theoretical prerogatives and constraints
discussed at the workshop.
A final, potentially major source of demand for benefit transfers comes from international
aid organizations such as the World Bank and the U.S. Agency for International Development.
These groups are responsible for capturing the environmental effects of their lending in
developing countries, but with very few exceptions (Wbittington et aJL, 1989), no original studies
of the benefits of environmental improvements in these countries exist Here, protocols for
benefit transfer mat take into account different personal and market characteristics are
2At fee valuation stage, this approach involves moDetizkgiiiqttcts(e.g^ acme bcaltb effects) ndter than me
Mirhi
changes in environmental quality, implicitly including ta^iacu, to tbeuttot out iwltvidiak are aware
-------
particularly important, as differences in incomes, institutions, cultures, climate, and resources,
for example, are surely far larger between a developed and developing country than among states
in die U.S. (in the case of social costing of electricity). The existence of widespread subsidies on
energy and other commodities greatly distorts relative prices, adding die identification of shadow
prices to die long list of challenges to benefit transfer.
Researchers even debate whether benefit transfer is legitimate for certain types of
nonmarket commodities affected by programs in developing countries. The basic tenet of
individual sovereignty underlying benefit estimation may not be applicable in societies that
emphasize group welfare. And die profound influence of poverty in developing countries on
willingness to pay raises questions about whether any benefit transfer technique involving U.S.
income elasticities of demand can be justified.
WHICH BENEFITS CAN BE TRANSFERRED NOW?
Benefits can be characterized into four groups by their effects on die following: health,
output, economic assets, and environmental assets—with my subjective ratings on die ease with
which benefit transfers can be conducted, given die existing state of die original research
literature, die characteristics of die commodity being valued (e.g., its dependence on personal
characteristics, site and regional characteristics, and extent of die market questions), and die
degree of codification of die literature for benefit transfer. The perspective in making these
judgments is diat of die PUC evaluating die mediods used to provide estimates of social costs. It
is recognized diat die scope of die task requires some degree of "quick and dirty" analysis, rather
dian die courtroom-proof reliability of natural resource damage assessment estimates.
Two of die four categories can be pretty much ignored: damage to output and to
economic assets. Damages to output, for example crop damage from air pollution or damages to
commercial fishing from a spill, are easy enough to cy|Jmafr- using original research and
gatiiering market price and supply and demand elasticities, for example, for die products, as
warranted. On die otiier hand, damages to economic assets cannot reliably be estimated in
original studies, let alone in a benefit transfer. Materials inventories are still lacking, and no
major modeling efforts for valuing die complex behavioral linkages necessary for a defensible
materials benefit estimate have been nn«v»rtaV»f> in many years.
Probably die health effects category is die easiest for making credible benefit transfers.
Once atmospheric or odier natural processes are taken into account (e.g., when estimating die
effect of reduced emissions on ambient air quality), die researcher can presume to a first
approximation, diat die heahh effects and die values people place on avoiding these effects are
I
V
I
I
I
I
I
I
t
I
J
J
1
f
-------
Suppose that the study population distribution is defined over attributes X and choices
j e C. This is a joint distribution, which can be decomposed as a conditional distribution times a
marginal distribution:
= PGIX)p(X)
(1)
Now, if the study sample happened to be truly representative of the study population, the
likelihood function for the individual choices observed in the sample would be given as follows
(where vy = 1 if individual i chooses alternative j and yy « 0 otherwise):
£ =
f(),Xi)Yij
(2)
ij p(Xj)
This calculation results in a formula for the log-likelihood given by
log £ = Ii IjeC yij log P(j I Xi, P) + £ log p(XO
(3)
By exploiting the decomposition of the joint distribution into a conditional distribution times a
marginal distribution, the log-likelihood function in Eq. (3), to be maximized over the unknown
parameters p, consists of a sum of two components. The second component does not depend on
the parameters P, so it can be ignored, and the optimization of log £ can proceed simply by
maximizing the first term in Eq. (3). Weights are unnecessary.
However, most benefits assessments require voluntary participation of members of the
affected study population in the survey necessary to gamer the data. In RUM models.
researchers now generally acknowledge that nonparticipation should be included as a relevant
choice along with specific site choices conditional on participation. Whether contacted
individuals opt to comply by completing their questionnaire or interview will determine their
presence in the final estimating sample for the study.3 Nonparticipants in the associated
recreational activity are typically less likely to be interested in the survey and hence less likely to
appear in the final sample. Because of this tendency, most modern RUM applications involve
fundamentally choice-based samples.
Ben-Akiva and Lennan (198S) demonstrate that unweighted MLE is still feasible for the
standard multinomial logit specifications typically used to estimate RUM models, providing the
^Intended observations can end up being omitted from the estimating sample because of item nomesponse or
complete nonttsponsc.
-------
choice model has a full set of J-l alternative-specific constants (i.e., site-specific dummy
variables plus a nonparticipation dummy variable). Exogenous information concerning the true
*•
study population distribution of attributes X is still required for the process of adjusting the
estimated probabilities after the estimation process. Manski and Lerman (1977) call this
approach "exogenous sample maximum likelihood" (ESML),
However, in benefits transfer applications, the last thing a researcher wants in the model
for the original study sample is a set of site-specific dummy variables, for the following reason.
Using these dummy variables is akin to estimating entity-specific fixed effects in a panel data
model for pooled time-series and cross-section data. Providing no new entities appear in the data
set for which a policy forecast is desired, these fixed effects are fine. But if new entities will
appear, the researcher will have no fixed effects to use for them. Random-effects models for the
study sample are preferred under these conditions.
Benefits transfer exercises require, by definition, that models calibrated for one set of site
choices be applied to different sites (or at different time periods). This feature precludes using
ESML estimation for RUM models destined for transfer exercises. A formal choice-based
sample maximum likelihood estimator is clearly indicated in this context. Unfortunately, th's
estimator is somewhat intractable. A consistent estimator for P that represents a tractable
alternative is the WESML estimator.
The WESML estimator is typically implemented by partitioning the estimating sample
into G groups (or "cells") defined over intervals of the values of some subset of the exogenous
variables. The group-specific weights, wg, are given by **?*/£, where the numerator is the
population relative frequency of individuals in group g, and the denominator is the sample
relative frequency of individuals in group g. With Ng designating the number of sample
observations in group g. the WESML log likelihood function is given by
log£
(4)
Proving that this estimator for f) is consistent under very general conditions is daunting.
Furthermore, the WESML estimator is not fully efficient even asymptotically, so its variance-
covariance matrix is more complex than that of a true n1 •*'""?*" likelihood estimator (see
Manski and Lerman, 1977). Even its corrected variance-covariance matrix (outlined in Ben-
AMva and Lerman (1985, p. 239) does not attain the Cramer-Rao lower bound. Thus these are
compromise estimators; computational tractability is gained at the expense of full statistical
efficiency. They are nevertheless highly practical.
-------
reasonably similar across locations. The extent of the market is clear: people living in the air
basin in which the postulated air quality change occurs.
Codification has proceeded for many years. Estimates of the value of a statistical life
taken from summary reviews and specific studies are widely used, multiplied by expected deaths
"delayed" to obtain the mortality benefits from a particular program, investment, or other
exogenous change in baseline conditions. A similar protocol is followed in using the literature
on the values of avoiding acute health effects to estimate the benefits of baseline pollution
reductions (see Hall et al., 1989; Krupnick and Portney, 1991; and National Economic Research
Associates, 1990) for benefit transfer studies for improving air quality in Los Angeles that
include estimates of mortality and morbidity benefits). Indeed, "spreadsheet" models are
available that first match estimates of changes in air pollution concentrations to dose-response
functions for a wide variety of health effects and then match these to unit values for avoiding
these effects to obtain health benefit estimates for environmental improvements.
Yet, the benefit transfers are of the crudest type: they use unit values and unaided
judgment to combine the different values obtained from the literature. Few of the spreadsheets
use valuation functions in the benefit transfer, for example, of the land arising from regression
analysis explaining variation in willingness-to-pay (WTP) responses. The methods for
establishing error bounds and best estimates are ad hoc and heterogeneous across benefit transfer
studies.
The original studies do not always lend themselves to transfers. Virtually the entire
mortality risk valuation literature addresses accidental deaths in prime-age adults, a setting
inappropriate for all environmental mortality except perhaps accidental toxic waste releases and
similar catastrophes. One study (Mitchell and Carson, 1986) addresses the latency issue so
important to valuing deaths due to cancer bat is silent on the effect of prior health status and age
on valuation. These issues are important in environmentally related deaths to those with heart
dispajf. and chronic lung disease. Further, researchers trying to use this study to value noncancer
related deaths may find that it postulated risk changes outside the risk changes associated with
power plant emissions. Also no reliable studies are available to value life-years saved (except in
occupational accidents) even though this health endpoint can be estimated by health scientists.
The most problematic area for benefit transfer is damage to environmental assets,
although there is some differentiation among these subcategories. Benefit transfer of recreation
values or demand functions presents one of the greatest challenges. Accounting for regional
factors (such as the range and quality of substitute sites) and site-specific factors (such as
T
-------
congestion) is likely to be difficult Furthermore no acceptable procedures exist for determining
the "spatial extent of the market" That is, debate is still lively on methods for determining the
size of the population that would be or is affected by a recreation quality or quantity change.
Codification of the chain of effects from concentration change to valuation is absent, with
the exception of the Type A model noted above. Because benefit transfers have generally
followed the procedure of using unit-day values, these values exist in great profusion for all types
of uses and environments (Walsh, Johnson, and McKean, 1988). But applying these values to
specific sites is problematic, more so than applying unit values to health because of the
presumption that WTP to avoid health effects is less influenced by region and site variables than
WTP for recreation. Codification of recreational fishing damages from oil spills in the Type A
model represents a useful prototype for the future development of portable, PC-based models for
use in benefit transfer. However, this particular model uses a unit-day value approach for the
valuation step.
Likewise, the recreation literature is of somewhat limited usefulness in estimating social
costs because the majority of the literature focuses on changes in the availability of resources not
on changes in their quality. Few studies incorporate explanatory variables that map back into
readily measured physical quantities, such as water turbidity, nutrient concentrations, and the
like. Most of the literature values catch rate changes.
Benefit transfer for valuing visibility also presents formidable challenges because of the
sensitivity of values to region, site, and personal characteristics. Characterizing the policy and
study site is particularly difficult for visibility benefit transfers. Although visual range can be
characterized in a relatively straightforward way, the vista being affected is particularly difficult
to characterize, beyond "urban," "rural," and "recreational area," which is unlikely to be
sufficient In addition, the extent of the market problem is even more difficult than that for
recreation because "use" as a function of distance to me site can be observed for recreation, but
not for some visibility problems (e.g., urban visibility).
The literature on visibility benefits is fairly conducive to benefit transfer (see Chestnut
and Rowe, 1992). Studies of visibility values in multiple cities (Tolley et al, 1988) are available,
which then permit examination of city-specific factors affecting values and derivation of
functional relationships to predict WTP, given the baseline visual range and the size of the
change (National Acid Precipitation Assessment Program, 1989). A number of examples of
benefit transfers involving visibility (Rowe, Chestnut, and Skumanich. 1990; Chestnut and
Rowe, 1988) are available. The Electric Power Research Institute (EPRI) (1991), which
I
1
-------
examines benefits from improved visibility in the eastern U.S. from reductions in SO2 emissions,
is a particularly good example of a benefit transfer where all the steps of the damage function
approach were linked together (i.e., emissions to concentrations, concentrations to optics, optics
to perceptions, and perceptions to value).
The major problem with benefit transfer in this category is the original studies.
Significant debate surrounds protocols for eliciting values in contingent valuation studies. For
example, the size of photographs shown to respondents appears to influence WTP. Concerns
about joint valuation of visibility and health (i.e., that visibility is used as a proxy for health
effects) and about embedding are also important From the perspective of the social costs of
electricity issues, research efforts have concentrated too much on national parks in the southwest
and not enough on valuing visibility effects at more mundane locations, both rural and urban.3
The literature on nonuse values for environmental assets clearly cannot yet support
benefit transfers associated with social costing of electricity, because most of the studies are for
non-marginal changes in unique environments (species extinction, loss of an ecosystem) while
the effects of a single power plant on any species or ecosystem is likely to be small and on
unique areas or species (after compliance with the Endangered Species Act and other federal
legislation) negligible. An exception might be nonuse values for visibility at national parks, such
as the Grand Canyon, associated with power plant emissions (Decisions Focus, Inc., 1990).
Admitting nonuse values into the benefit transfer exercise has the potential for
complicating matters enormously. For instance, in the presence of altruism about people's
health, the "extent of the market" issue, which is so easy to dismiss when researchers are
considering only "use" values, must be addressed anew.
For social costing of electricity, the bottom line is that environmental benefit transfers are
most feasible and reasonable for the health benefit category (although some serious problems
remain) and are not needed for crop damage estimation. Recreation damage estimation
associated with a new power plant is, generally, beyond our abilities, not because die economics
isn't op to it but because of gaps in the science and the lack of baseline recreation participation
information specific to reference environments of interest Visibility damages fail for similar
reasons—scientific linkages between emissions and changes in visual range are absent. Nonuse
value estimation studies for marginal changes in resource quality or quantity are virtually
nonexistent Given these problems, researchers must conclude that estimates of damages
resulting from benefit transfers are not sufficient or reliable enough to support more than a rank
3Califoraia cities and Denver have also been the subject of multiple benefit studies.
-------
ordering of new generation technology options on the basis of social costs. That is, reliance on
benefit transfers to support social cost dispatch or social cost pricing of electricity is probably
4
pushing benefit transfer (and original study) techniques beyond their capabilities.
PROTOCOLS
Researchers confronting the need to estimate the benefits of environmental improvements
but who, for one reason or another, cannot conduct original research to estimate such benefits,
currently either must rely on simplistic protocols for conducting their benefit transfer study or
find no guidance, except from what they can glean from other examples of benefit transfers. For
instance, the U.S. Forest Service sanctions the use of "unit-day values" for estimating recreation
benefits. But such values are averages over a wide range of site characteristics and policy
scenarios (most examining the value of recreation at a site rather the change in value associated
with a change in site quality) that may be inappropriate for the study site.
Reliance on existing benefit transfer studies is also risky because such studies are not
designed for educating the practitioner on how a reasonable benefit transfer should be (or was)
done, making communication about such protocols dependent on the often haphazard and
incomplete reporting of such procedures. Further, as different benefit transfer studies use
different protocols, die researcher is left with the task of sorting them out This task should be a
subject of a generalizable research effort not reinvented every time by each researcher.
The papers published in Water Resources Research as well as the participants in the
workshop are in close agreement on general protocols for using existing studies, so I do not need
to recount them in detail here. The care and effort used in conducting a benefit transfer—indeed,
whether researchers should attempt it at all—depend on the commodity being valued; differences
in regional, site, and personal characteristics; and the nature of the original literature being relied
on for the benefit transfer. Given that a benefit transfer is called for, much emphasis is placed on
using demand or value functions where possible, as opposed to using average unit values—be
they for a day of recreation or a day of coughing avoided Using the function approach puts
some additional burden on the researcher (data must be gathered on the variables at the study site
found by the original study to affect WTP, for instance); indeed, without careful reporting of
results in the original study, this approach may be impossible.
•
Nevertheless, in the practical application of these broad guidelines, many choices are
available with few guidelines to follow. What does the researcher do when the valuation
literature is based on changes in physical effects (e.g., catch rates) but no link exists from catch
rates to fish populations or changes in water quality? When the underlying science is poor,
^
8
-------
Up to this point we have only considered the CV associated with a change in the price of
a day at the site. However, often we need to estimate or approximate the CV associated with a
change in the characteristics of a site. How much of the argument above generalizes to cases
involving changes in characteristics?
To begin our investigation of this question, again consider our Coke-drinking thought
experiment, but now determine how much you would pay to have the number of calories in the
ith Coke that you drink reduced by 50 percent (with all other product characteristics unchanged).
Without loss of generality, Til denote your answer cq. This amount, otj, is your CV for this
calorie reduction for the ith Coke drank. Contrast this amount with a per-Coke CV for this
calorie reduction that is independent of the number of Cokes you choose to drink. A per-Coke
CV independent of the number of Cokes you choose to drink only exists if 04 = a V i, in which
case is the per-Coke CV. In cases such as our first thought experiment where the change is
solely a price change, ai = a = the price reduction V i, but typically oq * a V i if the change
involves a change in the characteristics of the commodity. For example, how much I would pay
to have the calories reduced in the last Coke I drink will increase as I drink more Cokes.
By definition, oj = a V i only if how you value the change in monetary terms is
independent of the number of Cokes you choose to drink. This must be true for a change in the
price of a Coke, but what would be sufficient to make it true for a change in the characteristics of
the Coke? A sufficient but not necessary condition is a world with the following properties:7
• Assume a world of only three commodities: two activities, drinking a Coke and not
drinking a Coke, and a numeraire good.
• Both activities take all day.
* If you choose not to drink a Coke you spend all of your income for the day on the
numeraire. Otherwise you allocate to the numeraire your income for the day minus the
price of the Coke.
• How much utility you receive on a day is only a function of whether you drink a Coke
that day, the amount of the numeraire consumed that day, and the characteristics of
Coke. If these four conditions hold, you will always have a CV per Coke drank for a
change in the price and/or characteristics of Coke, which is independent of the number
of Cokes you choose to drink.
Now consider a similar thought experiment for a change in the characteristics of a
recreational site. As before, assume a world of three commodities: two types of activities, days
at the recreational site and days at home, and a numeraire good that can be consumed anywhere.
7The reason I choose tbese properties rather than some other set of sufficient conditions will become dear.
-------
What is the maximum amount an individual would pay each and every time he or she spends a
day at the site to have characteristics of the site be C1 rather than C°? As our last thought
experiment indicates, the individual will not in general be able to answer this question because
there is no such amount For changes in the characteristics of a site, a constant CV for a day of
use, a, does not usually exist
In a world with a recreational site but no Coke, 04 = a V i only if the manner in which the
individual values the change in the site in money terms is independent of the number of days he
or she spends at the site, at = a V i must be true for a change in the price of a day at the site but
does not have to be true for changes in the characteristics of the site.
A constant CV for a day of use will exist for an individual in pur world of three
commodities only if we make the additional assumption that the utility the individual receives on
a day is only a function of whether he or she spends that day at the site, the amount of the
numeraire consumed that day, and the characteristics of the site. In this case, oq = a V i for any
change in the price of a day and/or change in the characteristics of the site. Note that when this
assumption is made a price change always exists that would make the individual indifferent
between that price change and the proposed change in the characteristics, and this price change is
independent of the number of days spent at the site. We might denote this price change as die
quality-equivalent price change. Therefore, when we adopt the additional assumption outlined
above, any change in the characteristics of a site can be converted into its quality-equivalent
price change, and Theorem 1 and all the approximation results apply in terms of this price
change. The fact that a characteristics change can be converted into an equivalent price change
when these restrictive assumptions bold makes Theorem 1 and the approximation results
particularly relevant to discrete-choice models of recreational demand.
The assumption that the utility the individual receives on a day is only a function of
whether he or she spends that day at a site, the amount of the numeraire consumed that day, and
the characteristics of the site is the basic assumption of many discrete-choice models of
recreational demand. Therefore a constant CV for a day of use can be derived for changes in
both prices and characteristics from most discrete-choice models of recreational demand.8
Consider a simple dichotomous Logit or Probit model designed to predict the probability that an
individual will visit a particular site on a given day. Such models are based on two conditional
indirect utility functions. One function specifies the utility received for the day if the site is
visited, and the other function specifies the utility received for the day if the site is not visited.
*Fbr example, see die earlier references to Bocfcstael. Hanemmn. and Strand (1984) and Canon. Hanemann, and
Wegge (1987).
10
1
1
•
1
1
J
J
1
k
-------
II: 1.
2.
S: 1.
2.
II:
G:
R:
TABLE 1. NUSAP DATA ENTRY FORM EXPLANATION
Enter the number, notation, variable name, or note about practice.
Enter the measure for the number, upper and lower bound, or variable (e.g.. pounds).
Also enter the time period for the entry (e.g., per hour).
Enter the statistic which the number or variable is (e.g., mean, median, no
distribution).
Enter the degree of confidence of the spread. Use 90 percent whenever possible for
standardization.
Enter the upper and lower bound or±% range, ± standard deviations, or factor of
variation of the spread.
Enter the assessment ratings for each applicable category (i.e., H, M, or L). Enter
N/A for not applicable.
Assess the informative value based on spread. That is, assess the extent to which the
entry narrows the spread of plausible values over what was known before the study
that produced the entry was conducted (prior).
L: Many prior plausible values exist in spread.
M: Spread is a fair amount narrower than range of prior plausible values.
H: Spread is much narrower than range of prior plausible values.
Assess the informative value based on the foreseen application for the entry. That is,
how informative are the results of calculations with this entry expected to be given
the current persisting (posterior) uncertainty about the entry. ^2 nere i* a firsl guess,
to be refined when the particular application is considered.)
L: The existence of other posterior plausible values (i.e., values in spread) matters
for application.
M: The existence of other posterior plausible values matters marginally.
H: The existence of other posterior plausible values does not matter.
Assess the generalizability of the entry to other applications, locations, or sample
spaces different from the application for which it was origina
; originally generated.
L: does not generalize to other applications
M: can be generalized with limitations
H: easily generalized
Assess the entry's robustness over time.
L: highly perishable
M: moderately perishable
H: time independent
(continued)
11
-------
TABLE 1. NUSAP DATA ENTRY FORM EXPLANATION (CONTINUED)
£: Enter the pedigree ratings for the applicable categories (i.e., 1 to 5). Enter N/A for
any inapplicable pedigree category.
T: Assess the theoretical basis of the entry and the tcnahility of the theory's application
to produce the entry,
1: no theory or concepts
2: weak theory or concepts, controversial empirical support
3: weak theory, good empirical support
4: good theory, but one of competing theories
5: well-understood and accepted theory
D: Assess the quality of the data inputs used to generate the entry.
1: unacceptable
2: poor
3: fair
4: good
5: excellent
E: Assess the estimation methods used to generate the entry.
1: unacceptable
2: poor
3: fair
4: good
5: excellent
M: Assess the estimation metric (i.e., proxy or indicator for what we want to measure.)
1: unacceptable
2: poor
3: fair
4: good
5: excellent
Comments: Eater any comments about the NUSAP categories. The level of spread may require explanation such
as "confidence level corresponds to +/-2 standard errors corresponding to multiplication or division by
a factor of 1.7 for the upper and lower bound." Ibe reasons why assessment ratings and pedigree
rating were received should be explained bete.
12
-------
i
1
t
establishing error bounds. Smith and Kaoru (1990) performed one of the first meta-analyses of
the environmental benefits literature analyzing 77 studies of recreation demand. Nevertheless,
because the authors' purpose was to see if methodological choices made a difference in value
rather than to explain differences for reasons of site, regional, or personal characteristics, this
study is not particularly useful for a benefit transfer.
These approaches need not be confined to the valuation step. Morton and Krupnick
(1988) obtained original data on ozone dose-response studies conducted in four laboratories. By
combining the samples and accounting for differences in protocols, the authors were able to
estimate a composite dose-response function for use in EPA's ozone Regulatory Impact
Analysis.
Although researchers should not be overly optimistic that data or uncited reports
underlying previously published benefit studies are available and researchers are willing to pan
with them, an effort to collect (for payment) and analyze old but useful databases and reports
could pay off, particularly for studies that did not estimate or report on variables or analyses
capturing mediating factors on WTP. Contingent valuation studies that report central tendency
WTP values but not regression results explaining these values or studies that use linear
functional forms that result in mediating factors dropping out for marginal valuation would be
good candidates, provided data on mediating factors were collected in the first place.
Improve the Quality of Original Studies
Undoubtedly better original studies will make for more credible benefit transfers. In the
context of social costing, a "better" study is one that makes explicit linkages between its
valuation starting point and the science endpoint. The case of recreation quality change is the
classic case, where much of the recreation literature uses "catch rate" as a starting point, while
the scientific literature ends with water quality changes or changes in fish populations. Only the
NAPAP studies (Engtin et al., 1991) explicitly account for all of the linkages, from emissions to
concentrations to impacts to values, in its analysis of the recreational benefits of SO2 emissions
reductions. In the contingent valuation literature, a study that makes very clear the commodity
being valued is not only a better study than one that is unclear about the commodity, but the
former is likely to make for a more reliable benefit transfer.
As I noted in several places above, improvements in protocols for natural science studies
are needed if benefit transfers are to be broadly successful. Better protocols would include
designing endpoints for the studies that map into economic starting points. For instance, in the
health area, much of the literature on the acute effects of air pollution measures lung function.
13
-------
primarily because it is easy to measure and is "scientific." However, no one values a change in
lung function; people need to know what this means in terms of their everyday health, A focus
on symptoms effects is an improvement.
Change Original Study Reporting/Designs for Use in Benefit Transfer
If researchers engaged in original benefit analysis would consider how the results of their
study will be used, other researchers would benefit enormously. At a minimum, reporting of
results would be affected. Many articles omit mean values for independent variables and the
equations used to estimate changes in consumer surplus, but this type of information would help
enormously in a sophisticated benefit transfer exercise. Even if journal space limitations
preclude publishing such information, journals such as the Journal of Environmental Economics
and Management (JEEM) could require that a diskette with the data and/or key regression results
(if these are unpublished) be submitted as a condition for publication. Or EPA could monitor
article publication and request such data.
Studies* designs could also change, focusing much more on site, region, and person-
specific variables that might influence valuations and using functional forms or interactive terms
that permit examining confounding factors on marginal valuations. In addition, most studies
examine the benefits of environmental improvements rather than the WTP to avoid further
environmental degradation. The former is certainly more germane to analyses supporting
environmental policy analyses. But, for social costing, the premise is that the environment will
worsen, at least in some dimensions (absent tradable permit programs, for instance). In general,
we have no reason to expect that the benefits of a given environmental or health improvement are
equal but opposite in sign to the damages from an equivalent decrement in environmental quality
or health.
•f
r
Conduct Major Benefits Studies for Use in Benefit Transfers
Because commodity characteristics and regional, site, and personal characteristics are
likely, a priori, to affect WTP, designing studies from the bottom up would be helpful to capture
these differences, investigate which factors matter most, and report results to facilitate benefit
transfer.
For instance, in the health area, valuation studies that provide estimates of WTP for
reductions in premature mortality risks of the type associated with environmental exposures—
presence of latency periods, effects on the elderly and the very young, allowance for values to
differ by cause of death, for instance—would reduce reliance on the largely inappropriate
J
J
.1
14
-------
accidental death hedonic wage/contingent valuation literature. The effects of age and sex on
such values are particularly important to establish. Work on estimating WTP for life-years saved
directly would supplant the ad hoc approaches currently used to modify the current average
lifetime valuation literature in benefit transfer exercises.
Morbidity studies, primarily using contingent valuation, are out-of-date and not risk-
based. Changes in health risks are often so small that the approach of calculating number of days
of effects and multiplying by a unit value per certain day of effect may seriously mislead
researchers. In addition, most studies seek values for single symptoms of types of effects rather
than illness complexes or episodes. Studies that provide values on the latter would help in
aggregating values over multiple acute health effects, although the health science literature
provides little guidance as yet on the relationship between health episodes and air pollution.4
Taking a broader view, studies that seek WTP estimates for a multiple set of effects, such as
acute and chronic effects or chronic disease and mortality risks (while being explicit on the
effects being valued, unlike property value studies), would also aid benefit transfer, while
obviously being important in their own right.
In the recreation area a promising, if expensive, approach would be to conduct national
studies of recreation benefits from site-quality changes that consider as much as possible regional
differences in site availability and baseline site qualities, as well as the relevant personal
characteristic variables with a regional dimension (such as recreator experience). Recreation
benefit analysis has a tradition of examining the benefits of large changes in quality, for example
an improvement in stream quality from fishable to swimmahlc. Such analyses have their uses,
but the changes in quality associated with social costing of electricity are much smaller than this.
Continuing the pioneering work of NAPAP researchers on the linkages between pollution
concentrations and changes in catch rates and the generalization of this work into portable
computer models would also greatly facilitate benefit transfers in involving this category of
benefits.
T
Develop Researcher Incentives
Professional academic economists will not conduct benefit transfers or go out of their
way to make their work more helpful and accessible for others to conduct benefit transfers unless
it is in their interest to do so. Professional journals put a premium on original research and, as in
any other endeavor, opportunity costs and risks of preemption of making data available may be
seen as large. Participants at the workshop recommended developing a new peer-reviewed
4A new study by Resources for the Future is taking ibis tact in its epidcminlngical analysis of panel data on a sample
of Taiwanese.
15
-------
journal for presenting benefit transfer results and methods. Making additional funds available to
increase the usefulness and accessibility of contract and grant-based research to the broader
0
research community would obviously .help in inducing cooperation. However, for this strategy to
work, government needs to have in place a system for accepting data, unpublished reports, and
unpublished results for easy cataloging, retrieval, and use.
*
REFERENCES
Bernow, S., B. Biewald, and D. Marron. 1991. "Full-Cost Dispatch: Incorporating
Environmental Externalities in Electric System Operation." Electricity Journal (March).
Chestnut. L.G., and R.D. Rowe. 1988. Ambient Paniculate Matter and Ozone Benefit Analysis
for Denver. Draft report prepared for U.S. Environmental Protection Agency, Denver,
Colorado, (January).
Chestnut, L.G., and R.D. Rowe. 1992. "Visibility Valuation: Acid Rain Provisions of the Clean
Air Act." Paper presented at the 1992 AERE Benefits Transfer: Procedures, Problems.
and Research Needs Workshop, Snowbird, Utah. June 3-5.
Cohen, S.D., J.H. Eto. C.A. Goldman, J. Beldock, and G. Crandall. 1990. "Environmental
Externalities: What State Regulators are Doing." Electricity Journal (July):24-34.
Decisions Focus Inc. September 1990. Development and Design of a Contingent Value Survey
far Measuring the Public's Value for Visibility Improvements at the Grand Canyon
National Park. Revised draft report, Los Altos, California.
Desvousges, W.H., M.C. Naughton, and G.R. Parsons. 1992. "Benefit Transfer: Conceptual
Problems in Estimating Water Quality Benefits Using Existing Studies." Water
Resources Research 28:675-683.
Englin, JE. et al. January 1991. Valuation of Damages to Recreational Trout Fishing in the
Upper Northeast Due to Acidic Deposition. Report prepared for the National Acidic
Precipitation Assessment Program under a Related Services Agreement with the U.S.
Department of Energy.
Electric Power Research Institute (EPRI) and Decision Focus Inc. January 1991. Analysis of
SO2 Reduction Strategies. Final report prepared for EPRI, Palo Alto, California.
Funtowicz, S., and J. Ravetz. Uncertainty and Quality in Science far Policy. Dordrecht, The
Netherlands: Kluwer Academic Publishers.
Hall. J.V. et al. June 1989. Assessment of the Health Benefits from Improvements in Air Quality
in the South Coast Air Basin. Final report to South Coast Air Quality Management
District, Contract No. 5685, California State University Fullerton Foundation, Fulleiton,
California.
Jones, Carol Adaire. 1992. "Recreational Fishing Valuation: Application of the Type A
Model." Paper presented at the 1992 AERE Benefits Transfer: Procedures, Problems.
and Research Needs Workshop, Snowbird, Utah, June 3-5.
J
16
f
-------
Krupnick, A.J., and P.R. Portney. 1991. "Controlling Urban Air Pollution: A Benefit-Cost
Assessment." Science 252:522-528.
Mitchell, R.C., and R.T. Carson. 1986. Valuing Drinking Water Risk Reductions Using the
Contingent Valuation Method: A Methodological Study of Risks from THM and Giardia.
Report prepared for the U.S. Environmental Protection Agency, Washington, DC.
Morton, B., and A.J. Krupnick. 1988. "Ozone Acute Health Dose-Response Functions:
Combined Results from Four Clinical Studies." unpublished manuscript.
National Acid Precipitation Assessment Program (NAPAP). September 1989. "Acid
Deposition: State of Science and Technology." Report 27 in Methods of Valuing Acidic
Deposition and Air Pollution Effects, Washington, DC.
National Economic Research Associates, Inc. March 1990. Benefits of the 1989 Air Quality
Management Plan for the South Coast Air Basin. Prepared for the California Council of
Environment and Economic Balance.
Ottinger, R. et al. 1990. Environmental Costs of Electricity. White Plains, New York: Pace
University Center for Environmental Legal Studies.
Rowe, R.D., L.G. Chestnut, and M. Skumanich. 1990. Controlling Wintertime Visibility
Impacts at the Grand Canyon National Park: Social and Economic Benefit Analysis.
Report prepared for the U.S. Environmental Protection Agency. Boulder, CO:
RCG/Hagler, Bailly, Inc.
Smith, V.K., and Y. Kaoru. 1990. "Signals or Noise: Explaining the Variation in Recreation
Benefit Estimates. American Journal Agricultural Economics 72:19-433.
Tolley, G.M. et al. 1988. The Economic Value of Visibility. Mount Pleasant, Michigan.
U.S. Environmental Protection Agency. 1985. Costs and Benefits of Reducing Lead in
Gasoline, Final Regulatory Impact Analysis. Washington. DC.
Walsh, R.G., D.M. Johnson, and J.R. McKean. 1988. "Review of Outdoor Recreation
Economic Demand Studies with Nonmarket Benefit Estimates, 1968-1988." (June).
Whittington, D. et al. 1989. "Paying for Urban Services: A Study of Water Vending and
Willingness to Pay for Water in Onisha, Nigeria." Washington, DC: World Bank, Urban
Development Division.
17
-------
I
1
f
.1
I
1
"J
J
J
\
-------
FUNDAMENTAL ISSUES IN BENEFIT TRANSFER AND
NATURAL RESOURCE DAMAGE ASSESSMENT
James J. Opaluch and Marisa J. Mazzotta*
ABSTRACT
In this paper, we address three pertinent questions for benefit transfer. Can we reliably
measure benefits within the original study context? To what extent can benefit measures that
are reliable within the study context be transferred to provide reliable estimates for the policy
site? How can we improve benefit transfers to make them more reliable under a wider set of
conditions? Researchers must establish that the benefits estimates they are transferring are
defensible themselves. Researchers should also test the adequacy of benefit transfers by
quantifying their accuracy. Finally, we need to improve our methods of transferring benefit
estimates, perhaps by developing a wider range of calibration variables.
I
I
t
I
Benefit transfer estimates values in a policy context using available information from
studies carried out in another context (the study context). For example, we may have an estimate
of the value of recreational fishing derived from a study of coho salmon fishing in Oregon and
attempt to transfer this result to estimate the value of king salmon fishing in Alaska.
Participants at the AERE workshop agree that, for practical reasons, benefit transfer is a
necessary component of policy analysis. In many situations the expense of carrying out an
original study cannot be justified, or the funds or time simply aren't available. Yet some
information is needed to support decision making.
In a sense, even site-specific studies are a form of information transfer, where data from
the sample is transferred to a more general population. In many cases, the same lands of issues
arise (e.g., Loomis, 1987). For example, researchers must be careful that the sample is
representative of the larger population. In cases where the sample is not representative,
researchers frequently adapt results, often using socioeconomic characteristics of the sample and
the population.
Therefore, the relevant concern for economists is not whether to do benefit transfer;
instead, we suggest three pertinent questions for benefit transfer. The first is whether we can
reliably measure benefits within the original study context We certainly shouldn't consider
transferring benefit estimates that are unreliable even within their own context. The second is to
what extent benefit measures that are reliable within the study context can be transferred to
"University of Rhode Island, Department of Resource Economics.
-------
provide reliable estimates within the policy context Reliability will depend on the extent to
which values vary between the study and policy site, the extent to which we can explain and
*
correct for these differences, and the standard of accuracy for benefit estimation. The third
question is how we can improve benefit transfers to make them more reliable under a wider set
of conditions.
The answers to these questions will depend on the context of the benefit transfer, where
different, standards might be applied in different contexts. In many arenas, our institutions have
set differing standards of accuracy or burdens of proof for different kinds of social decisions.
For example, society has established the most rigorous burden of proof for criminal cases,
requiring the evidence to prove the case "beyond a reasonable doubt" This standard applies
independent of associated penalty and holds for criminal fines, as well as loss of personal
freedom through prison sentences or the death penalty.
A weaker standard has been placed on other cases, such as in civil suits, where the
standard is preponderance of the evidence. Here, the judge or jury will side with the stronger
case. This standard will differ to some degree for cases that include a rebuttable presumption
where a result is assumed to be correct unless a preponderance of evidence to the contrary exists.
Finally, the weakest standard of proof exists for policy decisions, where the agency
making the decision only needs to show that it is not being "arbitrary and capricious." An action
by an agency is considered arbitrary and capricious when the agency has
relied on factors which Congress has not intended it to consider, entirely failed to
consider an important aspect of a problem, offered an explanation for its decision that
runs counter to the evidence before the agency, or is so implausible that it could not be
ascribed to a difference in view or the product of agency experience. (Motor Vehicle
Manufacturers Association versus State Farm Mutual Insurance Company 463 U.S. 29,
43,103 S.Ct 2856,2867,77 L.Ed.2d 443,458 [1983])
In contrast, an agency's judgment will generally be accepted when it can show that it
Mexamine[d] the relevant data and articulate[d] a reasoned basis for its decision** (NRDC v.
Harrington 247 U.S. App. D.C. 340,370,768 F.2d 1355.1385 [1985]). Hence, in developing
regulations or in policy analysis, the agency is given considerable latitude for judgment and need
not demonstrate, for example, a "preponderance of the evidence** or an absence of a "reasonable
doubt**
These legal doctrines may provide one basis for establishing different standards of
accuracy or acceptability of benefit transfer within different contexts. An alternative standard
may be provided by a form of benefit-cost analysis whereby a higher standard of accuracy might
-------
be required when the costs of making a bad decision are high. A lower standard of accuracy
might be acceptable when costs are lower, such as when the information from the benefit transfer
is only one of a number of sources of information, or when benefit transfer is used as a screening
device for the early stages of a policy analysis.
Hence, the acceptability of benefit transfer depends not only oh how appropriate the
estimated value but also on the institutional context A number that may be "good enough" to be
used as part of agency judgment for a screening study may be judged to be inadmissible in a
criminal case or even in a civil case, such as in litigation surrounding a natural resource damage
assessment (NRDA).
However, even for screening studies, we need to apply sensible standards of accuracy.
We cannot allow ourselves to accept the proposition that "some number is better than no
number," particularly because an unreliable number may be given undue credibility. Benefit
transfer will obstruct, rather than facilitate, rational planning and will lose credibility if we apply
misinformation. In some cases we are better off acknowledging that we have no reliable estimate
for a particular factor, and we can then either collect information to estimate this factor or
account for it in qualitative terms.
FRAMEWORK FOR EVALUATING BENEFIT TRANSFER
The basic goal of benefit transfer is to estimate benefits for one context by adapting an
estimate of benefits from some other context Consider the example discussed above, where we
have an estimate of the value of recreational fishing for coho salmon in Oregon, and we attempt
to transfer this result to estimate the value of king salmon fishing in Alaska. The estimate of the
value of coho fishing in Oregon may not accurately measure the value for king salmon fishing in
Alaska for at least three reasons:
• The preferences of participants in Alaska may differ from the preferences of
participants in Oregon.
• The characteristics of the king salmon fishing experience in Alaska may differ from the
characteristics of the coho salmon fishing experience in Oregon.
• The estimated value of coho fishing in Oregon may not measure the true value of coho
fishing in Oregon.
The first two reasons imply that the value of fishing in Oregon differs from the value of fishing
in Alaska, while the third implies that the estimated value in Oregon is incorrect
-------
Figure 1 formalizes these three sources of variation. Individual variation denotes
variation across people that might arise because of differences in preferences. The value of
Fixed
Component
Random
Component
"Mean"
Value
U
Individual
Variation
W
ei
Commodity
Variation
HC
EC
-Other"
Variation
HO
eo
Figure 1. Framework for Benefit Transfer
Alaskan king salmon fishing may differ from that of Oregon coho fishing because anglers in
Alaska may have different preferences than anglers in Oregon. Some of mis variation may be
due to fundamental differences in tastes across individuals, while other components may be due
to differences in socioeconomic characteristics like income or age.
Commodity variation denotes variation across commodities that might arise because of
differences in their characteristics. The value in Alaska may differ from the value in Oregon
because of differences in the two experiences. For example, on average, differences may exist in
catch rates, scenery, congestion, size of fish, or other characteristics.
The third source of variation is meant to capture any variation that is independent of
preferences or the commodity and thus includes "bias** or "error" in measuring value. Various
sources of bias and error have been recognized in the literature. For example, using an incorrect
functional form can bias regression results and subsequent value estimates. Similarly this form
of variation may arise when survey respondents do not correctly express their values in a
contingent valuation survey or when people act in ways that do not allow us to infer their true
values through revealed preference methods.
The issue of bias has been most carefully considered for contingent valuation (e.g.,
Mitchell and Carson, 1989), but important biases may also exist for revealed preference methods,
such as the travel cost approach (e.g., Bockstael, 1984; Smith, 1989). Many studies have
4
-------
1
I
f
I
attempted to test reliability and validity of benefit estimation techniques within the context of a
specific study (e.g., Bishop and Heberlein, 1979; Loomis, 1989). To our knowledge, Cummings,
Brookshire, and Schulze (19S6) are the only one's who have taken a broad look at the issue of
accuracy. We are aware of two studies that look at accuracy within the context of benefit
transfer (Loomis, 1992; Downing and Ozuna, 1992).
In Figure 1, each source of variation is composed of "fixed" components and random
components. The fixed components are the components of variation that in some cases can
potentially be estimated and corrected for, while the random components cannot be explained.
For example, the value of recreational fishing may vary systematically over individuals
because of differences in age or income. By including age and income as explanatory variables
in the demand function, these sources of differences in value between the study and policy
contexts can be explained, predicted, and corrected. However, tastes may differ randomly and
unexplainably across individuals between the study context and the policy context.
The Oregon estimate will misrepresent Alaskan values if the distribution of these random
components of participants' tastes is different in Alaska than in Oregon, after correcting for
identifiable differences in the populations (e.g., age, income). In some cases we may be able to
place confidence intervals on this variation. For example, our Oregon data may allow us to
estimate the variance in tastes over participants. However, we cannot guarantee that our policy
site will fall within the confidence intervals because sources of variance may exist between the
study and policy sites that we cannot observe within the data for the study site only. Thus, to the
extent that values in the study and policy contexts differ because of random differences in tastes,
we may be unable to adjust our benefit estimates to reflect these differences.
A similar problem may arise if the contribution of "identifiable" factors differs across the
two sites. That is, using the Oregon coho study, we may be able to estimate how age and income
affect the value of fishing in Oregon. However, these characteristics may have different effects
on the value of fishing in Alaska. For example, if the weather is colder during the fishing season
in Alaska, age may be a more significant factor in Alaska. Similarly, two regions may have
cultural differences, so that in one region participants in some age group would not be "caught
dead" fishing, while this attitude may not be a factor in the other region. Again information
obtained in one region may not be transferable to another region. Furthermore, if no studies are
available in the policy region, we may have no systematic means of identifying or measuring
these sources of variance.
-------
The three sources of variation may be viewed somewhat differently in terms of three
components of benefit transfer. The first component consists of the "knowns," or the differences
between the study context and policy context for which information is available and that can be
estimated. For example, we may be able to estimate how the value of fishing varies with income,
age, and catch rates. Using the known (or knowable) information, such as demographic
information and characteristics of the activity, we can adjust the estimates of value obtained for
the study context to produce estimates for the policy context
The second component consists of the "known-unknowns," which might include the
random components of the individual or commodity variation. For example, preferences may
differ in ways that we cannot explain, but we may nevertheless be able to estimate the variance
. due to these effects at the study site. Thus these random components may be accounted for by
using confidence intervals on the estimates. Alternatively, known-unknowns might also arise
because of variables that are known to affect value but for which no data are available at the
policy site. We may know how income affects value, but we may not have data on income of
participants at the policy site. One way of accounting for these effects would be to use
sensitivity analysis to place plausible upper and/or lower bounds on these variables.
Finally, "unknown-unknowns" could arise in the original study from the "other" sources
of variation discussed above, or from unobservable or unknown differences between the two
populations and/or commodities. Limiting the magnitude of the unknown-unknowns is crucial to
the success of the benefit transfer. However, the magnitude of these sources of variance is, by
definition, not known to the researcher, and the researcher generally has no way to quantitatively
account for them in the transfer, short of carrying out a study at the policy site.
In the case of unknown differences between the study context and policy context, one
possible approach is to use a pilot study to assure that results appear to be transferable. Of
course, the cost of carrying out such a study may negate much of the cost savings that may come
from benefit transfer.
Thus, the answers to the first two questions posed earlier—whether we can reliably
measure benefits in the original study context and whether we can reliably transfer benefit
estimates—depend on the types of variation that occur and their magnitudes and on how well we
can identify and measure them. For benefit measures to be suitable for transfer, unexplained
variation must be limited to an "acceptable" level. The margin of error that is "acceptable" will
depend on the appropriate reliability standard, as described above.
J
-------
TESTING AND IMPROVING BENEFIT TRANSFERS
I
I
t
I
We might judge our confidence in the soundness of a benefit transfer in several ways.
Generally, statistical tests are used to evaluate the original study. These tests include statistical
significance of explanatory variables, the equation R2, and the size of prediction intervals. The
significance of important explanatory variables, such as the travel cost coefficient, is necessary
but not sufficient proof of the validity of value estimates. In addition, the model must have
acceptable explanatory power, which is indicated by the R2 and prediction intervals. Typically,
economists focus on statistical significance and tend to ignore R2 and prediction intervals.
These statistics suggest the relative magnitudes of the knowns and the known-unknowns,
but they will not measure the effects of the unknown-unknowns. Therefore, we can only really
be certain about negative test results. If a model has poor explanatory power for the study site.
we will not have much confidence in its soundness for benefit transfer. However, a model may
have good explanatory power, but extrapolation of its results outside of the original sample may
imply that we cannot measure important components of the variance—the unknown-
unknowns—so that the model may not provide good benefit estimates for the policy site.
Given the above problems, researchers should place greater emphasis on testing and
calibrating benefit transfers. Socioeconomic variables, which are typically used as calibrating
variables in transfers, often have very low explanatory power, implying that they are not good
calibrating variables for benefit transfer. Consequently, we need to be more creative in the
variables used for transfer and focus research efforts on finding variables that better explain
variations in preferences.
For example, attitude statements about the importance of an activity or the experience
level of participants might be one type of variable mat would improve the explanatory power of
models and transfers. Travel cost and contingent valuation surveys could include a series of
attitude statements, answered on a scale of one to ten, about the importance of an activity or the
respondent's experience level. If these variables have good explanatory power they could be
used to calibrate the transfer.
One reason that socioeconomic variables are generally used to calibrate transfers, despite
their low explanatory power, is that these are the variables for which data are easily available.
To use other calibrating variables researchers might conduct a small "calibration" survey for the
policy site, where respondents are only asked to answer the same series of attitude/experience
questions asked in the original study. The results from the original study could then be weighted
by the attitude/experience values for the policy site to calibrate the transfer. Again, we should be
-------
concerned that the need for such a study could negate much of the potential cost savings that may
arise from benefit transfer.
ir
However, these approaches are not particularly effective in addressing the issue of
"unknown-unknowns." In the absence of studies for the policy context, our only guide may be
our knowledge of the two contexts so that we begin to understand how they differ. This situation
implies that benefit transfer should not be done at "arm's length," but that our judgment needs to
be educated by knowledge of both the study and policy contexts.
Another strategy that may be useful in reducing all three sources of variance, including
"unknown-unknowns," is to transfer estimates from study sites that are most similar to the policy
site, hence minimizing differences between the sites. For example, using an estimate of long
salmon fishing in Alaska or marlin fishing in the Caribbean to estimate the value of warm water
fishing in Mississippi would be completely inappropriate. The activities are just so different that
the estimates would probably not provide any useful information.
This difference suggests that the approach of using the mean of the largest possible
number of estimates for recreational fishing, for example, may be misguided. Rather, a better
approach may be to identify a smaller number of estimates from study sites that are most similar
to the policy site, which will tend to minimize differences between the two contexts, including
unknown sources of variation. A larger number of value estimates cannot substitute for
judgment, and to apply judgment, the researcher must have sufficient knowledge of the study and
policy contexts to determine how well the benefit estimates fit This requirement reinforces the
notion that benefit transfer should not be done at "arm's length," because we need to know which
study contexts are most like the policy context While some differences are obvious (e.g., marlin
fishing versus warm water fishing), similar but less obvious differences might also exist for
different regions or different cultures.
BENEFIT TRANSFER FOR NRDA
. One area of research that has become increasingly important for economists is NRDA.
Because conducting an original study for every damage assessment can be prohibitively
expensive, benefit transfer must often be used. This fact is recognized in the legislation, which
requires developing simplified damage assessment approaches for relatively minor incidents.
Furthermore, the law states that trustees may only recover "reasonable" costs of damage
nent Without using some form of benefit transfer, the costs of damage assessment could.
.
in many cases, easily exceed the calculated damages. For example, the NRDA for the Area
8
-------
I
I
I
f
Anchorage oil spill cost nearly $250,000 to conduct, but natural resource damages were
estimated at $32,000 (Washington State Department of Ecology, 1987).
In recognition of these issues, researchers have developed a variety of structures for
benefit transfer. For example, the Department of Interior (DOI) has developed the Natural
Resource Damage Assessment Model for Coastal and Marine Environments (NRDAM/CME),
which is a structure for benefit transfer based on a computer model that simulates the physical
fates of a spilled substance, the biological effects of this spill, and the resultant economic
damages (e.g., Grigalunas, Opaluch, French and Reed, 1989; Jones, 1992). Alaska and
Washington State have developed more ad hoc approaches for simplified damage assessment
that use damage indexes based on the properties of the substance spilled and the environment in
which the spill occurs. In formulating the regulations for the Oil Pollution Act (OPA), the
National Oceanic and Atmospheric Administration (NOAA) is considering using compensation
tables, the NRDAM/CME model, and other means of "expedited" damage assessment Many
damage assessments have been based on more "traditional" applications of benefit transfer,
where available estimates of impacts, such as body counts or lost beach days, are combined with
available estimates of the values of the resources to estimate damages (e.g., Washington State
Department of Ecology, 1987).
Most of the NRDA work by economists attempts to calculate the value of lost services
due to a spill. This approach is based on the usual definition of Hicksian compensation:
C =
- E(P,NRl,UO)
0)
where C is monetary compensation required to make the individual whole, £(*) is the
expenditure function, P is a vector of market prices, NR° is the without-spill vector of natural
resources, NR1 is the with-spill vector of resources, and U° is the without-spill level of utility.
Thus, monetary compensation is the difference between the with-spill and without-spill levels of
expenditure needed to achieve the fixed level of utility. The aggregate level of compensation
required can be calculated by aggregating over all individuals. This level is often calculated by
estimating compensation required by a "representative" individual and then multiplying by the
size of the affected population.
However, under CERCLA and OPA, a strong preference is expressed for making the
public whole by restoring injured natural resources rather than providing monetary compensation
(e.g., Mazzotta. Opaluch, and Grigalunas, 1992). Additionally, all funds collected, including
•
i
-------
money collected for lost values over the period prior to restoration, must be applied to restoring
resource services.
Consequently, individuals do not receive monetary compensation in the amount C.
Instead, received compensation must be used to restore injured resources. Thus, Hicksian
compensation (the amount C in Eq. [ 1]) does not tell us how much money is needed to make the
public whole. Rather, the amount of compensation required to make the public whole is the least
amount of money sufficient to provide a vector of resources that results in utility equal to the
level of utility provided by the without-spill vector of resources. Consequently, Hicksian
compensation could be far greater than, or far less than, the amount of money required to make
the public whole through resource restoration, depending on how expensive restoration programs
are.
Furthermore, under the law, the cost of restoration must not be "grossly disproportionate
to" the value of the resources provided. Therefore, we need to identify restoration programs that
are sufficient to make the public whole, that are cost effective, and whose costs are not grossly
disproportionate to the value of the resource. If economists wish to continue to play a significant
role in NRDA, we must focus our tools on assessing restoration alternatives and assuring that
benefit transfer structures provide a defensible framework for evaluating resource restoration.
Because we often cannot define restoration as replacing precisely what was lost, the first
step in this process is to develop economic definitions of restoration. The second step is to
develop economic methods for assessing and choosing among restoration alternatives. We can
then explore how to use these methods to produce information that can be transferred from one
context to another.
We can define the amount of restoration, R, required to make the public whole as
U(P, Y". NRl + R)
U(P, YO. NRO) • U(P,
(2)
where U(*> is the utility for a ''representative" individual, and R is a vector of resources restored.
Therefore, R is a restoration program that maintains the social value of the natural resource
assets. Note that no direct link occurs between the cost of a restoration program R and the
amount of Hicksian compensation C. Therefore, the amount C does not measure compensation
required to make the public whole through resource restoration, although it is critical for
determining whether restoration costs are grossly disproportionate to benefits.
J
10
-------
The cost-effective restoration program that makes the public whole is defined by
Min C(R)
R
s.t.
U(P, Y, NRO) = U(P, Y, NR1 + R)
C(R) < FGp • [E(P,NR°,U<>) - E(P,NRl,U°)]
(3)
where C(R) is the cost associated with restoration program R, E is the expenditure function, and
FGP is a factor of gross proportions, as described below. The first constraint requires that the
public be made whole through resource restoration, R, and the second constraint requires that the
cost of restoration not be "grossly disproportionate" to the value of the resource. This sort of
constraint is implicit in the Ohio Decision, where the court suggests "the rule might for instance
hinge on the relationship between restoration cost and use value (e.g., damages are limited to
three-times the amount of use value)" (U.S. Court of Appeals, 1989, footnote 7, p.21). Thus, the
Court's suggestion for grossly disproportionate would be based on a factor of gross proportions
(Fop) of 3.
Equation system (3) is equivalent to the traditional expenditure minimization problem of
utility theory with two exceptions. First, OPA and CERCLA's restriction that the funds must be
used to "replace, restore, rehabilitate or acquire the equivalent" implies that the resulting
expenditure function is restricted in the commodities that can be purchased. This restriction is
reflected in the fact that the minimization is over R, not over all possible commodities. Second,
the purchases are constrained to those sets that are not "grossly disproportionate" to the value of
the resource.
In practical terms, the solution to this problem would progress in stages. For example,
researchers could first identify a number of feasible restoration plans and estimate the time path
of recovery for various resources under each plan.
Next, researchers could identify "equivalent" resources to restore, in terms of social
preferences. Here, researchers could use standard discrete choice models, where a sample of
respondents are presented with alternative programs for restoration, described in terms of the
resources and time frame for each. The respondents would then be asked to choose the most
preferred restoration programs or to rank alternative programs. Standard methods of discrete
choice analysis (McFadden, 1973) could then be applied to determine the levels of restoration for
11
-------
different programs that maintain indifference for a representative individual. That is, we would
identify levels of alternative restoration programs, {Ri, R2, £3, . .. }, that achieve the following:
U(P, Y, NRl + RI) = U(P, Y, NRl + R2) - U(P, Y, NRU R3) * ... (4)
U(P, Y,
For example, we may have a spill that kills 700 seabirds. However, restoring these
seabird populations may not be technically feasible, so researchers could find tradeoffs that
respondents are willing to make between alternative resource restoration programs. We could
conduct a discrete choice study that finds a (hypothetical) program that restores 700 seabirds is
indifferent to one that restores 200 waterfowl or one that restores 800 salmon.
Next, researchers would determine the "equivalent" restoration program that is least
costly. This step identifies the most cost-effective restoration plan that makes the public whole.
Finally, researchers would make sure that this plan is not grossly disproportionate by comparing
the cost of carrying out this plan with the value of the resources restored.
The question of whether this method of "in-kind" compensation is appropriate for transfer
depends on whether the resulting measures of "equivalent" resources can be transferred from one
context to another. Perhaps researchers could construct a hedonk function that balances public
preferences for a variety of resources and includes other explanatory variables that are important
determinants of preferences for resources.
The definitions and methods suggested here are based on the idea of making the public
whole in terms of maintaining the utility derived from resource services, or maintaining the value
to the public of the supply of resources. Thus, this view of restoration is anthropocentric. In
contrast, Hanemann (1992) argues that measuring damages using restoration cost is based on a
deontological philosophy — the idea that nature has innate value independent of people. He
interprets Congressional intent as making the natural environment whole "regardless of cost and
regardless of whether anybody cares about the injury" (p. 573).
Whether Congress based the law on a deontological principle is questionable, however,
or whether they simply felt that economic methods may not capture the full value of injuries so
that restoration costs would provide a more equitable measure. This issue is discussed in the
Ohio decision, where the court interpreted that Congress* intention was not to forego efficiency
but that
Congress was skeptical of the ability of human beings to measure the true 'value' of a
natural resource. Indeed, even the common law recognizes that restoration is the proper
remedy for injury to property where measurement by some other method will fail to
f
12
-------
I
1
f
compensate fully for the injury. Congress" refusal to view use value and restoration cost
as having equal presumptive legitimacy merely recognizes that natural resources have
value that is not readily measured by traditional means. (U.S. Court of Appeals, 1989,
p. 51)
This quote suggests that Congress* intention was not to suggest that full restoration
always be carried out, "regardless of cost and regardless of whether anybody cares," but to make
sure that the value of resources are not systematically understated.
Although the intentions of Congress are not stated clearly, an alternative to Hanemann's
(1992) interpretation would allow for an anthropocentric approach such as that presented above,
where the objective is to make the public whole in terms of maintaining the value to the public of
the stock of resources and associated services, rather than to make the environment whole
regardless of public values. This interpretation is also consistent with the idea of gross
proportions.
Duffield also addresses this issue, discussing the fact that requiring full restoration of
injured resources is based on an equity goal and is likely to result in losses in economic
efficiency (Ward and Duffield, 1992). He states that Congress expressed a preference for full
restoration, which will often be economically inefficient Yet, the DOI proposed regulations
under CERCLA do not require that damages be calculated as the cost of full restoration but that
some combination of restoration and compensation be chosen. A "reasonable number" of
alternatives must be considered, and these alternatives must include the possibility of "no action-
natural recovery." Although restoration is still the preferred goal, combining both economic
efficiency and the idea of restoration is possible under the proposed regulations.
Hanemann (1992) and others see restoration as based on the deontological principle so
that "economic analysis plays only a minor role, associated with calculating restoration costs and
cost-effectiveness. It is under the anthropocentric approach mat economics moves to center
stage" (p. 574). However, the notion of restoration as in-kind compensation is useful for framing
the restoration problem as one of compensating the public in a cost-effective manner, while
remaining compatible with the expressed Congressional preference for restoring damaged
resources. If we view restoration from this anthropocentric viewpoint, economists play a critical
role.
SUMMARY AND CONCLUSIONS
Benefit transfer is a necessary and important economic tool for practical policy analysis.
However, to establish and improve the credibility of benefit transfer we need to place greater
13
-------
emphasis on three areas. First, and most fundamentally, benefit transfer requires valid and
reliable underlying studies. Specifically, we must be very careful to establish that the benefit
*
numbers we are attempting to transfer are themselves defensible. Thus, efforts to establish and
improve the accuracy of our benefit measurement techniques within different contexts will also
improve benefit transfer results.
Most economists would agree that, all else equal, economic valuation techniques are most
credible when individuals are very familiar with the commodity and have repeated experience
with purchase and decision making. As commodities become less familiar and as purchase and
decision making experience becomes more infrequent, estimates are likely to become less
accurate. However, this assessment is purely qualitative, with the exception of the work of
Cummings, Brookshire, and Schulze (1986), which is the only attempt to date to quantify the
expected degree of accuracy of benefit estimates in a broad context Additional efforts are
needed in this area.
Second, we must emphasize testing the adequacy of benefit transfers. Experiments that
attempt to quantify the accuracy of benefit transfers could be quite easily constructed by
comparing the results of site-specific studies to those obtained from transfers. We are aware of
two such studies (Loomis, 1992; Downing and Ozuna, 1992). Additional studies of this nature
could provide a broader basis for evaluating the adequacy of benefit transfers and could help
formulate standards for acceptability in different contexts.
Finally, we need to improve our methods for transferring benefit estimates. One
approach is to improve the calibration of benefit transfers. Traditionally, economists have used
socioeconomic characteristics as calibration variables. However, we typically find that
socioeconomic variables have poor explanatory power, which suggests that they are not good
variables for calibration. Thus, we need to be more creative in developing variables for benefit
transfer. We recommend considering a wider range of calibration variables.
Some current efforts in benefit transfer are focused on developing an extensive catalog of
study results. Researchers may be tempted to use mean values from a large number of studies to
estimate values at the policy site, but these efforts may be misguided if important sources of
variation exist between value estimates, particularly if we cannot identify or explain important
sources.
For example, an estimate of the value of martin fishing in the Caribbean may not provide
any useful information for warm water fishing in Mississippi but may instead provide
misinformation. To the extent that misinformation is used, benefit transfer detracts from reliable
I
I
I
I
1
*
1
1
-^
J
14
-------
benefit estimation and will likely lose credibility as a policy tool. This potential effect suggests
that we should focus our attempts at benefit transfer on obtaining a smaller number of studies
v
that fit the policy context most closely, rather than obtaining the largest number of studies
possible, where many may not fit the policy context.
Other interesting issues regarding benefit transfer arise within the context of NRD A.
Because the statutes express a strong preference for restoration of resources over monetary
compensation, we need to develop methods to evaluate restoration alternatives. More
specifically, we need to identify restoration programs that will make the public whole in the least
costly manner and whose costs are not grossly disproportionate to the value of the injured
resources. We also need to determine the extent to which these measures of compensation are
transferable across contexts.
Bockstael, Nancy E. 1984. "Valuing Natural Resource and Environmental Amenities: Can
Economic Valuation Techniques be Made Defensible." Journal of the Northeast
Agricultural and Resource Economic Association 13(2).
Bishop, Richard C., and Thomas A. Heberlein. 1979. "Measuring Values of Extra-Market
Goods: Are Indirect Measures Biased?" American Journal of Agricultural Economics
61(5):926-930.
Cummings, Ronald G., David S. Brookshire. and William D. Schulze. 1986. Valuing
Environmental Goods: An Assessment of me Contingent Valuation Method. Totawa,NJ:
Rowman and Allanheld.
Downing, Mark, and Teofilo Ozuna, Jr. 1992. "Testing the Feasibility of Recreation Benefit
Function Transferability." Paper presented at AAEA Annual Meeting, August, 1992,
Baltimore, Maryland.
Grigalunas, Thomas A., James J.Opaluch, Deborah French, and Mark Reed. 1988. "Measuring
Damages to Marine Natural Resources from Pollution Incidents Under CERCLA:
Application of an Integrated Ocean Systems/Economic Model" Journal of Marine
Resources 5(1).
Hanemann, W. Michael. 1992. "Natural Resource Damages for Oil Spills in California." In
Natural Resource Damages: Law and Economics, Kevin M. Ward and John W. Duffield,
eds. New York: Wiley Law Publications.
Jones, C. A. 1992. "Recreational Fishing Valuation: Application of the Type A Model." Paper
presented at the 1992 AERE Benefits Transfer Procedures, Problems, and Research
Needs Workshop, Snowbird, UT, June 3-5.
Loomis, John B. 1987. "Expanding Contingent Value Sample Estimates: Current Practices and
Proposed Solutions." Land Economics 63(4):396-402.
15
-------
Loomis, John B. 1989. 'Test-Retest Reliability of the Contingent Valuation Method: A
Comparison of General Population and Visitor Responses." American Journal of
Agricultural Economics 71(l):76-84.
Loomis, John B. 1992. "The Evolution of a More Rigorous Approach to Benefit Transfer:
Benefit Function Transfer." Water Resources Research 28(3):701-705.
Mazzotta, Marisa J., James J. Opaluch, and Thomas A. Grigalunas. 1992. "Economics of
Restoration and Natural Resource Damage Assessment" University of Rhode Island,
Department of Resource Economics Staff Paper.
McFadden, D. 1973. "Conditional Logit Analysis of Qualitative Choice Behavior." In
Frontiers of Econometrics, P. Zarcmbka, ed. Academic Press: New York.
Mitchell, Robert C., and Richard T. Carson. 1989. Using Surveys to Value Public Goods: The
Contingent Valuation Method Resources for the Future.
Motor Vehicle Manufacturers Association versus State Farm Mutual Insurance Company. 1983.
463 U.S. 29,43,103 S.Ct 2856,2867,77 L£d.2d 443,458.
NRDC versus Harrington. 1985. 247 U.S. App. D.C. 340,370,768 F.2d 1355,1385.
Smith, V. Kerry. 1989. 'Taking Stock of Progress with Travel Cost Recreation Demand
Methods: Theory and Implementation/' Marine Resource Economics 6(4):279-310.
U.S. Court of Appeals, D.C. Circuit 1989. State of Ohio v. U.S. Department of the Interior, 880
F.2d.432.
U.S. Department of Interior, Notice of Proposed Rulemaking, Federal Register, April 29.1991.
Washington State Department of Ecology. January 1987. "Marine Resource Damage
Assessment Report for the Arco Anchorage Spill." Rept No. 87-4.
Ward, Kevin M., and John W. Duffield. 1992. Natural Resource Damages: Law and
Economics, New York: Wiley LAW Publications.
J
.1
* -
j
16
-------
BENEFITING BENEFITS TRANSFER:
INFORMATION SYSTEMS FOR COMPLEX SCIENTIFIC DATA
Martin H. David*
ABSTRACT
In this paper, I suggest reorganizing the science on which we build benefits estimates. I
advocate developing a system for sharing data, creating support for archiving scientific
measurements, and making data easier to use.
This paper advocates reorganizing the science on which we build estimates of benefits.
Reorganization implies four imperatives:
1. Build an effective system for sharing data, an Information System for Complex Data
(ISCD).
2. Create necessary support for archiving scientific measurements.
3. Begin now. Deploy existing computer and software capabilities to reduce learning
time for secondary use of data, to increase scope of questions mat can be addressed to
existing data, and to anticipate the arrival of new generations of software and
hardware.
4. Change incentives.
WHY BUILD ISCD?
Positive Reasons
Benefits transfer is applied science, statistical science, implying the following:
• Estimates of benefit are simulations based on empirical fact Observations and models
of those observations are used to simulate out-of-sample forecasts for benefits transfer.
• The procedure for benefits transfer must be reproducible. Reproducibility keeps the
estimates out of court and away from accusations of fraud.
• Bounds on error in estimates are needed to tell us how good the estimate is. We have
more certainty about the value of the salmon fishery, governed by market prices, than
we have about the value of wilderness, whose nonmarket externalities preserve options,
species, and perhaps even climate.
• Learnin
scopeo
ig-by-doing. Over time we expect the error of benefit estimates to decline. The
if estimates will increase as the range and quality of measurements increases.
1
f
*University of Wisconsin-Madison, Department of Economics.
-------
Accepting these ideas suggests a need for better organization of facts and models based on facts.
I call that better organization an ISCD.
Negative Reasons
The way researchers currently handle data relevant to benefits transfer is deficient for the
following reasons:
* Data are seldom "recycled" and distributed for secondary analysis.
• Data are seldom exposed to sufficiently broad scientific examination to locate
significant errors and systematic problems.
• Methods for archiving and retaining data for later use are unsatisfactory.
• The meaning of measurements changes over time.
• Underlying physical, biological, and behavioral data are seldom integrated.
Positive and negative reasons imply that benefits transfer will benefit from improved information
systems for access to data about benefits—that system is ISCD.
DAVID'S DISCOVERIES—ISCD AND NECESSARY SUPPORT
Over the past 8 years my colleagues at the University of Wisconsin and I have developed
an innovative management system for data (David and Robbin, 1990.1992; Flory, 1989; Robbin,
1992). The system was directed to reducing the cost of learning and the cost of access to a
complex body of data. The ISCD differs from earlier efforts at data dissemination and data
libraries in the same way that the telephone instrument differs from die telephone system. In the
past data disseminators have supplied the instrument (data plus codebook) as a static package.
Useis were not integrated into a network or research community. The ISCD is founded on bigh-
capacity communication that links users to data, to tools for interpreting data, to the producer, to
an expert, and to each other. Figure 1 conceptualizes the system.
An ISCD links four capabilities: U links the capabilities of the data collector to users
through two intermediaries, a computer-based knowledge resource and a human expert (David
and Robbin, 1990). The ISCD is described by communication links among these four
capabilities; die links between users and expert, users and the knowledge resource, and the expert
and the knowledge resource are particularly important How the knowledge resource is
organized depends on the skills and knowledge of users and the ability of the expert to design
intelligence into the knowledge resource. Every dataset has a collector or data producer; many
users can exploit the public good that results from data collection; the expert is what computer
-------
f
Producer
Documentation
Feedback
Data Enhancements
User Interfaces
Research Queries
Knowledge
Resource
Data Release
Enhancements
Expert
Consultation
Other communication
High-capacity data link
Figure 1. An Information System for Complex Data
scientists call a domain specialist, a diagnostician who can solve problems in interpreting data in
one-on-one consultation or who can alter the computational capability to teach many users to
solve the problem for themselves. (This keeps cost down.)
The knowledge resource integrates software and data in a system mat can respond to
inquiries from users. Pre-programmed artificial intelligence and carefully designed data structure
(the database schema) speed the recovery of frequently used data and its description. The
knowledge resource includes several critical elements: a body of data accessible for statistical
analysis (e.g., a S AS system file), descriptions of the data that provide necessary support, an
archive of reports generated from the data, and bibliographic databases mat can be searched for
citations, data sources, and subject are crucial to the process of teaming about and using data. At
Wisconsin many of these capabilities were organized in a relational database management
system (RDBMS). The power of those systems is explained in the Appendix.
-------
Experience at Wisconsin with an ISCD for Census data (the Survey of Income and
Program Participation) demonstrates cost-effectiveness. The ISCD reduced the "learning time"
required for investigators to produce socially useful output The system pooled discoveries about
data. Errors identified by staff and users led to revisions in most cases and to systematically
indexed error-messages in the remaining cases. The system was productive and generated over
two-thirds of the research output not authored by the data purveyors (or their contractors,
including the Wisconsin "SIPP-ACCESS," Census, and the Michigan ICPSR). The system was
timely; SIPP-ACCESS staff were able to explore dynamics before the Census Bureau. The
system was open-ended and permitted users to deposit geo-coded data about labor markets and
welfare for general use.
What aspects of the successful SIPP-ACCESS are suited to the many small-scale datasets
that measure benefits? Thinking about the cost of access to data collected by others helps us to
think about transferability.
MAKING DATA EASIER TO USE
Secondary users of data complain about the data provided. Typically, they complain that
the data are not adequately "documented." Also users typically complain about "dirty data."
What do these complaints signify? Secondary users do not have access to the laboratory
notebooks of the collector, and they cannot interrogate the collector to determine what
inconsistencies and anomalies resulted from die execution of the scientific design. What David
(1991) calls necessary support provides critical information about the scientific design.
What would we really like to know? Computer scientists call information about data
"meta-data," A scheme for organizing meta-data follows easily from the scientific method. We
need to know how the scientific design is executed. We need to know how the design covers a
population of interest and how well the instruments used measure reality. For every dataset we
need to know:
• Provenance of each datum: Where do the numbers come from? What questionnaires,
• forms, and computer-assisted data capture were used? What algorithms are used for
processing and transforming responses?
• Interpretation of values:
- Lateral level—When are numbers substitutesif or text? When are numbers
nondecimal, as numeric values for dates or time?
- Contextual level—What apparent logical inconsistencies are present because of
errors in response and processing?
-------
- Conceptual level — When are the numbers real? When are numbers imputed?
When are numbers randomly altered, "fuzzed" (to limit disclosure)?
*
• Consequences of using each datum: What unusual interpretations? When are zeroes
null? When are values truncated?
* Inference from the data: How can honest inferences be made from the data? Does
selection bias inference from reports? Does variance in the measurement process mask
real world phenomena? Is the measurement process biased by moral hazard?
All of this information can be embedded in and linked to the RDBMS that contains the
measurements. Table 1 contains a summary of the lands of information that should be stored in
an ISCD to provide necessary support for the secondary data researcher.
WHAT PAYOFFS DO ISCD AND NECESSARY SUPPORT CREATE FOR BENEFITS
TRANSFER?
»
Three kinds of payoff follow from organizing data on economic benefits in an ISCD:
discovering the state of the an of benefit measurement will be less costly, synthesizing benefit
measurements will be easier, and incorporating superior methodology in estimating models on
larger sets of data will be possible.
State-of-the-Art
Garner's bibliographic database (circulated for the AERE conference) represents an
important step towards an ISCD for benefit measurements. Citations to reports about what we
already have discovered are accessible. The perspective of the ISCD, says Gamer, should add
one element to the bibliographic database— a tide for each dataset exploited in each report.
Titling and citing datasets are the only ways to establish the empirical foundation for any
analysis. The title concept is implemented in Roistacher et at (1980) but has not found its way
into accepted referencing for scientific publication or into bibliographic databases. The
perspective of an ISCD implies that the database is complemented by electronically stored files
containing each of die reports and articles cited. Electronic preservation of reports allows any
user to review any document cited. Archiving scientific work in this way assures permanence for
die published record.
Because many excellent datasets are collected to pursue contractual obligations, reports
containing important datasets are difficult to find— even in die contractor's archives. This makes
electronic archiving of reports critical.
-------
TABLE 1. IMPORTANT CAPABILITIES FOR META-DATA (TYPE OF META-
DATA INDICATED IN PARENTHESES)
• Access to general language descriptions of variables (text)
• Browsing both with text search and controlled vocabulary
* Comparability of meta-data over time for any repeated measurement
* Descriptions of the semantic principles for data "tables" (text)
* Access to "forms" used to collect the data (graphics)
* Access to underlying "rules" for completing the form, that is links to legal databases or
interviewer manuals (text)
- In the case of IRS forms, rules include code, administrative code, informal rulings, and
judicial precedent More general conventions apply to principal business activity,
occupation, and product codes.
Capacity to refer to analyses of data. This capacity requires two steps: a link to
bibliographic databases (text) and retrieval of the analyses (text and graphics). Analyses
clarify the following:
- entities present or the universe of analysis
- implications of inconsistencies
- implications of arbitrary processing decisions
- correct use of sample weights
• Meta-data that provide statistics on truncated variables (text and aggregate data)
• Imputation models and models used to generate noise to mask data (text and code specific
to a computer language)
• Reports on error rates (text and aggregate data)
Implications For Data
Some effort to "normalize" the data to third normal form
Data on higher moments for partially aggregated data (numeric)
Partially aggregated data, particularly on processes monitoring data quality (numeric)
-------
i
1
i
f
I
I
Synthesizing Measurements
Smith and Huang (1991) undertook to synthesize measures of willingness to pay for air
quality from data on the housing market in areas with differing levels of air quality. Their
analysis searched over fifty statistical studies spanning two decades of observations. The
response that Smith and Huang seek to estimate is the marginal rate of substitution between price
paid for housing and paniculate deposition. Underlying that measure of response is a model of
the price paid by individuals for residential housing units in various urban areas of the U.S.
Smith and Huang develop an excellent "meta-analysis" of models fit to the underlying
data. Their work synthesizes past investigation but cannot recover much of the variability in
underlying data (because each model is a projection of the underlying data into a small number
of dimensions). Furthermore, the technique fails to recover any direct information about the
dynamics of willingness to pay; all human response to pollution is inferred from differences in
prices paid by similar people buying housing in different places.
Crippling Problems
This study faced extreme difficulties and I admire the authors for their perseverance:
• Assembly. The study required assembling 26 journal articles. 5 unpublished papers, 5
dissertations, and 1 edited volume.
• Incomplete estimates of response to air quality. Many studies did not contain responses
to ozone, SO2. and other indicators of air pollution. The meta-analysis is confined to
understanding response to particulates (arguably the most obvious aspect of air quality).
Models that did not include particulates had to be excluded.
• Reuse of data. The same data were used to estimate several models, both within and
between research teams. For that reason models estimated are not independent
• Incomplete documentation. Some papers failed to describe either the data or estimating
method in sufficient detail to permit meta-analysis. Contact with one researcher filled
some lacunae. In other cases pollution data were augmented; still other models could
not be included in the meta-analysis.
Limitations of the Approach
Some of the studies are based on measures of house value and air pollution that are
aggregated over space (e.g., Census tracts). A second problem is that the studies use a variety of
measures of willingness to pay: samples of sales data, samples of FHA-mortgaged properties,
samples derived from the Census self-reported house values, and the Annual Housing survey.
-------
Each technique has its own problems of coverage, bias, reporting error, and relation to
underlying willingness to pay.
How could ISCD reduce the cost and limitations of Smith and Huang's investigation?
The comment on Gamer's bibliographic database applies. Smith and Huang would have been
well served by bibliographic databases (BDB) that can be searched on the basis of sources of
data, where the titles of datasets are controlled by bibliographic conventions. Electronic images
of every publication cited would obviate the task of assembly.
Meta-analysis requires a database of information about the scientific design used in
estimation. Smith and Huang carefully assembled such a database. Their task would have been
vastly easier had each article been archived with a database containing necessary support—the
scientific design, including information about instrumentation (questionnaires), data processing,
error profiles, and meanings of values encoded in the datasets (see Table 1).
Reanalysls of Pooled Data
Reanalysis of a collection of datasets is the most powerful capability of the ISCD. Had
an ISCD containing house values and indices of air pollution been available, Smith and Huang
would certainly have preferred to examine the original measurements. Access to data and
interpretation of measurements would be assured in the ISCD. Examining the data would make
it possible to specify a single model that transcends numerous measurement systems repeated at
intervals. The ISCD also makes incorporating new estimation methods that were not known at
the time of the original data collection possible. Those methods can assist in resolving problems
created by changing precision of instruments, missing data, and outliers (Manski, 1991).
Lastly an ISCD provides a template that can be used to add new data to an existing
database. This template lowers the cost of extending analysis in time. Smith and Huang (1991)
note mat the last study of the relationship of house value to air pollution refers to 1980; with a
lower cost of access to data, a decade might not have elapsed between replicated models.
INCENTIVES
How can we move to implement ISCD and
ry support for benefit measurements
and other data pertaining to the environment? The present disincentives to share data can be
convened into positive incentives. The institutions that fund data collection can insist on data
sharing. The NSF follows this stipulation in its grants; EPA and other contractors would be wise
to follow. This requirement implies that every set of measurements assembled must be
-------
I
I
I
t
"documented" and "titled." Funding institutions can also require the deposit of completed
publications in a data-oriented library. Funding institutions are in a position to proyide necessary
resources and to enforce their agreements with data collectors.
Funding institutions cannot proceed without support from the professions. Journals need
to require citation of data sources. They also need to require datasharing that permits replication
of published findings. Some journals have already adopted this point of view (AER, JHR, and
JEEM),
These changes in institutions and journals are easy. Change in our own professional
conduct is also needed. While the computational capability to create low-cost datasharing is
available on most desktops, social conventions need to be forged to support effective
datasharing. Just as we have conventions to drive on the right and stop at the red signal, we need
conventions on a common system for organizing shared data. Up to now, we have been unable
to specify completely what is required for datasharing. Concepts from computer science clarify
what is needed at the same time that those concepts forged the technology that we can now use to
organize data.
Experience at Wisconsin shows the efficacy of ISCD. Computer science has given us the
relational data model, which organizes data and necessary support for data in a common
framework. The geniuses of Silicon valley have given us technology that makes the
computational costs of ISCD trivial in comparison to the cost of professional time. So long as
we do not implement ISCD, much professional time will be wasted in searching for the right
data, the right model, and implementing the simulation required for benefits transfer. Do we
really want to waste scarce resources for learning about the environment?
Casella, George, and Edward L George. 1992. "Explaining the Gibb's Sampler." American
Statistician 46(3): 167-174.
Codd.E.F. 1985. "Is Your Database ReaUy Relational?' Computerworld IDMD9. Also:
"Does your Database Run by the Rules?' Computerworld 45-49ff.
Date.CJ. 1987. An Introduction to Database Systems. VoL 1.4th Ed. Reading MA: Addison-
Wesley Publishing Co.
David, Martin H. 1991. The Science of Data Sharing: Documentation.** ID Sharing Social
Science Data: Advantages and Challenges. Joan E. Sieber, ed., Newbury Park, CA:
Sage Publications.
-------
David, Martin H,, and Alice Robbin. 1990. "Computation Using Information Systems for
Complex Data." Proceedings of the Conference on Advanced Computing for the Social
Sciences, April! 0-12,1990. Oak Ridge, TN: Oak Ridge National Laboratories.
David, Martin H., and Alice Robbin. January 1992. Building New Infrastructures for the Social
Science Enterprise: Final Report to the NSF on the SIPP ACCESS Project, November
1984-December] 991. 2Vols. Madison, WI: Institute for Research on Poverty.
Flory, Thomas S. 1989. "Commercial Relational Database Management Systems in a Social
Research Environment: Expectations and Evaluation." ZA-Information 24 (May):
128-138.
Manski, Charles. 1991. "Regression." Journal of Economic Literature 29(1):34-50.
Robbin, Alice. 1992. "Social Scientists at Work on Electronic Research Networks." Electronic
Networking: Research, Applications, and Policy (forthcoming).
Roistacher, R., S. A. Dodd, B.B. Noble, and A. Robbin. 1980. A Style Manual for Machine-
Readable Data. Report SD-T-3, NCH62766. Washington DC: Department of Justice,
Bureau of Justice Statistics.
Smith, V. Kerry, and Ju Chin Huang. 1991. "Can Hedonic Models Value Air Quality? AMeta-
Analysis." Discussion paper QE92-06. Washington, DC.
10
-------
APPENDIX A
RELATIONAL DATABASE MANAGEMENT SYSTEMS—RDBMS
In the past dozen years computer scientists have discovered important principles for
storing and retrieving large volumes of numerical data and smaller volumes of text. Their ideas
culminated in relational database management systems (RDBMS) technology that is now
available to every PC owner for about the same price as a spreadsheet program.
RDBMS denotes a system with several essential features, starred in Table A-l. The
systems were designed to meet needs for commercial "transactions processing," whose
requirements are somewhat different from scientific statistical processing, although the
commonalities are much greater than roost social scientists understand. The systems are
designed for multiple users—both multiple data suppliers (i.e.. points of data entry) and multiple
researchers. RDBMS are designed to support interactive use of the data at all times and maintain
an unambiguous outcome for statistics (reports in the RDBMS jargon) that are generated at any
point in time. This feature is called data concurrency.
A mandatory requirement for RDBMS is dynamic independence. Adding new data to the
system without restructuring the existing data must always be possible. For example, successive
measures of pollution control and abatement expenditures (PACE) can be loaded into the system
without knowing about or interfering with older data. Contextual data can be added to the
system without determining the attributes used to link those data to individuals in advance. Thus
interview data obtained from households can be loaded without knowing that their report of
industry affiliation might subsequently be used to assign worker exposure to safety risks or that
geography might later be used to assign prevalence of radon exposures.
Data entry is controlled by logical rules that can draw on any part of the existing data to
enforce consistency; consistency may be applied to individuals, households, firms, activities
(other entities), and combinations of entities. Consistency rules are called integrity constraints
on the database. Referential integrity implies that adjustments to the database do not leave
garbage in the system. For example, if an individual is found to be associated with the wrong
address, all traces of that individual are dissociated from that address when the address is
corrected.
A-l
-------
TABLE A-l. PREREQUISITES FOR AND CHARACTERISTICS OF AN RDBMS
Essential Prerequisites for an RDBMS (Codd 1985, Date 1987)
Dynamic independence*
Referential integrity8
Query languages that are logically complete8
Support for numerical, text, and symbolic data types8
Characteristics of RDBMS that Enhance Data Quality, Researcher Productivity, and
Program Documentation
• Data Entry
Integrity constraints8
• Storage
Flat files
Indexed retrieval
• Analysis and Reports
Integration of logically different units (e.g., jobs and families)
Discovery of anomalies
• Housekeeping
Unioue nomenclature in large databases8
Meta-data driven loading8
Data concurrency8
Joumaling
Security
Platform independence
Distributed databases
• Interfaces
Statistical processors8
•Features of RDBMS that are critical to research data p
ling.
A-2
-------
For researchers the most important property of RDBMS is its "query language."
Requests for information are written in the query language, which has a simple structure derived
from concepts of mathematical logic. Query languages support any logical operation on any
mathematical or lexical function of the attributes or variables in the database. Query languages
are compact, and SQL has been adopted as an industry standard that will be supported by all
database vendors.1
RDBMS provide permanent housekeeping that is essential when multiple points of entry
and multiple users must be accommodated. Finally, the RDBMS support sophisticated security
and reporting. Users can be restricted from access to particularly sensitive data. Operations can
be monitored continuously by reports on the capture of interviews, error-rates, outliers, and
interviewer comments.
The logic of RDBMS results in "flat files," rectangular arrays that are easy to move
outside of the RDBMS environment Furthermore, the RDBMS encompass two capacities that
aid a complex data collection through a nation-wide system. The databases support "distributed
databases" whose parts may reside on different computers. For example, the database required
for sampling can be separated from the data generated by interviewing. The second capability is
"platform independence" that assures the system operates in the same manner on all hardware
using identical programs or applications.
The most important feature of RDBMS for a complex data collection is that it maintains a
vocabulary of names for each measurement, each transformation, and each relationship
encompassed in the database, no matter how many users are proceeding to make independent
uses of the data.2
Table A-l lists aspects of RDBMS that are critical for successful processing of scientific
data pertaining to the environment
1 RDBMS qjplyvtulcialiDielugeiiceio minimizing te coa Therefore, execution of
particular requests does not proceed in me procedunDy defined mauier of scientific programming languages and
statistical processors. Tins feature implies mat embedding scientific programming languages in me database (and
an RDBMS support such capabilities) causes poor performance. Undemanding the strengths of me RDBMS.
however, allows as to design interfaces to statistical processors (e»g.. SAS and SPSS) mat permit me RDBMS to
locate required data efficiently while permitting the aggregation of data aooa entities to |*oceed equally
efficiently. Tlie merit of such interface* is mat menlckncies of dau management and stoia^
processors can be eliminated without yinnmaii^g the finely toned calculation of estimates mat those processors
support
2 Thll Capability nMfcPt J! pMliHt fr p***"**"' *"" ^ntM"l« *"" ** «i«tfa»g metadata that Af*r*n*. the tarvey
instrument and automate the evolution of me database as each pawl proceeds and as new panels are created. The
result is greater productivity in manipulating cm-going change to the strortiire of me database and greater clarity
in the documentation and diagnostics produced for me documentation of all data-processing steps.
A-3
-------
-------
1
t
APPENDIX A
ECONOMIC ANALYSIS AND
RESOURCE BRANCH
ENVIRONMENTAL BENEFITS
DATABASE
1
-------
-------
1983. Contingent Valuation Surveys for Evaluating Environmental Assets, Natural Resources Journal.
1984. Facts and Values in Risk Analysis in Environmental Toxicants, Risk Analysis.
1983. Second Annual Symposium on Environmental Epidemiology, Environmental Health Perspectives.
1989. Biological Habitat Reconstruction, ed. Buckley, G. P., Belhaven Press, London.
1984. Estimating Willingness to Pay to Reduce the Risk of Infertility: An Expolratory Inquiry, prepared by Charles
River Associates, Inc.
1983. Benefits of Preserving Cultural Materials From Damages Associated with Acidic Deposition, prepared by
Charles River Associates, Inc.
1984. Assessing Cost-Benefit Assessments. Journal of Water Pollution Control Federation.
1986. An Economic Assessment of Marine Recreational Fishing in Southern California, prepared by National
Marine Fisheries Service.
1991. Valuing the Environment Six Case Studies, ed. Barde, J.P and Pearce, D.W., Eanhscan Pub. Ltd., London.
AERE. 1987. Environmental Monitoring and Enforcement: Theory and Practice, prepared by AERE, EPA, NOAA
for Association of Environmental and Resource Economists.
AERE. 1990. Natural Resource Market Mechanisms, prepared by AERE, EPA, NOAA, USDA for Association of
Environmental and Resource Economists.
AERE. 1985. Recreation Demand Modeling, prepared by AERE, EPA, NOAA for Association of Environmental
and Resource Economists.
AERE. 1991. The Management of Non-Point Source Pollution, prepared by AERE, EPA, NOAA, USDA for
Association of Environmental and Resource Economists.
AERE. 1988. Fourth Annual AERE Workshop - Marine and Sport Fisheries - Economic Valuation and Management
- Papers, Association of Environmental and Resource Economists.
Abdalla, Charles. 1990. Measuring Economic Losses firm Groundwater Contamination: An Investigation of
Household Avoidance Cost, Water Resources Bulletin, Vol. 26, No. 3, pp. 451-463.
Abel, Fred H., Dennis P. Tehansky, and Richard G. Walsh. 1975. National Benefits of Water Pollution Control,
prepared by Washington Environmental Research Center for US Environmental Protection Agency (ORD).
AbelsoD, Peter W. 1979. Property Prices and me Value of Amenities. Journal of Environmental Economics and
Management, VoL 6. pp. 11-28.
Abt Associates, Inc. 1984. Air Pollution Damages to Cultural Materials.
Adamowicz, W. L.. and W. R Phillips. 1983. A Comparison of Extra Market Benefit Evaluation Techniques,
Qn«Hian Journal of Agricultural Economics, Vol. 31, pp. 401-412.
Adamowicz, Wilder L., Jerald J. Fletcher, and Theodore Graham-Tomasi. 1989. Functional Form and the Statistical
Properties of Welfare Measures, American Journal of Agricultural Economics, VoL 71, No. Z pp. 414-421.
Adams, R. L., R. C. Lewis, and B. H. Drake. 1973. An Economic Analysis of Outdoor Recreation, prepared by
Bureau of Outdoor Recreation for US Department of Interior, Washington, DC.
A-l
-------
Adams, Richard M., Scott A. Hamilton, and Brace A. McCarl. 1986. The Benefits of Pollution Control: The Case of
Ozone and US Agriculture, American Journal of Agricultural Economics, Vol. 68, No. 4, pp. 886-893.
*•
Adams, Richard M., Scott A. Hamilton, and Bruce A. McCarl. 1984. The Economic Effects of Ozone on
Agriculture, prepared by EPA Environmental Research Laboratory for US Environmental Protection Agency.
Adams, Richard M., Thomas D. Crocker, and Richard W. Katz. 1983. The Value of Yield Response Information in
Economic Assessments of Pollution Impacts on Managed Ecosystems: A Methodology with Illustrations,
prepared by EPA Office of Policy Analysis for US Environmental Protection Agency, University of
Wyoming, National Center for Atmospheric Res.
Adams, Richard M., Thomas D. Crocker, and Richard W. Katz. 1984. Assessing the Adequacy of Natural Science
Information: A Bayesian Approach, Review of Economics and Statistics, Vol. 66, No. 4, pp. 568-575.
Adler, Kenneth I., Robert C. Anderson, Zena L. Cook, and Roger C. Dower, el al. 1982. The Benefits of Regulating
Hazardous Waste Disposal: Land Values as an estimator, prepared by EPA Office of Solid Waste and
Emergency Response for US Environmental Protection Agency, Public Interest Economics Center,
Washington, DC.
Adler, Kenneth J., William Desvousges, M. P. Lynch, and K. L. McDonald. 1987. Evluating the Potential Economic
Benefits of Estuarine Water Quality Improvements: A Cross-estuary Comparison, prepared by Environmental
Protection Agency for The Coastal Society, Bethesda, MD.
Adrangi, Bahrain. 1982. Changes in Economic Efficiency Resulting from Allocation of Oregon National Forest to
Skiing, University of Oregon.
Agnello, R. J. 1988. Economic Valuation of Marine Recreational Fishing.
Agthe, D. £.. and R. B. Billings. 1980. Dynamic Models of Residential Demand. Water Resources Research, Vol.
16, pp. 476-480.
Alig, Ralph J. 1983. Impacts of Forest Land Conversion: An Overview, Renewable Resources Journal, Vol. 2, No.
1, pp. 8-13.
Allen, P. Geoffrey, Thomas H. Stevens, and Scott Banco. 1981. The Effects of Variable Omission in me Travel
Cost Technique. Land Economics, Vol. 57, No. 2, pp. 173-180.
Allen, P. G, and T. H. Stevens. 1979. The Economics of Outdoor R
Congestion: A Case Study of
Camping, Journal of Northeastern Agricultural Economics Council, Vol. 8, pp. 13-16.
AIlsup, Jerry R. 1987. Effect of Low Levels of Lead and Alternate Additives to Lead on Engines designed to operate
on Leaded Gasoline. US Environmental Protection Agency.
Amfrfami, P., R. Narayanan, B. Bishop, and D. Larson. 1984. A Methodology for E«ti|iiarinE Instream Flow Values,
prepared by Utah Water Resources Lab for Utah Stale University, Logan, UT.
Amoroso, Frank L., and Linda R. Keenan. 1991. Liability for Restoration is Looming, National Law Journal, pp. 19-
23.
Amocoso, Frank U and Linda R. Keenan. 1990. Natural Resource Damage Recovery Actions: A Legal Tidal Wave
Looming, Westcbester Bar Journal, Vol. 17. No. 2. pp. 109-127.
Anderson, Elizabeth. 1983. Quantitative Approaches in Use to Assess Cancer Risk, Risk Analysis.
Anderson, F. J., and N. C. Bonsor. 1974. Allocation, Congestion, and me Valuation of Recreational Resources, Land
Economics, Vol. 50, pp. 51-57.
A-2
-------
*
•
f
Anderson, L. G. 1980. Estimating the Benefits of Recreation Under Conditions of Congestion: Comments and
Extension, Journal of Environmental Economics and Management, Vol. 7, pp. 401-496.
Anderson, Lee. 1983. The Demand Curve for Recreational Fishing with an Application to Stock Enhancement
Activities, Land Economics, Vol. 59, No. 3, pp. 279-286.
Anderson, R. C, and R. C. Dower. 1980. Land Price Impacts of the Adirondack Park Land Use and Development
Plan, American Journal of Agricultural Economics, Vol. 62, No. 3, pp. S43-S48.
Anderson, R. J . 1981 . A Note on Option Value and the Expected Value of Consumer Surplus, Journal of
Environmental Economics and Management, Vol. 8, pp. 187-191.
Anderson, R. J., and T. Crocker. 1972. Air Pollution and Property Values: A Reply, Review of Economics and
Statistics, Vol. 54, No. 4, pp. 470-473.
Anderson, R. J., and T. Crocker. 1971. Air Pollution and Residential Property Values, Urban Studies, Vol. 8, pp.
171-180.
Anderson, Robert, and Ban Ostro. 1983. Benefits Analysis and Air Quality Standards, Natural Resources Journal.
Angelo, R. J., and L. G. Anderson. 1984. The Value of Fish and Fishing Days: A Partial Solution to Managing
Recreational Fisheries with Stock Externalities, University of Delaware.
Amdorfer, David J., and Nancy Bockstael. 1986. Estimating the Effects of King Mackerel Bag Limits on Charter
Boat Captains and Anglers, prepared by Environmental Resources Management, North Central Inc. for
NMFS Southeast Fisheries Center.
Arrow, Kenneth, and Anthony C. Fisher. 1974. Environmental Preservation. Uncertainty, and Irreversibility,
Quarterly Journal of Economics, Vol. 88, pp. 312-319.
Asako, Kazumi. 1979. Environmental Pollution in an Open Economy, Vol. 55, No. 151, pp. 359-367.
Ashford, Nicholas A., and Christopher T. Hill. 1982. Analyzing the Benefits of Health, Safety, and Environmental
Regulations, prepared by Massachusetts Institute of Technology for US Environmental Protection Agency.
Ashford, Nicholas, C. W. Ryan, and C. C. Caldard. 1983. A Hard Look at Federal Regulation of Formaldehyde: A
Departure from Reasoned Detisionmaking, Harvard Environmental Law Review.
Assaf. George B., Brent C. Kroetch, and Suboodh C. Matfaur. 1986. Non-Market Valuation of Accidental Oil Spills:
A Survey of Economic and Legal Principles, Marine Resource Economics, Vol. 2, No. 3, pp. 211-238.
Atkinson, S. 1983. Marketable Pollution Permits and Acid Rain Externalities,
Journal of Economics.
Atkinson, Scott E., Thomas D. Crocker, and Herbert L. Needteman. 1983. The Economic Consequences of Elevated
Body Lead in Children: A Proposed Study Framework, prepared by EPA Economic Analysis and Research
Branch for US Environmental Protection Agency.
Atkinson, Scott, and T. H. Tietenberg. 1984. Approaches for Reaching Ambient Standards in Non-Attainment
Areas: Financial Burden and Efficiency Considerations, Land Economics.
Australia. 1992. RAC Forest and Timber Inquiry Final Report, Volume 2A, prepared by Resource Assessment
Commission for Canebera, AGPS, pp. E20-E22.
Avol, A. E., W. S. Linn, and T. G. Venet 1983. Acute Respirator Effects of Los Angeles Smog in Continously
Excersizing Adults, Journal of Air Pollution Control Association.
Bailey, M. J. 1982. Risks, Costs and Benefits of Flourocarbon Regulation, American Economic Review.
A-3
-------
Bailey, Martin J. 1982. Externalities, Rents, and Optimal Rules, prepared by Economics Department, University of
Maryland for College Park, MD.
Baker, J. 1983. Selective Effects of Insecticides on Within Species Variation: The Lesson to be Learnt when
Considering the Environmental Effects of Pollution, Agriculture and Environment.
Baker, M.D. 1986. Property Values and Potentially Hazardous Production Facilities: A Case Study of the Kanawha
Valley, West Virginia, prepared by Florida State University for Tallahassee, Fla.
Balco, John J. 1981. Assessing Wetlands Values - Evaluation Dilemmas, Wetland Values & Management
Conference, pp. 421-429.
Balkan. Erol, and James R. Kahn. 1988. The Value of Changes in Deer Hunting Quality: A Travel Cost Approach,
. Applied Economics, Vol. 20, No. 4, pp. 533-539.
Balson, William E., and Jennie S. Rice. 1990. Estimating Uncertainty Bounds on the Public Benefit of Visual-range
Improvements Attributable to Sulfur-Dioxide Emission Reductions in the Eastern United States, prepared by
Decision Focus, Inc.
i
Baram, Micahel. 1984. Charting the Future Course for Corporate Management of Health Risks, American Journal of
Public Health.
Barbera, Anthony J., and Virginia McConnell. 1990. The Impact of Environmental Regulations on Industrial
Productivity: Direct and Indirect Effects, Journal of Environmental Economics and Management, Vol. 18,
No. 1, pp. 50-65.
Barbier. E. B. 1989. Economic Valuation of Tropical Wetland Resources: Application in Central America,
University College, London, London.
Barbier, EJJ. 1989. The Economic Value of Ecosytsems: 1 - Tropical Wetlands, prepared by London Environmental
Economics Centre.
Barbier, E.B. 1991. The Economic Value of Ecosytsems: 2 - Trpoical Forests, prepared by London Environmental
Economics Centre.
Barbier, E.B., W.M. Adams, and K. Kimmage. 1991. Economic Valuation of Wetland Benefits: the Hadejia-
Jama'are Floodplain, Nigeria, prepared by London Environmental Economics Centre for London.
Bardecki, Michael J. 1984. What Value Wetlands?, Journal of Soil and Water Conservation, pp. 166-169.
Barnes, David. 1983. Backdoor Cost-Benefit Analysis Under a Safety-First Clear Air Act. Natural Resources
Journal
Banes, R. A., G. S. Parkinson, and A. E. Smith. 1983. The Com and Benefits of SidphiirOxkleCtatioL Journal of
Air Pollution Control Association.
Bamett, W. A. 1977. Pollack and Wachter on the Household Production Function Approach. Journal of Political
Economy, Vol. 85, No. 5, pp. 1073-1086.
Bamett, Andy H. 1980. The Pigouvian Tax Rule Under Monopoly. American Economic Review, VoL 70, No. S. pp.
1037-1041.
Barr, Nathaniel. 1983. The Role of Environmental Risk Analysis in the Cost-Effective Development and Operation
of Emerging Energy Technologies. Annals of Nuclear Energy.
A-4
-------
•
T
i
Barrick, Kenneth A. 1986. Option Value in Relation to Distance Effects and Selected User Characteristics for the
Washakie Wilderness, Northeast Wyoming, prepared by Intennountain Research Station, Ogden, UT for US
Forest Service, pp. 411 -422.
Barry, Gorgon. 1983. Law and Economics: Its Application to Air Quality Management in the United States, Israel
Ecological Society/et al Ecology & Environmental Qual.
Bartelmus, P., C. Stahmer, and J. van Tongeren. 1989. Integrated Environmental and Economic Accounting -
Framework for a SNA Sattelite System, Review of Income and Wealth.
Bartik, Timothy 1.1987. The Estimation of Demand Parameters in Hedonic Price Models, Journal of Political
Economy, Vol. 95, No. 1, pp. 81-88.
Bartik, Timothy J. 1987. Estimating Hedonic Demand Parameters Using Single-Market Data: The Problems Caused
by Unobserved Tastes, Review of Economics and Statistics, Vol. 69, No. 1, pp. 178-180.
Bartik, Timothy J. 1988. Measuring the Benefits of Amenity Improvements in Hedonic Price Models, Land
Economics, Vol. 64, pp. 172-183.
Bartik, Timothy J., and V. Kerry Smith. 1987. Urban Amenities and Public Policy, North Holland Publishing
Company, Amsterdam.
Bartik, Timothy 1.1988. Evaluating the Benefits of Non-Marginal Reductions in Pollution Using Information on
Defensive Expenditures, Journal of Environmental Economics and Management, Vol. IS, No. 1, pp. 111-127.
Bartik, Timothy J. 1988. The Effects of Environmental Regulation on Business Location in the United States,
Growth and Change, Vol. 19, No. 3. pp. 22-44.
Bartlett, R. V., and W. F. Baber. 1987. Matrix Organization Theory and Environmental Impact Analysis: A Fertile
Union?, Natural Resources Journal, Vol. 27. No. 3, pp. 605-615.
Bateman, I. J., K. C. Willis, G. Garrod, and P. Dofctor, et al. 1992. A Contingent Valuation Study of the Norfolk
Broads, prepared by Environmental Appraisal Group, University of East Anglia for National Rivers
Authority, England.
Batie, S. S. 1985. Economics: Nonpoinl Source Pollution Impacts, National Conference of Nonpoint Source
Pollution, pp. 229-231.
Batie, S. S., and J. R. Wilson. 1979. Economic Value Attributable to Virginia's Coastal Wetlands and Inputs in
Oyster Production, prepared by Department of Agricultural Economics for Virginia Polytech Institute,
Blacksburg, VA.
Bane, S. S., R. B. Jensen, and L. G. Hogue. 1976. A Lancasterian Approach for Specifying Derived Demands for
Recreational Activities. Southern Journal of Agricultural Economics. Vol. 8, pp. 101-107.
Batie. Sandra S.. and James R. Wilson. 1978. Economic Values Attributable to Virginia's Coastal Wetlands as Inputs
in Oyster Production, Southern Journal of Agricultural Economics, pp. 111-118.
Batie, Sandra S.. and Leonard A. Shabman. 1981 Estimating the Economic Value of Wetlands: Principles, Methods.
and Limitations, Coastal Zone Management Journal, VoL 10. No. 3, pp. 255-278.
Batie, Sandra, and Carl Mabbs-Zeno. 1985. Opportunity Costs of Preserving Coastal Wetlands: A Case Study of a
Recreational Housing Development, Land Economics. Vol. 61. No. 1.
Baumol, William J. 1980. Theory of Equity in Pricing for Resource Conservation, Journal of Environmental
Economics and Management, Vol. 7, pp. 308-320.
A-5
-------
Baumol, William I., and Wallace E Gates. 1988. The Theory of Environmental Policy (2nd Ed.), Cambridge
University Press, New York and Sydney.
Beanlands, G. E., and P. N. Dunker 1984. An Ecological Framework for Environmental Impact Assessment,
Journal of Environmental Management
Beavis, Brian, and Martin Walker. 1983. Acneiving Environmental Standards with Stochastic Discharges, Journal of
Environmental Economics and Management
Beggs, Cardell, Hausman. 1981. Assessing the Potential Demand for Electric Cars, Journal of Econometrics, Vol.
16, pp. 1-19.
Bell, Frederick W. 1983. Consumer Avooidance Damages and Legal Liability: Some Reports from the Kepone
Caper.
Bell, Frederick W. 1980. Commercial Fishing and Trapping: An Economic Analysis of the Atcbafalaya River Basin,
US Fish and Wildlife Service.
Bell, Frederick W. 1989. Application of Wetland Valuation Report Theory to Florida Fisheries, Honda Sea Grant
Program.
Bell, Frederick W. 1980. Recreational Benefits for the Atchafalaya River Basin, US Fish and Wildlife Service.
Bell, Frederick W. 1989. Main Quarry Hypothesis and Salmon Angling, Marine Resource Economics, Vol. 6, pp.
71-82.
Bell, Frederick W., Philip E. Sorenson, and Veraon R. Leeworthy. 1982. The Economic Impact and Valuation of
Saltwater Recreational Fisheries in Florida, Florida Sea Grant Program.
Bell, Frederick W., and Vernon R. Leeworthy. 1986. An Economic Analysis of the Importance of Saltwater Beaches
in Florida, Florida Sea Grant Program.
Bdl, Frederick W., and Vernon R. Leeworthy. 1990. Recreational Demand by Tourists for Saltwater Beach Days,
Journal of Environmental Economics and Management, Vol. 18, pp. 189-205.
Belzer, Richard B., and Albert L. Nichols. 1988. Economic Incentives to Encourage Hazardous Waste Minimization
and Safe Disposal prepared by Energy and Environmental Analysis, Kennedy School for US Environmental
Protection Agency.
Bennet J. W. 1984. Using Direct Questioning to Value the Existence Benefits of Preserved Natural Areas,
Australian Journal of Agricultural Economics, VoL 28, No. 2, pp. 136-152.
Berger. Mark C., and ETAL. 1987. Valuing Changes in Health Risks: A Comparison of Alternative Measures.,
Southern Economic Journal. Vol. 53, No. 4, pp. 967-984.
Bergland, Oliver, and Alan Randall. 1984. Operation Techniques for Calculating the Exact Hicksian Variations from
Observable Data, prepared by Department of Agricultural Economics for University of Kentucky.
Bergman, Lars. 1991. General Equilibrium Effects of Environmental Policy: A CGE-Modelling Approach,
Environmental and Resource Economics, VoL 1, No. 1, pp. 43*61.
Bergstrom, John C. 1990. Concepts and Measures of the Economic Value of Environmental Quality: A Review,
Journal of Environmental Management, Vol. 31, pp. 215-228.
Bergstrom, John C., B. L. DUman, and John R. Stoll. 1985. Public Environmental Amenity Benefits of Private Land:
The Case of Prime Agricultural Land, Southern Journal of Environmental Economics, Vol. 17, No. l.pp.
139-149.
A-6
-------
Bergstrom, John C, and H. Ken Cordell. 1990. An Analysis of the Demand for and Value of Outdoor Recitation in
the United States, University of Georgia.
Bergstrom, John C., and H. Ken Cordell. 1989. Household Market Demand and Supply Comparisons for Outdoor
Recreation, prepared by Southeastern Forest Experiment Station, Athens, GA
Bergstrom, John C., and John R. Stoll. 1986. Structure, Conduct and Performance in Contingent Markets, prepared
by Texas A&M University for Texas A&M University.
Bergstrom, John C., and John R. Stoll. 1989. Application of Experimental Economics Concepts and Precepts to
CVM Reid Survey Procedures, Western Journal of Economics, Vol. 14, No, 1, pp. 98-109.
Bergstrom. John C., John R. Stoll, and Alan Randall. 1990. The Impact of Information on Environmental
Commodity Valuation Decisions. Journal of Agricultural Economics Research, Vol. 72, No. 3, pp. 614-621.
Bergstrom, John C., John R. Stoll and Alan Randall. 1989. Information Effects in Contingent Markets. American
Journal of Agricultural Economics, Vol. 71, No. 3, pp. 685-691.
Bergstrom, John C., John R. Stoll, John P.Titre, and Vernon Wright 1990. Economic Value of Wetlands-Based
Recreation, Ecological Economics, Vol. 2, pp. 129-147.
Bergstrom, John C., and John R Stoll. 1987. A Test of Contingent Market Bid Elicitation Procedures for Piecewise
Valuation, Western Journal of Economics, Vol. 12, No. 2.
Bianchi, Dennis. 1969. The Economic Value of Streams for Fishing, prepared by Water Resources Institute for
University of Kentucky.
Billings, Bruce R., and Donald E. Agtbe. 1980. Price Elasticities for Water A Case of Increasing Block Rate, Land
Economics, Vol. 56, No. 1, pp. 73-84.
Bingham, Taylor, and Luanne Lohr. 1984. A Preliminary Assessment of the Benefits of Reducing Formaldehyde
Exposures - Draft Rpt. prepared by Research Triangle Institute, Inc. for US Environmental Protection Agency
(EARB), Research Triangle Institute, Research Triangle Park, NC.
Bingham, Taylor, Donald Anderson, and Phillip Cooley. 1987. Distribution of the Generation of Air Pollution,
Journal of Environmental Economics and Management, Vol. 14, No. 1, pp. 30-40.
Bmkley, Clark S., and W. Michael Hanemann. 1978. The Recreation Benefits of Water Quality Improvement:
Analysis of Day Trips in an Urban Setting, US Environmental Protection Agency.
Biosystems Analysis. 1984. Methods for Valuation of Environmental Costs and Benefits of Hydroelectric Facilities:
A Case Study of the Sultan River Project, The Office of Power and Resource Management, Oregon.
Bishop, R. C. 1978. Endangered Species and Uncertainty: The Economics of a Safe Minimum Standard, American
journal of Agricultural Economics, VoL 60, No. 1, pp. 10-18.
Bishop, R. C. 1979. Endangered Species, Irreversibflity and Uncertainty: A Reply. American Journal of Agricultural
Economics, Vol. 61, No. 2, pp. 376-379.
Bishop, R. C. 1981. Option Value and the Great Lakes: AFirst Assessment- Draft
Bishop, R. C. 1982. Option Value: An Exposition and Extension, Land Economics, Vol. 58, pp. 1-15.
Bishop, R. C. 1986. Resource Valuation under Uncertainty: Theoretical Principals for Empirical Research, JAI
Press, Inc., Greenwich. CT. pp. 133-158.
A-7
-------
Bishop, Richard C., and Kevin J. Boyle. 1985. The Economic Value of Illinois Beach State Nature Preserve,
prepared by Heberlein-Baumgartner.
Bishop, Richard C., Kevin J. Boyle, and Michael P. Welsh. 1987. Toward Total Economic Valuation of Great Lakes
Fishery Resources, American Fisheries Society, Vol. 116, pp. 339-345.
Bishop, Richard C., Kevin J. Boyle, Michael P. Welsh, and R. M. Baumgarmer. et al. 1987. Glen Canyon Dam
Releases and Downstream Recreation: An Analysis of User Preferences and Economic Values, US Bureau of
Reclamation, Salt Lake City, Utah.
Bishop, Richard C., and Thomas A. Heberlein. 1979. Measuring Values of Extramarket Goods: Are Indirect
Measures Biased?, American Journal of Agricultural Economics, Vol. 61, No. 5, pp. 926-930.
Bishop, Richard C., and Thomas A. Heberlein. 1980. Simulated Markets, Hypothetical Markets, and Travel Cost
Analysis: Alternative Methods of Estimating Outdoor Recreation Demand, prepared by Department of
Agricultural Economics for University of Wisconsin, Madison, Wl.
Bishop, Richard C.. and Thomas A. Heberlein. 1979. Travel Cost and Hypothetical Valuation of Outdoor
Recreation: Comparisons with a Hypothetical Market, prepared by Department of Agricultural Economics for
University of Wisconsin, Madison, Wl.
Bishop, Richard C., and Thomas A. Heberlein. 1984. Contingent Valuation Methods and Ecosystem Damages from
Acid Rain. Madison. Wl. No. 217.
Bishop, Richard C., and Thomas A. Heberlein. 1986. Does Contingent Valuation Work?. Rowan and Allanbeld,
Totowa,NJ.
Bishop, Richard C., Thomas A. Heberlein, Daniel W. McCoUum, and Michael P. Welse. 1988. A Validation
Experiment for Valuation Techniques, prepared by Department of Agricultural Economics for University of
Wisconsin.
Bishop, Richard, T. Heberlein, and Mary Jo Kealy. 1983. Contingent Valuation of Environmental Assets:
Comparisons with a Simulated Market, Natural Resources Journal.
Bisson, Brian T. 1986. Insurance Against Losing Groundwater Supplies From Contamination, Journal of the New
England Waterworks Association, Vol. 100, No. 2, pp. 157-161.
Biswaa H., and B. A. Bell. 1984. A Method for F.ttihlishing Site-Specific Design Flows for Wastelaod Allocation,
Journal of Water Pollution Control Federation.
Blackorby. C.. D. Donaldson, and D. Moloney. 1984. Consumer's Surplus and Welfare Change in a Simple Dynamic
Model Review of Economic Studies.
Blake-Hedges, Lynne, K. E. McCoonell, and 1. E Strand. 1990. Modelling Catch Rates in the RUM Framework,
University of Maryland.
Blarney, Russell K. 1991. Contingent Valuation and Fraser Island.
Blank, Frederic M.. David Brookshire, T. Crocker, and Ralph D' Arge, et al. 1977. Valuation of Aesthetic
Preferences: A Case Study of the Economic Value of Visibility, Electric Power Institute, University of
Wyoming, Resource & Environ Economics Lab.
Biasing, T. J, S. B. McLaughlin, and L. K. Mann. 1983. Effects of Acid Rain and Gaseous Pollutants on Forest
Productivity: A Regional Scale Approach, Journal of Air Pollution Control Association.
Bkanquist, Glenn C.. Mark C. Berger, and John P. Hoebn. 1988. New Estimates of Quality of Life in Urban Areas,
American Economic Review, Vol. 78, No. 1, pp. 89-107.
A-8
-------
I
Blumenfeld, Karen, and Teresa M. Lynch. 1989. EPA's Economic Analyses and Regulatory Decision Making,
prepared by Alliance Technologies Corporation.
Bockstael, Nancy E., and Catherine L. Kling. 1988. Valuing Environmental Quality: Weak Complementarity with
Sets of Goods, American Journal of Agricultural Economics, Vol. 70, No. 3, pp. 654-662.
Bockstael, Nancy E., and I. E. Strand. 1985. Distribution Issues and Non-Market Benefit Valuation, Western Journal
of Agricultural Economics, Vol. 10, pp. 162-169.
Bockstael, Nancy E., and I. E. Strand. 1987. The Effect of Common Sources of Regression Error on Benefit
Estimates, Land Economics, Vol. 63. pp. 11-20.
Bockstael, Nancy E., 1. E. Strand, and W. Michael Hanemann. 1987. Time and the Recreational Demand Model,
American Journal of Agricultural Economics, Vol. 69, No. 2, pp. 293-302.
Bockstael, Nancy E., I. E. Strand, and W. Michael Hanemann. 1984. Time and Income Constraints in Recreation
Demand Analysis, University of Maryland.
Bockstael, Nancy E., and K. E. McConnell, 1981. Theory and Estimation of the Household Production Function for
Wildlife Recreation, Journal of Environmental Economics and Management, Vol. 8, pp. 199-214.
Bockstael, Nancy E., and K. E. McConnell. 1980. Calculating Equivalent and Compensating Variation for Natural
Resource Facilities, Land Economics, Vol. 56, No. 1, pp. 56-63.
Bockstael, Nancy E., and K. E. McConnell. 1988. Welfare Effects of Changes in Quality: A Synthesis. University at
Maryland.
Bockstael, Nancy £., and K. E. McConnell. 1983. Welfare Measurement in the Household Production Function
Framework, American Economic Review, Vol. 73, pp. 806-814.
Bockstael, Nancy E., and K. E. McConnell. 1984. Implicit Market Methods for Benefit Estimation - Vol. I, prepared
by University of Maryland for US Environmental Protection Agency (EARB).
Bockstael, Nancy E., Kenneth McConnell, and Ivar E. Strand. 1987. Benefits from Improvements in Chesapeake
Bay Water Quality - VoLHJ, prepared by University of Maryland for US Environmental Protection Agency
(EARB), US Government Printing Office, Washington, DC.
Bockstael Nancy E., Kenneth McConnell and Ivar E. Strand. 1989. Measuring the Benefits of Improvements in
Water Quality: Hie Chesapeake Bay. Marine Resource Economics, Vol. 6. No. 1, pp. 1-18.
Bockstael Nancy E., W. M. Hanemann, and I. E. Strand, Jr. 1984. Measuring the Benefits of Water Quality
Improvements Using Recreation Demand Models - Vol. H, prepared by University of Maryland for US
Environmental Protection Agency (EARB).
Bockstael Nancy E., W. Michael Hanemann, and Catherine L. Kling. 1985. Modeling Recreational Demand in a
Multiple Site Framework, Boulder, CO.
Bockstael Nancy E., W. Michael Hanemann, and Catherine L. Kling. 1987. An Evaluation of Multiple Site Model
Development, University of Maryland, Water Resources Research, Vol. 3, pp. 951-960.
Bockstael Nancy E., W. Michael Hanqmmm and Catherine L. Kling. 1987. Estimating the Value of Water Quality
Improvements in a Recreational Demand Framework, Water Resources Research, Vol. 23, No. 5, pp. 951-
960.
Bohm, P. 197 1 . An Approach to die Problem of Estimating Demand for Public Goods, Swedish Journal of
Economics, Vol. 73, No. 1, pp. 94-105.
A-9
-------
Bohm, P. 1971. Option Demand and Consumer Surplus: Comment, Quarterly Journal of Economics, Vol. 85, pp.
523-527.
Bohm, Peter. 1984. Revealing Demand for an Actual Public Good, Journal of Public Economics, Vol. 24, No. 2, pp.
135451.
Bohm, P. 1991. Deposit-refund Systems: Theory and Application to Environmental, Conservation, and Consumer
Policy, prepared by Resources for the Future for Johns Hopkins Press.
Bohm, Peter, and Clifford S. Russell. 1985. Comparitive Analysis of Alternative Policy Instruments, ed. Kneese,
Allen V. and Sweeney, James L., North-Holland, Amsterdam, pp. 395-460.
Boland, James. 1986. Economic Instruments for Environmental Protection in the United States, prepared by OECD
for Paris.
Boaneville Power Adm. 1986. Estimating Environmental Costs and Benefits for Five Generating Resources,
BonnevUle Power Administration, Oregon.
Booneville Power Adm. 1987. Generic Coal Study: Qualification and Valuation of Environmental Impacts,
Bonneville Power Administration. Oregon.
Boaneville Power Adm. 1984. Environmental Costs and Benefits Case Study: Nuclear Power Plant, prepared by
Nero and Associates, Inc. for Bonneville Power Administration, Oregon.
Botmicksen, T. M., and R. H. Becker. 1983. Environmental Impact Studies: An Interdisciplinary Approach for
Assigning Priorities, Environmental Management.
Bord, Richard J., Donald J. Epp, and Robert E. O'Connor. 1989. Achieving Greater Consistency Between Subjective
~" * .prepared by Pennsylvania State University for US Environmental Protection Agency
and Objective Risks,]
(EARB).
VErfVItD;.
Bouwes, Nicolaas W., and R. Schneider. 1979. Procedures in Estimating Benefits of Water Quality Change,
American Journal of Agricultural Economics. Vol. 61, No. 3. pp. 535-539.
Bower, Blair, and ETAL. 1981. Incentives for Water Quality Management: France and the Ruhr Area, prepared by
Resources for the Future for Washington, D.C.
Bowers. J. K_ 1983. Cost-Benefit Analysis of Wetland Drainage, Environment and Planning, Vol. 15, No. 2.
Bowers, J. K. 1988. Cost-Benefit Analysis in Theory and Practice: Agricultural Land Drainage Projects.
Bowes, M., sod J. B. Loomis. 1980. A Note on the Use of Travel Cost Methods with Unequal Zonal Populations,
Land Economics, Vol. 56, No. 4, pp. 465-470.
Bowes. Michael D. 1982. A Note on the Use of Travel Cost Models with Unequal Zonal Populations: Reply, Land
Economics, Vol. 58, No. 3, pp. 408-410.
Bowes, Michael D.. and John V. KrutiUa. 1989. Multiple-Use Management: The Economics of Public Forestlands,
prepared by Resources for the Future for pp. 177-247.
Bowker, J. M., and John R. Stoll. 1988. Use of Dichotomous Choice Nonmarket Methods to Value the Whooping
Crane Resource, American Journal of Agricultural Economics, Vol. 70, No. 2. pp. 372-381.
Boyce, Steve. 1990. An Analysis of Travel Costs in Discrete Choice Travel Cost Models, University of California at
Cam TWon
San Diego.
A-10
-------
I
Boyd, Robert, and Noel D. Uri. 1991. Hie Cost of Improving the Quality of the Environment, Journal of Policy
Modeling, Vol. 13, No. 1, pp. 114-140.
Boyle, Kevin. 1989. Commodity Specification and Framing Contingent Valuation Questions, Land Economics, Vol.
65, No. 1, pp. 57-63.
Boyle, Kevin. 1985. Essays on the Valuation of Non-market Resources: Conceptual .Issues and Empirical Case
Studies, University of Wisconsin.
Boyle, Kevin J., Michael P. Welsh, and Richard C. Bishop. 1988. Using Scenarios of Unexperienced Environmental
Condition in Contingent Valuation Studies.
Boyle, Kevin J., Michael P. Welsh, and Richard C. Bishop. 1988. Validity of Welfare Measures of Welfare Change:
Comment, Land Economics, Vol. 64, No. 1, pp. 94-98.
Boyle, Kevin J., Michael P. Welsh, and Richard C. Bishop. 1985. Starting Point Bias in Contingent Valuation
Bidding Games, University of Maine, Land Economics, Vol. 61, No. 2, pp. 188-194.
Boyle, Kevin J., Michael P. Welsh, and Richard C. Bishop. 1990. Contingent Valuation of Unexperienced
Environmental Conditions, University of Maine, Journal of Environmental Economics and Management.
Boyle, Kevin J., and Richard C. Bishop. 1987. Valuing Wildlife in Benefit-Cost Analyses: A Case Study Involving
Endangered Species, University of Maine, Water Resources Research, Vol. 23, No. 5, pp. 943-950.
Boyle. Kevin J., and Richard C. Bishop. 1988. Welfare Evaluations Using Contingent Valuation: A Comparison of
Techniques, American Journal of Agricultural Economics, Vol. 70, No. 1, pp. 20-28.
Boyle, Kevin J., and Richard C. Bishop. 1979. Toward the Total Valuation of Great Lakes Fishery Resources, Water
Resources Research, Vol. 5, pp. 943-990.
Boyle. Kevin J., and Richard C. Bishop. 1984. Economic Benefits Associated with Boating and Canoeing on the
Lower Wisconsin River, prepared by Department of Agricultural Economics for University of Wisconsin.
Boyle, Kevin J., and Richard C. Bishop. 1985. The Total Value of Wildlife Resources: Conceptual and Empirical
Issues.
Boyle, Kevin J., Stephen D. Reiling. and Marcia L. Phillips. 1990. Species Substitution and Question Sequencing in
Contingent Valuation Surveys Evaluating the Hunting of Several Types of Wildlife. University of Maine,
Leisure Science. Vol. 12, pp. 103.
Boyle, Kevin J., and Trisb Heekin. 1989. Benefits and Costs in Natural Resource Planning, prepared by Western
Regional Research.
Braccio, Ralph, Dan Pyne, Jean Tilly, and B. Hendricks. 1988. Results of Ac Preliminary Analysis of the Proposed
Second Third Land Disposal Restrictions Rule, prepared by ICF, inc.
Braden, John B., and Randolph M. Lyon. 1985. Pollution Control Cost Analysis, prepared by University of Illinois
for US Environmental Protection Agency (EARB), Pollution.
Braden, John B.. and Charles D. Kolstad. 1991. Measuring the Demand for Environmental Quality. North-Holland.
Bradford, D. P.. and G. G. HikJebrandL 1977. Observable Preferences for Public Goods, Journal of Political
Economy, Vol. 8. pp. 111-132.
Brady, G. L., and R. E. Morrison. 1984. Emissions Trading: An Overview of the EPA Policy Statement,
International Journal of Environmental Studies.
A-ll
-------
Brajer, Victor, Jane V. Hall, and Robert Rowe. 1991. The Value of Cleaner Air: An Integrated Approach,
Contemporary Policy Issues, Vol. 9, pp. 81-90.
Bratton, David M., and Gary L. Rutledge. 1990. Pollution Abaitment and Control Expenditures, 1985-88, Survey of
Current Business, pp. 32-38.
Breen, Barry. 1989. Citizen Suits for Natural Resource Damages: Dosing a Gap in Federal Environmental Law,
Wake Forest Law Review. Vol. 24, pp. 851-880.
Bromley, D. 1982. Land and Water Problems, American Journal of Agricultural Economics.
Bromley, David. 1986. Natural Resource Economics: Policy Problems and Contemporary Analysis, Kluwcr-Nijboff
Publishing, Boston, MA.
Brooks, R. 0.1979.The Economic Impacts of the Wild and Scenic Rivers Act on the Middle Fork of the Clearwater
River, University of Idaho.
Brooks, Robert. 1988. The Net Economic Value of Deer Hunting in Montana, prepared by Montana Department of
Fish, Wildlife & Parks for Montana Department of Fish, Wildlife and Parks, Helena, MT.
Brookshire, D. S., A. Randall, and J. Stoll. 1977. Methodological Experiments in Valuing Wildlife Resources: Phase
I - Interim Report to the U.S. Fish and Wildlife Service, US Fish and Wildlife Service.
Brookshire. D. S., and A. Randall. Economic Valuation of Wildlife, US Fish and Wildlife Service, University of
Wyoming, Resource & Environ Economics Lab.
Brooksbire, D. S.. and A. J. Randall. 1978. Public Policy Alternatives, Public Goods and Contingent Valuation
Mechanism, Western Economics Association, Hawaii.
Brookshire, D. S., B. C. Ives, and W. D. Schulze. 1976. The Valuation of Aesthetic Preferences, Journal of
Environmental Economics and Management. Vol. 3, No. 4, pp. 325-346.
Brookshire, D. S., M. A. Tbayer. W. D. Scbulze. and R. C d1 Arge. 1982. Valuing Public Goods: A Comparison of
Survey and Hedonk Approaches, American Economic Review, Vol. 72. No. 1. pp. 165-177.
Brookshire. D. S.. R. D'Arge, W. Schulze, and M. A. Thayer. 1979. Methods Development for Assessing Trade-offs
in Environmental Management, US Environmental Protection Ageucy. EPA, Washington, DC.
Brookshire. D. S., Rah* C. D'Arge, William Scbulze, and Mark Thayer. 1982. Experiments in Valuing Public
Goods. In Advances in Applied Microeconomics, ed. Smith, V. K., JAI Press, Inc., Greenwich, CT, pp. 123-
172.
Brookshire, D. S.. Ronald G. Comings. Morteza PptmmHan, and William D. Schulze, et al. 1982. Experimental
Approaches for Valuing Environmental Commodities.
Brookshire, D. S.. and T. Crocker. 1981. The Advantages of Contingent Valuation Methods for Benefit Cost
Analysis. Public Choice, VoL 36, No. 2, pp. 235-252.
Brookshire, D. S., and T. Crocker. 1979. The Use of Survey bu
s in Determining the Economic Valuation of
Environmental Goods, prepared by Rocky Mountain Forest and Range Land Experiment Station for US
Forest Service, Fort Collins, CO.
Brookshire. David S.. Alan Randall, and John R. Stoll. 1980. Valuing Increments and Decrements in Natural
Resource Service Flows, American Journal of Agricultural Economics, Vol. 62. No. 3, pp. 478-488.
Brookshire, David S.. Larry Eubanks, and Cindy Sorg. 1986. Existence Values and Normative Economics:
Implications for Valuing Water Resources. Water Resources Research, VoL 22. No. 11, pp. 1509-1518.
A-12
-------
i
T
Brookshire, David S., Larry S. Eubanks, and Alan Randall. 1983. Estimating Option Prices and Existence Values for
Wildlife Resources, Land Economics, Vol. 59, No. 1, pp. 1-15.
Brookshire, David S., and V. Kerry Smith. 1987. Measuring Recreation Benefits: Conceptual and Empirical Issues,
Water Resources Research, Vol. 23, No. 5. pp. 931-935.
Brooksbire, D. S., M.A. Tnayer, J. Tscherhart, and WD. Scbulze. 1985. A Test of the Expected Utility Model:
Evidence from Earthquakes, Journal of Political Economy, Vol. 93, No. 2, pp. 369-389.
Brookshire, David D., and Don L. Coursey. 1987. Measuring the Value of a Public Good: An Empirical Comparison
of Elicitation Procedures, American Economic Review, Vol. 77, No. 4, pp. 554-566.
Brown, G. M.f and H. O. Pollakowski. 1977. Economic Valuation of Shoreline, Review of Economics and Statistics,
Vol. 59, pp. 272-278.
Brown, Gardner Jr., and Jon H. Goldstein. 1984. A Model for Valuing Endangered Species, Journal of
Environmental Economics and Management, Vol. 11, No. 4, pp. 303-309.
Brown, Gardner, Jr., and Ralph Johnson. 1984. Pollution Control by Effluent Charges: It Works in Germany, Why
Not the U.S.?, Natural Resources Journal, Vol. 24, No. 4, pp. 929-966.
Brown, Gardner, Jr., J. John Charbonneau, and Michael J. Hay. 1978. The Value of Wildlife Estimated by the
Hedonic Approach, US Fish and Wildlife Service, Washington, DC.
Brown. Gardner, Jr., J. John Charbonneau, and Michael J. Hay. 1979. Estimating the Values of Wildlife: Analysis of
the 1975 Hunting and Fishing Survey, prepared by Division of Program Planning for US Fish and Wildlife
Service, Washington, DC.
Brown, Gardner, Jr., and Mark Plummer. 1979. Recreation Valuation: An Analysis of the Nontimber Uses of
Forestland in the Pacific Northwest, Washington State University.
Brown, Gardner, Jr.. and Michael J. Hay. 1987. Net Economic Recreation Values for Deer and Waterfowl Hunting
and Trout Fishing, 1980, prepared by USFWS Division of Policy and Directives Management for US Fish
and Wildlife Service.
Brown, Gardner, Jr., and Robert Mendelsohn. 1984. The Hedonic Travel Cost Method, University of Washington,
Department of Economics, Review of Economics and Statistics, Vol. 66, No. 3, pp. 427-433.
Brown, W. G., C. Sorhus, B. Chou-Yang, and J. Richards. 1983. Using Individual Observations to Estimate
Recreation Demand Functions: A Caution. American Journal of Agricultural Economics, Vol. 65, No. 1. pp.
154-157.
Brown, William, Colin Sorhus, and Kenneth Gibbs. 1980. Estimated Expenditure by Sport Anglers and Net
Economic Values of Salmon and Steelhead for Specified Fisheries in the Pacific Northwest, prepared by
Department of Economics for Oregon State University.
Brown, William, D. M. Larson, R. S. Johnston, and R. J. Wanle. 1979. Improved Economics Evaluation of
Commercially and Soon Caught Salmon and Steelbead of the Columbia River, prepared by Department of
Economics for Oregon State University.
Brown, William, and F. Nawas. 1975. Impact of Aggregation on the Estimation of Outdoor Recreation Demand
Function, American Journal of Agricultural Economics. Vol. 55. pp. 246-249.
Brown, William, F. Nawas, and Joe Stevens. 1973. The Oregon Big-Game Resource: An Economic Evaluation,
prepared by Agricultural Experimental Station for Oregon State University.
A-13
-------
Brown, James N., and Harvey S. Rosen. 1982. On the Estimation of Structural Hedpnic Price Models,
Econometrica, Vol. 50, No. 3, pp. 765-768.
Buc, Lawrence, and Curtis Haymore. 1983 Regulating Hazardous Waste Incinerators Under the Resource
Conservation and Recovery Act, Natural Resources Journal.
Buerger, Robert, and James R. Kahn. 1989. New York Value of Chesapeake Striped Bass, Marine Resource
Economics, Vol. 6, pp. 19-25.
Burness, H., R. Cummings, A. Mehr, and M. Walben. 1983. Valuing Policies which Reduce Environmental Risk,
Natural Resources Journal.
Burrows, T. 1983. Pesticide Demand and Integrated Pest Management: A Limited DependentVariable Analysis,
American Journal of Agricultural Economics.
Bun, O., and D. Brewer. 1971. Estimates of Net Social Benefits for Outdoor Recreation, Econometrica, Vol. 39, No.
5, pp. 813-827.
CEQ. 1989. Environmental Trends.
Cabelli, V. J.. A. P. Dufour, L. J. McCabe, and M. A. Levin. 1983. A Marine Recreational Water Quality Criterion
Consistent with Indicator Concepts and Risk Analysis, Journal of Water Pollution Control Federation.
Cameron, Trudy A. 1989. Contingent Valuation Assessment of The Economic Damages of Pollution to Marine
Recreational Fishing, prepared by University of California at Los Angeles for US Environmental Protection
Agency.
Cameron, Trudy Ann. 1989. The Effects of Variations in Gamefish Abundance on Texas Recreational Fishing
Demand: Welfare Estimates, prepared by University of California at Los Angeles for US Environmental
Protection Agency.
aron, Trudy Ann. 1988. A New Paradigm for Valuing Non-market Goods Using Referendum Data: Maximum
Likelihood Estimation by Censored Logistic Regression, Journal of Environmental Economics and
Management. Vol. 15, No. 3, pp. 353-379.
Cameron, Trudy Ann, and Daniel Huppen. 1989. OLS versus ML Estimation of Non-Market Resource Values with
Payment Card Interval Data, University of California at Los Angeles, Journal of Environmental Economics
and Management, Vol. 17. No. 3, pp. 230-246.
Cameron, Trudy Ann, and M. D. James. 1987. Efficient Estimation Methods for "Closed-ended' Contingent
Valuation Surveys, Review of Fcononrics and Statistics. VoL 69, No. 2. pp. 269-276.
Cameron, Trudy Aim, and M. D. James. 1987. The Determinants of Value for * Recreational Fishing Day: Estimates
from a Contingent Valuation Survey, prepared by O""1""* Government for University of California at Los
Angeles.
(^pbell,TbomasA,,aixlRaiidaUA.Luthi. 1991-Nann^ Another
Perspective, Natural Resources Law Newletter, Vol. 22, No. 4, pp. 1-6.
Carlson. Cynthia. 1988. Making CERGLA Natural Resource Damage Regulations Work The Use of the Public
Trust Doctrine and Other State Pfm*qfi Environmental Law Reporter, Vol. 18, No. 8, pp. 10299-307.
Carlson, J. Lou, G. Johnson, and T. Ukn. 1984. Economic Analysis of Illinois' New Hazardous Waste Law, Natural
Resources Journal.
A-14
-------
f
Carruthers, Carrey, Frank A. Ward, Mary Libbin, and Edgardo Kalaw. 1981. A Review of Models for Assessing the
Economic Impact of Alternative Water and Sponfishery Strategies, prepared by New Mexico Game and Fish
Department.
Carson, Richard T., and Robert Cameron Mitchell. The Value of Diamonds and Water, University of California at
San Diego and Clark University.
Carson, Richard T., J. Horowitz, and M. Machina. 1987. Discounting Mortality Risks Paper ALSO: Discounting
Statistical Lives (March 1988), prepared by University of California at San Diego for US Environmental
Protection Agency.
Carson, Richard T., and Robert Cameron Mitchell. 1984. Tlie Value of Clean Water The Public's Willingness to
Pay for Boatable, Fishable, and Swimmable Quality Water.
Carson, Richard T., and Robert Cameron Mitchell. 1987. Economic Value of Reliable Water Supplies for
Residential Water Users in the State Water Project Service Area, prepared by Resources for the Future for
Metropolitan Water District of Southern California.
Carson, Richard, and Peter Navarro. 1988. Fundamental Issues in Natural Resource Damage Assessment, Natural
Resources Journal, Vol. 28, No. 4, pp. 815-836.
Caswell, Margriet F., and K. E. McConnell. 1980. Simultaneous Estimation of Jointly Dependent Participation
Functions, Journal of Environmental Economics and Management, Vol. 7, No. 1, pp. 65-73.
Caudill, James, and John P. Hoehn. 1989. A Benefit Cost Framework for Groundwater Contamination Risks.
CanUctns. Peter P. 1982. An Empirical Study of the Recreational Benefits Generated by a Water Quality
Improvement, University of Wisconsin.
Caulkins, Peter P., Richard C. Bishop, and Nicolaas W. Bouwes. 1985. Omitted Cross-Price Variable Biases in the
Linear Travel Cost Model: Correcting Common Misperceptions, Land Economics, Vol. 61. No. 2, pp. 182-
187.
Caulkins, Peter P., Richard C. Bishop, and Nicolaas W. Bouwes. 1984. A Comparision of Two Travel Cost Models
for Valuing Lake Recreation, University of Wisconsin, Dept of Agricultural Econ, Madison, WI.
Caulkins, Peter P., Richard C. Bishop, and Nicolaas W. Bouwes, Sr. 1986. The Travel Cost Model for Lake
Recreation: A Comparison of Two Methods for Increasing Site Quality and Substitution Effects, American
Journal of Agricultural Economics, VoL 68, No. 2, pp. 291-297.
Cave, Jonathan A. K. 1988. Age. Time, and the Measurement of Mortality Benefits, prepared by Rand Corporation
for US Environmental Protection Agency.
Cerda, A. A., and R. M. Adams. 1990. Estimating Recreational Demand and Benefits from Salmon and Steelbead
Sport Fishing in the Central and Northeast Regions of Oregon, Santiago, Cb"?
Cesario, F. J. 1980. Congestion and the Valuation of Recreation Benefits, Land Economics. Vol. 56, pp. 329-338.
Cesario, F. J. 1976. Value of Time in Recreation Benefit Studies. Land Economics, Vol. 52, pp. 32-41.
Cesario, F. I., and J. F. Knetsch. 1970. Tune Bias in Recreation Benefits Estimation Models. Water Resources
Research, Vol. 6, pp. 700-704.
Chappie, M., and L. Lave. 1982. The Health Effects of Air Pollution: A Reanalysis, Journal of Urban Economics.
Vol. 12, pp. 346-376.
A-15
-------
Charbonneau, J. John, and Micbael Hay. 1979. A Hedonic Approach to Recreation Values: A Waterfowl Hunting
Application.
Charbonneau, J. John, and Michael Hay. 1978. Determinants and Economic Values of Hunting and Fishing,
Transactions of the North American Wildlife and Natural..., Vol. 43, pp. 391-403.
Chase, David £., and Perry M. Lindstrom. 1986. Compensation for Medical Expenses and Lost Wages of the
Chronically HI, prepared by American Petroleum Institute.
Chesapeake Exec Cod. 1989. Chesapeake Bay Oyster Management Plan - Agreement Committee Report.
Chess, Caron, and Billie Jo Hance. 1988. Alerting the Apathetic and Reassuring me Alarmed: Communicating
About Radon Risk in three Communities, prepared by Rutgers University for US Environmental Protection
Agency (EARB).
Chestnut. Lauraine G., and Daniel M. Vtolette. 1990. The Relevance of Willingness-to-Pay Estimates of the Value
of a Statistical Life in Determining Wrongful Death Awards, prepared by EPA Economic Analysis and
Research Branch for US Environmental Protection Agency.
Chestnut, Lauraine G., and Daniel M. Violent. 1984. Estimates of Willingness to Pay for Pollution-Induced
Changes in Morbidity: A Critque for Benefit-Cost Analysis of Pollution Regulation, prepared by Energy and
Resource Consultants, Inc. for US Environmental Protection Agency (OPA).
Chestnut, Lauraine G.. R.D. Rowe, J. J. MacKenzie.andM. T.El-Ashry. 1989. Economic Measures of the Impacts
of Air Pollution on Health and Visibility, Yale University, pp. 316-342.
Chestnut, Lauraine G.. S. Colome, L. R. Keller, and W. Lambert, et al. 1988. Heart Disease Patients' Averting
Behavior, Costs of Illness, and Willingness to Pay to Avoid Angina Episodes - Final Report, prepared by
University of California at Irvine; RCG/Hagler. BaUly, Inc. for US Environmental Protection Agency (OPA),
TTC rt^...«CT.m»nt Prinrino Office. Washington. DC.
University 01 uauimiua *u. u ••»>»., .-»^_, „ ,
US Government Printing Office, Washington, DC.
Chestnut, Lauraine G., and Daniel M. Violette. 1986. Estimates of Willingness to Pay for Pollution-Induced
Changes in Morbidity: A Critique for Benefit-Cost Analysis of Pollution RegulationENVIRONMENTAL
BENEFITS ANALYSIS SERIES append, prepared by EPA Economic Analysis and Research Branch for US
Environmental Protection Agency.
Christensen, Barbara. 1985. Can Pollution Contaminate Value?. Real Estate Appraiser & Analyst. Vol. S3. No. 3,
pp. 53-55.
Ciccneni. C. I., A. C. Fisher, and V. K. Smith. An Econometric Evaluation of a Generalized Consumer Surplus
Measure: Mineral King Controversy, Econometric^ Vol. 44, pp. 1259*1276.
Cicchetti, C. J., and V. K, Smith. 1973. Congestion, Quality Deterioration and Optimal Use: Wilderness Recreation
in toe Spanish Peaks Primitive Area, Social Science Research. Vol. 2, pp. 15-30.
Cicbetti, C .J., and A. M. Freeman. 1971. Option Demand for Dnsumer Swpliis: Fiurn^ Oxnment, Quarterly
Journal of Economics, Vol. 85. pp. 528-539.
Cicbetti, Charles J, and V. K. Smith. 1976. The Cost of Congestion, Ballinger Publishing Company. Cambridge,
MA.
Coiacy-Wantrup, S. V., and Richard Bishop. 1985. ^Common Property' as a Concept in Natural Resources Policy,
Natural Resources Journal, Vol. 15, No. 4, pp. 713-727.
Clark, David E., and James R. Kahn. 1989. The Two-stage Hedonic Approach: A Methodology for the Valuation of
Environmnetal Amenities, Journal of Environmental Economics and Management, Vol. 16, No. 2, pp 106-
120.
A-16
-------
•
r
Clark, Lyman H. 1988. The Impact of Forthcoming Environmental Regulations upon Small Businesses.
Clark, Lyman H, 1988. The Small Business Sector Study: Impacts of Environmental Regulations on Small
Businesses, US Environmental Protection Agency.
Clarke, Harry C., and W. J. Reed. 1990. Land Develoipmenl and Wilderness Conservation Policies Under
Uncertainty: A Synthesis, Natural Resources Modelling, Vol. 4, No. 4, pp. 11-37.
Clawson, M., and J. L. Knetsch. 1966. Economics of Outdoor Recreation, Resources for the Future, Washington,
DC.
Clawson, Marion. 1959. Methods of Measuring the Demand for and the Value of Outdoor Recreation, Resources for
the Future, Washington, DC.
Cocbeba, Donald J., and William A. Langford. 1981. Direct Willingness to Pay Questions: An Analysts of their Use
for Quantitatively Valuing Wildlife, Journal of Leisure Research, Vol. 13, pp. 311-322.
Cocheba, Donald J., and William A. Langford. 1978. Wildlife Valuation: The Collective Good Aspect of Hunting,
Land Economics, Vol. 54, No. 4, pp. 490-504.
Cochran, H., and P. Huszar. 1984. Economics of Timing Storm Drainage Improvements, Water Resources Research.
Cohen, Bernard. 1983. Probabilistic Risk Assessment of Wastes Buried in the Ground, Risk Analysis.
Colby, Bonnie G. 1989. Alternative Approaches to Valuing Water Rights. Appraisal Journal pp. 180-1%.
Colby, Bonnie G., T. Sargent, and H. H. Gary. 1988. Market Clearing Prices and Market Clearing Price Plus
Consumer Surplus Values for Water Flowing Through National Forests, prepared by Forest Trust, Santa Fe.
NM for US Forest Service.
Colby, Michael E. 1989. The Evolution of Paradigms of Environmental Management in Development, prepared by
World Bank.
Coilinge, R., and W. E. Oates. 1982. Efficiency in Pollution Control in the Short and Long Runs: A System of
Rental Emission Permits, ranadfap Journal of Economics.
Coilinge, Robert, and M. Bailey. 1983. Optimal Quasi-Market Choice in the Presence of Pollution Externalities,
Journal of Environmental Economics and Management.
Common, M. S. 1983. Implications of Discounting Under Varying Technological and Prefernece Regimes,
International Journal of Environmental Studies.
Commons. M.S. 1973. A Note on the Use of the Clawson Method for the Evaluation of Recreation Sight Benefits,
Regional Studies, Vol. 7. pp. 401-406.
Connelly, Nancy A., and Tommy L. Brown. 1988. Fsrimatrs of Nonconsumptive Wildlife Use on Forest Service and
BLM Lands, prepared by USFS Wildlife & Fisheries Management Staff for US Forest Service.
Conrad, J. M. 1980. Quasi-Option Value and the Expected Value of Information, Quarterly Journal of Economics,
Vol. 94. pp. 813-820.
Conservation Fnd. 1984. State of the Environment An Assessment it Mid-Decade (Risk Assessment and Risk
Control), prepared by Conservation Foundation for Conservation Foundation Report.
Cook, Zena L., Allen R. Ferguson, and Julia I. Leighton. 1984. An Economic Assessment of the Benefits of the
Proposed Effluent Limitation Guidelines for Organics and Plastics Manufacturers, prepared by EPA Office of
A-17
-------
Policy Analysis for US Environmental Protection Agency, Public Interest Economics Foundation,
Washington, DC.
Cooper, Joseph, and John Loomis. 1990. Pooled Time-Series Cross-Section Travel Cost Models: Testing Whether
Recreation Behavior is Stable over Time, Leisure Science. Vol. 12, pp. 161-171.
Cooper, M. Harris. 1984. The Integrative Research Review: A Systematic Approach, Sage Publications, Beverly
Hills, CA.
Cordell, H. Ken. 1984. Criteria for Outdoor Recreation Pricing Policies.
Cordell, H. Ken. 1981. Pricing for Allocating Low Density Recreational Use Between Private and Commercial
Users of Natural Areas.
Cordell, H. Ken, and John C. Bergstrom. 1989. Theory and Techniques for Assessing the Demand and Supply of
Outdoor Recreation in the United States, prepared by USES South-East Forest Experimental Station for US
Forest Service.
Cordell, H. Ken, John C. Bergstrom, Donald B.K. English, and Deborah Klinko, et al. 1990. Economic Value and
Impacts of Outdoor Recreation at Western North Carolina Lakes under Different Management Alternatives,
prepared by USFS South-East Forest Experimental Station for US Forest Service.
Cordray, David S. 1987. Strengthening Causal Interpretations of Non-Experimental Data: The Role of Meta-
Analysis, Sage Publications, Menlo Park, CA.
Comes, Richard, and Todd Sandier. 1986. The Theory of Externalities, Public Goods, and dub Goods, Cambridge
University Press, New York and Sydney.
Cory, D. C, and B. C. Satiba. 1987. Requiem for Option Value, Land Economics, Vol. 63, No. 1, pp. 1-10.
Cory, Dennis C. 1988. Uncertain Recreation Quality and Wildlife Valuation: Are Conventional Benefit Measures
Adequate?, Western Journal of Agricultural Economics, Vol. 13, No. 2, pp. 153-162.
Cory, Dennis C., and J. Dauben. 1981. Evaluating Direct Benefit Estimation Procedures for Inagation Projects: A
Simplified Approach, Water Resources Research, Vol. 17. No. 6, pp. 1066-1070.
Cory, Dennis C. and William E. Martin. 198S. Valuing Wildlife for Efficient Multiple Use: Elk Versus Cattle,
Western Journal of Agricultural Economics, Vol. 10, No. 2. pp. 282-293.
Costanza. Robert, and Charles Perrings. 1990. A Flexible AssumiaBowlmgSysiem for Improved Eovironmental
Management. Ecological Economics, Vol. 2, pp. 57-75.
f>flpiCTtt JhAfst and Stephen T Faitiar IQltt Thanriat Mill MfeihotU of Valuation of Natural Systems! A
Comparison "f Wi""ignfSMo-Pay and *M*ffgy Analyst P>«*fl Approaches, Man, Environment, Space and
Time, Vol. 4. No. 1. pp. 1-38.
Costanza. Robert, Steven Farber, and Judith Maxwell. 1989. Valuation and Management of Wetland Ecosystems,
Ecological Economics, Vol. 1, pp. 335-361.
Coulson, A. H, N. Duan, E. Korn, and W. G. Manning. 1985. Effects of Air Pollution on Health Outcomes,
prepared by Rand Corporation for US Environmental Protection Agency.
Coulson, Amy, Mark Dickie, Shelby Gating, and William Schulze. 1985. Improving Accuracy and Reducing Costs
of Environmental Benefit Assessments, Volume HI: Fff*«"qatfifE me Benefits of Reducing Community Low-
Level Ozone Exposure. Draft, prepared by University of Colorado for US Environmental Protection Agency
(OPA), University of Colorado. Center for Economic Analysis, Boulder, CO.
A-18
-------
1
Courant, P. N., and R. C. Porter. 1981. Averting Expenditures and the Cost of Pollution, Journal of Environmental
Economics and Management, Vol. 8, pp. 321-329.
Coursey, Donald L., John L. Hovis, and William D. Schulze. 1987. The Disparity Between Willingness to Accept
and Willingness to Pay Measures of Value, Quarterly Journal of Economics, Vol. 102, pp. 679-690.
Coursey, Donald L., and William D. Schulze. 1986. The Application of Laboratory Experimental Economics to the
Contingent Economics to the Contingent Valuation of Public Goods. Public Choice. Vol. 49, No. 1, pp. 47-
68.
Cowan, J. H., R. E. Turner, and D. R. Cahoon. 1988. Marsh Management Plans in Practice: Do They Work in
Coastal Louisiana, USA?, Environmental Management, Vol. 12, No. 1.
Crandall, R. W.. T. E. Keeler, and L. B. Lave. 1982. The Cost of Automobile Safety and Emissions Regulations to
the Consumer: Some Preliminary Results, American Economic Review.
Crandall, Robert W. 1981. Pollution Controls and Productivity Growth in Basic Industries. In Productivity
Measurement in Regulated Industries, ed. Cowing, T. and Stevenson, R., Academic Press, New York. pp.
347-368.
Crocker, T. D., and R. L. Horst Jr. 1977. Oxidam Air Pollution and Work Performance of Citrus Harvest Labor, US
Environmental Protection Agency.
Crocker, T. D., and R. L. Horst Jr. 1981. Hours of Work, Labor Productivity, and Environmental Conditions: A Case
Study, Review of Economics and Statistics, Vol. 43, No. 3.
Crocker, Thomas D. 1985. On the Value of the Condition of a Forest Stock, Land Economics, Vol. 61, No. 3, pp.
244-254 p.
Crocker, Thomas D. 1989. On the Value of the Condition of a Forest Stock: Reply, Land Economics, Vol. 65, No. 1,
pp. 73-75.
Crocker, Thomas D., and Henry J. Vaux, Jr. 1983. Some Economic Consequences of Ambient Oxidant Impacts on a
National Forest, prepared by EPA Economic Analysis and Research Branch for US Environmental Protection
Agency.
Crocker, Thomas D., and Ronald G. Cummings. 1983. On Valuing Air-Pollution induced Materials Damages: A
Methodological Inquiry, prepared by EPA Office of Policy Analysis for US Environmental Protection
Agency, University of Wyoming; University of New Mexico.
Croke, Kevin. 1986. Estimating the Value of Natural Open Space Preservation in an Urban Area, Environmental
Management, Vol. 23, No. 4, pp. 317-324.
Crompton, J. L. 1984. How to Establish a Price for Park and Recreation Services. Trends: User Fees and Charges.
Vol. 21, No. 4. pp. 12-21.
Cronin, Francis J. 1982. Valueing Non-Market Goods Through Contingent Markets, Battelle Memorial Institute.
Cropper, M. L. 1988. A Note on the Extinction of Renewable Resources, Journal of Environmental Economics and
Management, Vol. 15, No. 1, pp. 64-70.
s: A Survey. Journal of Economic Literature,
Cropper. M. L., and Wallace E. Oates. 1992. Environmental E
Vol. 30. No. 2, pp. 675-740.
Cropper, M. L., and Amalia Arriaga-Salinas 1980. Inter-City Wage Differentials and the Value of Air Quality,
Journal of Urban Economics, Vol. 8, No. 2, pp. 236-254.
A-19
-------
Cross, Frank B. 1989. Natural Resource Damage Valuation, VanderbUt Law Review. Vol. 42, No. 2, pp. 269-341.
Crosson, Bent. 1990. Supplying the Environmental Values of Agriculture, prepared by Resources for the Future for
Washington. DC. Resources, Vol. 98.
Crutcbfield. James A., and Kurt Scbelle. 1978. An Economic Analysis of Washington Ocean Recreational Salmon
Fishing, prepared by Department of Economics for University of Washington.
Cuddington, John T., F. Reed Johnson, and Jack L. Knetsch. 1981. Valuing Amenity Resources in the Presence of
Substitutes, Land Economics, Vol. 57, pp. 526*535.
Cullen, Ross. 1985. Rationing Recreation Use of Public Land, Journal of Environmental Management, Vol. 21, No.
3, pp. 213-224.
Cummings, Ronald G., David S. Brookshire. and William Scbulze. 1986. Valuing Environmental Goods: An
Assessment of the Contingent Valuation Method, Rowan and AUanheld, Totowa, NJ.
Cuds, T. D., and F. W. Shows. 1984. A Comparitive Study of Social Economic Benefits of Artificial Beach
Nourishment - Civil Works in Northeast Florida, prepared by Florida Department of Natural Resources.
Cuds, T. D., and F. W. Shows. 1982. Economic and Social Benefits of Artificial Beach Nourishment - Civil Works
at Debay Beach, prepared by Florida Department of Natural Resources.
D'Arge, R. C., W. D. Schulze. and D. S. Scbulze. 1982. Carbon Dioxide and Intergenerational Choice, American
Economic Review.
D'Arge, Ralph. 1985. Environmental Quality Benefits Research For me Next Five Years: Some Observations and
Recommendations, University of Wyoming, Resource & Environ Economics Lab.
D'Arge, Ralph C. and Jason Snogren. 1985. Water Quality Benefits: Analysis of the Lakes at Okoboji, Iowa,
prepared by University of Wyoming for US Environmental Protection Agency (EARB).
ion. Journal of Leisure
Daniels, Steven E. 1987. Marginal Cost Pricing and Efficient Provision of Public R
Research. Vol. 19, No. 1, pp. 22-34.
Daniels, Steven E. 1985. Methods of Outdoor Recreation Demand Analysis, prepared by USFS Soutb-East Forest
Experimental Station for US Forest Service.
Daniels, Steven E. 1986. Efficient Provision of Recreational Opportunines: The Case of U.S. Forest Service
Campgrounds, Duke University.
Danielson, Leon E. 1977. Estimation of Residential Water Demand, North Carolina University at Raleigh.
Data Decision Group. 1988. Southeast Alaska Recreation Values, Southeast Alaska Marketing Council.
Daubert. J. T* and R. A. Young. 1981. Recreational Demands for Maintaining Instream Flows: A Contingent
Valuation Approach, American Journal of Agricultural Economics. Vol. 63, No. 4. pp. 666-676.
David, E L. 1968. Lake Shore Property Values: A Guide to Public uvestment in Recreation, Water Resources
Research, Vol. 4, No. 4. pp. 697-707.
David, Martin, and FJi/abcth David. 1983. Cost-Effective Regulatory Options for Water Quality Limited Streams.
Water Resources Bulletin.
Davidson, Paul. F. Gerard Adams, and Joseph Seneca. 1966. Toe Social Value of Water Recreational Facilities
Resulting from an Improvement in Water Quality: toe Delaware Estuary. In Water Research, ed. Kneese, A.
and Smith. S.. Johns Hopkins University Press. Baltimore. MD.
A-20
-------
l
f
Davies, J. C. 1983. TSCA's Impact on Society and Chemical Industry (Overall Costs and Benefits), ACS
Symposium Series.
Davis, Robert K. 1963. The Value of Outdoor Recreation: An Economic Study of the Maine Wooods, Harvard
University Press, Cambridge, MA.
Davis, Robert K.. and Diane Lim. 1987. On Measuring the Economic Value of Wildlife, Westview Press, Boulder,
CO.
Dawson, G. W. 1983. Risk Management and the Landfill in Hazardous Waste Disposal, Journal of Hazardous
Materials.
Decker, Daniel J., and Gary R. Goff. 1987. Valuing Wildlife: Economic and Social Perspectives, Westview Press,
Boulder, CO.
Delvin, Patrick J. 1985. Quantifying Recreation: Allocation of Value to the Collection and Burning of Firewood in
Colorado, Colorado State University.
Dendrinos, D. S., and H. Mullally. 1983. Optimum Control in Nonlinear Ecological Dynamics of Metropolitan
Areas, Environment and Planning.
Dennis, R., T. Stewart. P. Middleton, and M. Downton, et al. 1983. Integration of Technical and Value Issues in Air
Quality Policy Formulation: A Case Study, Socio-Economic Planning Sciences.
Desvousges, W., and K. Pate. 1982. Alternative Methods for Estimating Recreation and Related Benefits of Water
Quality Improvement: A Case Studt of the Monongabela River, Research Triangle Institute, Research
Triangle Park, NC.
Desvousges, W. H., R. W. Dunford, and J. L. Domanico. 1989. Measuring Natural Resource Damages: An
Economic Appraisal, prepared by Research Triangle Institute, Inc. for American Petroleum Industry.
Desvousges, William H. 198S. The Varying Parameter Model: In Perspective. Association of Environmental and
Resource Economists.
Desvousges, William H., and V. Kerry Smith. 1983. The Benefits of Hazardous Waste Management Regulations -
Technical Report, prepared by Research Triangle Institute; University of North Carolina for US
Environmental Protection Agency (EARB).
Desvousges. William H.. and V. Kerry Smith. 1984. The Travel Cost Approach for Valuing Improved Water
Quality: Additional Considerations • Draft Report, prepared by Research Triangle Institute, Inc. for US
Environmental Protection Agency (EARB).
Desvousges, William H., and V. Kerry Smith. 1983. The Benefits of Hazardous Waste Management Regulations,
prepared by Research Triangle Institute, Inc. for US Environmental Protection Agency.
Desvousges, William H., and V. Kerry Smith. 1983. Benefit-Cost Assessment Handbook for Water Programs,
Volume I (draft), prepared by Research Triangle Institute, Inc.
Desvousges, William H.. V. Kerry Smith, and Matthew P. McGivney. 1983. A Comparision of Alternative
Approaches for Estimating Recreation and Related Benefits of Water Quality Improvements, prepared by
Research Triangle Institute, Inc. for US Environmental Protection Agency (EARB). Research Triangle
Institute, Research Triangle Park. NC.
Desvousges, William, V. Kerry Smith, and Ann Fisher. 1985. Option Price Estimates for Water Quality
Improvements: A Contingent Valuation Study for the Monongahela River, US Environmental Protection
Agency, Research Triangle Institute. Research Triangle Park. NC.
A-21
-------
Desvousges. William, V. Kerry Smith, and Ann Fisher. 1987. Option Price Estimates for Water Quality
Improvements: A Contingent Valuation Study for the Monongahela River, Journal of Environmental
Economics and Management, Vol. 14, No. 3, pp. 248-267.
Desvousges, William H., and V. Kerry Smith. 1983. An Overview: The Benefits of Hazardous Waste Management
Regulations, prepared by RTI for US EPA.
Desvousges, William H., F. Reed Johnson, Richard W. Dunford, and Kevin J. Boyle, etal. 1992. Measuring Nonuse
Damages Using Contingent Valuation: An Experimental Evaluation of Accuracy, prepared by Research
Triangle Institute for Exxon.
Dewees, Donald. 1982. Cost of Risk Avoidance. Royal Society of Canada/et al Risk Assessment & Perception S.
Deyak, T. A., and V. K. Smith. 1978. Congestion and Participation in Outdoor Recreation: A Household Production
Function Approach, Journal of Environmental Economics and Management. Vol. 5, pp. 63-80.
Deyak, Timothy A,, and V. Kerry Smith. 1978. Residential Property Values and Air Pollution: Some New Evidence.
Quarterly Review of Economics and Business, Vol. 14. No. 4, pp. 93-100.
Dickens, William. 1984. Differences Between Risk Premiums in Union and Non-Union Wages and the Case for
Occupational Safety Regulations, American Economic Review.
Dickie. M., S. Gerking, D. Brookshire, and D. Coursey, et al. 1987. Reconciling Averting Behavior and Contingent
Valuation Benefit Estimates of Reducing Symptoms of Ozone Exposure - Draft, prepared by University of
Chicago for US Environmental Protection Agency (OPPE), Zappa.
Dickie. M., S. Gerking, W. Scbulze, and A. Coulson, et al. 1986. Value of Symptoms of Ozone Exposure: An
Application of the Averting Behavior method, prepared by EPA Economic Analysis and Research Branch for
US Environmental Protection Agency.
Dickie. Mark, Ann Fisher, and Shelby Gerking. 1987. Market Transactions and Hypothetical Demand Data: A
Comperitive Study, Journal of the American Statistical Association, Vol. 82, pp. 68-75.
Dickie, Mark, and Shelby Gerking. 1989. Benefits of Reduced Morbidity from Air Pollution Control. In Valuation
Methods and Policy Making in Environmental Economics, ed. Folmer, H. and Van ledand, E., Elsevier
Science Publishers, Amsterdam, pp. 105-122.
Dickie, Mark, Shelby Gerking, and Mark Agee. 1990. Stratospheric Ozone Depletion, Skin Damage Risks, and
Protective Action. (Draft), US Environmental Protection Agency.
Dinan, Terry, Craig Simons, and Roger Lloyd. 1988. Agricultural Sector Study: Impacts of Environmental
Regulations on Agriculture, prepared by EPA Office of Policy Analysis for US Environmental Protection
Agency.
Dixoo, John A., and M. M. Hufschmidt. 1986. Economic Valuation Techniques for the Environment: A Case Study
Workbook. Johns Hopkins University Press, Baltimore, MD.
Domestic Policy CncL 1987. Benefit Analysis of Sulfur Dioxide Emission Control Strategies.
Donagi, A. E* J. R. Goldsmith, A. L Goren, and R. Toeplitz. 1983. Epidemiological Monitoring near a Coal-Fired
Power Plant in Israel, Journal of Air Pollution Control Association.
Donnelly. Dennis M., John B. Loomis, Cindy F. Sorg, and Lotus J. Nelson. 1983. The Net Economic Value of
Recreation Steelbead Fishing in Idaho, prepared by Rocky Mountain Forest and Range Land Experiment
Station for US Forest Service. Fort Collins, CO.
A-22
-------
f
Donnelly, Dennis M., and Louis I. Nelson. 1986. Net Economic Value of Deer Hunting in Idaho, prepared by Rocky
Mountain Forest and Range Land Experiment Station for US Forest Service.
Dooley, Peter. 1984. Consumer's Surplus: Marshall and His Critics, Canadian Journal of Agricultural Economics.
Dower, Roger C., and Paul F. Scodari. 1987. Compensation for Natural Resource Injury: An Emerging Federal
Framework, Marine Resource Economics, Vol. 4, No. 3, pp. 155-174.
Dower, Roger C., T. Henderson, M. J. Marvin, and P. Reed, et al. 1983. A Technical Analysis of Economic-Based
Regulation in Environmental Rulemaking, prepared by Environmental Law Institute for US Environmental
Protection Agency (EARS).
Dower, Roger, and Paul R. Portney. 1984. Enhancing the Efficiency of Environmental Regulation, prepared by
Environmental Law Institute; Resources for the Future.
Dowlatabadi, Hadi. and Winston Harrington. 1989. The Effects of Uncertainty on Policy Instruments: The Case of
Electricity Supply and Environmental Regulations, prepared by Resources for the Future for Resources for
the Future, Washington, D.C.
Downing, T. E., and R. W. Kates. 1982. The International Response to the Threat of Cblorofluorocarbons to the
Atmospheric Ozone, American Economic Review.
Doyle, J. K., S. R. Elliot, and P. A. Locke. 1990. Economics and Psychology Policy Research for Environmental
Management: An Evaluation of Strategies for Solving the Radon Problem, (draft), prepared by University of
Colorado; Environmental Law Institute for US Environmental Protection Agency.
Dragun, Andy K. 1991. An Economic Study of Commercial and Recreational Fishing Conflicts in Port Phillip Bay
and Western Port. Victoria, Australia, prepared by Department of Legal Studies, La Trobe University for
Victorian Department of Conservation and Environment.
Dresser, Richard C., Stuart L. Fribush, and William M. Mendez. 1983. Regulatory Impact Analysis for New
Chemical Reporting Alternatives under Section 5 of TSCA, prepared by ICF. Inc. for US Environmental
Protection Agency.
Driver, B. L. 1985. Specifying What is Produced by Management of Wildlife by Public Agencies, Leisure Sciences,
Vol. 7, No. 3, pp. 281-295.
DuMouchel, William, and I. Harris. 1983. Bayesian Methods for Combining the Results of Cancer Studies in
Humans and Other Species, Journal of American Statistical Association.
Duan, N., T. Hayashi, A. H. Carlson, and E. Keeler. et al. 1987. Sbort-Term Health Effects of Air Pollution,
prepared by Rand Corporation for US Environmental Protection Agency, Air.
Duffield, John H. 1984. Travel Cost and Contingent Valuation: A Comparative Analysis, JAI Press, Inc.,
Greenwich, CT.
Duffield. John H. 1989. RPA Values for Recreation: Theory and Practice, Public Land Law Review, Vol. 10, pp.
105-129.
Duffield, John H. 1988. The Net Economic Value of Elk Hunting in Montana, prepared by Montana Department of
Fish, Wildlife &. Parks for Montana Department of Fish, Wildlife and Parks. Helena, MT.
Duffield, John W., and David Patterson. 1991. Inference and Optimal Design for a Welfare Measure in Dicbotomous
Choice Contingent Valuation. Land Economics, Vol. 67, No. 2, pp. 225-239.
A-23
-------
Duffield, John W., and S. Allen. 1988. Contingent Valuation of Montana Trout Fishing by River and by Angler
Subgroup, prepared by Montana Department of Fish, Wildlife & Parks for Montana Department of Fish,
Wildlife and Parks, Helena, MT.
Duffield, John, John Loomis, Rob Brookes, and Joe HoUiman, et al. 1987. The Net Economic Value of Fishing in
Montana, University of Montana, Missoula, MT.
Dumsday, R. G., K. M. Jakobsson, and S. Ransome. 1991. Applying Contingent Valuation Estimates to Statewide
Assessment of a Proposal to Protect River Segments in Victoria. Victorian Department of Conservation and
Environment.
Burden, Garey, and Jason F. Shogren. 1988. Valuing Non-market Recreation Goods: An Evaluative Study of the
Literature on the Travel Cost and Contingent Valuation Methods, Review of Regional Studies, Vol. 18, No. 3,
pp. 1-15.
Durojaiye. Bamidele O., and Anthony E. Ipke. 1988. The Monetary Value of Recreational Facilities in a Developing
Economy: A Case Study of Three Centers in Nigeria, Natural Resources Journal. Vol. 28, No. 2, pp. 31S-328.
•
Dutta, Nilima. 1984. The Value of Recreational Boating and Fishing in the Central Basin of Ohio's Portion of Lake
Erie, prepared by Sea Grant Program for Ohio State University, Ohio Slate University.
Dutton, Ronald A., and Robert D. Rowe. 1984. The Value of Groondwater Protection, US Environmental Protection
Agency.
Dwyer, Jf, J. R. Kelly, and M. D. Bowes. 1977. Improved Procedures for the Valuation of the Contribution of
Recreation to National Economic Development, University of Illinois, Champaigti-Urbana, IL.
Dwyer, John F., George L. Peterson, and Alexander J. Darragh. 1983. Estimating Value of Urban Forests Using the
Travel Cost Method, Journal of Arboriculture. Vol. 9. pp. 182-185.
Dwyer. John, and Michael D. Bowes. 1978. Concepts of Value for Marine Recreational Fishing. American Journal
of Agricultural Economics, Vol. 60. No. 5. pp. 1008-1012.
Dyson, Pamela J. 1984. Recreational Water Availability in the United States: The Impact of Pollution Control,
prepared by Resources for toe Future.
EPRI. 1984. Review of Proprosed Dry-Storage Concepts Using Probabilistic Risk A
Easter, K. William, Keith Koztoff, Theodore GnJum-Tomasi, and Ian Moore, etal. 1990. An Evaluation of Options
for Micro-Targeting Acquisition of Cropping Rights to Reduce Nonpoint Source Water Pollution, prepared
by EPA Office of Policy, Planning and Evaluation for US Environmental Protection Agency, University of
Minnesota.
Edwards. Steven. 1988. Option Prices for Gioundwater Protection, prepared by Woods Hole for Journal of
Environmental Economics and Management. Vol. IS. No. 4. pp. 475-487.
Edwards, Steven F. 1989. On Estimating the Demand for Outdoor Recreation from Property Values: An
Exploration. Northeastern Journal of Agricultural and Resource Economics. Vol. 18, No. 2, pp. 140-148.
Edwards, Steven F. 1990. An Economics Guide to Allocation of Fish Stocks Between Commercial and Recreational
Fisheries, National Oceanographic m$ Atmospheric Administration.
Edwards, Steven F. 1984. An Analysis of me Net Economic Benefits of Protecting Salt Pond Water Quality in
Southern Rhode Island, University of Rhode Island.
Edwards, Steven F., and Glen D. Anderson. 1987. Overlooked Biases in Contingent Valuation Surveys: Some
Considerations, Land Economics. Vol. 63, No. 2, pp. 168-178.
A-24
-------
Eikmeyer, Leo. 1985. Pollution: An Effect on Value, Canadian Appraiser, Vol. 29, No. 1, pp. 21-23.
Ellis, F., K. Klima, R. Wycoff, and L. McKay. 1984. Integrating Water Quality and Construction Grants
Management, Journal of Water Pollution Control Federation.
Ellsaesser, H. W. 1982. Should We Trust Models or Observations?, Atmospheric Environment.
Englin, Jeffrey, and Robert Mendelsohn. 1991. A Hedonic Travel Cost Analysis for Valuation of Multiple
Components of Site Quality: The Recreation Value of Forest Management, prepared by Battelle Pacific
Northwest Laboratories, Portland, Oregon for US Forest Service, Academic Press, Inc.. New York, NY,
Journal of Environmental Economics and Management, Vol. 21, pp. 1-16.
Ess, Terry, and C. S. Shih. 1982. Perspectives of Risk Assessment for Uncontrolled Hazardous Waste Sites, EPAfct
al Mgmt of Uncontrolled Hazardous Waste Sites Nau'ona.
Eubanks, L. S., and D. Brookshire. 1980. Household Production and Non-Market Valuation, AERE.
Eubanks, L. S., and D. S. Brookshire. 1981. Methods for Valuing Non-Market Resources: Theoretical and Empirical
Compahsions, Econometric Society.
Evans, David S. 1985. An Analysis of the Differential Impact of EPA and OSHA Regulations across Finn and
Establisment Sizes in The Manufacturing Industries, prepared by CERA.
Farber, Stephen. 1990. Final Assurance, Risk and Underground Injection of Hazardous Waste: Levelling the Playing
Field.
Farber, Stephen C. 1988. The Value of Coastal Wetlands for Recreation: An Application of Travel Cost and
Contingent Valuation Methododlogies, Journal of Environmental Management, Vol. 26, No. 4, pp. 299-312.
Farber, Stephen C. 1987. The Value of Coastal Wetlands for the Protection of Property Against Hurricane Wind
Damage, Journal of Environmental Economics and Management, Vol. 14, No. 2, pp. 143-151.
Farber, Stephen, and Robert Costanza. 1987. The Economic Value of Wetland Systems, Journal of Environmental
Management, Vol. 24, No. 1, pp. 41-51.
Farawoitb, E. G. 1983. Synthesis of Ecological and Economic Theory Towards More Complete Valuation of
Tropical Moist Forests, International Journal of Environmental Studies.
Farnworth, Edward G., Thomas H. Tidrick, Webb M. Smathers, and Carl F. Jordan. 1983. A Synthesis of Ecological
and Economic Theory Toward More Complete Valuation of Tropical Moist Forests, Journal of
Environmental Studies, Vol. 21, pp. 11-28.
Farrow, Scott. 1990. Modelling me Importance of Oceans art Ettoraries, prepared by Caraegie Melton for Woods
Hole Oceanographic Institution. Marine Policy Center.
Farzin, Y. Hossein. 1984. The Effect of the Discount Rate on Depletion of Exhaustible Resources, Journal of
Political Economy.
Federal Register. 1984. Environmental Protection rules; Tuesday, March 20,1984.
Fedkiw, J. 1987. Coming Back to Market Value and Valuation for the Great Lakes Fisheries, Transactions of the
American Fisheries Society.
Feenberg, Daniel, and Edwin S. Mills. 1980. Measuring the Benefits of Water Pollution Abatement,
Press, Inc., New York, NY.
A-25
-------
Fcinennan, £., and K. Knapp. 1984. Benefits from Groundwater Management: Magnitude, Sensitivity, and
Distribution, American Journal of Agricultural Economics.
Fesenmaier, D.R. 1985. Modeling Variation in Destination Patronage for Outdoor Recreation Activity, Journal of
Travel Research, Vol. 23, No. 2, pp. 17-23.
Findeis, J. L., and E.L Micbalson. 1984. The Demand for and Value of Outdoor Recreation in UK Targhee National
Forest, Idaho, prepared by Idaho Agricultural Experimentation Station, College of Agr for University of
Idaho, Moscow, ID.
Findlater, P. A., and I. A. Sinden. 1982. Estimation of Recreation Benefits from Measured Utitlity Functions,
American Journal of Agricultural Economics, Vol. 64, pp. 102-109.
Findley, M., V. Rose, G. Cutter, and R. Windsor. 1983. An Assessment of EPA's Asbestos Hazard Evaluation
Algorithm, American Journal of Public Health.
Fiore, John, and Frank A. Ward. 1987. Managing Recreational Water Resources to Increase Economic Benefits to
Anglers in the Arid Southwest, prepared by Ag Expermnt Slam, College of Ag & Home Economics for New
Mexico State University, Las Cruces, NM.
Fischer, A., J. Krutilla, and C. Cicbetti. The Economics of Environmental Protection: A Theoretical and Empirical
Analysis.
Fischer, Anthony, M. Hannemann, and J. Harte. 1986. Economic Valuation of Aquatic Ecosystems, US
Environmental Protection Agency.
Fiscboff, Banco, and Lite Furby. 1988. Measuring Values: A Conceptual Framework for Interpreting Transactions
with Special Reference to Contingent Valuation Visibility, Journal of Risk and Uncertainty, Vol. 1, No. 2, pp.
147-184.
Fischoff, Baruch, Paul Skmc, and Sarah Lichtenstein. 1980. Knowing What You Want: Measuring Labile Values,
Lawrence Erlbaum Associates. Hillsdale, NJ.
Fisher, A., L. G. Chestnut, and D. M. Violette. 1989. The Value of Reducing Risks of Death: A Note on New
Evidence, Journal of Policy Analysis and Management, Vol. 8, No. 1, pp. 88-100.
Fisher, A. C., and J. V. Krutilla. 1972. Determination of Optimal Capacity of Resource-Based Recreation Facilities,
Natural Resources Journal, Vol. 12, pp. 417-444.
Fisher. A. C., and W. M. Hanemann. 1990. Option Value: Theory and Measurement, European Review of
Agricultural Economics, Vol. 17. No. 2, pp. 167-180.
Fisher. A. C, and W.M.H
1986. Option Value and the Extinction of Species, pp. 169-190.
Fisher, A. C.. and W. M. Hanemann. 1987. Quasi-Option Value: Some Misconceptions Dispelled, Journal of
Environmental Economics and Management, VoL 14, No. 2, pp. 183-190.
Fisher, A. C., and W. M. Haveonan. 1984. Endangered Species: The Economics of Irreversible Damage.
Fisher, Ann, Lauraine Chestnut, and Daniel Violette. 1989. The Value of Reducing Risks of Death: A Note on New
Evidence, Journal of Policy Analysis and Management. VoL 8, No, 1.
Fisher. Ann, and Robert Raucber. 1983. Intrinsic Benefits of Improved Water Quality: Conceptual and Empirical
Perspectives, prepared by EPA Office of Policy Analysis for US Environmental Protection Agency.
Fisher, Warren. 1982. Travel Cost and Contingent Value Estimates Explored, prepared by Eastern Economics
Association for Washington, DC.
A-26
-------
f
Forster, B. A. 1983. Pollution Variability and the Shape of the Dose-Response Curve, Journal of Air Pollution
Control Association.
Forster, Bruce. 1984. The Backward Incidence of Pollution Control: A Dual Approach, Journal of Environmental
Economics and Management.
Foster. Henry S., and Bruce R. Beanie. 1979. Urban Residential Demand for Water in the United States, Land
Economics, Vol. 55. pp. 475^87.
Fraas, Arthur G., and Vincent G. Munley. 1989. Economic Objectives within a Bureaucratic Decision Process:
Setting Pollution Control Requirements under the Clean Water Act, Journal of Environmental Economics and
Management, Vol. 17, No. 1, pp. 35-53.
Fraas, Arthur, and Vincent Munley. 1984. Municipal Wastewater Treatment Cost, Journal of Environmental
Economics and Management
Freedman, S. M., and D. B. Rosenberg. 1984. An Ecological Succession Model Applied to Environmental
Management, International Journal of Environmental Studies.
Freeman, A. Myrick. 1984. The Sign and Size of Option Value, Land Economics.
Freeman, A. Myrick. 1981. Notes on Defining and Measuring Existence Values, Bowdoin College, Department of
Economics.
Freeman, A. Myrick. 1982. On Measuring Public Good Demands from Market Data. In Advances in
Microeconomics, ed. Smith, V. K., JAI Press, Inc., Greenwich, CT, pp. 13-29.
Freeman, A. Myrick. 1979. The Economics of Environmental Improvement, Johns Hopkins University Press,
Baltimore, MD.
Freeman, A. Myrick. 1979. Approaches to Measuring Public Good Demands, American Journal of Agricultural
Economics, Vol. 61, pp. 915-920.
Freeman, A. Myrick. 1979. Hedonic Prices, Property Values and Measuring Environmental Benefits: A Survey of
the Issues, Scandinavian Journal of Economics, VoL 81, No. 2, pp. 154-173.
Freeman, A. Myrick. 1974. Air Pollution and Property Values: A Methodological Comment, Review of Economics
and Statistics, Vol. 53, No. 4, pp. 415-416.
Freeman, A. Myrick. 1974. Air Pollution and Property Values: A Further Comment, Review of Economics and
Statistics, Vol. 56, No. 4, pp. 554-556.
Freeman, A. Myrick. 1974. On Estimating Air Pollution Control Benefits from Land Value Studies, Journal of
Environmental Economics and Management, Vol. 1, No. 1, pp. 74-83.
Freeman, A. Myrick. 1979. The Benefits of Air and Water Pollution Control: A Review and Synthesis of Recent
Frjflimatp*
Freeman, A. Myrick. 1985. The Sign and Size of Option Value: Reply, Land Economics.
Freeman, A. Myrick. 1984. The Quasi-Option Value of Irreversible Development, Journal of Environmental
Economics and Management, Vol. 11, No. 3.
Freeman, A. Myrick. 1984. Depletable Externalities and Pigouvian Taxation, Journal of Environmental Economics
and Management, Vol. 11, No. 2, pp. 1973-1979.
A-27
-------
Freeman, A. Myrick. 1991. Valuing Environmental Resources under Alternative Management Regimes, Ecological
Economics, Vol. 3, pp. 247-256.
Freeman, A. M., and R. H. Haveman, 1977. Congestion Quality Deterioration and Heterogeneous Tastes, Journal of
Public Economics, Vol. 8, pp. 225-232.
Freeman, m, A. Myrick, and Winston Harrington. 1990. Measuring Welfare Values of Productivity Changes,
Southern Economic Journal, Vol. 56, No. 4, pp. 892-904.
Freeman, A. Myrick. 1992. The Economics of Valuing Marine Recreation: A Review of the Empirical Evidence, US
EPA.
Bench. Michael T., and Lisa D. McNeilly. 1987. Economic Impact Analysis for the Dry Cleaning NESHAP - Draft,
prepared by Research Triangle Institute, Inc.
Fritz, W. R., and S. C. Helle. 1979. Cypress Wetlands for Tertiary Treatment, prepared by EPA Office of Water for
US Environmental Protection Agency.
Furby, Lita. 1988. Specifying Subjective Evaluations • A critique of Dickie et al.'s interpretation of their contingent
valuation results for reduced minor health symptoms, prepared by Eugene Research Institute for US
Environmental Protection Agency (EARB).
Gaffey, William. 1984. The Methodology of Risk Assessment. Risk Analysis.
Gallop, Frank, and Mark Roberts. 1983. Environmental Regulations and Productivity Growth: The Case of Fossil-
Fueled Electric Power Generation. Journal of Political Economy.
Gamble. H.B., and RJi. Downing. 1981. Effects of the Accident at Three Mile Island on Residential Property
Values and Sales, prepared by Institute for Research on Land and Water Res. (Penn State) for Nuclear
Regulatory Commission.
Garcia. Philip, Bruce L. Dixon, James W. Mjelde, and Richard M. Adams. 1986. Measuring the Benefits of
Environmental Change Using a Duality Approach: The Case of Ozone and Illinois Cash Grain Farms, Journal
of Environmental Economics and Management, Vol. 13, No. 1. pp. 69-80.
Gardner. R. L,, and R. A. Young. 1985. An Economic Evaluation of the Colorado River Basin Salinity Control
Program, Western Journal of Agricultural Economics, Vol. 10. No. 1, pp. 1-12.
Garifio, Susan. 1984. Distortions to Travel Cost Derived Demand Curves: Water Quality and Durable Goods,
prepared by Department of Agricultural and Resource Economists for Oregon State University.
Coking, Shelby D., and Linda R. Stanley. 1986. An Economic Analysis of Air Pollution and Health: The Case of
St Louis, Review of Economics and Statistics. VoL 68. No. 1. pp. 115-121.
Gating, Shelby D. Linda R. Stanley, and William N. Wetrick. 1983. An Economic Analysis of Air Pollution and
Health: The Case of St. Louis, prepared by EPA Office of Policy Analysis for US Environmental Protection
Agency, University of Wyoming.
Gianessi, Leonard P., Henry M. Peskin, and Edward N. Wolff. 1979. The Distributfonal Effects of Uniform Air
Pollution Policy in the United States, Quarterly Journal of Economics. VoL 93, No. 2. pp. 281-301.
Gianessi. Leonard P., and Henry M. Peskin. 1980. The Distribution of the Q»ts of Federal Water Pollution Control
Policy, Land Economics, Vol. 56. No. 1, pp. 85-102.
Gibbons, Diana. 1986. The Economic Value of Water, prepared by Resources for the Future for Washington, DC.
A-28
-------
I
I
Gibbs, K. C., and J. P. Conner. 1973. Components of Outdoor Recreational Values: Kissimmee River Basin, Florida,
Southern Journal of Agricultural Economics, Vol. 25, pp. 239-244.
Gibbs, Kenneth C. 1978. Price Variable in Residential Water Demand Models. Water Resources Research, Vol. 14.
No. 1, pp. 15-18.
Gilbert, Alphonse H., Daniel W. McCollum, and George L. Peterson. 1988. A Comparison of Valuation Models
Using Cross-Country Skiing Data From Colorado and Vermont, prepared by Rocky Mountain Forest and
Range Land Experiment Station for US Forest Service, Fort Collins, CO.
Gilbert, C., and V. Kerry Smith. 1985. Role of Economic Adjustment for Environmental Benefits Analysis,
Econometric Society Meetings, New York.
Gilliard. M. W. 1984. A Conceptual Framework for Environmental Protection. Environmental Management,
Gisser, Micha. 1983. Groundwater: Focusing on the Real Issue, Journal of Political Economy.
Glasure, Y. U. 1987. An Evaluation of the Florida Saltwater Resident Recreational Fishery. Florida State University.
Gordon, Peter, and Harry W. Richardson. 1983. Intercommunity Distribution of Tax and Land Value Impacts of Air
Pollution, Urban Ecology, Vol. 7, No. 2, pp. 125-135.
Graham, D. A. 1981. Benefit Cost Analysis under Conditions of Uncertainty, American Economic Review, VoL 71,
pp. 715-725.
Graham, Daniel. 1984. Cost-Benefit Analysis under Uncertainty: Reply, American Economic Review.
Graham-Tomasi, Theodore. 1986. The Economic Value of Damages to the Pigeon River, Michigan, Due to an
Abnormal Discharge of Sediment from the Lansing Club Dam. prepared by School of Natural Resources for
University of Michigan.
Graves, Phil, J. Murdoch, M. Thayer, and D. Waldman. 1985. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments - Vol. VI Improving the Accuracy of Hedonic Price Methods:
Econometric Analysis of Existing Data Sets, prepared by University of Colorado for US Environmental
Protection Agency (EARB).
Graves, Phil, James C. Murdoch, Mark A. Thayer. and Don Waldman. 1988. The Robustness of Hedonic Price
Estimation: Urban Air Quality, Land Economics, Vol. 64, No. 3, pp. 220-233.
*
Graves, Phil, Jim Murdoch, Mark Thayer. and Don Waldman. 1985. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments: Volume IV, Improving the Accuracy of Hedonic pricing Methods:
Econometric Analysts of Existing Studies - Draft, prepared by EPA Office of Policy Analysis for US
Environmental Protection Agency, University of Colorado, Center for Economic Analysis, Boulder, CO.
Green, Trellis G. 1986. Specification Considerations for the Price Variable in Travel Cost Demand Models:
Comment. Land Economics, Vol. 62. No. 4, pp. 416-418.
Green, Trellis G. 1984. Compensating and Equivalent Variation of the Florida Saltwater Tourist Fishery, Florida
State University.
Green. Trellis G. 1989. Economic Implications of Red Drum Catch Rate Elasticity in die Gulf of Mexico, National
Marine Fisheries Service, MARFIN.
Greenberg, M., M. Burrington, and C. Smith. 1984. Guidelines and an Illustration of the Use of Ecological Data for
Seeking Clues to Excess Risk, Socio-Economic Planning Sciences.
A-29
-------
Greenley, D. A., and R. G. Walsh. 1985. Option Value: Empirical Evidence from a Case Study of Recreation and
water Quality: Reply, Quarterly Journal of Economics, Vol. 98. pp. 295-299.
Greenley, D. A., R. G. Walsh, and R. A. Young. 1981. Option Value: Empirical Evidence from a Case Study of
Recreation and water Quality, Quarterly Journal of Economics, Vol. 96, No. 4, pp. 657-673.
Gregory, Robin. 1986. Interpreting Measures of Economic Loss: Evidence from Contingent Valuation and
Experimental Studies, Journal of Environmental Economics and Management, Vol. 13, No. 14, pp. 325-337.
Gregory, Robin S., Baruch Fischboff, Liu Furby, and Jack Knetsch, el al. 1985. Measures of Consumer Surplus:
Interpreting the Disparity in Views, prepared by University of Wyoming for US Environmental Protection
Agency (EARB).
Gregory, Robin, and Lata Furby. 1987. Auctions, Experiments, and Contingent Valuation, Public Choice, Vol. 55,
No. 3, pp. 273-289.
Gregory, Robin, and T. McDaniels. 1987. Valuing Environmental Losses: What Promises does the Right Measure
Hold?", Policy Science. Vol. 20, No. 1, pp. 11-26.
Grether, D., and C. Plott 1979. Economic Theory of Choice and the Preferences Reversal Phenomenon, American
Economic Review. Vol. 69, pp. 623-638.
Grieg, P. J. 1983. Recreation Evaluation Using a Characteristics Theory of Consumer Behavior, American Journal
of Agricultural Economics, Vol. 65. No. 1. pp. 90-97.
Griffen. Ronald, and John Stoll. 1983. The Enhanced Role of Water Conservation in the Cost-Benefit Analysis of
Water Projects, Water Resources Bulletin.
Griffin. R. C., and D. Bromley. 1982. Agricultural Runoff as a Nonpoint Externality: A Theoretical Development,
American Journal of Agricultural Economics.
Grigalunas, Thomas A., and James J. Opalach. 1988. Assessing Liability for Damages Under CERCLA: A New
Approach for Providing Incentives for Pollution Avoidance". Natural Resources Journal, Vol. 28, No. 3. pp.
509-533.
Grigalunas, Thomas A., lames J. Opalach, Deborah French, and Mark Reed. 1988. Measuring Damages to Marine
Natural Resources from Pollution Incidents tpnter CERCLA: An Application of an Integrated Ocean
Systems/Economic Model, Marine Resource Economics. Vol. 5. No. 1, pp. 1-22.
»
Grima. Angelo P. 1972. Residential Water Demand: Alternative Choices for Management, University of Toronto
Press.
Grossman, M. 1982. Government and Health Outcomes. Ai
i Economic Review.
Groves, T.. and D. Ledyaud. 1977. Optimal Allocation of Public Goods: A Solution to the Free-rider Problem,
Economevica, Vol. 45, pp. 784-809.
Grubb, W. Norton, M. Humphries, and D. Whiumgton. 1982. The Use (and Abuse) of Cost-Benefit Analysis in the
Federal Government: The Implementation of Executive Order 12291, prepared by Association of Public
Policy and Management
Guildin. R. 1980. Wilderness Costs in New England, Journal of Forestry, Vol. 78, pp. 548-552.
Guldin, Richard W. 1989. An Analysis of the Water Situation in the United Slates: 1989-2040: A Technical
Document Supporting the 1989 USDA Forest Service RPA Assessment, prepared by US Forest Service for
US Forest Service.
A-30
-------
1
I
t
1
Gum, R., and W. E. Martin. 1975. Problems and Solutions in Estimating the Demand for the Value of Rural Outdoor
Recreation, American Journal of Agricultural Economics, Vol. 57, pp. 558-566.
Gum, R. L. 1985. Assessment of Wildlife Values, prepared by USFS Wildlife Values Assessment Committee for US
Forest Service.
Gum, Russell L. 1986. Recreation Use of Wildlife Resources: An Economic Inquiry, prepared by Department of
Agricultural Economics for University of Arizona.
Hahn, Robert W. 1987. Jobs and Environmental Quality. Some Implications for Instrument Choice, Policy Science,
Vol. 20, No. 4, pp. 289-306.
Haigh, J., D. Harrison, and A. Nichols. 1984. Benefit-Cost Analysis of Environmental Regulation: Case Studies of
Hazardous Air Pollutants, Harvard Environmental Law Review.
Haimes, Yacov. 1984. Integrated Risk and Uncertainty Assessment in Water Resources Within a Multiobjective
Framework, Journal of Hydrology.
Hall, Charles H. 1985. An Estimate of the Potential Costs of Guidelines limiting Public Exposure to Radiofrequency
Radiation from Broadcast Sources, prepared by Livermore National Laboratory for US Environmental
Protection Agency.
Halvorsen, Robert, and Tim R. Smith. 1984. On Measuring Natural Resource Scarcity, Journal of Political
Economy.
Hammennesh, D., and F. Hammennesh. 1983. Does Perception of Life Expectancy Reflect Health Knowledge,
American Journal of Public Health.
Hammond, P. Brett, and Rob Coppock. 1990. Valuing Health Risks, Costs, and Benefits for Environmental Decision
Making - Report of a Conference, National Academy Press.
Hanemann, W. H. 1984. Entropy as a Measure of Concensus in the Valuation of Recreation Site Quality, Journal of
Environmental Management.
Hanemann. W. Michael. 1987. Welfare Evaluations in Contingent Experiments with Discrete Responses: Reply,
American Journal of Agricultural Economics, Vol. 69, No. 1, pp. 185-186.
Hanemann, W. Michael. 1984. Welfare Evaluations in Contingent Experiments with Discrete Responses, American
Journal of Agricultural Economics, VoL 66, pp. 332-341.
Hanemann, W. Michael. 1984. Discrete/Continuous Models of Consumer Demand, Econometrica, VoL 52. pp. 541-
561.
Hanemann, W. Michael. 1989. Information and the Concept of Option Value, Journal of Environmental Economics
and Management, Vol. 16, No. 1, pp. 23-37.
Hanemann, W. Michael. 1983. Welfare Evaluations with Simulated and Hypothetical Market Data: Bishop and
Heberlein Revisited, University of California at Davis. Calif. Agricultural Exp. Stat
Hanemann. W. Michael. Willingness to Pay and Willingness to Accept How Much Can They Differ?, University of
California at Berkeley, American Economic Review.
Hanemann. W. Michael, Richard T. Carson, Russell Gum, and Robert Mitchell. 1988. Northcentral Alaska Sport
Fishing Economic Study, prepared by Jones and Stokes Associates, Inc. for Alaska Department for Fish and
Game, Anchorage, AK.
A-31
-------
Hanemann. W. M. 1980. Quality Changes, Consumer Surplus, and Kedonic Prices Indexes, University of California
at Berkeley, Ag. and Resource Econ.
Hanemann, W. M. 1981. Water Quality and the Demand for Recreation, University of California at Berkeley, Ag.
and Resource Econ.
Hanemann, W. M., and Edward Morey. 1992. Separability, Partial Demand Systems, and Consumer's Surplus
Measures, Journal of Environmental Economics and Management, Vol. 22, No. 3, pp. 241-2S8.
Hanley, N. D. 1989. Valuing Non-market Goods Using Contingent Evaluation, Journal of Economic Surveys, Vol.
3, No. 3, pp. 235-252.
Hanley, N. D. 1989. Valuing Rural Recreation Benefits: An Empirical Examination of Two Approaches, Journal of
Agricultural Economics, Vol. 40, No. 3, pp. 361-374.
Hanley, Nicholas. 1988. Using Contingent Valuation to Value Environmental Improvements, Applied Economics,
Vol. 20, No. 4, pp. 541-549.
Hanley, Nick. 1990. The Economics of Nitrate Pollution, European Review of Environmental Economics, Vol. 17,
No. 2. pp. 129-151.
Hanna, J. Ed., Peter Victor, and Daniel R. Talhelm. 1985. The Product Travel Cost Approach: An Application for
Estimating Acid Rain Damage to Sports Fishing in Ontario.
Hansen, LeRoy T. 1991. National Estimates of me Recreational Value of Streamflow, Water Resources Research,
Vol. 27, No. 2, pp. 167-175.
Hanson, Jeffrey E. The Economic Value of Fish Contamination Advisories to Individual Great Lakes Sport Anglers.
Harding, C. M.. and J. R. Eisner. 1984. Characterizing the Perceived Risks and Benefits of Some Health Issues, Risk
Analysis.
Harford, J., and S. Ogura. 1983. Pollution Taxes and Standards: A Continuum of Quasi-Optimal Solutions, Journal
of Environmental Economics and Management.
Harford, Jon. 1984. Averting Behavior and the Benefits of Reduced Soiling, Journal of Environmental Economics
and Management.
Harrington, Leigh. 1984. The Valuation of the Life Shortening Aspects of Risk, prepared by Everest Consulting for
US Environmental Protection Agency (EARS). Everest Consulting Associates, Inc., Arlington. VA.
Harrington, W., and P. R. Portney. 1982. Valuing the Benefits of Improved Human Health, Resources for the Future,
Washington, DC.
Harrington, W., Alan J. Krupnick, and Henry M. Peskin. 1985. Policies for Non-point Soure Water Pollution
Control, prepared by RFF for Journal of Soil and Water Conservation. Vol. 40, pp. 27-32.
Harris, A. N. 1982. Hedonic Techniques and Valuation of Environmental Quality. In Advances in Microeconomics.
ed. Smith. V. IL. JA1 Press. Inc., Greenwich. CT. pp. 31-49.
Harris. B. S. 1984. Contingent Valuation of Water Pollution Control, Journal of Environmental Management, Vol.
19, pp. 199-208.
Harris, Charles C. B. L. Driver, and William Mclaughlin. 1989. Improving the Contingent Valuation Method: A
Psychological Perspective, Journal of Environmental Economics and Management, Vol. 17. No. 3, pp. 213-
229.
A-32
-------
*
Hams, Thomas, and Chauncey Ching. 1983. Economic Resource Multipliers for Regional Impact Analysis, Water
Resources Bulletin.
Harrison, D., and D. L. Rubinfeld. 1978. Distribution of Benefits from Improvements in Urban Air Quality, Journal
of Environmental Economics and Management, Vol. 5, pp. 313-333.
Harrison, D., and D. L. Rubinfeld. 1978. Hedonic Housing Prices and toe Demand for Clean Air, Journal of
Environmental Economics and Management, Vol. 5, pp. 81-103.
Harrison, D., and D. L. Rubinfeld. 1978. Air Pollution and the Property Value Debate: Some Empirical Evidence,
Review of Economics and Statistics, Vol. 60, pp. 635-639.
Harrison, David, Jr. 1984. Research and Demonstration of Improved Methods for Carrying out Benefit-Cost
Analyses oflndividual Regulations, Vols, I.II.III, & IV, prepared by Harvard University for US
Environmental Protection Agency (EARB).
Harrison, David, Jr., and Daniel L. Rubinfeld. 1978. The Air Pollution and Property Value Debate: Some Empirical
Evidence, Review of Economics and Statistics, Vol. 4, pp. 635-638.
Harrison, David, Jr., and J. H. Stock. 1984. Hedonic Housing Values, Local Public Groups, and the Benefits of
Hazardous Waste Cleanup, prepared by Energy and Environmental Policy Center, Harvard University for
Cambridge, MA.
Hartman, Richard, and Mark Plummer. 1987. Option Value under Income and Price Uncertainty, Journal of
Environmental Economics and Management, Vol. 14, No. 3, pp. 212-225.
Haspel, A. E., and F. R. Johnson. 1982. Multiple Destination Trip Bias in Recreation Benefit Estimation, Land
Economics, Vol. 58, pp. 364-372.
Hay, Michael J. 1988. Net Economic Recreation Values for Deer. Elk and Waterfowl Hunting and Bass Fishing, US
Fish and Wildlife Service.
Hay, Michael J. 1988. Net Economic Values of Nonconsumptive Wildlife, US Fish and Wildlife Service.
Hay, Michael, and Kenneth McConnell. 1984. Harvesting and Nonconsumptive Wildlife Recreation Decisions, Land
Economics.
Hayden, F. Gregory. 1989. Survey of Methodologies for Valuing Externalities and Public Goods, prepared by
University of Nebraska for US Environmental Protection Agency (OPPE).
Hayes, Karen Marie. 1987. An Analysis of the Benefits of Improving Water Quality in Narragansett Bay: An
Application of the Contigent Valuation Method.
Hazilla, Michael, and Raymond J. Kopp. 1985. The Social Cost of Environmental Regulations: A General
Equilibrium Analysis, prepared by Resources for the Future for US Environmental Protection Agency
(OAQPS).
Heaney, lames, and Wayne Huber. 1984. Nationwide Assessment of Urban Runoff Impact on Receiving Water
Quality, Water Resources Bulletin.
Heberlein. Thomas A., and Richard C. Bishop. 1985. Assessing the Validity of Contingent Valuation: Three Field
Experiments.
Heck. W., W. Cure, J. Rawlings, and L. Zaragoza, et al. 1984. Assessing Impacts of Ozone on Agricultrual Crops: I
Overview, Journal of the Air Pollution Control Association.
A-33
-------
Heintz, H. T., A. Horshafu and G. C. Horak. 1976. National Damages of Air and Water Pollution, prepared by EPA
Office of Research and Development for US Environmental Protection Agency.
Henry, C. 1974. Option Values in the Economics of Irreplacable Assets, Review of Economic Studies, pp. 89-104.
Herman, Stewart W. 1977. The Health Costs of Air Pollution: A Survey of Studies Published between 1967 and
1977, prepared by American Lung Association.
Hildebrandi, G. G., and T. D. Trcgartbe. 1979. Observing Preferences for Educational Quality: The Weak
Complementarity Approach, Scandinavian Journal of Economics, Vol. 8, No. l.pp. 188-197.
Hinckley, Dexter. 1990. Classification and Quantification of Ecosystem Services, prepared by EPA Office of Policy,
Planning and Evaluation for US Environmental Protection Agency.
Hoehn, John P. 1991. Valuing the Multidimensional Impacts of Environmental Policy: Theory and Methods,
American Journal of Agricultural Economics, pp. 289-299.
Hoehn, John P. 1988. Quality Adjusted Prices and Weak Complementarity: A New Method for Estimating the
Demand for Environmental Services, Resources and Energy, Vol. 10, pp. 337-354.
Hoehn, John P., and Alan Randall. 1987. A Satisfactory Benefit-Cost Indicator from Contingent Valuation, Journal
of Environmental Economics and Management, Vol. 14, No. 3, pp. 212-225.
Hoehn, John P., and Alan Randall. 1989. Too Many Proposals Pass the Benefit Cost Test, American Economic
Review, Vol. 79, No. 3, pp. 544-551.
Hoehn, John P., and J. Caudill. 1989. Economic Evaluation of Groundwater Policy Risks in Incomplete Markets,
Mirhioan State University.
*
Michigan State University.
*T*»**Ug««o *,__ _ f
Hoehn, John P., Mark C. Berger, and Glenn C. Bloomquist. 1987. A Hedonic Model of Interregional Wages, Rents,
and Amenity Values, Journal of Regional Science, Vol. 27. No. 4, pp. 605-620.
Hof, John G., and David A. King. 1982. On the Necessity of Simultaneous Recreation and Demand Equation
Estimation, Land Economics, Vol. 52, pp. 547-552.
Hof, John G., and David A. King. 1983. On the Necessity of Simultaneous Recreation and Demand Equation
Estimation: A Reply, Land Economics, Vol. 59, No. 4, pp. 459-460.
Hof, John G., and H. F. Kaiser. 1983. Long-term Outdoor Recreation Participation Projects for Public Land
Management Agencies, Journal of Leisure Research, Vol. 15, No. 1, pp. 1-14.
Hof, John G., and John B. Loomis. 1983. A Recreation Optimization Model Based on the Travel Cost Method,
Western Journal of Agricultural Economics, Vol. 8, No. 1, pp. 76-85.
Hof. John G., R. D. Lee, A. A. Dyer, and B. M. Kent 1985. An Analysis of Joint Costs in a Managed Forest
Ecosystem, Journal of Environmental Economics and Management. Vol. 12, No. 4, pp. 338-352.
Hen, H. 1975. Revealed Preferences for Public Goods, American Economic Review, Vol. 65. pp. 978-991.
Horowitz, loel L. 1985. Inferring Willingness to Pay for Housing Amenities from residential Property Values,
prepared by University of Iowa for US Environmental Protection Agency (EARB).
Horn, Robert L. Jr., Ernest H. Manuel Jr., Jerome T. Bentley, and Alexander R. Stankunas. 1984. Economic
Benefits of Reduced Acidic Deposition on Common Building Materials: Methods Assessment, prepared by
«*-.l_. »_„
Mathtec, Inc.
A-34
-------
Howe, Charles, and Dwight Lee. 1983. Priority Pollution Rights: Adapting Pollution Control to a Variable
Environment, Land Economics.
Howiu, R. £., T. W. Gossard, and R. M. Adams. 1984. Effects of Alternative Ozone Concentrations and Response
Data on Economic Assessments: The Case of California Crops, Journal of the Air Pollution Control
Association.
Huang, Cbung-Huang. 1986. The Recreation Benefits of Water Quality Improvement in Selected Lakes in
Minnesota, University of Minnesota.
Huber, Peter. 1984. The Market for Risk, Regulation.
Huetb, D., and E. J. Strong. 1985. A Critical Review of the Travel Cost, Hedonic Travel Cost, and Household
Production Models, prepared by Department of Agricultural and Resource Economists for Oregon State
University.
Hufschmidt, M., D. James, A. Meister, and B. Bower. 1983. Environment, Natural Systems, and Development: An
Economic Valuation Guide.
Hufschmidt, Maynard. 1982. New Approcahes to Economic Analysis of Natural Resources and Environmental
Quality.
Hull, R. Bruce IV, and Gregory Buyhoff. 1983. Distance and Scenic Beauty: A Nonmonotonic Relationship,
Environment and Behavior.
Huppert, D. D., and C. L. Thomson. 1984. Demand Analysis of Party Boat Angling in California Using the Travel
Cost Method, National Marine Fisheries Service, La JoUa, CA.
Huppert, Daniel D. 1990. Economic Benefits from Commercial Fishing, prepared by A. Kearney Centaur, Inc..
Garry Brown.
Huppert, Daniel D. 1988. An Examination of Nonresponse Bias and Divergence Among Value Concepts: An
Application to Central California Anadromous Fish Runs, Association of Environmental and Resource
Economists.
Huppert, Daniel D. 1989. Measuring the Value of Fish to Anglers: Application to Central California Anadramous
Species, Marine Resource Economics, Vol. 6, pp. 89-107.
Huppert, Daniel D. 1983. NMFS Guidelines on Economic Valuation of Marine Recreational Fishing, National
Oceanographic and Atmospheric Administration.
Hushak, L. J., I. M. Winslow, and N. Dutta. 1984. Economic Value of Lake Erie Sport Fishing to Private-Boat
Anglers, prepared by Sea Grant Program for Ohio Stale University, Ohio State University.
Hushak, L. J., J. M. Winslow, and N. Dutta. 1988. Economic Value of Great Lakes Sportfishing: The Case of Private
Boat Fishing in Ohio's Lake Erie, Transactions of the American Fisheries Society. Vol. 117, pp. 363-373.
Hyatt, William A. 1984. An Angler Survey and Economic Study of the Farmington River Fishery Resource,
prepared by Connecticut Department of Environmental Protection.
Hyde, W. E, A. Dickerman, and D. Stone. 1982. Development Versus Preservation in the Snake River Birds of Prey
Conservation Area, American Journal of Agricultural Economics.
ICF, Inc. 1989. Economic, Environmental, and Coal Market Impacts of SO2 Emissions Trading Under Alternative
Acid Rain Control Proposals, prepared by ICF, Inc.
A-35
-------
ICF, Inc. 1990. Economic Impact Analysis for the Proposal for a Self-Insurance Test for Government Entitities to
Demonstrate Financial Responsibility for Underground Storage Tanks.
ICF, Inc. 1988. The Unified Title III Economic Analysis: Subject Paper on Facilities: Vol. I.
JEc, Inc. 1989. Comparitive Analysis of State Hazardous Waste Management.
Imber, David, Gay Stevenson, and Leanne Wilks. 1991. A Contingent Valuation Survey of the Kakadu Conservation
Zone, prepared by Resource Assessment Commission for Australian Government. Resource Assessment
Commission, Australia.
Isaac, R. Mark, Kenneth F. McCue, and Charles R. PlotL 1985. Public Goods Provision in Experimental Economics,
Journal of Public Economics, Vol. 26, No. 1, pp. 51-74.
Jackson, John E. 1983. Measuring the Demand for Environmental Quality with Survey Data, Journal of Political
Economy, Vol. 45, No. 2, pp. 335-350.
Jaworski, Eugene, and C. N. Raphael 1981. Results of Specific Wetlands Value Study in Michigan. Wetland Values
& Management Conference, pp. 445-451.
Jennings, A. A., and R. Scholer. 1984. Hazardous Waste Disposal: Network Analysis, Journal of Environmental
Engineering Div-ASCE.
Johansson, Per-Olov. 1987. The Economic Theory and Measurement of Environmental Benefits, Cambridge
University Press, New York and Sydney.
Johansson, Per-Olov, and B. Kristom. 1988. Measuring Values for Improved Air Quality from Discrete Response
Data: Two Experiments, Journal of Agricultural Economics Research, VoL 39, No. 3, pp. 439-445.
Johansson, Per-Olov, B. Kristom, and Karl Maler. 1989. Welfare Evaluations in Contingent Valuation Experiments
with Discrete Response Data: Comment, American Journal of Agricultural Economics, Vol. 71, No. 4, pp.
1054-1056.
Johnson. Donn M., and Richard G. Walsh. 1987. Economic Benefits and Costs of the Fish Stocking Program at Blue
Mesa Reservoir, Colorado, prepared by Colorado Water Resources Research Institute for Colorado State
University, Fort Collins, CO.
Johnson, Donn M., and Richard G. Walsh. 1988. Economic Benefits of Low, Medium, and High Skilled Anglers.
Johnson, F. Reed. 1984. Market Disruption, Regulated Health Risks and the Media: The Case of EDB - (draft),
prepared by EPA Economic Analysis and Research Branch for US Environmental Protection Agency.
Johnson, F. Reed. 198S. Does Risk Information Reduce Welfare? Evidence from the EDB Food Contamination
Scare.
Johnson, M. B. 1966. Travel Time and the Price of Leisure, Western Economics Journal, VoL 4. pp. 135-145.
Johnson, R, and R. M. Adams. 1989. On the Marginal Value of a Fish: Some Evidence from a Steelhead Fishery,
Marine Resource Economics, VoL 6. pp. 43-55.
Johnson, Reed, John Krutilla, Michael Bowes, and Elizabeth WUman. 1981. A Methodology for Estimating the
Consequences of Forest Management on Recreation Benefits, prepared by Resources for the Future for
Washington, DC.
Johnson, T. 1984. Measuring the Cost of Tune in Re
Agricultural Economics.
cation Demand Analysis: Comment, American Journal of
A-36
-------
L
Johnson, R.L., N.S. Brezenger, and B. Shelby. 1990. Contingent Valuation Question Formats: Dichomotous Choice
vs. Open-Ended Responses, in Economic Valuation of Natural Resources: Issues, Theory, and Application,
ed. Johnson, RX. and Johnson, G.V., Westview Press, Boulder, Colorado.
Jones, Anne, Lynn Knight, and David Meyers. 1989. Regulatory Impact Analysis of the Proposed Regulations for
Sewage Sludge Use and Disposal, prepared by Eastern Research, Inc. for US Environmental Protection
Agency.
Jones, Carol A., and Yusen D. Sung. 1991. Valuation of Environmental Quality at Michigan Recreational Fishing
Sites: Methodological Issues and Policy Applications, prepared by University of Michigan, School of Natural
Resources for US Environmental Protection Agency.
Jones & Stokes Assoc. 1988. Development and Application of a Predictive Model to Analyze the Economic Effects
of Species Availibility, prepared by Jones and Stokes Associates, Inc. for National Coalition for Marine
Conservation, Pacific Region.
Jorgenson, Dale W., and Peter J. Wilcoxen. 1990. Environmental Regulation and U.S. Economic Growth, Harvard
University. Rand Journal of Economics, Vol. 21, No. 2, pp. 314-340.
Joyce, T. J., Michael Grossman, and Fred Goldman. 1989. An Assessment of the Benefits of Air Pollution Control:
The Case of Infant Health, Journal of Urban Economics, Vol. 25, No. 1, pp. 32-51.
Just, R., D. Huetb, and A. Schmitz. 1982. Applied Welfare Economics, Prentice Hall, Englewood Cliffs, NJ.
Kagawa, J. 1984. Heatlh Effects of Air Pollution and Their Management, Atmospheric Environment.
Kann, James R. 1987. Measuring the Economic Damages Associated with Terrestial Pollution of Marine
Ecosystems, Marine Resource Economics, Vol. 4, pp. 193-209.
Kabn, James R. The Economic Value of Long Island Saltwater Recreational Fishing, prepared by New York Sea
Grant/New York State Department of Env. Cons, for Slate University of New York at Binghamton.
Kann, James R., and Michael W. Kemp. 1985. Economic Losses Associated with a Decline in the Ecosystem: The
Case of Submerged Aquatic Vegetation in the Chesapeake Bay, Journal of Environmental Economics and
Management, Vol. 12, No. 3, pp. 246-263.
Kahneman, D., and J. Knetsch. 1992. Valuing Public Goods: The Purchase of Moral Satisfaction, Journal of
Environmental Economics and Management, Vol. 22, pp. 57-70.
Kaoru, Yoshiaki. 1989. Valuing Improvements in Estuarine Quality for Marine Recreational Fishing: A Discrete
Choice Benefit Analysis, Vanderbilt University. Knoxville, TN.
Kasabach, H. F., and W. F. Althoff. 1983. An Overview of New Jersey's Groundwater Quality Program,
Groundwater.
Kavanaugh, Michael, and Robert M. WolcotL 1982. Economically Efficient Strategies for Preserving Groundwater
Quality, prepared by Public Interest Economics for US Environmental Protection Agency.
Kealy, Mary Jo, John F. Dovidio, and Mark L. Rockel. 1988. Accuracy in Valuation is a Matter of Degree, Land
Economics, Vol. 64, No. 2, pp. 158-171.
Kealy, Mary Jo, and Richard C. Bishop. 1986. Theoretical and Empirical Specifications Issues in Travel Cost
Demand Studies, American Journal of Agricultural Economics, VoL 68, No. 3, pp. 660-667.
Kealy, Mary Jo, John F. Dovidio. and Mark L. Rockel. 1987. Willingness to Pay to Prevent Additional Damages to
the Adirondack* from Acid Rain, Regional Science Review.
A-37
-------
Kealy, Mary Jo, Mark Montgomery, and John F. Dovidio. 1990. Reliability and Predictive Validity of Contingent
Values: Do the Characteristics of the Good Matter?, Journal of Environmental Economics and Management,
Vol. 19, pp. 244-263.
Kealy, Mary Jo. 1991. Economic Quantification of Leisure Benefits, in Benefits of Leisure, ed. B.L., Driver and PJ.,
etal. Brown, Venture Publishing, Inc.. State College, PA, pp. 431-438.
Keddy, P. A. 1983. Freshwater Wetlands Human-Induced Changes: Indirect Effects Must Also Be Considered,
Environmental Management
Keith, John E. 1980. Snowmobiling and Cross-country Skiing Conflicts hi Utah: Some Initial Research Results,
University of Minnesota, pp. 57-63.
Keith, John E., and Kennith S. Lyon. 1985. Valuing Wildlife Management: A Utah Deer Herd, Western Journal of
Agricultural Economics, Vol. 10, No. 2, pp. 216-222.
Keith, John, Phil Halverson, and Lon Furnworth. 1982. Valuation of a Free Flowing Riven the Salt River, Arizona,
prepared by Utah State University for US Army Corps of Engineers, Logan.
Kellert, Stephen R. 1984. Assessing Wildlife and Environmental Values in Cost-Benefit Analysis, Journal of
Environmental Management, Vol. 18, pp. 355-363.
Kent, D. C., J. W. Nancy, and F. E. Witz. 1982. Prediction of Economic Potential for Inagab'on Using a
Groundwater Model, Groundwater, Vol. 20, No. 5, pp. 577-578.
Ketkar, Kusum. 1983. Pollution Control and Inputs to Production, Journal of Environmental Economics and
Management.
King, D., and J. Hof. 1986. Experiential Commodity Definition in Recreation Travel Cost Models, Forest Science,
Vol. 31, pp. 519-529.
King, David A., Deborah J. Bugorsky, and William W. Shaw. 1987. Contingent Valuation: An Application to
Wildlife, prepared by Department of Resource Recreation Management, O. State Univ for Oregon State
University.
King, David A., and J. A. Sinden. 1988. Influence on Soil Conservation on Farm Land Values, Land Economics,
Vol. 64, No. 3, pp. 242-255.
King, David, and Ann Walka. 1980. A Market Analysis of Trout Fishing on Fort Apache Indian Reservation,
prepared by School of Renewable Natural Resources for University of Arizona.
Kitabatake, Y., and T. Miyazaki. 1983. Toe Location of Sewage Treatment Plants on a Continuous Space:
Theoretical and Empirical Analyses, Environment and Planning.
itionbythe
Klemperer, David W., Paul S. Verbyla, and Linda D. Joyner. 1984. Valuing White-Water River R
Travel Cost Method, pp. 709-719.
Kling, Catherine L. 1988. The Reliability of Estimates of Environmental Benefits from Recreation Demand Models,
American Journal of Agricultural Economics, Vol. 70, No. 4, pp. 892-901.
Kling, Catherine L. 1988. Comparing Welfare Estimates of Environmental Quality Changes from Recreational
Demand Models, Journal of Environmental Economics and Management. VoL IS, No. 3, pp. 331-340.
Kling, Catherine L. 1987. A Simulation Approach to Comparing Multiple-Site Recreation-Demand Models Using
Chesapeake Bay Survey Data, Marine Resource Economics, Vol. 4, No. 2. pp. 95-109.
A-38
-------
1
I
T
i
i
Kling, Catherine L., and R.J. Sexton. 1990. Bootstrapping in Applied Welfare Analysis, American Journal of
Agricultural Economics, Vol. 72, pp. 406-418.
Kneese, A. V., and J. L. Sweeney. 1985. Handbook of Natural Resource and Energy Economics, North Holland
Publishing Company, Amsterdam.
Kneiscb, J. F., and F. J. Cesario. Some Problems in Estimating the Demand for Outdoor Recreation: Comment,
American Journal of Agricultural Economics, Vol. 58, No. 3, pp. 596-597.
Knetsch, Jack L.. and J. A. Sinden. 1984. Willingness to Pay and Compensation Demanded: Experimental Evidence
of an Unexpected Disparity in Measuring Values, Quarterly Journal of Economics.
Knetsch, Jack L., and Robert K. Davis. 1966. Comparison of Methods for Recreation Evaluation. In Water
Research, ed. Kneese, A. and Smith, S., Johns Hopkins University Press, Baltimore, MD.
Kobn, Robert E. 1985. A General Equilibrium Analysis of the Optimal Number of Firms in a Polluting Industry,
Canadian Journal of Economics, Vol. 18, No. 2, pp. 347-354.
Kolb, Jeff, and Joel D. Scheraga. 1988. The Choice of Discount Rates for Public Projects: A Comment and Further
Evidence, American Economic Review.
Kolstad, Charles D. 1986. Empirical Properties of Economic Incentives and Command-and-Control Regulations for
Air Pollution Control, Land Economics, Vol. 62, No. 3, pp. 250-268.
Kolstad, Charles D. 1987. Uniformity vs. Differentiation in Regulating Externalities., Journal of Environmental
Economics and Management, Vol. 14, No. 4, pp. 386-399.
Kopp, R. J., W. J. Vaughan, and M. Hazilla. 1984. Agricultural Sector Benefits Analysis for Ozone: Methods
Evaluation and Demonstration, US Environmental Protection Agency.
Kopp, Raymond J., and V. Kerry Smith. 1989. Benefit Estimation Goes to Court: The Case of Natural Resource
Damage Assessments, Journal of Policy Analysis and Management, Vol. 8. No. 4, pp. 593-612.
Kopp, Raymond J., and Alan J. Krupnick. 1987. Agricultural Policy and the Benefits of Ozone Control, prepared by
Resources for the Future for American Journal of Agricultural Economics, Vol. 69, No. 5, pp. 956-962.
Kostow, L. P., and J. F. Kowalczyk. 1983. A Practical Emission Trading Program, Journal of the Air Pollution
Control Association.
Kovan, Jessica, and Lawrence Ubby. 1987. An Analysis of Selected Policy Alternatives to Alleviate Groundwater
Contamination in Michigan. Michigan State University.
Krouse, M. It, L. Antle, G. Goicoechea, and A. Goldman, et al. 1983. An Approach to Risk and Uncertainty in
Benefit-Cost Analysis of Water Resources Projects, Water Resources Research.
Krupnick. A., and R. Kopp. 1988. The Health and Agricultural Benefits of Reductions in Ambient Ozone in the
United States, Resources for the Future, Washington, DC.
Krupnick, A. J., W. Harrington, and B. Ostro. 1987. Air Pollution and Acute Health Effects: New Evidence,
Resources for the Future, Washington, DC.
Krupnick, A. J., W. Harrington, and B. Ostro. 1990. Ambient Ozone and acute Health Effects: Evidence from daily
data, Journal of Environmental Economics and Management, Vol. 18, pp. 1-18.
Krupnick, Alan J. The Valuation of Environmental Externalities: Guidance Document, prepared by Resources for
the Future.
A-39
-------
Krupnick, Alan J. 1988. Reducing Bay Nutrients: An Economic Perspective, prepared by Resources for the Future
for Maryland Law Review, Vol. 47, No. 2, pp. 452-480.
Krupnick, Alan J., and M. L. Cropper. 1989. Valuing Chronic Morbidity Damages: Medical Costs, Labor Market
Effects, and Individual Valuations, prepared by Resources for the Future for US Environmental Protection
Agency (OPPE).
Krupnick, Alan, W. Oates, and E. Van De Verg. 1983. On Marketable Air-Pollution Permits: Tbe Case for A System
of Pollution Offsets, Journal of Environmental Economics and Management.
KrutiUa, John F., and Anthony C. Fisher. 1985. Economics of Natural Environments: Studies in the Valuation of
Commodity and Amenity Resources, prepared by Resources for the Future for Resources for the Future,
Washington, DC.
Kurz, M. 1974. An Experimental Approach to the Determination of the Demand for Public Goods, Journal of
Political Economy, Vol. 3, pp. 329-348.
Kushman, John. 1987. Welfare Evaluations in Contingent Valuation Experiments with Discrete Responses:
Comment, American Journal of Agricultural Economics, Vol. 69, pp. 182-184.
an, T. R. 1983. A Multiregional Model of the Economy, Environment, and Energy Demand in the United
States, Economic Geography.
Lapbam, Sandra, and Stuart Castle. 1984. Survey of Health Department-Based Environmental Epidemiology
Programs, American Journal of Public Health.
Lave, L. B. 1982. Mitigating Strategies for Carbon Dioxide Problems, American Economic Review.
Lave, Lester B. 1986. The Cost of Abating Sulfur, Nitrogen, and Ozone Air Pollutants, American Journal of
Agricultural Economics, Vol. 68, No. 2, pp. 473-478.
Lawless, Edward W. 1988. Risk Ass
Methodology for Hazardous Waste Management, prepared by MRJ for
US Environmental Protection Agency.
Lee, Dwight 1984. Tbe Economics of Enforcing Pollution Taxation, Journal of Environmental Economics and
Management.
Lee. L. K., and E. G. Nielsen. 1987. The Extent and Costs of Groundwater Contamination by Agriculture, Journal of
Soil and Water Conservation, VoL 42. No. 4. pp. 243-248.
Leewortby, Veroon R. 1991. Recreational Use Value for John Pennekamp Coral Reef State Park and Key Largo
National Marine Sanctuary, prepared by Strategic Assessment Branch for National Oceanographic and
Atmospheric Administration.
Leeworthy, Yemen R. 1990. An Economic Allocation of Fishery Stocks Between 1te?ffmtinmfl and Commercial
Fishermen: The Case of King Mackeral, Florida State University.
Leewortby, Vernon R.. and Peter C, Wiley. 1991. Recreational Use Value Cor Island Beach State Park, prepared by
Strategic Assessment Branch for National Oceanographic and Atmospheric Administration.
Leipert, Christian, and Udo Ernst Simonis. 1987. Environmental Damage - Environmental Expenditures: Statistical
Evidence on the Federal Republic of Germany. International Journal of Sociological Economics, Vol. IS, No.
7, pp. 37-52.
Leonard, H. Jeffrey. 1988. Pollution and the Struggle for the World Product, Cambridge University Press,
Cambridge, England.
A-40
-------
•
Leuscbner, William A., Phillip S. Cook, Joseph W. Roggenbuck, and Richard G. Oderwald, 1987. A Comparative
Analysis for Wilderness User Fee Policy, Journal of Leisure Research. Vol. 19, No. 2, pp. 101-114.
Leuscbner, William A., and Rodney L. Young. 1978. Estimating the Southern Pine Beetle's Impact on Reservoir
Campsites, Forest Science. Vol. 24, No. 4, pp. 527-537.
Levin, Ronnie. 1986. Reducing Lead in Drinking Water: A Benefit Analysis, prepared by EPA Office of Policy,
Planning and Evaluation for US Environmental Protection Agency.
Liao, David S. 1988. Proceedings of the Symposium of Demand and Supply of Sport Fishing, prepared by American
Fishing Tackle Manufacturers.
Libby, Lawrence W., John P. Hoebn, James Caudill, and David Walker. Water and the Michigan Economy:
Estimating the Economic Value of Michigan's Fresh Water, prepared by Institute of Water Research for
Michigan State University.
Library of Congress. 1979. The Status of Environmental Economics: An Update, Environmental Policy Division,
Library of Congress.
Lichtenberg, E., and D. Zilberman. 1986. The Econometrics of Damage Control: Why Specification Matters,
American Journal of Agricultural Economics, Vol. 68, No. 2, pp. 261-273.
Lichtenberg, Erik, David Zilberman, and Kenneth T. Bogen. 1989. Regulating Environmental Health Risks under
Uncertainty: Groundwater Contamination in California, Journal of Environmental Economics and
Management, Vol. 17, pp. 22-34.
Lieu, T. S. 1986. Impacts of Air Pollution Control Costs: An Input Output Approach, Annals of Regional Science,
pp. 55-^5.
Light. J. Richard, and David B. Pillemer. 1984. Summing Up: The Science of Reviewing Research, Harvard
University Press, Cambridge, MA.
Lindell, M. K., and T. C. Earle. 1983. How Close is Close Enough: Public Perceptions of the Risks of Industrial
Facilities, Risk Analysis.
Under, Stephen, and Mark McBride. 1984. Enforcement Costs and Regulatory Reform: The Agency and Finn
Response, Journal of Environmental Economics and Management.
Lindsay, C. M. 1969. Option Demand and Consumer Surplus, Quarterly Journal of Economics. Vol. 83, pp. 344-
346.
Lipfert, F. W. 1979. On the Evaluation of Air Control Benefits, National Commission on Air Quality.
Lipfert, F. W. 1977. The Association of Air Pollution with Human Mortality: Multiple Regression Results for 136
Cities, 1969.
Lipfert, F. W. 1984. Air Pollution and Mortality: Specification Searches using SMS A-based data. Journal of
Environmental Economics and Management, Vol. 11, pp. 208-243.
Lipfert, F. W. 1980. Sulfur Oxides, Particulates, and Human Mortality: Synopsis of Statistical Correlations, Journal
of the Air Pollution Control Association, Vol. 30, pp. 366-371.
Lapton, Douglas W., and Ivar E. Strand. 1989. The Effect of Common Property on the Fishing Industry, Journal of
Environmental Economics and Management.
A-41
-------
Liu, C. M., W. A. Leuschner, H. E. Burkhan, and J. E. de Steiguer. 1990. A Model for Assessing the Impact of
Atmospheric Deposition on Regional Forest Inventories, Journal of Environmental Management, Vol. 30, pp.
337-351.
Loehman. E. T., S. V. Berg, A. A. Arroyo, and R. A. Hedinger. et al. 1979. Distributional Analysis of Regional
Benefits and Costs of Air Quality Control, Journal of Environmental Economics and Management, Vol. 6, pp.
222-243.
Loehman, Edna T., and David BoldL 1990. Valuing Gains and Losses in Visibility and Health with Contingent
Valuation, prepared by Purdue University; U. New Mexico for US Environmental Protection Agency.
Loftis, J. C., R. C. Ward, and G. M. Smillie. 1983. Statistical Models for Water Quality Regulation, Journal of Water
Pollution Control Federation.
Loomis, John B. 1987. Expanding Contingent Value Sample Estimates to Aggregate Benefit Estimates: Current
Practices and Proposed Solutions, Land Economics, Vol. 63, No. 4, pp. 396-402.
Loomis, John B. 1989. Test-Retest Reliability of the Contingent Valuation Method: A Comparison of General
Population and Visitor Responses, University of California at Davis, American Journal of Agricultural
Economics, Vol. 71, No. 1, pp. 76-84.
Loomis, John B. 1989. Consumptive and Nonconsumptive Values of a Game Animal: The Case of California Deer,
Washington, DC, Transactions of the 54th North American Wildlife and Natural.
Loomis, John B. 1989. A More Complete Accounting of Costs and Benefits from Timber Sales, University of
California at Davis, Journal of Forestry, Vol. 87, No. 3, pp. 19-23.
Loomis, John B. 1980. Monetizing Equity-efficient Trade-offs in Rationing Wilderness Use, prepared by Western
Energy and Land Use Team for US Fish and Wildlife Service.
Loomis, John B. 1982. Use of Travel Cost Models for Evaluating Lottery Rationed Recreation: Application to Big
Game Hunting, Journal of Leisure Research, VoL 14, No. 2. pp. 117-124.
Loomis, John B. 1982. Effect of Non-price Rationing on Benefit Estimates From Publicly Provided Recreation,
Journal of Environmental Management, Vol. 17, No. 2. pp. 283-289.
Loomis, John B. 1987. Balancing Public Trust Resources of Mono Lake and Los Angeles' Water Right: An
Economic Approach, University of California at Davis, Water Resources Research, Vol. 23, No. 8, pp. 1449-
1456.
Loomis. John B. 1989. A Bioecooomk: Approach to Estimating the Economic Effects of Watershed Disturbance on
Recreational and Commercial Fisheries, University of California at Davis, Journal of Soil and Water
Conservation, Vol. 44. No. 1, pp. 83-87.
Loomis. John B. 1989. Estimation of and Variation in Site Specific Marginal Values for Recreational Fisheries,
University of California at Davis, Journal of Environmental Management, Vol. 29, pp. 183-191.
Loomis, John B. 1988. The Bioeconomic Effects of Timber Harvesting on Recreational and Commercial Salmon
and Steelbead Fishing: A Case Study of the Siuslaw National Forest, University of California at Davis,
Marine Resource Economics, VoL 5. pp. 43-6U
Loomis. John B. 1980. Monetizing Benefits under Alternative River Recreation Use Allocation Systems, Water
Resources Research, VoL 16. No. 1. pp. 117-124.
Loomis, John B., Dennis M. Donnelly. Cindy F. Sorg, and Lloyd Oldenburg. 1985. Net Economic Value of Hunting
Unique Species in Idaho: Bighorn Sheep, Mountain Goat, Moose, and Antelope, prepared by Rocky
Mountain Forest and Range Land Experiment Station for US Forest Service.
A-42
-------
Loomis, John B., and Joseph Cooper. 1990. Comparison of Environmental Quality-Induced Demand Shifts Using
Time-Series and Cross-Section Data, Western Journal of Agricultural Economics, Vol. 15, No. 1, pp. 83-90.
Loomis, John B., and Joseph Cooper. 1988. The Net Economic Value of Antelope Hunting in Montana, prepared by
Montana Department of Fish, Wildlife & Parks for Montana Department of Fish, Wildlife and Parks.
Loomis, John B., and Joseph Cooper. 1990. Economic Benefits of Instream Flow to Fisheries: A Case Study of
California's Feather River, Rivers, Vol. 1. No. 1, pp. 23-30.
Loomis, John B., Michael Creel, and Joseph Cooper. 1989. Economic Benefits of Deer in California: Hunting and
Viewing Values, prepared by Division of Environmental Studies for University of California at Davis.
Loomis, John B., and Richard G. Walsh. 1986. Assessing Wildlife and Environmental Values in Cost-Benefit
Analysis: State of the Art, Journal of Environmental Management, Vol. 22, No. 2, pp. 125-131.
Loomis, John, and Cindy Sorg. 1982. Critical Summary of Empirical Estimates of Values of Wildlife, Wilderness
and General Recreation Related to National Forests, prepared by Department of Economics for Colorado
State University.
Loomis, John, Cindy Sorg, and Dennis Donnelly. 1986. Evaluating Regional Demand Models for Estimating
Recreation Use and Economic Benefits: A Case Study, Water Resources Research, Vol. 22, pp. 431-438.
Loomis. John, Cindy Sorg, and Dennis Donnelly. 1986. Economic Losses to Recreational Fisheries due to Small-
head Hydro-Power Development A Case Study of the Henry's Fork in Idaho, Journal of Environmental
Management, Vol. 22, pp. 85-94.
Luken, Ralph A. 1987. Setting National Standards for Inorganic Arsenic Emissions from Primary Copper Smelters,
US Environmental Protection Agency.
Luken, Ralph A. 1985. Benefit-Cost Analysis at EPA, Environmental Forum.
Luken, Ralph. 1988. Municipalities, Small Business, and Agriculture: The Challenge of Meeting Environmental
Responsibilities, prepared by Environmental Protection Agency for US Environmental Protection Agency
(EARB), Sector Studies.
Lynne, Gary D.. Patricia Conray, and Frederick J. Prochaska. 1981. Economic Valuation of Marsh Areas for Marine
Production Processes, Journal of Environmental Economics and Management, Vol. 8, No. 2, pp. 175-186.
Madariaga, Bruce, and Kenneth McConneU. 1987. Exploring Existence Value, Water Resources Research, Vol. 23.
No. 5, pp. 936-942.
Maddock, Thomas m. 1976. Cost Functions for Additional Groundwater Development, Water Resources Bulletin,
Vol. 12, No. 3, pp. 539-545.
Magat, Wesley A., and W. Kip Viscusi. 1987. Economic Efficiency of Enforcement and Enforcement - Related
Monitoring: Project Period 2 Report, prepared by Duke University for US Environmental Protection Agency
(EARB).
Magat, Wesley A., and W. Kip Viscusi. 1990. Effectiveness of the EPA's Regulatory Enforcement: The Case of
Industrial Effluent Standards, prepared by Duke University and Northwestern University for US
Environmental Protection Agency (EARB), Law and Economics. Vol. 33, No. 2, pp. 331*360.
Magat, Wesley A., W. Kip Viscusi, and Joel Huber. 1988. Paired Comparison and Contingent Valuation Approaches
to Morbidity Risk Valuation, Journal of Environmental Economics and Management. Vol. 15, No. 4, pp. 395-
411.
A-43
-------
Magau Wesley A., Alan J. Krupnick, and Winston Harrington. 1986. Rules in the Making: A Statistical Analysis of
Regulatory Agency Behavior, prepared by Resources for the Future for Resources for the Future,
Washington, D.C.
Magnuson, Paula L. 1983. The Uses of Different Kinds of Aquifers, prepared by National Science Foundation for
pp. 151-169.
Magnussen, Kristin. 1992. Valuation of Reduced Water Pollution Using the Contingent Valuation Method. In
Country Reports for the CVM Network., prepared by Norwegian Agricultural Economics Resaercb Institute
for Oslo, Norway.
Mahoney, James R. 1990. National Acid Precipitation Assessment Program - Key Results, prepared by NAPAP for
National Acid Precipitation Assessment Program.
Majid, I., A. Sinden, and Alan Randall. 1983. Benefit Evaluation of Increments to Existing Systems of Public
Facilities, Land Economics.
Majid. L., J. Sinder, and A. Randall. The Benefits of Increments in Existing Park Systems, Land Economics.
Maler, K.G. 1974. Environmental Economics: A Theoretical Enquiry, Johns Hopkins University Press, Baltimore,
MD.
Maler, K. G. 1977. A Note on the Use of Property Values in Estimating Marginal Willingness to Pay for
Environmental Quality, Journal of Environmental Economics and Management. Vol. 4, No. 4, pp. 355-369.
Maler, K. G. 1971. A Method for Estimating Social Benefits from Pollution Control, Swedish Journal of Economics
Vol.75.
Maloney, Michael, and Bruce Yandel. 1984. Estimation of the Cost of Air Pollution Control Regulation, Journal of
Environmental Economics and Management.
Mann, Grady E. 1981. Minnesota Priorities in Planning for Wildlife Values, Wetland Values & Management
Conference, pp. 35-49.
Maonesto, Gregory, and John Loomis. 1991. Evaluation of Mail and In-person Vomingent Valuation Surveys:
Results of a Study of Recreational Boaters, Journal of Environmental Management, Vol. 32, pp. 177-190.
Manton, K. G., E. Stallard, and H. D. Tolley. 1983. The Economic Impact of Health Policy Interventions, Risk
Analysis.
Marits, J. S., and R. G. JarretL 1983. The Use of Statistics to Examine the Association Between Flouride in Drinking
Water and Cancer Death Rates, Applied Statistics.
Markstrom, Donald C., and Donald H. RosenthaL 19S7. Demand and Value of Firewood Permits as Determined by
the Travel Cost Model, Western Journal of Applied Forestry, Vol. 2, No. 2. pp. 48-50.
Marshall, John. 1984. Gambles and the Shadow Price of Death, Ai
•n Economic Review.
Martin, W. E., and T. Archer. 1971. Cost of Pumping Irrigation Water in Arizona, 1891 to 1967. Water Resources
Research, Vol. 7. No. 1. pp. 23-31.
Martin, W. E., and J. F. Thomas. 1986. Policy Relevance in Studies of Urban Residential Water Demand, Water
Resources Research, Vol. 22, pp. 1735-1741.
Martin, William E., Frank H. Bollman, and Russell L. Gum. 1982. Economic Value of Lake Mead Fishing,
Fisheries, Vol. 7. No. 6, pp. 20-24.
I
I
J
J
J
1
1
A-44
-------
Maihai, C. V., R. D. Rowe, L. G. Chestnut, and L. Deck. 1989. Controlling Wintertime Visibility impacts at toe
Grand Canyon National Park: Preliminary Benefit Cost Analysis. In Visibility and Fine Particles, Air and
Waste Management Association and EPA, pp. 625-638.
, Inc. 1987. Economic Impact Analysis of Alternative Lead National Ambient Air Quality Standards,
prepared by Mathtec, Inc. for US Environmental Protection Agency.
Mauskopf, Josepbina A., Donald Anderson, and Pamela Jacobs. 1987. Improving RJAs: Suggestions for (be
Analysts of Hazardous Waste Regulations, prepared by Research Triangle Institute, Inc. for US
Environmental Protection Agency (EARB).
Mauskopf, Josephine A., Anne Forrest, and William Desvousges. 1990. Hazardous Substances in our Environment:
A Citizen's Guide to Understanding Health Risks, prepared by EPA Office of Policy, Planning and Evaluation
for US Environmental Protection Agency. Research Triangle Institute, Research Triangle Park, NC.
Maxey, Margaret. 1982. A Bioethical Perspective on Risk Assessment Models for Managing Toxic Wastes,
Radioactive or Non-Radioactive.
McCabe, Patrick. A Brief Overview of Natural Resource Valuation Issues and Methodologies, prepared by EPA
Science Policy Branch for US Environmental Protection Agency.
McCafferty, D., and F. L. Hall. 1982. The Use of Multinomial Logit Analysis to Model the Choice of Time to
Travel, Economic Geography.
McCarthy, Patrick S., Richard Tay. and Jerald J. Fletcher. Estimating the Value of Water Quality Improvements
from a Fully Discrete Model of Recreational Fishing, prepared by InsL for Research in the Behavioral, Econ.
and Man. Sci. for Purdue University.
McClelland, Gary H., Wm. Schulze, and D. Coursey. 1987. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments, prepared by EPA Office of Policy Analysis for US Environmental
Protection Agency.
Nfraesky
, G. W. 1972. Problems and Benefits in Groundwater Management, Groundwater, Vol. 10. No. 2, pp. 2-5.
vfcCollum. Daniel W., Richard C. Bishop, and Michael P. Welsh. 1988. A Probabilistic Travel Cost Method,
prepared by Department of Agricultural Economics for University of Wisconsin.
fcConnell, K. 1984. Measuring the Cost of Time in Recreation Demand Analysis: A Reply, American Journal of
Agricultural Economics.
cConnell, K. E. 1980. Valuing Congested Recreation Sites, Journal of Environmental Economics and
Management, Vol. 7, pp. 185-195.
£onnell, K. E. 1979. Values of Marine Recreational Fishery: Measurement and Impact of Measurement,
American Journal of Agricultural Economics, Vol. 61, No. 5, pp. 921-925.
Council, K. E. 1975. Some Problems in Estimating the Demand for Outdoor Recreation, American Journal of
Agricultural Economics, Vol. 57, No. 2, pp. 330-334.
s, Vol. 66. No. 2. pp.
K. E. 1990. Double Counting in Hedonic and Travel Cost Models, Land £
I2M27.
oanell, K. E. 1991. Consumer Surplus from Discrete Choice Models, University of Maryland.
«neli K. E. 1985. The Economics of Outdoor Recreation, Elsevier Science Publishers, New York, NY.
1
I
A-45
-------
Mitchell, Robert Cameron, and Richard T. Carson. 1982. Comment on Option Value: Empirical Evidence from a
case study of Recreation and Water Quality, Resources for the Future, Washington, DC.
Mitchell, Robert Cameron, and Richard T. Carson. 1986. Valuing Drinking Water Risk Reductions Using the
Contingent Valuation Method: A Methodological Study of Risks from THM and Giardia, prepared by EPA
Office of Policy Analysis for US Environmental Protection Agency, Resources for the Future, Washington,
DC.
Mitchell, Robert Cameron, and Richard T. Carson. 1981. An Experiment in Determining Willingness to Pay for .
National Water Quality Improvements, prepared by Resources for the Future for US Environmental
Protection Agency.
Mjelde, J. W., R. M. Adams, B. L. Dixon, and P. Garcia. 1984. Using Farmer's Actions to Measure Crop Loss Due
to Air Pollution, Journal of the Air Pollution Control Association.
Moench, A. F., and A. J>. Visocky. 1971. A Preliminary Least Cost' Study of Future Groundwater Development
Costs in Northeastern Illinois.
Moench, A. F., R. J. Schicht, and A. P. Visocky. 1972. A Systems Study of Future Groundwater Development Costs
in the Chicago Region, Water Resources Bulletin, Vol. 8, No. 2, pp. 328.
Molburg, J. C, and E. S. Rubin. 1983. Air Pollution Control Costs for Coal-To-Etectricity Systems, Journal of the
Air Pollution Control Association.
Moll, Amy. 1985. Regulatory Impact Analysis of the Final Rule for Non-Substation PCB transformers, prepared by
Putnam, Hayes, and Bartleu, Inc. for US Environmental Protection Agency.
Moocur, J. £. 1975. Estimating the Value of Alternative Outdoor Recreation Facilities Within a Small Area, Journal
of Leisure Research, Vol. 7, pp. 303-311.
Morey, Edward R. 1981. Demand For Site Specific Recreational Activities: A Characteristic Approach, Journal of
Environmental Economics and Management, Vol. 8, pp. 345-371.
Morey, Edward R. 1985. Characteristics, Consumer Surplus, and New Activities: A Proposed Ski Area, Journal of
Public Economics, Vol. 26, No. 2. pp. 221-236.
Morey, Edward R. 1984. Confuser Surplus, American Economic Review.
Morey, Edward R. 1991. What Are "Per-trip" Consumer's Surplus Measures? And What, If Anything, Do They Tell
Us About Consumer's Surplus?, University of Colorado.
Morey, Edward R. 1985. The Logit Model and Exact Expected Consumer's Surplus Meaures: Valuing Marine
Recreational Fishing, Association of Environmental and Resource Economists.
Morey, Edward R., Robert D. Rowe, and Michael Watson. 1991. An Extended Discrete-Choice Model of Atlantic
Salmon Fishing: With Theoretical and Empirical Comparisons to Standard Travel-Cost Models, prepared by
RCG/Hagler, Bailly, Inc. for University of Colorado and Bangor Hydro-Electric Company.
Morey, Edward R., W. Douglass Shaw, and Robert D. Rowe. 1988. Repealed Discrete Choice. Expected Consumer's
Surplus and Option Price: Valuing Site-Specific Activities, University of Colorado.
Morey, Edward R.. W. Douglass Shaw, and Robert D. Rowe. 1991. A Discrete-Choice Model of Recreational
Participation, Site Choice, and Activity Valuation When Complete Trip Data Are Not Available, prepared by
RCG/Hagler. BatUy, Inc. for University of Colorado and Vassar, Academic Press, Inc., New York, NY,
Journal of Environmental Economics and Management, Vol. 20, pp. 181-201.
A-49
-------
f
Navrud, Stale. 1989. Estimating Social Benefits of Environmental Improvements from Reduced Acid Depositions:
A Contingent Valuation Survey. In Valuation Methods and Policy Making in Environmental Economics, ed.
Folmer, H. and van lerland, £., Elvesier, Amsterdam: Holland.
Neil, Jon R. 1988. Another Theorem on Using Market Demands to Determine Willingness-to-Pay for Nontraded
Goods, Journal of Environmental Economics and Management, Vol. 15, No. 2, pp. 224-232.
Nelson, J. P. 1982. Measuring Benefits of Environmental Improvements: Aircraft Noise and Hedonic Prices. In
Advances in Microeconomics, ed. Smith, V. K.. JAI Press, Inc., Greenwich, CT, pp. 71-75.
Nelson, J. P. 1977. Accessibility and the Value of Time in Commuting, Southern Economic Journal, Vol. 43, pp.
1321-1329.
Nelson, J. P. 1979. Residential Choice, Hedonic Prices, and the Demand for Urban Air Quality, Journal of Urban
Economics, Vol. 5. pp. 357-369.
Nelson, R. W., and E. C. Weller. 1984. A Better Rationale for Wetland Management, Environmental Management
Nelson, J. P. 1981. Three Mile Island and Residential Property Values: Empirical Analysis and Policy Implications,
Land Economics, Vol. 57, No. 3, pp; 363-372.
Newey, Whitney K., and Kenneth D. West. 1987. A Simple, Positive Semi-definite, Heteroscedasticity and
Autocorrelation Consistent Covariance Matrix, Econometrica, Vol. 55, pp. 703-08.
Nichols, A. L. 1982. The Importance of Exposure in Evaluating and Designing Environmental Regulation: A Case
Study, American Economic Review.
Nichols, L. M., M. Bowes, and J. F. Dwyer. 1978. Reflective Travel Tune in Travel-Cost-Based Estimates of
Recreation Use and Value, University of Illinois, Champaign-Urbana, IL.
Nielsen, Elizabeth G., and Linda K. Lee. 1986. The Magnitude and Costs of Groundwater Contamination from
Agricultural Chemicals: A National Perspective, prepared by USFS National Resource Economics Division
for US Forest Service.
Noll, R. G. 1982. Implementing Marketable Emissions Permits, American Economic Review.
Norton, V., T. Smith, and I. Strand. 1983. Stripers: The Economic Value of the Atlantic Coast Commercial and
Recreational Striped Bass Fisheries, prepared by Sea Grant Program for University of Maryland, College
Park,MD.
Nowell, Clifford, Marc A. Evans, and Lyman McDonald. 1988. Length-biased Sampling in Contingent Valuation
Studies, Land Economics, Vol. 64. No. 4, pp. 367-371.
CXHagan, Ciaran. 1984. Pigouvian Taxes, Polluter Subsidies, Regulation, and the Size of a Polluting Industry: A
Note, Canadian Journal of Economics.
CWeil, William, Martin David, C. Moore, and E. Joeres. 1983. Transferable Discharge Permits and Economic
Efficiency:The Fox River, Journal of Environmental Economics and Management
Gates, Wallace. 1984. Markets for Pollution Control, Challenge.
Gates, Wallace E.. and Albert M. McGartland. 1985. Marketable Pollution Permits and Acid Rain Externalities-. A
Comment and Some Further Evidence, Canadian Journal of Economics, Vol. 18, No. 3. pp. 668-675.
Ochs, J., and R. Thorn. 1984. Measuring the Site-Specific Recreation Benefits Resulting from Improved Water
Quality: An Upper-Bound Approach, Water Resources Bulletin.
A-51
-------
Ofiara, Douglas &., and Joseph J. Seneca. 1990. Review and Assessment of Economic Techniques Used to Assess
Economic Impacts and Losses Attributable to Degradations in Water Quality in the Marine Environment,
prepared by US Army Corps of Engineers, Economic and Social Analysis for Rutgers University.
Olson, Eric D. 1989. Natural Resource Damages in the Wake of the Ohio and Colorada Decisions: Where Do We
Go From Here?, Environmental Law Reporter, Vol. 19, pp. 10551.
Osier, Gerry, Graham A. Colditz, and Nancy L. Kelly. 1984. The Economic Costs of Smoking and Benefits of
Quitting, Lexington Books, D .C.; Heath and Company.
Ostro, B. D. Air Pollution and Morbidity Revisited: A Specification Test, Journal of Environmental Economics and
Management.
Ostro, B. D., and S. Rothschild. 1989. Air Pollution and Acute Respiratory Morbidity: An Observational Study of
Multiple Pollutants, Environmental Research, Vol. 50, pp. 238-247.
Ostro. Ban. 1983. Urban Air Pollution and Morbidity: A Retrospective Approach, Urban Studies.
Ostro, Baa 1984. A Search for a Threshold in the Relationship of Air Pollution to Mortality: a Reanalysis of Data
on London Winters, Environmental Health Perspectives. Vol. 58, pp. 397-399.
Ostro, Bart 1983. The Effects of Air Pollution on Work Loss and Mortality, Journal of Environmental Economics
and Management, Vol. 10, pp. 371-382.
PCAO. 1987. Americans Outdoors: The Legacy, the Challenge, prepared by President's Commission on Americans
Outdoors for Island Press, Washington, DC.
Padgett, J., and H. Richmond. 1983. The Process of Establishing and Revising National Ambient Air Quality
Standards, Journal of the Air Pollution Control Association.
Palm, Roger C., and Stephen P. Malvestuto. Relationships Between Economic Benefit and Sport-fishing Effort on
West Point Reservoir, Alabama-Georgia.
Pahnqirist, Raymond B. 1988. Welfare Measurement for Environmental Improvements Using Hedonic Model: The
Case of Nonparametric Marginal Prices, Journal of Environmental Economics and Management, Vol. 15, No.
3, pp. 297-312.
Pahnquist, Raymond B. 1982. Measuring Environmental Effects on Property Values without Hedonic Regressions,
Journal of Urban Economics, Vol. 11. pp. 333-347.
Park, Timothy, John B. Loomis, and Michael Creel 1991. Confidence Intervals for Evaluating Benefits Estimates
from Dicbotomous Choice Contingent Valuation Studies. Land Economics, Vol. 61. No. 1. pp. 64-73.
Park, W. M., and L. A. Shabman. 1982. Distributional Constraints on Acceptance of Nonpoint Pollution Controls,
American Journal of Agricultural Economics.
Park, W. M., and S. S. Batie. 1979. Methodological Issues Associated with Estimation of the Economic Value of
Coastal Wetlands in Improving Water Quality, prepared by Department of Agricultural Economics for
Virginia Polytech Institute.
Parsons. George R. 1989. On the Value of the Condition of a Forest Stock: Comment, Land Economics, Vol. 65.
No. 1, pp. 68-72.
Parsons. George R.. and Mary Jo Kealy. 1990. Measuring Water Quality Benefits Using a Random Utility Model of
Lake Recreation in Wisconsin, University of Delaware.
A-52
-------
I
4
Parsons, George, and Mary Jo Kcaly. 1988. Measuring the Benefits of Water Quality Improvements: A Preliminary
Report Using Wisconsin Survey Data, US Environmental Protection Agency.
Parsons, George R. 1991. The Participation Decision in a Random Utility Model of Recreation.
Parsons, George, and Mary Jo Kealy. 1992. Randomly Drawn Opportunity Sels in a Random Utility Model of Lake
Recreation, Land Economics, Vol. 68, No. 1, pp. 422-434.
Pashigan, Peter B. 1984. The Effect of Environmental Regulation on Optimal Plant Size and Factor Snares, Vol. 27,
No. 1, pp. 1-28.
Pashigian, B. Peter. 1983. How Large and Small Plants Fare Under Environmental Regulation, Regulation.
Pasurka, Carl Jr. 1984. The Soon-Run Impact of Environmental Protection Costs on U.S. Product Prices, Journal of
Environmental Economics and Management
Payne, B. A., S. J. Olshansky, and T. E. Segel. 1987. The Effects on Property Values of Proximity to a Site
Contaminated with Radioactive Waste, Natural Resources Journal, Vol. 27, pp. 579-590.
Pearce, D. W. 1900. Tbe Valuation of Pollution Damage: Noise Nuisance. In Readings in Environmental
Economics, ed. Butler, J., Allen and Unwin, London.
Pearce, David. 1983. Ethics, Irreversibility, Future Penetration and the Social Rate of Discount, International Journal
of Environmental Studies.
Pearce, David, and Anil Markandya. 1987. The Benefits of Environmental Policy: An Appraisal of the Economic
Value of Environmental Improvement and the Economic Cost of Environmental Damage, prepared by
OECD.
Pearce, David, and R. K. Turner. 1990. Economics of Natural Resources and the Environment, Harvester
Wbeatsheaf, Hemel Hempstead.
Pendse, D., and J. B. Wycoff. 1974. Scope for the Valuation of Environmental Goods. Land Economics, Vol. 50.
No. 1, pp. 89-92.
Perl, L. J., and F. C. Dunbar. 1982. Cost Effectiveness and Cost-Benefit Analysis of Air Quality Regulations,
American Economic Review.
Peskin, Henry. 1990. Review of the Economic Assumptions in the report "Policy Options for Stabilizing Global
Climate", prepared by Edgevale Associates.
Peskin, Henry M. 1989. Accounting for Natural Resource Depletion and Degradation in Developing Countries,
prepared by Environment Department for World Bank.
Peterson, D. G., R. D. Rowe, W. D. Schulze, and G. W. Russell, et al. 1987. Valuation of Visual Forest Damages
from Ozone, prepared by EPA Office of Policy, Planning and Evaluation for US Environmental Protection
Agency.
Peterson, Donald C, Robert D. Rowe, and W. Schulze. 1987. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments: Valuation of Visual Forest Damages from Ozone, prepared by
University of Colorado for US Environmental Protection Agency (EARB).
Peterson, George L., and Cindy R. Sorg. 1987. Toward the Me
at of Total Economic Value, prepared by
Rocky Mountain Forest and Range Land Experiment Station for US Forest Service.
Peterson, George L., and J. Ross Arnold. 1987. The Economic Benefits of Mountain Running the Pike's Peak
Marathon, Journal of Leisure Research, Vol. 19, No. 2, pp. 84-100.
A-53
-------
Peterson, George L., John F. Dwyer, and Alexander J. Darragh. 1983. A Behavioral Urban Recreation Site Choice
Model, Leisure Science, Vol. 6, No. 1, pp. 61-81.
Peterson, George L., Richard G. Walsh, and John R. McKean. 1988. The Discriminatory Impact of Recreation Price,
prepared by Rocky Mountain Forest and Range Land Experiment Station for US Forest Service.
Phillips, Carl V., and Richard J. Zeckhauser. 1989. Contingent Valuation of Damage to Natural Resources: How
Accurate? How Appropriate?, BNA, Toxics Law Reporter, pp. 520-529.
Pickles, J. H. 1982. Air Pollution Estimation Error and What it Does to Epidemiologies! Analysis, Atmospheric
Environment.
Pindyck, R. S. 1984. Uncertainty in the Theory of Renewable Natural Resources, Review of Economic Studies.
Piper, S., R. S. Magleby, and C. E. Young. 1989. Economic Benefit Considerations in Selecting Water Quality
Projects: Insights from the Rural Clean Water Program, prepared by USFS Economic Research Service, Res.
and Tech. Div. for US Forest Service, Washington, DC.
Piper, Steven. 1989. Benefit and Cost Insights from the Rural Clean Water Program, Journal of Soil and Water
Conservation, Vol. 44, No. 3, pp. 230-206.
Pitcher, Hugh. 1989. IQ and Lead Exposure: Analytic Issues Arising in the Water Lead and Gasoline Lead
Standards, prepared by EPA Office of Policy, Planning and Evaluation for US Environmental Protection
Agency, American Journal of Agricultural Economics, Vol. 71, No. 2, pp. 475*479.
Pitcher, Hugh, and Janice D. Longstreth. 1991. Melanoma Mortality and Exposure to Ultraviolet Radiation: An
Empirical Relationship, Environment International, Vol. 17, pp. 7-21.
Plummer, Mark. 1985. The Sign and Size of Option Value: Comment, Land Economics.
Plummer, Mark L. 1986. Supply Uncertainty; Option Price and Option Value: An Extension. Land Economics, Vol.
62, No. 3, pp. 313-318.
Plummer, Mark L., and Richard C. Hartman. 1986. Option Value: A General Approach, Economic Inquiry, Vol. 24,
No. 3, pp. 455-472.
Polinsky, A. Mitchell, and Steven Shavell. 1978. Amenities and Property Values in a Model of an Urban Area: A
Reply, Journal of Public Economics, Vol. 9, No. l.pp. 111-112.
Polinsky, A. Mitchell, and Steven ShaveU. 1976. Amenities and Property Values in a Model of an Urban Area,
Journal of Public Economics, VoL 5, No. 1. pp. 119-129.
Polinsky, A. Mitchell, and Steven ShaveU. 1975. The Air Pollution and Property Value Debate, Review of
Economics and Statistics, Vol. 57, No. 1, pp. 100-104.
Pollack. R. A. 1971. Conditional Demand Functions and Implications of Direct Seperable Utility, Southern
Economic Journal, Vol. 57.
Pollack, R. A, and M. L. Wacter. 1975. The Relevance of the Household Production Function and Its Implications
for the Allocation of time. Journal of Political Economy, VoL 83, No. 2, pp. 255-277.
Pollinsky, A. M., and D. L. Rubinfeld. 1977. Property Values and the Benefits of Environmental Improvements:
Theory and Measurement In Public Economics and me Quality of Life, cd. Wingo, and Evans,.
Pope. C. Arden m, and Jeffrey W. Jones. 1987. Value of Wilderness Designation in Utah, Brigham Young
University, Department of Ag Economics.
A-54
-------
Poppiui, J., and T. Dietz. 1983. Social Impact Evaluation of the U.S. Resource Conservation and Recovery Act,
Environmental Management.
Portney, P. R. Economics and the Clean Air Act, Journal of Economic Perspectives.
Pormey, P. R., and J. Mullahy. 1983. Ambient Ozone and Human Health: An Epidemological Analysis - Draft,
prepared by EPA Office of Air Quality Planning and Standards for US Environmental Protection Agency.
Portney, Paul R. 1981. Housing Prices, Health Effects, and Valuing Reductions in Risk of Death, Journal of
Environmental Economics and Management, Vol. 8, No. 1, pp. 72-78.
Pormey, Paul R., and John Mullahy. 198S. Ambient Ozone and Human Health: An Epidemiological Analysis -
Vol.1, n, m, and Executive Summary, prepared by Resources for the Future for Thomas Walton.
Preston, Valerie, S. Martin Taylor, and David Hodge. 1983. Adjustment to Natural and Technological Hazards: A
Study of an Urban Residential Community, Environment and Behavior.
Prince, Raymond. 1988. Estimating Recreation Benefits Under Congestion. Uncertainty, and Disequilibrium, James
Madison University, Department of Economics.
Putnam, Hayes. 1984. Analysis of Cost per Health Risk avoided in Selected EPA standards and regulations,
prepared by Putnam, Hayes, and Bartlett, Inc.
Queenan, Charles J., and Michael M. Schnitzer. 1982. Regulatory Impact Analysis of the Proposed Rule for PCB-
containing Electrical Equipment, US Environmental Protection Agency.
Quiggan, John. 1988. Murray River Salinity - An Illustrative Model, American Journal of Agricultural Economics,
Vol. 70, No. 3, pp. 635-645.
Quinn, R. J., and M. C. Weinstein. 1983. Psychological Considerations in Valuing Health Risk Reductions, Natural
Resources Journal.
RCG/Hagler, Bially. 1991. Valutaion of Other Externalities: Air Toxics, Water Consumption, Waste Water, and
Land Use. New England Power Service Company.
RTI. 1992. The Benefits to Diversionary Users of Improved Water Quality: Phase I Assessment, prepared by RTI
for US Environmental Protection Agency (OPPE).
RTI, Inc. 1984. Contingent Valuation Survey to Estimate Benefits of Hazardous Wastes Management Regulations:
Held Interviewer Manual prepared by Research Triangle Institute, me.
Raab. R. L., and D. N. Steinnes. 1980. An Econometric Model of Success toA the Demand for Recreational Angling.
Minnesota Sea Grant.
I
Rae, D. 1982. The Value to Visitors of Improving Visibility at Mesa Verde and Great Smoky Mountain Natural
Parks, Charles River Associates, pp. 217-234.
Rae, Douglas A^ and Frederick W. Ljpfen. 1987. Estimating Damage to Historic Structures Due to Acid Deposition.
Randall, A., J. P. Hoehn, and G. S. Tolley. 1982. The Structure of Contingent Markets: Some Results of a Recent
Experiment, American Economic Review.
Randall, A., and J. Stoll. 1983. Existence Value in a Total Valuation Framework, Westview Press, Boulder, CO.
Randall, A, B. C. Ives, and C. Eastman. 1974. Bidding Games for Evaluation of Environmental Improvement,
Journal of Environmental Economics and Management, Vol. 1, No. 2, pp. 132-149.
A-55
-------
Randall, Alan, and Glenn C. Bloomquist, 1985. National Aggregate Benefits of Air and Water Pollution Control -
Volumes 1 & II, prepared by University of Kentucky for US Environmental Protection Agency (EARB),
Pollution.
Randall, Alan, and Glenn C. Bloomquist 1985. National Aggregate Benefits on Air and Water Pollution Control-
Lead in Gasoline, prepared by University of Kentucky for US Environmental Protection Agency.
Randall, A., J. P. Hoehn, and D. S. Brookshire. 1983. Contingent Valuation Surveys for Evaluating Environmental
Assets.
Raphael, C. Nicholas, and Eugene Jaworski. 1979. Economic Value of Fish. Wildlife, and Recreation in Michigan's
Coastal Wetlands, Coastal Zone Management Journal, Vol. 5, No. 3, pp. 181-194.
Raucber, Robert L. 1987. Estimating the Replacement Costs as an Eligibility Criterion for Critical Aquifer
Protection Areas: Issues and Evidence from the Estimation of Household Water Costs, prepared by Energy
and Resource Consultants, Inc. for US Environmental Protection Agency (OP A and ODW).
Raucber, Robert L. 1986. The Benefits and Costs of Policies Related to Groundwater Contamination, prepared by
EPA Office of Policy, Planning and Evaluation for US Environmental Protection Agency, Land Economics,
Vol. 62, No. 1, pp. 33-45.
Ravenscraft, D J, M. D. Bowes, and J. F. Dwyer. 1978. Dealing with the Capacity Constraints in Estimating
Recreation Benefits. In Forestry Research Report, University of Illinois, Champaign-Urbana, IL.
Ravenscraft, David I., and John F. Dwyer. 1978. Reflecting Site Attractiveness in Travel Cost-Based Models for
Recreation Benefit Estimation, University of Illinois, Department of Forestry, Champaign-Urbana, IL.
Ready, R. C., and R. Bishop. 1988. The Value to Consumers of a Great Lakes Commercial Fishery.
Reed, Krista, and C. Edwin Young. 1983. Impacts of Regulatory Delays on the Costs of Waste water Treatment
Plants, Land Economics.
Regens, James L., Frederick W. Cuddage, and Donald G. Hodges. 1989. Greenhouse Gases, Climate Change, and
U.S. Forest Markets, Environment, Vol. 31, No. 4, pp. 2+.
Reich, M., E. Scwartz, W. Rae, and E. Seagull et al. 1983. Environmental Politics and Science: The Case of PBB
Contamination. American Journal of Public Health.
Retches, Nancy, and John Wilkins DI. 1983. MuMvariate Statistical Relationships Between Water Plant Data and
Trihalometnanes-Implicabons for Studies of Human Health. Water Research.
Rdling, Stephen D., Kevin J. Boyle, Marda L. Phillips, and Mark Anderson. 1990. Temporal Reliability of
Contingent Values, Land Fconomirs, VoL 66, No. 2, pp. 128-134.
Retting. Stephen. M. Anderson, and K. Gibbs. 1983. Measuring The O)ste of PiWic^Siipplied Outdoor Recreation
Facilities: A Methodological Note, Journal of Leisure Research.
Reopen. R. T. 1981. Wetland Values. Concepts and Methods for Wetlands Evaluation, Wetland Values &
Management Conference, pp. 385-393.
Resource Assess Com. 1991. Commentaries on the Resource Assessment Commission's Contingent Valuation
Survey of the Kakadu Conservation Zone, prepared by Resource Assessment Commission for Resource
Rettig, R. B., and B. A. McCarl. 1982. Potential and Actual Benefits from Commercial Fishing Activities, National
Oceanographic and Atmospheric Administration, Portland, Oregon.
A-56
-------
f
Rhoads, R, G. 1983. Regulatory Application of Air Quality Simulation Models, Journal of the Air Pollution Control
Association.
Ribaudo, Marc 0.1989. Offsite Water Quality Benfits from the Conservation Reserve Program- (Draft), prepared by
USFS Economic Research Service, Soil & Water for US Forest Service.
Ribaudo, Marc, and Donald Epp. 1984. The Importance of Sample Discrimination in Using the Travel Cost Method
to Estimate the Benefits of Improved Water Quality, Land Economics.
Richards, Martin T., D. Brent Wood, and David A. Coyler. 1985. Sport Fishing at Lees Ferry, Arizona: User
Differences and Economic Values, prepared by N. Ariz Univ Organized Research Committee, Scb of Forestry
for University of Northern Arizona.
Richards, Martin T., and David A. King. 1982. An Economic Measure of Nonconsumptive Wildlife Values,
prepared by Experiment station, St Paul for University of Minnesota.
Ridker, Ronald G., and John A. Henning. 1976. The Determinants of Residential Property Values With Special
Reference to Air Pollution, Review of Economics and Statistics, Vol. 49, pp. 246-257.
Riely, Patricia Lavin, and David B. Rocklahd. 1988. A Review of Fisheries Economic Valuation Methods, prepared
by Sport Fishing Institute.
Roberts, Kenneth J., Mark E. Thompson, and Perry W. Pawlyk. 1985. Contingent Valuation of Recreational Diving
at Petroleum Rigs, Gulf of Mexico, Vol. 114, No. 2, pp. 214-219.
Roberts, Richard, and J. W. Gebb JU. 1984. Hazardous Material Management- A Local Approach, Journal of
Environmental Health.
Robertson, B. E., R. J. Woods, and D. W. Larson. 1982. Societal Valuation of Human Life as Manifested by Use or
Non-Use of Air Heat Exchangers, ISES/ Solar Energy Society of Canada Energex 8th Conference.
Rockel, Mark L., and Mary Jo Kealy. 1991. The Value of Nonconsumptive Wildlife Recreation in the United States,
Land Economics, Vol. 67, No. 4, pp. 422-434.
Rockland, David B. 1983. An Economic Analysis of Delaware's Recreational/Commercial Fisheries Conflict,
University of Delaware.
Rolston, m, Holmes. 1985. Valuing WUdlands, Environmental Ethics. Vol. 7, pp. 23-48.
Rose, Adam. 1983. Modeling the Macroeconomic Impact of Air Pollution Abatement, Journal of Regional Science.
Rosen, S. 1974. Hedonic Prices and Implicit Markets: Product Differentiation in Price Competition, Journal of
Political Economy, Vol. 82, pp. 35-55.
Rosenthal, Donald H. 1987. The Necessity for Substitute Prices in Recreation Demand Analyses, American Journal
of Agricultural Economics. VoL 69, No. 4, pp. 828-837.
Rosenthal, Donald H., and H. Kenneth Cordeil. 1984. Pricing River Recreation: Some Issues and Concerns, pp. 272-
284.
Rosenthal, Donald H., John B. Loomis, and George L. Peterson. 1984. The Travel Cost Model: Concepts and
Applications, prepared by Rocky Mountain Forest and Range Land Experiment Station for US Forest Service.
Rosenthal, Donald H.. and Richard G. Walsh. 1986. Hiking Values and the Recreation Opportunity Spectrum, Forest
Science, Vol. 32, No. 2, pp. 405-415.
A-57
-------
Rowe, Robert D., E. R. Morey, A, D. Ross, and W. D. Shaw. 1985. Valuing Marine Recreation on the Pacific Coast,
National Marine Fisheries Service, La Jolla, CA.
Rowe, Robot D., and Lauraine G. Chesmut. 1986. Addendum to Oxidants and Asthmatics in Los Angeles: A
Benefit Analysis, prepared by EPA Office of Policy. Planning and Evaluation for US Environmental
Protection Agency.
Rowe, Robert D., and Lauraine G. Chestnut. 1983. Valuing Environmental Commodities: Revisited. Land
Economics.
Rowe. Robot D., and Lauraine G. Chestnut 198S. Economic Assessment of the Effects of Air Pollution on
Agricultural Crops in the San Joaquin Valley, Journal of the Air Pollution Control Association, Vol. 35, No.
7, pp. 728-734.
Rowe, Roben D., and Lauraine G. Chestnut 1986. Oxidants and Asthmatics in Los Angeles: A Benefits Analysis:
Addendum, prepared by Energy and Resource Consultants, Inc. for US Environmental Protection Agency.
.Rowe. Robert D., and Lauraine G. Chestnut 1986. Oxidants and Asthmatics in Los Angeles: A Benefits Analysis,
prepared by Energy and Resource Consultants. Inc. for US Environmental Protection Agency (OPPE).
Rowe, Roben D., and Lauraine G. Chestnut 1981. Visibility Benefits Asses
prepared by Abt Associates.
Guidebook - Interim Report.
Rowe. Roben D., R. A. Dutton, and L. G. Chestnut 1985. The Value of Groundwater Protection: Miami Case Study
Design and Pretest, prepared by Energy and Resource Consultants, Inc. for US Environmental Protection
Agency (OPA), Washington, DC.
Rowe. Roben D.. Ralph C. D'Arge, and David S. Brookshire. 1980. An Experiment on the Economic Value of
Visibility, Journal of Environmental Economics and Management, Vol. 7, pp. 1-19.
Rowe; Roben D., An M. Micbelson. and Edward R. Morey, 1989. Fishing for Atlantic Salmon in Maine: An
Investigation into Angler Activity and Management Options, prepared by RCG/Hagler, Bially Inc. for Bangor
Hydro-Electric Company (Fred Ayer).
, prepared by Royal Society Study Group for Royal Society Study
Royal Soc Study Grp. 1983. Risk Asses;
Group Report.
Range, Carlisle. 1983. Risk Assessment and Environmental Benefits Analysis, Natural Resources Journal.
Russell, Clifford S. 1982. Publicly Vested Values for Fish and Wildlife: Comment, Land Economics, Vol. 58, No. 4.
pp. 559-562.
Rutledge, Gary L.. and Nicolas A. Stergioulas. 1988. Plant and Equipment expenditures by Business for Pollution
Abatement, 1987 and Planned 1988, Survey of Current Business, pp. 26-29.
SMS Research. 1983. Experimental Valuation of Recreation Fishing in Hawaii, prepared by SMS Research for
National Marine Fisheries Service. Hawaii
Saliba, Bonnie Colby, and David B. Bush. 1987. Do Water Market Prices Appropriately Measure Water Values?,
Natural Resources Journal, Vol. 27, pp. 617-651.
Samples, K. C. 1980. An Economic Analysis of Integrated Fisheries Management The Case of the Lake Michigan
Alewife/SalmoQid Fisheries, University of Wisconsin.
Samples, K. C., and R. Bishop. 1985. Estimating the Value of Variations in Anglers' Success Rates: An ApUication
of the Multiple-Site Travel Cost Method, Marine Resource Economics, Vol. 2, pp. 155-174.
A-58
-------
Sassone, P., and W. Schaffer 1978. Cost-Benefit Analysis: A Handbook, Academic Press, Inc., New York, NY.
Sawicki, D. S., and L. B. ludd. 1983. Nonpoim Source Water Pollution Abatement and the Feasibility of Voluntary
Programs, Environmental Management.
Scberaga, Joel D. 1989. Supplemental Guidelines on Discounting in the Preparation of Regulatory Impact Analyses,
prepared by EPA Office of Policy, Planning and Evaluation for US Environmental Protection Agency.
Scheraga, Joel D., and Neil A. Leary. 1992. Improving the Efficiency of Policies to Reduce CO2 Emissions, Energy
Policy, pp. 394-404.
Schmalensee, R. 1972. Option Demand and Consumer's Surplus: Valuing Price Changes under Uncertainty,
American Economic Review, pp. 813-829.
Schmalensee, R. 1990. Option Value and Consumer Surplus: A Reply, American Economic Review, Vol. 65, pp.
737-739.
Schmalensee, R., R. Ranianathan, W. Ramm, and D. Smallwood. 1975. Measuring the External Effects of Solid
Waste Management, US EPA.
Schroth, Peter. 1982. The Impact of Envrionmentalism on Land-Use Control, American Journal of Comparative
Law.
Scbulze. W. D., D. S. Brookshire, E. G. Walter, and K. Kelley. 1981. The Benefits of Preserving Visibility in the
National Parklands of the Southwest: Volume 8 of Methods Development for Environmental Control Benefits
Assessment, US Environmental Protection Agency, University of Wyoming, Resource & Environ Economics
Lab.
Scbulze, William D., David Brookshire, and Mary Thayer. National Parks and Beauty: A Test of Existence Values,
US Environmental Protection Agency.
Schulze, William D., G. McClelland, P. Locke, and S. Elliott, et al. 1990. An Evaluation of Strategies for Promoting
Effective Radon Mitigation, prepared by University of Colorado for US Environmental Protection Agency
(EARS).
Schulze, William D., Ralph C. D'Arge, and David S. Brookshire. 1981. Valuing Environmental Commodities: Some
Recent Experiments, Land Economics, Vol. 57, No. 2, pp. 151-172.
Scbulze, William. David Brookshire. Eric G. Waltber, and Karen Kelly MacFariand, et al. 1983. Economic Benefits
of Preserving Visibility in the National Parklands of the Southwest, Natural Resources Journal, Vol. 23, No.
1, pp. 149-173.
Scbulze, William, G. McClelland, D. Brookshire, and D. Coursey. 1985. Improving Accuracy and Reducing Costs
of Environmental Benefit Assessments: VoL V Experimental Approaches for Measuring the Value of
Environmental Goods, prepared by University of Colorado for US Environmental Protection Agency
(EARB).
Scbulze, William, G. McClelland, D. Coursey, and B. Hard. 1987. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments - Risk Communication for Superfund Sites - An Analysis of Problems
and Objectives, prepared by University of Colorado for US Environmental Protection Agency (EARB).
Scbulze, William, Gary McClelland, Brian Hurd, and Joy Smith. 1985. Improving Accuracy and Reducing Costs of
Environmental Benefit Assessments, Volume IV: Estimating the Benefits for Toxic Waste Management: An
Application of the Property Value Method. Draft, prepared by University of Colorado for US Environmental
Protection Agency (OPA), University of Colorado, Center for Economic Analysis, Boulder, CO.
A-59
-------
Schulze, William, Thomas Crocker, Shaul Ben-David, and Allen Kneese. 1979. Methods Development for
Assessing Air Pollution Control Benefits. Vols 1,11,111, IV, & V, prepared by University of Wyoming;
University of New Mexico for US Environmental Protection Agency (EARB).
Schwartz, Joel, Vic Ha&selblad, and Hugh Pitcher. 1988. Air Pollution and Morbidity: A Further Analysis of the Los
Angeles Student Nurses Data, Journal of the Air Pollution Control Association, Vol. 38, pp. 158-162.
Sector Studies. 1987. Sector Study Profiles, prepared by Sector Studies.
Seigneur, Christian, and Pradeep Saxena. 1986. The SO2/Sulfate Relationship in the Northeastern United States,
prepared by Systems Applications, Inc. for US Environmental Protection Agency (EARB).
Sellar, Christine, John R. Stoll, and Jean-Paul Cbavas. 1985. Validation of Empirical Measures of Welfare Change:
A Comparison of Nonmarket Techniques, Land Economics, Vol. 61, No. 2, pp. 156-175.
Sellar, Christine, John R. Stoll, and Jean-Paul Chavas. 1986. Specification of the Logit Model: The Case of
Valuation of Non-market Goods, Journal of Environmental Economics and Management, VoL 13, pp. 382-
390.
Selvin, S.. D. MerriL L. Wong, and S. T. Sacks. 1984. Ecological Regression Analysis and the Study of the
Influence of Air Qualityon Mortality, Environmental Health Perspectives.
Seskin, Eugene, R. Anderson, and R. Reid. 1983. An Empirical Analysis of Economic Strategies for Controlling Air
Pollution, Journal of Environmental Economics and Management
Setia, P.. and R. Magleby. 1988. Measuring Physical and Economic impacts of Controlling Water Pollution in a
Watershed, Lake Reservoir Management. Vol. 4, No. 1, pp. 63-71.
, L., and Benelson. 1979. The Use of Development Value Estimates for Coastal Wetland Permit
Decisions, Land Economics, Vol. 55, No. 2.
L., S. S. Balk, and C. Mabbs-Zeno. 1979. The Economics of Wetland Preservation in Virginia, Journal of
the Norm Eastern Agricultural Economics Council, Vol. 8, No. 2.
Shabman, Leonard A., and Michael K. Bertelson. 1979. The Use of Development Value Estimates for Coastal
Wetland Permit Decisions, Land Economics, VoL 55. No. 2, pp. 213-222.
Shabman, Leonard A., and Sandra S. Batie. 1978. Economic Value of Natural Coastal Wetlands: A Critique, Coastal
Zooe Management Journal, VoL 4. No. 3. pp. 231-247.
Shapiro. Michael, and Ellen WarhiL 1983. Marketable Permits: The Case of CMorofluorocarbons, Natural
Resources Journal.
Shapiro, P. and T. Smith. 1982. Preferences for Non-Market Goods Revealed through Market Demands. In
Advances in Microeconomics, ed. Smith, V. 1C, JAI Press, Inc. Greenwich, CT. pp. 105-122.
Sharefkm, M., M. Sbecter, and A. Kneese. 1984. Impacts. Costs, and Techniques for the Mitigation of Contaminated
Groundwater. Water Resources Research.
Snarefkm. Mark F.
p^^fi^ frpm Control of Major Et""*nnr"ffttl F|WKH*ff,
Resources for the Future for US Environmental Protection Agency (EARB).
Shaw, Daigee. 1988. On-site Samples' Regression: Problems of Non-negative Integers, Truncation, and Endogenous
Stratification. Journal of Econometrics, Vol. 37. pp. 21 1-24.
Shaw, Douglas W. 1989. Valuing the Effect of Acidification Damages on the Adirondack Fishery: Comment,
American Journal of Agricultural Economics, pp. 217-220.
I*
I
1
1
1
I
1
f
3
3
J
1
A-60
-------
t
Shechter, Mordechai. 198S. Economic Aspects in the Investigation of Groundwater Contamination Episodes,
Groundwaier, Vol. 23, pp. 190.
Shechter, Mordechai. 1985. An Anatomy of a Groundwater Contamination Episode, University of Haifa, Journal of
Environmental Economics and Management, Vol. 12, pp. 72-88.
Shelby, Bo. 1980. Crowding Models for Backcountry Recreation, Land Economics, Vol. 56, pp. 43-55.
Shopley, J. B., and R. F. Fuggle. 1984. A Comprehensive Review of Current Environmental Impact Assessment
Methods and Techniques, Journal of Environmental Management.
Sbortle. James S., and Keith D. Willett 1986. The Incidence of Water Pollution Control Costs: Partial vs. General
Equilibrium Computations, Growth and Change, Vol. 17, No. 2, pp. 32-43.
Shulstad, R. N., and H.H Stovener. 1978. The Effects of Mercury Contamination in Pheasants on the Value of
Pheasant Hunting in Oregon, Land Economics, Vol. 54, pp. 39-48.
Sbupe, Steven J., and Nancy Zeilig. 1988. Face to Face - Water Marketing: An Overview, prepared by Shupe &
Associates, Santa Fe, NM for American Water Works Association, Journal of American Water Works
Association, Vol. 80, No. 3, pp. 18-26.
Sim. Steven R. 1991. How Environmental Costs Impact DSM (Demand Side Management), Public Utilities
Fortnightly, Vol. 128, pp. 24-27.
Sinden, J. A., and G. K. Windsor. 1981. Estimating the Value of Wildlife for Preservation: A Comparison of
Approaches, Journal of Environmental Management, Vol. 12.
Sinden, J. A. 1991. Valuation of Unpriced Benefits and Costs of River Management, prepared by Office of Water
Resources, Victorian DepL of Conservation for Melbourne, Australia.
Singh, Jasbinder, Raffael Stein, and Brett Snyder. 1988. Municipal Sector Study: Impacts of Environmental
Regulations on Municipalities, US Environmental Protection Agency.
Sloggett, Gordon. 1981. Prospects for Groundwater Irrigation: Declining Levels and Rising Energy Costs, US Forest
Service, pp. 43.
Sloggett, Gordon, and H. P. Mapp, Jr. 1984. An Analysis of Rising Irrigation Costs in the Great Plains, Water
Resources Bulletin, Vol. 20, No. 2, pp. 229-233.
Small Kenneth A. 1975. Air Pollution and Property Values: Further Comment, Review of Economics and Statistics,
Vol. 57, No. 1, pp. 105-107.
Smiu R. J.. and W. A. Chapin. 1983. Cost of Clean Water- Impact on Small Communities, Journal of Water
Pollution Control Federation.
Smith, A. E., and K. L. Bnibaker. 1983. Costs and Air Quality Impacts of Alternative National Ambient Air Quality
Standards for Paniculate Matter.
Smith, V. Kerry. Option Value: A Review and Assessment, Southern Economic JoumaL
Smith, V. Kerry. 1981. Congestion, Travel Cost Recreational Demand Models, and Benefit Evaluation, Journal of
Environmental Economics and Management, Vol. 8, No. 1, pp. 92-96.
Smith, V. Kerry. 1988. Resource Evaluation at the Crossroads, Resources for the Future, pp. 2-6.
Smith, V. Kerry. 1982. Introduction to Advances in Applied Microeconomics and Some Perspectives. In Advances
in Microeconomics, ed. Smith, V. K., JAI Press, Inc., Greenwich, CT, pp. 1-11.
A-61
-------
Smith. V. Kerry. 1987. Nonuse Values in Benefit Cost Analysis, Southern Economic Journal
Smith, V. Kerry. 1988. Selection and Recreation Demand, American Journal of Agricultural Economics, Vol. 70,
pp. 29-36.
Smith, V. Kerry. 1987. Benefit Estimation and Recreation Policy, Policy Studies Review, Vol. 7, No. 2, pp. 432-
442.
Smith, V. Kerry. 1975. The Estimation and Use of Models of the Demand for Outdoor Recreation. In Assessing the
Demand for Outdoor Recreation, National Academy of Sciences, Washington, DC.
Smith, V. Kerry. Hie Role of Site and Job Characteristics in Hedonic Wage Models, Journal of Urban Economics.
Smith. V. Kerry. 197S. Travel Cost Demand Models for Wilderness Recreation: A Problem of Non-Tested
Hypothesis, Land Economics, Vol. 51, pp. 103-111.
Smith, V. Kerry. 1976. The Economic Conseqi
MA.
of Air Pollution, Ballinger Publishing Company, Cambridge,
Smith, V. Kerry. 1989. Recreational Benefits Transfer Project, prepared by North Carolina State University for US
Environmental Protection Agency (EARS).
Smith, V. Kerry. 1987. Uncertainty, Benefit-Cost Analysis, and the Treatment of Option Value, Journal of
Environmental Economics and Management, Vol. 14, pp. 283-292.
Smith, V. Kerry. 1984. A Bound for Option Value, Land Economics.
Smith, V. Kerry, Edward R. Morey, and Robert D. Rowe. 1985. AERE Workshop on R
Demand Modeling,
prepared by Association of Environmental ft Resource Economists for US Environmental Protection Agency.
Smith, V. Kerry, and John V. Knitilla. 1976. Structure and Properties of a Wildmess Travel Simulator, Johns
Hopkins University Press, Baltimore, MD.
Smith, V. Kerry, and John V. Knitilla. 1979. Endangered Species. Irreversibihoes, and Uncertainty: A Comment,
American Journal of Agricultural Economics, Vol. 61, No. 2, pp. 371*375.
Smith. V. Kerry, and Raymond B. Palmquist. 1989. The Value of Recreational Fishing on the Ablemarle and
Palmico Estuaries, US Environmental Protection Agency.
Smith, V. Kerry, and Raymond J. Kopp. 1980. The Spatial Limits of toe Travel Cost Recreation Demand Model,
Land Economics, Vol. 56, No. 1. pp. 64-72.
Smith, V. Kerry. W. H. Desvousges, and A. M. Freeman. 1983. The Benefits of Hazardous Wastes Management
Regulation Using Contingent Valuation: A Cooperative Agreement Application.
Smith, V. Kerry, and William H. Desvousges. 1986. The Value of Avoiding a Lulu: Hazardous Waste Disposal
Sites, Review of Economics and Statistics, Vol. 68, No. 2, pp. 293-299.
Smith, V. Kerry, and William H. Desvousges. 1987. An Empirical Analysts of toe Economic. Value of Risk
Changes, US Environmental Protection Agency, Journal of Political Economy, VoL 95, No. 1.
Smith, V. Kerry, and William H. Desvousges. 1985. The Generalized Travel Cost Model and Water Quality
Benefits: A Reconsideration. Southern Economic Journal, Vol. 52, No. 2, pp. 371-381.
Smith, V. Kerry, and William H. Desvousges. 1986. Measuring Water Quality Benefits, Kluwer-Nijhoff Publishing,
Boston, MA.
A-62
-------
f
Smith, V. Kerry, William H. Desvousges, and A. Myrick Freeman. 1985. Valuing Changes in Hazardous Waste
Risks: A Contingent Valuation Analysis - Vol. I, II, ni, prepared by Vanderbilt University; Research
Triangle Institute for US Environmental Protection Agency (EARB).
Smith, V. Kerry, William H. Desvousges, and Ann Fisher. 1986. A Comparison of Direct and Indirect Methods for
Estimating Environmental Effects, American Journal of Agricultural Economics, Vol. 68, No. 2, pp. 25-37.
Smith, V. Kerry, William H. Desvousges, and Matthew McGivney. 1983. Estimating Water Quality Benefits: An
Econometric Analysis, Southern Economic Journal, Vol. 50, No. 2, pp. 422-437.
Smith, V. Kerry, William H. Desvousges, and Matthew McGivney. 1983. The Opportunity Cost of Travel Time in
Recreation Demand Models, Land Economics, Vol. 59, No. 3, pp. 259-278.
Smith, V. Kerry, and Yosjaki Kaoru. 1988. Signals or Noise? Explaining the Variation in Environmental Benefit
Estimates, prepared by Department of Economics for North Carolina State University.
Smith, V. Kerry, and T. Deyak. 1975. Measuring the Impact of Air Pollution on Property Values, Journal of
Regional Science. Vol. 15, No. 3, pp. 277-288.
Smith, V. Kerry. 1989. Taking Stock of Progress with Travel Cost Recreation Demand Models: Theory
Implementation, Marine Resource Economics, Vol. 6, pp. 279-310.
Smith, V. Kerry, Raymond B. Palmquist, and Paul Jakus. 1990. Combining Farrel Frontier and Hedonic Travel Cost
Models for Valuing Estuarine Quality.
Smith, Anne £., Anders R. Gjerde, Lynn I. DeLain, and Ray R. Zhang. 1992. CO2 Trading Issues, prepared by
Decision Focus, Inc. for US EPA - Global Climate Change Division.
Smith, V. Kerry. 1992. Environmental Costing for Agriculture: Will it be Standard Fare in the Farm Bill for 2000?,
prepared by Resorces for the Future.
Sorenson, Jay. 1984. The Assurance of Reasonable Toxic Risk?, Natural Resources journal.
Sorg, Cindy F. 1982. Valuing Increments and Decrements of Wildlife Resources: Further Evidence, University of
Wyoming, Department of Economics.
Sorg, Cindy F., and John Loomis. 1984. Empirical Estimates of Amenity Forest Values: A Comparative Review,
prepared by Rocky Mountain Forest and Range Land Experiment Station for US Forest Service, Fort Collins,
CO.
Sorg, Cindy F., and John Loomis. 1986. Economic Value of Idaho Sports Fisheries with an Update on Valuation
Techniques, North American Journal of Fisheries Management, Vol. 6, pp. 494-503.
Sorg, Cindy F., John Loomis. Dennis Donnelly, and George L. Peterson, et al. 1985. Net Economic Value of Cold
and Warm Water Fishing in Idaho, prepared by Rocky Mountain Forest and Range Land Experiment Station
for US Forest Service, Fort Collins. CO.
Sorg, Cindy F., and Louis J. Nelson. 1986. Net B
Value of Elk Hunting in Idaho, prepared by Rocky
Mountain Forest and Range Land Experiment Station for US Forest Service.
St Oair. Ann, M. McCtoskey, and J. Sherman. 1982. Development of a Framework for Evaluating Cost-
Effectiveness of Remefial Actions at Uncontrolled Hazardous Waste Sites.
Stankey. George H. 1972. A Strategy for the Definition and management of Wilderness Quality. In Natural
Environments, ed. Krutilla, J., Johns Hopkins University Press, Baltimore, MD.
A-63
-------
Steinnes, Donald, and Raymond Raab. 1981. The Economics of a "Happening": Spring Smelt Fishing on Lake
Superior.
Stenstrant, M.. H. Vazirinejad, and A. Ng. 1984. Economic Evaluation of Upgrading Aeration Systems, Journal of
Water Pollution Control Federation.
Stevens, Joe. 1984. Satisfaction with Environmental Change: An Empirical Analysis of Attitudes toward Air Quality
by Recent Interstate Migrants. Journal of Environmental Economics and Management
Stevens, Joe. 1966. Angler Success as a Quality Determinant of Sports Fishery Recreational Values, Transactions of
the American Fisheries Society, Vol. 95, pp. 357-362.
Stevens, Thomas H., and P. G. Allen. 1980. Estimating the Benefits of Recreation Under Conditions of Congestion,
Journal of Environmental Economics and Management, Vol. 7, pp. 395-400.
*
Stoll. John R., and L. A. Johnson. 1984. Concepts of Value, Nonmarket Valuation and the Case of the Whooping
Crane, prepared by Wildlife Management Institute for Wildlife Management Institute, Washington, DC. pp.
382-393.
Stoll, John R.. Lansing!) S. Freeman, John C. Bergstrom, and Jim Henderson. 1991. Annotated Bibliography for
Regional Recreational Demand Models, prepared by US Army Corps of Engineers for US Government
Printing Office, Washington, DC.
Strait. Randy, and William Battye. 1989. Cost Assessment of Regulatory Alternatives for lead national ambient air
quality standards, US Environmental Protection Agency.
Strauss, R. O., and G. D. Hughes. 1976. A New Approach to the Demand for Public Goods. Journal of Political
Economy. Vol. 6. No. 3. pp. 191-204.
Stroeh. Dietrich J. 1986. After Drought - Does Price Reflect the Value of Water, American Water Works
Association.
Strong. Elizabeth J. 1983. A Note on the Functional Form of Travel Cost Models With Zones of Unequal
Populations. Land Economics, Vol. 59, No. 3, pp. 342-349.
Stynes, D. J., George L. Peterson, and Donald H. Rosenmal. 1986. Log Transformation Bias in Estimating Travel
Cost Methods, Land Economics. Vol. 62, No. 1, pp. 94-103.
Supalla, R. J. 1982. Economic Evaluations of Groundwater Policy Alternatives in the Northern Great Plains, NT1S.
Sussman, Frances. 1984. A Note on the Willingness to Pay Approach to the Valuation of Longevity. Journal of
Environmental Economics and Management
Sutherland, Ronald J. 1983. A Regional Recreation Demand and Benefits Model Los Alamos National Laboratory,
Los Alamos, NM.
Sutherland, Ronald J. 1982. The Sensitivity of Travel Cost Ffljpy*** of Recreation Demand to the Functional Form
and Definition of Origin Zones, Western Journal of Agricultural Economics, Vol. 7, No. 1. pp. 87-98.
Sutherland, Ronald J. 1982. A Regional Approach to Estimation of R
Benefits of Improved Water Quality,
Journal of Environmental Economies' and Management, VoL 9. No. 3, pp. 229-247.
Sutherland, Ronald J., and Richard G. Walsh. 1985. Effect of Distance on the Preservation Value of Water Quality.
Land Economics, Vol. 61, No. 3. pp. 281-291.
Swartzman, D., K. Croke, and S. Swibel. 1985. Reducing Aversion to Living near Hazardous Waste Facilities
through Compensation and Risk Reduction, Journal of Environmental Management, Vol. 20, pp. 43-50.
A-64
-------
Systems App, Inc. 1988. Economic Impacts of Acidic Deposition on Exterior Residential Surfaces: Comments on
Existing Studies and suggested Refinements, prepared by Systems Applications, Inc.
Tabler, S. K. 1984. EPA's Program for Establishing National Emission Standards for Hazardous Air Pollutants,
Journal of the Air Pollution Control Association.
Talbelm, Daniel R. 1978. A General Theory of Supply and Demand for Outdoor Recreation in Recreation Systems,
Michigan State University.
Talbelm, Daniel R. 1976. The Demand and Supply of Fishing and Boating on Inland Lakes in Michigan: Summary
Report, prepared by Michigan Department of Natural Resources.
Talhelm, Daniel R. 1987. Product Travel Cost Approach: Estimating Acid Rain Damage to Sport Fishing in Ontario,
Vol. 116,pp.42(W31.
Tanacredi, I. T. 1983. Coastal Zone Managenment Practices at an Urban National Park, Environmental
Management.
Terkla, David. 1984. The Efficiency Value of Effluent Tax Tegulations, Journal of Environmental Economics and
Management, Vol. 11, No. 2, pp. 107-123.
Thaler, H. T. 1984. Nonparametric Estimation of the Hazard Ratio, Journal of American Statistical Association.
Thayer, M. 1981. Contingent Valuation Techniques for Assessing Environmental Impacts: Further Evidence.
Journal of Environmental Economics and Management, Vol. 8, pp. 27-44.
Thayer, M., and W. Schulze. 1977. Valuing Environmental Quality: A Contingent Substitution and Expenditure
Approach, University of Southern California.
Thibodeau. F. R., and Ban Ostro. 1981. An Economic Analysis of Wetland Protection, Journal of Environmental
Management, Vol. 12, No. 1.
Thomas, John F., and Geoffrey J. Syme. 1988. Estimating Residential Price Elasticity of Demand for Water: A
Contingent Valuation Approach, Water Resources Research. Vol. 24, No. 11, pp. 1847-1857.
Thomas, John F., and William E. Martin. 1987. Mining of Aquifers Near Metropolitan Areas: Towards a General
Framework for Policy Analysis.
Thomas. V. 1983. Welfare Analysis of Pollution Control with Spatial Alternatives, Urban Studies.
Thompson, Mark A. 1987. Desalination of Groundwater in Suffolk. City of Suffolk, Virginia.
Thompson, Mark, and Herbert Stoevener. 1983. Estimating Residential Flood Control Benefits Using Implicit Price
Euations, Water Resources Bulletin.
Thomson, Cynthia J., and Daniel D. Huppen. 1987. Results of the Bay Area Sportfisb Economic Study (Bases),
National Oceanograpbic and Atmospheric Administration.
Tieraan, J. E. 1983. A Rational Evolution of Cancer Mortality Associations with Treated Drinking Water, Journal of
Environmental Health.
Titus, James G. 1986. Effects of Changes in Stratospheric Ozone and Global Climate Change.Volume 1 - an
Overview, prepared by EPA and UNEP for US Environmental Protection Agency.
Tobey, James A. 1990. The Effects of Domestic Environmental Policies on Patterns of World Trade: An Empirical
Test, Vol. 43, No. 2, pp. 191-209.
A-65
-------
Tolley, George, A. Randall, G. Bloomquist, and M. Brien, et al. 1986. Establishing and Valuing the Effects of
Improved Visibility in Eastern United States, prepared by University of Chicago for US Environmental
Protection Agency (OPA).
Tolley, George, and Lyndon Babcock. 1986. Valuation of Reductions in Human Health Symptoms and Risks - Vols.
I, II, HI, & IV. prepared by University of Chicago for US Environmental Protection Agency (EARB).
Traubennan, J. 1983. Statutory Reform of "Toxic Tons": Relieving Legal, Scientific, and Economic Burdens on the
Chemical Victim, Harvard Environmental Law Review.
Trijonis, John. 1987. National Relationship Between Visibility and NOx Emissions, prepared by Santa Fe Research
Corp. for US Environmental Protection Agency, Bloomington, MM.
Tschirhan, John, and Thomas D. Crocker. 1983. Integrating Ecosystems and Economics, prepared by EPA
Economic Analysis and Research Branch for US Environmental Protection Agency. University of Wyoming.
Turner. Monica, Eugene P. Odum, Robert Costanza, and Thomas Springer. 1988. Market and Nonmarket Values of
the Georgia Landscape, Environmental Management, Vol. 12. No. 2, pp. 209-217.
Turner, R. K.. D. Dent, and R. D. Hey. 1983. Valuation of the Environmental Impact of Wetland Rood Protection
and Drainage Schemes, Environment and Planning, Vol. 15, No. 4.
Tumovsky, Steven. 1969. The Demand for Water Some Empirical Evidence on Consumers' Response to a
Commodity Uncertain in Supply, Water Resources Research, VoL 5, No. 2. pp. 350-361.
US DOI. 1991. Natural Resource Damage Assessments; Notice of Proposed Rulemaking, prepared by US
Department of Interior for US Department of Interior, US Government Printing Office, Washington, DC, Vol.
56, No. 82. pp. 19752-19770.
US Senate. 1974. A Legislative History of the dean Air Amendments of 1970, prepared by US Senate.
US ACE. 1990. Preliminary Field Report on the Recreation Interviews Conducted in Support of the Rockaway Inlet
to Norton Point (Coney Island Area) Study. US Army Corps of Engineers.
USEPA. 1989. Estimating and Valuing Morbidity in a Policy Context. US Environmental Protection Agency.
USEPA. 1986. AERE Workshop on Marine Pollution and Environmental Damage Assessment - Final Report, US
Environmental Protection Agency.
USEPA. 1987. Economic Impact Analysis of Effluent Limitations Guidelines and Standards for the Organic
Chemicals, Plastics, and Synthetic Fibers Industry, US Environmental Protection Agency.
USEPA. 1987. Regulatory Impact Analysis of Proposed Hazardous Waste Disposal Restrictions for Class I Injection
of California List Wastes, prepared by Cadmus Group, Inc. for US Environmental Protection Agency.
USEPA. 1988. Preliminary Cost and Economic Impact Analysis for Listing Wood Preserving Wastes, US
Environmental Protection Agency.
USEPA. 1988. National Assessment of VCC CO, and NOx Controls, Emissions and Costs, prepared by EPA Office
of Policy, Planning and Evaluation for US Environmental Protection Agency.
USEPA. 1990. Environmental Accounting - (draft), prepared by EPA Economic Analysis and Research Branch for
US Environmental Protection Agency.
USEPA. 1989. Cost-Effectiveness Analysis of the Effluent Limitations Guidelines and Standards for be Notice of
Data Availability for Drilling Fluids and Drill Cuttings for the Offshore Oil and Gas Industry.
A-66
-------
USEPA. 1989. Improved Economic Efficiency in Environmental Rulemaking - Final Repon, prepared by
Environmental Law Institute for US Environmental Protection Agency (EARB).
USEPA. 1990. Markets for Selected Postconsumer Waste Paper Grades - Draft, US Environmental Protection
Agency.
USEPA. 1984. Cost-Benefit Analysis can be Useful in Assessing Environmental Regulations, Despite Limitations,
1 US Environmental Protection Agency.
USEPA. 1988. Regulatory Impact Analysis in Support of the Proposed Revisions to the National Oil and Hazardous
Substances Pollution Contingency Control Plan, US Environmental Protection Agency.
USEPA. 1990. Comparision of the Economic Impacts of of toe Acid Rain provisions of the Senate Bill (S. 1630)
I and the House Bill (S.1630). prepared by EPA Office of Policy Analysis for US Environmental Protection
i Agency.
I USEPA. 1989. Regulatory Impact Analysis in Support of a Permanent Reporting Threshold under Sections 311 and
312 of the Emergency Planning and Community Right-to-know Act, US Environmental Protection Agency.
USEPA. 1990. Regulatory Impact Analysis of the Land Disposal Restrictions for Third Third Scheduled Wastes-,
i Final Form, US Environmental Protection Agency.
USEPA. 1988. Regulatory Impact Analysis of the Land Disposal Restrictions on First Third Wastes - Final Report,
. US Environmental Protection Agency.
* USEPA. 1988. Regulatory Impact Analysis for the Proposed Rulemaking on Corrective Action For Solid Waste
Management Units - Draft, US Environmental Protection Agency.
I USEPA. 1987. Regulatory Impact Analysis in Support of Final Rulemaking Under Sections 311 and 312 of the
^_ Superfund Amendments and Reauthorizalion Act of 1986, US Environmental Protection Agency.
| USEPA. 1988. Regulatory Impact Analysis in Support of Final Rulemaking under Section 313 of Title HI of me
I Superfund Amendments and Reauthorizalion Act of 1986, US Environmental Protection Agency.
1 USEPA. 1987. Regulatory Impact Analysis in Support of Rulemaking Under Sections 302.303, and 304 of Tide m
of the Superfund Amendments and Reauthorization Act of 1986, US Environmental Protection Agency.
USEPA. 1988. Regulatory Impact Analysis of Controls on Asbestos and Asbestos Products, US Environmental
| Protection Agency.
USEPA. 1984. Analysis of Alternative Emission Reduction Strategies: Four/Eigbt/Twelve Million Ton Reductions
r and Ten/Twelve State Reductions, US Environmental Protection Agency.
' USEPA. 1984. Analysis of Alternative Sulfur Dioxide Ambient Standards, US Environmental Protection Agency.
| USEPA. 1984. Economic Evaluation of Achieving Environmental Objectives Through Sulfur Dioxide Emission
i Reductions, US Environmental Protection Agency.
USEPA. 1983. Regulatory Impact Analysis of the Revisions to the national and Hazardous Substances Pollution
Contigency Plan, US Environmental Protection Agency.
USEPA. 1985. Regulatory Impact Analysts of Reportabfc Quantity Adjustments under Sections 102 and 103 of the
Comprehensive Environmental Rsponse. Compensation, and Liability Act (three volumes), US
Environmental Protection Agency.
A-67
-------
USEPA. 1991. Potential Benefits of Preventing Contamination of Wetlands due to Releases from Hazardous Waste
Facilities: Analysis of Selected Sites, prepared by 1CF, Inc. for US Environmental Protection Agency
(OSWER).
USEPA. 1988. Regulatory Impact Analysis of Technical Standards for Underground Storage tanks. US
Environmental Protection Agency.
USEPA. 1986. Regulatory Analysis of Proposed Restrictions on Land Disposal of Certain Solid Wastes, US
Environmental Protection Agency.
USEPA. 1986. Regulatory Analysis of Restrictions on Land Disposal of Certain Dioxin-Containing Wastes,
prepared by EPA Office of Solid Waste and Emergency Response for US Environmental Protection Agency.
USEPA. 1985. Regulatory Impact Analysis of Proposed Regulation of Glycol Ethers, US Environmental Protection
Agency.
USEPA. 1983. Estimated Human Health Benefits from Chemical Plant Effluent Control in Two Regions, prepared
by EPA Economic Analysis and Research Branch for US Environmental Protection Agency, Cambridge, MA.
USEPA. 198S. Regulatory Analysis of Proposed Restrictions on Land Disposal of Hazardous Wastes, US
Environmental Protection Agency.
USEPA. 1985. A Methodological Approach to an Economic Analysis of me Beneficial Outcomes of Water Quality
Improvements from Sewage Treatment Plant Upgrading and combined Sewer Overflow Controls, prepared
by Meta Systems Inc. for US Environmental Protection Agency (OPA).
USEPA. 1987. Methodology for Valuting the Health Risks of Ambient Lead Exposure, prepared by Metatech, Inc.
for US Environmental Protection Agency.
USEPA. 1989. Improving Risk Communication, prepared by National Research Council for US Environmental
Protection Agency (EARB).
USEPA. 1985. Value of Groundwater in Regional Cases, prepared by Policy. Planning and Evaluation, Inc. for US
Environmental Protection Agency.
USEPA. 198S. Regulatory Impact Analysis of the Proposed Rule on the Use of Nitrites in Metalworking Fluids,
prepared by Putnam. Hayes, and Bartleu, Inc. for US Environmental Protection Agency.
USEPA. 1990. Ecosystems and Their Valuation, prepared by RCG/Hagler, Baflly, Inc. for US Environmental
Protection Agency (EARB).
USEPA. 1989. The Effects of Weight- or Vouhne-Based Pricing on Solid Waste Management, prepared by
Research Triangle Institute, me. for US Environmental Protection Agency (OPPE), Research Triangle
Institute, Research Triangle Park, NC.
USEPA. 1988. Economic Impact of Air Emissions Regulations: Waste Solvent Recycling, prepared by Research
Triangle Institute, Inc. for US Environmental Protection Agency.
USEPA. 1990. Mexico's Strategy on Ozone Layer Protection: A Case Study on the Costs of Implementing the
Montreal Protocol, prepared by Secretaria de DesarroUo Urbano y Ecologia (Mexico).
USEPA. 1981. Benefit Model for Pollution Effects on Material, prepared by TRC. Inc. for US Environmental
Protection Agency.
USEPA. 1982. Valuing the Benefits of Improvements in Drinking Water Quality: Theory and Practice * Draft,
prepared by Temple. Barker & Stoane for US Environmental Protection Agency.
A-68
-------
1
I
1
1
I
t
I
I
I
1
I
USEPA. 1987. Regulatory Impact Analysts of Proposed Revisions to Subtitle D Criteria for Municipal Solid Waste
Landfills, prepared by Temple, Barker & Sloane; ICF; Pope-Reid; AMS for US Environmental Protection
Agency.
USEPA. 1987. Unfinished Business: A Comparative Assessment of environmental Problems, prepared by
Environmental Protection Agency for US Environmental Protection Agency.
USEPA. 1989. Methods for Evaluating the Attainment of Cleanup Standards Volume 1: Soils and Solid Media,
prepared by EPA Office of Policy, Planning and Evaluation for US Environmental Protection Agency.
USEPA. 1989. Control of Air Pollution from New Motor Vehicles and New Motor Vehicle Engines: Evaporative
Emission Regulations for Gasoline and Methanol-fueled Light-Duty Vehicles, L-D Trucks, and H-D
Vehicles, US Environmental Protection Agency.
USEPA. 1987. Regulatory Impact Analysis of Restrictions on Land Disposal of California List Wastes, prepared by
EPA Office of Solid Waste and Emergency Response for US Environmental Protection Agency.
USEPA. 1990. Toxicity Characteristic Regulatory Analysis, prepared by EPA Office of Solid Waste and Emergency
Response for US Environmental Protection Agency.
•,
USEPA. 1987. RIA for Sections 322 and 323 of the Superfund amendments- Title m (incomplete), US
Environmental Protection Agency.
USEPA. The Cost of Clean Air and Water Report to Congress, prepared by DPRA, Inc. and Others for US
Environmental Protection Agency.
USEPA. 1985. Regulatory Impact Analysis of Proposed Standards for the Management of Used Oil, prepared by
Temple, Barker & Sloane for US Environmental Protection Agency.
USEPA. 1988. Regulatory Impact Analysis on the National Ambient Air Quality Standards for Sulfur Oxides,
prepared by EPA Air Quality Management Division for US Environmental Protection Agency.
USEPA. 1989. Review of the National Ambient Air Quality Standards for Lead: Exposure Analysis Methodology
and Validation, prepared by EPA Office of Air Quality Planning and Standards for US Environmental
Protection Agency.
USEPA. 1990. Analysis of the Economic and Environmental Effects of Compressed Natural Gas as a Vehicle Fuel
prepared by EPA Office of Mobile Sources for US Environmental Protection Agency.
USEPA. 1982. Regulatory Impact Analysis: Data Requirements for Registering Pesticides under the Federal
Insecticide, Fungicide, and Rodentcide Act, prepared by EPA Economic Analysis and Research Branch for
US Environmental Protection Agency.
USEPA, 1982. Economic Analysis of Regulations Implementing Certain Portions of FTFRA, section 3, Concerning
Registration of Pesticides, prepared by EPA Economic Analysis and Research Branch for US Environmental
Protection Agency.
USEPA. 1985. Costs and Benefits of Reducing Lead in Gasoline: Final Regulatory Analysis, prepared by EPA
Economic Analysis and Research Branch for US Environmental Protection Agency.
USEPA. 1982. Development Document for Effluent Limitations on Guidelines and Standards for the Pulp, Paper.
and Paperboard and the Builder's paper and Board Mills. Point Source Categories, US Environmental
Protection Agency.
USEPA. 1984. Guidance Manual for Paper, Pulp, and Paperboard and Builder's Paper and Board Mills Pretreatmem
Standards, US Environmental Protection Agency.
A-69
-------
USEPA. 1986. Ecoregions of the Pacific Northwest, US Environmental Protection Agency.
USEPA. 1989. Natural Resources for the 21st Century: An Evaluation of the Effects of Land Use on Environmental
quality, US Environmental Protection Agency.
USEPA. 1987. Development Document for Effluent Limitatitons Guidelines and Standards for the Organic
Chemicals, Plastics and Synthetic Fibers; Point Source Category, US Environmental Protection Agency.
USEPA. 1987. EPA's Use of Benefit-Cost Analysis: 1981-1986, prepared by EPA Office of Policy, Planning and
Evaluation for US Environmental Protection Agency.
USEPA. 1988. Guidelines for Performing Regulatory Analysis: with Appendices A.B.CS>, prepared by EPA Office
of Policy, Planning and Evaluation for US Environmental Protection Agency.
USEPA. 1988. Environmental Progress and Challenges: EPA's Update, prepared by EPA Office of Policy, Planning
and Evaluation for US Environmental Protection Agency.
USEPA. 1991. Using Environmental Indicators for Surface Water Quality Planning and Management: Region 5
Pilot Study, prepared by EPA Office of Policy, Planning and Evaluation for US Environmental Protection
Agency.
USEPA. 1976. National Benefits of Acheiving the 1977,1983, and 1985 Water Quality Goals, prepared by DPRA.
Inc. for US Environmental Protection Agency (ORD).
USEPA. 1982. Benefits Analysis of Alternative Secondary National Air Quality Standards for Sulfur Dioxide and
Total Suspended Particulates: Vols: I-VI, prepared by Mathtec, Inc. for US Environmental Protection Agency
(OAQPS).
USEPA. 1989. Control on Sulfur and Aromatics Contents of On-Highway Diesel Fuel (draft), prepared by EPA
Office of Mobile Sources for US Environmental Protection Agency.
USEPA. 1989. Analysis of the Economic and Environmental Effects of Using Methanol as an Automotive Fuel,
prepared by EPA Office of Mobile Sources for US Environmental Protection Agency.
USEPA. 1987. Regulatory Impact Analysis: Protection of Stratospheric Ozone layer, prepared by EPA Office of Air
and Radiation (SPP) for US Environmental Protection Agency.
USEPA. 1989. Hazardous Waste TSDF - Regulatory Impact Analysis for proposed RCRA Air Emission Standards,
prepared by EPA Office of Air Quality Planning and Standards for US Environmental Protection Agency.
USEPA. 1990. National Air Quality and Emissions Trends Report, 1988. prepared by EPA Office of Air Quality
Planning and Standards for US Environmental Protection Agency.
USEPA. 1982. Regulatory Impact Analysis of the National Ambient Air Quality Standards for Carbon Monoxide
(draft), prepared by EPA Office of Air Quality Planning and Standards for US Environmental Protection
Agency.
USEPA. 1987. Regulatory Impact Analysis: Listing of Surface Coal Mines for Source Review, prepared by EPA
Office of Air Quality Planning and Standards for US Environmental Protection Agency.
USEPA. 1QQQ Ffrial Regulatory Impact Analysis tnd Stirnffnupy *nfl Analyst of <>»mwim
Volatility Regulations, prepared by EPA Office of Mobile Sources for US Environmental Protection Agency.
USEPA. 1990. Regulatory Impact Analysis: Control of Sulfur and Aromatics Contents of On-highway Diesel Fuel,
prepared by EPA Office of Mobile Sources for US Environmental Protection Agency.
A-70
-------
f
USEPA. 1990. Final Regulatory Impact Analysis and Summary and Analysis of Comments on the NPRM - Interim
Control of Gasoline Volatility, prepared by EPA Office of Mobile Sources for US Environmental Protection
Agency.
USEPA. 1987. Draft RIA - Control of Gasoline Volatility and Evaporative Hydrocarbon Emissions from New Motor
Vehicles, prepared by EPA Office of Mobile Sources for US Environmental Protection Agency.
USEPA. 1985. Economic Impact Analysis of Proposed Regulations to Control Volatile Synthetic Organic Chemicals
in Drinking Water, prepared by EPA Office of Drinking Water for US Environmental Protection Agency.
USEPA. 1988. Estimating Exposure to 2,3,7,8 - TCDD (Draft), US Environmental Protection Agency.
USEPA. 1980. Regulatory Analysis and Environmental Impact of Final Emission Regulations for 1982 and 1983
model year High-Altitude Motor Vehicles, prepared by EPA Office of Mobile Sources for US Environmental
Protection Agency.
USEPA. 1979. Regulatory Analysis and Environmental Impact of Final Emission Regulations for 1984 and Later
Model Year Heavy Duty Engines, prepared by EPA Office of Mobile Sources for US Environmental
Protection Agency.
USEPA. 1987. Regulatory Impact Analysis: Worker Protection Standards For Agricultural Pesticides, prepared by
EPA Office of Pesticide Programs for US Environmental Protection Agency.
USEPA. 1981. Measuring and Comparing the Cost-effectiveness of EPA Regulatory Efforts to Control Toxics-
related Health Risks. Vol. I: Feasibility study. Vol. II. Cost-effectiveness Analysis of EPA Intermedia Prior,
prepared by EPA Office of Pesticides and Toxic Substances for US Environmental Protection Agency.
USEPA. 1987. Regulatory Impact Analysis in Support of Proposed Rulemaking Under Sections 322-323 of the
Superfund Amendments and the Reautborization Act of 1986. prepared by EPA Office of Pesticides and
Toxic Substances for US Environmental Protection Agency.
USEPA. 1985. Final Regulatory Impact Analysis - 40 CFR Pan 191: Environmental Standards for the Management
and Disposal of Spent Nuclear Fuel, High-level and Transuranic Radioactive wastes, prepared by EPA Office
of Radiation Programs for US Environmental Protection Agency.
USEPA. 1987. Low Level and NARM Radioactive Wastes • Draft Environmental Impact Staemem- Vol. 2, prepared
by EPA Office of Radiation Programs for US Environmental Protection Agency.
USEPA. 1990. Sites for Our Solid Waste - A Guidebook for Effective Public Involvement, prepared by EPA Office
of Solid Waste and Emergency Response for US Environmental Protection Agency.
USEPA. 1990. Toxic Chemical Release Inventory Reporting Form R and Instructions, prepared by EPA Office of
Toxic Substances for US Environmental Protection Agency.
USEPA. 1989. Economic Report on Terephtbalic Acid, prepared by EPA Office of Toxic Substances for US
Environmental Protection Agency.
USEPA. 1988. Regulatory Impact Analysis for Financial Responsibility Requirements for Petroleum Underground
Storage Tanks, prepared by EPA Office of Underground Storage Tanks for US Environmental Protection
Agency.
USEPA. 1989. Perspectives on Non-Point Source Pollution: Proceedings of a National Conference, prepared by
EPA Office of Water for US Environmental Protection Agency, Washington, DC.
USEPA. 1985. Pretreatmcnt Implementation Review Task Force: final report, prepared by EPA Office of Water for
US Environmental Protection Agency.
A-71
-------
USEPA. 1987. Regulatory Impact Analysis of the Effluent Guidelines Regulation for the Organic Chemicals,
Plastics, and Synthetic Fibers Industry, prepared by EPA Office of Water (OWRS) for US Environmental
Protection Agency.
USEPA. 1987. Regulatory Impact Analysis of the Effluent Limitation Guidelines Regulation for the Iron and Steel
Industry, prepared by EPA Office of Water (OWRS) for US Environmental Protection Agency, EPA,
Washington, DC.
USEPA. 1984. Regulatory Impact Analysis on the National Ambient Air Quality Standards for Paniculate Matter,
prepared by EPA Stationary Air Sources Division for US Environmental Protection Agency.
USEPA. 1989. Estimates of the Total Benefits and Total Costs Associated with Implementation of the 1986
Amendments to the Safe Drinking Water Act, prepared by Wade Miller Associates for US Environmental
Protection Agency (ODW).
USEPA. A Guide to Selected National Environmental Statistics in the US Government, USEPA, Office of Policy,
Planning and Evaluation, Washington, DC.
USFS. 1987. Resource Pricing and Valuation Guidelines for the 1990 RPA Program, prepared by US Forest Service
for US Forest Service.
USGAO. 1985. The Nation's Water Quality: Many Unanswered Questions - Draft, prepared by US General
Accounting Office for US General Accounting Office.
USGS. 198S. National Water Summary 1984 - Hydrological Events, Selected Water-Quality trends and Ground-
Water resources. (Water-Supply Paper 2275), prepared by US Geological Survey for US Geological Survey.
USOTA. 1988. The Feasibility and the Costs of More Stringent Mobile Source Emission Controls, prepared by
Sierra Research for US Office of Technology Assessment.
USOTA. 1984. Wetlands, Their Use and Regulation, prepared by US Office of Technology Assessment for US
Office of Technology Assessment. Washington, DC.
USWRC. 1983. Economic and Environmental Principles for Water and Related Land Studies, prepared by US Water
Resources Council for US Water Resources Council, Washington, DC.
Uhl Vincent W. 1979. A Standard Procedure for Cost Analysis of Pollution Control Operations: Volume II:
Appendices, prepared by Environmental Protection Agency for US Environmental Protection Agency.
Vaughan, W.J., C. S. Russell, and M. HazOla. 1982. ANote on the Uae of Tiavel Cost Models with Unequal Zone
Population.
Vaughan, William J., Charles Paulsen. Julie A. Hewitt, and ChffocdRusseU, 1985. The Estiination of Recreation-
Related Water Pollution Control Benefits: Swimming, Boating, and Marine Recreational Fishing, prepared by
Resources for the Future for US Environmental Protection Agency.
Vaughan. William I., and Clifford S. Russell. 1982. Valuing a Fishing Day: An Application of a Systematic Varying
Parameter Model, Land Economics, Vol. 58, No. 4, pp. 450-463.
Vaughan, William J., John Mullaby, Jufie A. Hewitt, and Michael Hazffla, et al. 1985. Aggregation Problems in
Benefits Estimation: A Simulation Approach, prepared by Resources for the Future for US Environmental
Protection Agency (EARB).
Vincent, Mary K., David A. Moser, and William J. Hansen. 1986. National Economic Development Procedures
Manual - Recreation. Volume 1: Recreation Use and Benefit Estimation Techniques, prepared by US Army
Corps of Engineers.
A-72
-------
f
Violette, Daniel. 1985. A Model Estimating the Economic Impacts of Current Levels of Acidification on
Recreational Fishing in the Adirondack Mountains, prepared by Energy and Resource Consultants. Inc. for
US Environmental Protection Agency (EARB).
Violeue, Daniel M., and Donald C. Peterson. 1982. Assessing the Benefits of Policies Designed to Reduce Acid
Deposition: A Decision Analytic Benefit-Cost Framework, prepared by EPA Economic Analysis and
Research Branch for US Environmental Protection Agency, Energy and Resource Consultants, Inc.
Violette, Daniel M., and Lauraine G. Chestnut. 1983. Valuing Reductions in Risks: A Review of the Empirical
Estimates - Summary and Final Report, prepared by Energy and Resource Consultants, Inc. for US
Environmental Protection Agency (EARB).
Violeue, Daniel M., and Lauraine G. Chestnut. 1983. Valuing Reductions in Risks: A Review of the Empirical
Estimates, prepared by EPA Office of Policy Analysis for US Environmental Protection Agency.
Violette, Daniel M., and Lauraine G. Chestnut. 1989. Valuing Risks: New Information on the Willingness to Pay
for Changes in Fatal Risks, prepared by Energy and Resource Consultants, Inc. for US Environmental
Protection Agency (OPA).
Viscusi, W. Kip. 1985. Review of the Literature on the Enforcement of EPA Standards, prepared by Duke
University for US Environmental Protection Agency.
Viscusi, W. Kip. 1983. Frameworks for Analyzing the Effects of Risk and Environmental Regulations on
Productivity, American Economic Review.
Viscusi, W. Kip. 1985. Environmental Policy Choice with an Uncertain Chance of Irreversibility, Journal of
Environmental Economics and Management
Viscusi, W. Kip. 1984. Phosphates and the Environmental Free Lunch, Regulation.
Viscusi, W. Kip, and Charles O'Connor. 1984. Adaptive Responses to Chemical Labeling: Are Workers Bayesian
Decision Makers?, American Economic Review.
Viscusi, W. Kip, and W. N. Evans. 1988. Utility Functions mat are Dependent on one's health status: estimates and
economics implications. University of Maryland, College Park, MD.
Viscusi, W. Kip, and Wesley A. Magat. 1986. Economic Efficiency of Enforcement and Enforcement - Related
Monitoring: Project Period One Report, prepared by Duke University for US Environmental Protection
Agency (EARB).
Viscusi, W. Kip, and Wesley A. MagaL 1986. Analysis of Economic Benefits of Improved Information, prepared by
Duke University for US Environmental Protection Agency.
Viscusi. W. Kip, Wesley A. Magat, and Joel Huber. 1986. Information regulation of consumer health risks: an
empirical evaluation of hazard warnings, Rand Journal of Economics.
Vladeck, Bruce. 1984. The Limits of Cost-Effectiveness, American Journal of Public Health.
Wade, William W., George M. McCollister, Richard J. McCann. and Grace M. Johns. 1988. Estimating Recreation
Benefits for Instream and Diverted Users of Waterflows of the Sacramento-San Joaquin Rivers Watershed.
Walker, D. 1982. A Damage Function to Evaluate Erosion Control Economics, American Journal of Agricultural
Economics.
Walker, David R-, and John P. Hoehn. 1990. The Economic Damages of Groundwater Contamination in Small Rural
Communities: An Application to Nitrates, North Central Journal of Agricultural Economics, Vol. 12, No. 1,
pp. 47-56.
A-73
-------
Walker, W. H. 1972. Comparative Costs of Producing Water, Illinois State University.
Walsh, G., D. A. Greenley, R. A. Young, and J. R. McKean, et al 1978. Option Values, Preservation Values, and
Reaeational Benefits of Improved Water Quality, US Environmental Protection Agency.
Walsh. R., D. Bjonback, D. Rosenthal, and R. Aiken. 1985. Public Benefits of Programs to Protect Endangered
Wildlife in Colorado, prepared by Thome Ecological Institute.
Walsh, R., and J. Olienylc. 1981. Recreation Demand Effects of Mountain Pine Beetle Damage to Quality of Forest
Recreation Resources in Colorado Front Range, prepared by Department of Economics for Colorado State
University, Fort Collins, CO.
Walsh, R., R. Gillman, and J. Loomis. 1982. Wilderness Resource Economics: Recreation Use and Preservation
Values, prepared by Department of Economics for Colorado State University, Fort Collins, CO.
Walsh, Richard G. 1989. Issues in Nonmarket Valuation and Policy Application: A Retrospective Glance, Western
Journal of Agricultural Economics, Vol. 14, No. 1, pp. 178-188.
Walsh, Richard G. 1986. Recreation Economic Decisions: Comparing Benefits and Costs. Venture Publishing, Inc.,
State College, PA.
Walsh, Richard G., Donn M. Johnson, and John R. McKean. 1988. Review of Outdoor Recreation Economic
Demand Studies with Non-Market Benefit Estimates, 1968-1988, prepared by USFS Western Regional
Project W-133 for US Forest Service.
Walsh, Richard G., and Gordon J. Davitt. 1983. A Demand Function for Length of Stay on Ski Trips to Aspen,
Journal of Travel Research, Vol. 22, pp. 23-29.
Walsh, Richard G., John B. Loomis, and Richard A. Gillman. 1984. Valuing Option, Existence, and Bequest
Demands for Wilderness, Land Economics. Vol. 60. No. 1, pp. 14-29.
Walsh, Richard G., and John Loomis. 1986. Contribution of Outdoor Recreation to National Economic
Development, The President's Commission on Americans Outdoors, Washington, DC, pp. 35-46.
Walsh. Richard G., Larry Sanders, and John Loomis. 1985. Wild and Scenic River Economics: Recreation Use and
Preservation Values, prepared by Dept of Ag. & Nat Resource Ecou.; Am Wilderness Alliance for Colorado
Stale University. American Wilderness Alliance, Englewood, CO.
Walsh, Richard G., Larry Sanders, and John R. McKean. 1987. The Value of Travel Time as a Negative Function of
Distance, prepared by Department of Ag and Resource Ecnoomics for Colorado State University.
Walsh, Richard G., and Lynde O. Gilliam. Benefit of Wilderness Expansion wiw Excess Demand for Indian Peaks,
Western Journal of Agricultural Economics.
Walsh, Richard G., Nicole P. Miller, and Lynde O. Gilliam. 1983. Congestion and Willingness to Pay for Expansion
of Skiing Capacity, Land Economics, Vol. 59, pp. 195-210.
Walsh, Richard G., Olga Radulaski. and Li-Chin Lee. 1984. Value of Hiking and Cross-country Skiing in Roaded
and Non-roaded Areas of a National Forest, prepared by International Union of Forestry Research
Organizations for Oregon State University.
Walsh, Richard G., Ray Ericson. Daniel Arosteguy, and Michael Hansen. 1980. An Empirical Application of a
Model for Estimating the Recreation Value of Instream Flow, prepared by Colorado Water Resource
Research for Colorado State University.
A-74
-------
f
Walsh, Richard G., Robert Aukennan, and Roben Milton. 1980. Measuring Benefits and the Economic Value of
Water in Recreation on High Country Reservoirs, prepared by Colorado Water Resource Research for
Colorado Slate University.
Walton, Thomas E. 1980. Economic Benefits of the Clean Lakes Program, prepared by OWRS for US
Environmental Protection Agency.
Ward, Frank. 1981. A Review of Problems and Solutions: The Use of the Travel Cost Method for Valuing
Recreation Resources. In Outdoor Recreation and the Public Interest, ed. Gum, A.
Ward, Frank. 1984. Specification Considerations for the Price Variable in Travel Cost Demand Models, Land
Economics.
Ward, Frank. 1984. Measuring the Cost of Time in Recreation Demand Analysis: Comment, American Journal of
Agricultural Economics.
Ward, Frank. 1982. The Demand for a Value of Recreational Use of Water in Southeastern New Mexico 1978-79,
prepared by Agricultural Experiment Station Research for New Mexico State University.
Ward, Frank, and John B. Loomis. 1986. The Travel Cost Demand Model as an Environmental Policy Assessment
Tool: A Review of Literature, Western Journal of Agricultural Economics. Vol. 11, pp. 164-178.
Ware. J. E., and R. M. North. 1967. The Price and Consumption of Water for Residential Use in Georgia, prepared
by School of Business Administration for Georgia State College.
Warren, John. Economics of Declining Water Supplies in the Ogallala Aquifer, Groundwater, Vol. 20, No. 1, pp. 73-
79.
Waterstone, Martin. 1983. Toxics and Groundwater. The Devlopment and Application of Net Risk Analysis,
Environmental Professional.
Webb, James R. 1980. Nuclear Power Plants: Effects on Property Value, Appraisal Journal, Vol. 48, pp. 230-235.
Weinstein, S., P. Hall. R. McMahon, and M. Hunsberger, et al. 1981. Economic Impact of Environmental
Regulations on Housing, Urban Systems Research and Engineering.
Weithman, Stephen, and Mark Haas. 1982. Socioeconomic Value of the Trout Fishery in Lake Tanneycomo,
Missouri, Vol. Ill, pp. 223-230.
Welle, P. 198S. Potential Economic Impacts of Acid Rain in Minnesota: The Minnesota Acid Rain Survey, prepared
by Department of Economics, Bemidji State University. Bemidji for Bemidji. MN.
Welsh, Richard P., and Richard C. Bishop. 1986. An Assessment of Nonmarket Valuation Methods with an
Emphasis on State, Local and Legal Applications.
Wener, Richard, and Robert Kaminoff. 1983. Improving Environmental Information: Effects of Signs on Perceived
Crowding and Behavior. Environment and Behavior.
Wetzastein, M. E., and J. G. McNeely. 1980. Specification of Errors and Influence in Recreation Demand Models.
American Journal of Agricultural Economics. Vol. 62. No. 4, pp. 798-800.
Wetzel, J. N. 1981. Congestion and Economic Valuation: A Reconsideration, Journal of Environmental Economics
and Management, Vol. 8, pp. 192-195.
Wetzel, S. N. 1977. Estimating the Benefits of Recreation Under Conditions of Congestion, Journal of
Environmental Economics and Management, Vol. 4, pp. 239-46.
A-75
-------
White, Halbert. 1980. A Heteroskedasiicity-consistent Covariance Matrix Estimator and a Direct Test for
Heteroskedasticity, Econometrica, Vol. 48, pp. 817-38.
Wbitebead, John C., and Glenn C. Bloomquist. 1990. Substitute Environmental Goods and Existence Value,
prepared by College of B usiness and Economics for University of Kentucky.
Wnittington, Dale, John Briscoe, Mu Shinming, and William Barron, el al. 1986. Estimating the Willingness to Pay
for Water Services in Developing Countries: A Case Study of the Use of Contingent Valuation Surveys in
Southern Haiti.
Wbittington. Dale, V. Kerry Smith, Apia Okorafor, and Augustine Okore, etal. 1992. Giving Respondents Time to
Think in Contingent Valuation Studies: A Developing Country Application, Journal of Environmental
Economics and Management, Vol. 22, No. 3, pp. 205-226.
Whyte, Anne. 1982. Probabilities, Consequences and Values in the Perception of Risk, Royal Society of Canada/el
al Risk Assessment & Perception S .
Wilkins, John HI. 1983. Exposure Assessment in Studies of Environmental Hazards: An Epidemiological
Perspective, Toxic Substances Journal.
Willard, D. E., and M. M. Swenson. 1984. Why not in Your Backyard? Scientific Data and Non-Rational Decisons
about Risk, Environmental Management
Wfflet, K. D., and J. S. Shuttle. 1986. The Incidence of Water Pollution Control Costs: Partial vs. General
Equilibrium, Growth and Change, Vol. 15, No. 3, pp. 32-43.
Willig, R. D. 1978. Incremental Consumer's Surplus and Hedonic Price Adjustments. Journal of Economic Theory,
Vol. 78, pp. 227-253.
Willis, K. G. 1990. Valuing Non-Market Wildlife Commodities: An Evaluation and Comparison of Benefits and
Costs. Applied Economics, Vol. 22, pp. 13-30.
Willis, K. G., and J. F. Benson. A Comparison of User Benefits and Costs of Nature Conservation at Three Nature
Reserves, Regional Studies, Vol. 22, No. 5. pp. 217-228.
Willis, K. G., and J. F. Benson. 1988. Valuation of Wildlife: A Case Study of the Upper Teesdale Site of Special
Scientific Interest and Comparison of Methods in Environmental Economics.
Willis, K. G.. J. F. Benson, and C. M. Saunders. 1988. The Impact of Agricultural Policy on the Costs of Nature
Conservation, Land Economics, Vol. 64, pp. 147-157.
Wiiman, Elizabeth A. 1980. The Value of Time m Recreation Benefit Studies, Journal of Environmental Economics
and Management, Vol. 7. pp. 272-286.
Wiiman,
A. 1982. Hedonic Prices and Beach Recreational Values. In Advances in Applied
Microeconomics, ed. Smith, V. K.. J Al Press, me., Greenwich, CT. pp. 77-104.
Wiiman. Elizabeth A. 1982. Recreation Benefits from Public Lands. Resources for the Future, Washington, DC.
A. 1987. A Simple Repackaging Model of Recreational Choices, American Journal of
Agricultural Economics, Vol. 69. pp. 603-612.
Wibnan. Elizabeth A. 1984. Benefits to Deer Hunters From Forest Management Practices Which Provide Deer
Habitat, Vol. 49. pp. 334-344.
Wilson. James H., Dale L. Keyes. and James W. Lent 1982. Cost and Economic Assessment of Regulatory
Alternatives for NO2 NAAQS - Draft, prepared by Energy and Environmental Analysts, Kennedy School.
A-76
-------
Winston, Harrington, Alan Knipnick, and Waller Spofford. 1985. The Benefits of Preventing an Outbreak of
Giardiasis Due to Drinking Water Contamination, prepared by Resources for tbe Future for US
Environmental Protection Agency, Water.
Wittman, D. 1982. Efficient Rules in Highway Safety and Sports Activity, American Economic Review.
Wolka, Keven K. 1988. Groundwater Contamination Benefit-Cost Analysis Methodology, Journal of Water
Resources Planning and Management, Vol. 114, No. 2, pp. 210-222.
Wong, S. T. 1972. A Model on Municipal Water Demand: A Case Study of Northeastern Illinois, Land Economics,
Vol. 48, No. 1, pp. 3444.
Woodard, Duane, and Michael Hope. 1990. Natural Resource Damage Litigation Under tbe Comprehensive
Environmental Reponse, Compensation, and Liability Act, Harvard Environmental Law Review, Vol. 14, No.
189, pp. 190-215.
World Bank. 1981. An Annotated Bibliography of Environmental Economics, prepared by Office of Environmental
Affairs for World Bank.
Yang, Edward J., Roger C. Dower, and Mark Menefee. 1984. The Use of Economic Analysts in Valuing Natural
Resource Damages, prepared by Environmental Law Institute for National Oceanographic and Atmospheric
Administration.
Yardas, David, Alan J. Krupnick, Henry M. Peskin, and Winston Harrington. 1984. Directory of Environmental
Asset Data Bases and Valutaion Studies, prepared by Resources for the Future for Washington, D.C.
Yocum, John E.. and Alexander R. Stankunas. 1980. A Review of Air Pollutant Damage to Materials, prepared by
TRCJnc.
Young, C. Edwin. 1984. Perceived Water Quality and the Value of Seasonal Homes, Water Resources Bulletin.
Young, John S., Dennis M. Donnelly, Cindy F. Sorg, and John B. Loomis, et al. 1987. Net Economic Value of
Upland Game Hunting in Idaho, prepared by Rocky Mountain Forest and Range Land Experiment Station for
US Forest Service.
Young, Robert A. 1973. Price Elasticity of Demand for Municipal Water. A Case Study of Tucson, Arizona, Water
Resources Research, Vol. 9. No. 4, pp. 1068-1072.
Zeckhauser, R., and A. Fisher. 1976. Averting Behavior and External Diseconomies.
Zeise, L., R. Wilson, and E. Crouch. 1984. Use of Acute Toxicity to Estimate Carcinogenic Risk. Risk Analysts.
Ziemer, R. F., W. N. Musser, and R. C. Hill. 1980. Recitation Demand Equations: Functional Form and Consumer
Surplus, American Journal of Agricultural Economics, pp. 136-41.
Zilberman. David, and Erik Lichtenberg. 1986. Framework - Case Study Design for a Risk Benefit Analysis of
Pesticides in the Special Review Process, prepared by Western Consortium for the Health Professions. Inc.
for US Environmental Protection Agency.
i
A-77
-------
-------
t
APPENDIX B
ATTENDEES LIST
-------
-------
•
f
Sergio Ardila
7605 Heatherfon Ln.
Potomac, MD 20854
Ross Arnold
U.S. Forest Service
2285 Double Eagle Ct.
Reston,VA 22091
John C. Bergstrom
The University of Georgia
Department of Agricultural
and Applied Economics
301 Conner Hall
Athens, GA 30602
Robert Berrens
Oregon State University
Department of Agriculture and
Resource Economics
Ballard Extension Hall. Room 322
Corvallis,OR 97331
Tayler Bingham
Research Triangle Institute
3040 Comwallis Rd.
Research Triangle Park, NC 27709
Kevin Boyle
University of Maine
Department of Agricultural
and Resource Economics
Room 207 Winslow Hall
Orono, ME 04469
David Brookshire
University of New Mexico
Department of Economics
1915 Roma St. NW
Albuquerque, NM 87131-1101
Trudy Ann Cameron
University of California-Los Angeles
Department of Economics
405 Highland Ave.
Los Angeles, CA 90024-1477
David Campbell
National Wildlife Foundation
1400 16th SL.NW
Washington. DC 20036-2266
J. Lon Carlson
Argonne National Laboratory
EID/900
9700 South Cass Ave.
Argonne, IL 60439
Laurainc Chestnut
RCG/Hugler, Bailly, Inc.
P.O. Drawer O
Boulder, CO 80306
Randy Childs
West Virginia University
College of Business and Economics
Morgamown, WV 26506-6108
Dr. Stephen Crutchfield
627 8th St. NE
Washington, DC 20002
Elizabeth David
Wisconsin ONR
2300 19th SL, NW
Washington. DC 20009
Martin David
Resources for the Future
1616PSL.NW
Washington, DC 10009
Lcland Deck
Abt Associates, Inc.
4800 Montgomery Lane, Suite 500
Bethesda.MD 20814
Mike Denning
Exxon
600 Jefferson St, Suite 947
Houston, TX 77095
William Desvousges
Research Triangle Institute
3040 Comwallis Rd.
Research Triangle Park, NC 27709
Jerry Diamanu'des
University of Rhode Island
lOOverhillRd.
Providence, RI 02906
B-l
-------
Mark Downing
Texas A&M
Agricultural and Resource Economics
Department
College Station, TX 77841
Richard Dubourg
University College-London
Centre for Social and Economic Research
on the Global Environment
Gower St
London WCffiGBT
England
Rick Dunford
Research Triangle Institute
3040 Comwallis Rd.
Research Triangle Park, NC 27709
Stephen Farber
Graduate School of Policy and International
Affairs
3E32 Forbes Quad
University of Pittsburgh
Pittsburgh. PA 15260
Jerald J. Fletcher
West Virginia University
Division of Resource Management
Morgantown.WV 26506-6108
A. Myrick Freeman m
Bowdoin College
Department of Economics
Brunswick, ME 04011
Don Gamer
Environmental Law Institute
1660P.SL.NW
Washington, DC 20036
Anne Grambsch
U.S. Environmental Protection Agency
401MSL.SW
PM-221,Rm.3220
Washington, DC 20460
Thomas A. Grigalunas
University of Rhode Island
Department of Resource Economics
LippittHall
Kingston, RI 02881
Michael Hanemann
University of California—Berkeley
16113 Scene Ave.
Berkeley, CA 94709
Christopher S. Hansen
U.S. Forest Service
21905 64th Ave., W
Mountlake Terrace, WA 98043
Ross HemphiU
Argonne National Laboratory '
EID/Bldg. 900
Argonne, IL 60439
John Hochn
Michigan State University
Department of Agricultural Economics
East Lansing, MI 48824
Thomas P. Holmes
U.S. Forest Service
Forestry Services Laboratory
3041 Comwallis Rd.
P.O. Box 12254
Research Triangle Park, NC 27709
Reed Johnson
Research Triangle Institute
3040 Comwallis Rd.
Research Triangle Park, NC 27709
Carol Jones
National Oceanic and Atmospheric
Administration
Damage Assessment Center
6001 Executive Blvd., Suite 425
Rockvillc, MD 20852
Susan B.Kask
Western Carolina University
Economics and Finance Department
School of Business
CuUowee,NC 28723-9034
MaryJoKealy
U.S. Environmental Protection Agency
Economic Analysis and Research Branch
PM223X
401 M St., SW
Washington, DC 20460
B-2
-------
f
f
Alan Krupnick
Resources for the Future
1616 P St., NW
Washington, DC 20036
Linda Langner
U.S. Forest Service
820N.Howard St.,#202
Alexandria, VA 22304
Donna Lawson
National Oceanic and Atmospheric
Administration
Damage Assessment Center
6001 Executive Blvd., Rm. 425
Rockville.MD 20852
Michael LeBlanc
U.S. Department of Agriculture
Economic Research Service
1301 New York Ave.,NW
Room 438
Washington, DC 20005-4788
Robert Leeworthy
National Oceanic and Atmospheric
Administration
N/ORCA1
6001 Executive Blvd., Rm. 220
Rockville,MD 20852
Wade Martin
Colorado School of Mines
Mineral Economics Department
Golden, CO 80401-1887
Kristy Mathews
Research Triangle Institute
3040CornwallisRd.
Research Triangle Park, NC 27709
Marisa Mazzotta
University of Rhode Island
Resource Economics
Lippitt Hall
Kingston, RI 02881
Daniel McCoUum
U.S. Forest Service
Rocky Mountain Station
240 West Prospect St
Ft Collins, CO 80526-2098
Norman Mcadc
National Oceanic and Atmospheric
Administration
Damage Assessment Center
6001 Executive Blvd., Rm. 425
Rockville, MD 20852
R. Gregory Michaels
Abt Associates, Inc.
4800 Montgomery Lane, Suite 500
Bethesda, MD 20814
Edward Morey
University of Colorado-Boulder
Department of Economics
Campus Box 256
Boulder, CO 80309-0256
Andrew Muller
McMuslcr University
Department of Economics
Hamilton, Ontario 28S4M4
Canada
Stale Navrud
Agricultural University of Norway
Noragric
P.O. Box 2, N-1432
Norway
James Opaluch
The University of Rhode Island
Department of Resource Economics
Kingston, RI 02881-0814
George Parsons
University of Delaware
College of Marine Studies
Newark, DE 19716
Spencer R. Pearce
Consultant
P.O. Box 417
Elephant Butte, NM 87935
Richard Ready
University of Kentucky
Department of Agricultural Economics
318 Agriculture Engineering Bldg.
Lexington. KY 40546-0276
B-3
-------
Marc O. Ribaudo
U.S. Department of Agriculture
Economic Research Service
1301 New York Ave., NW
Room 438
Washington, DC 20005-4788
Eirik Romstad
Agricultural University of Norway
Department of Economics
and Social Sciences
P.O. Box 33
N-1432 Aas
Norway
Don Rosenthal
U.S. Department of Energy
1000 Independence Ave., SW
Washington, DC 20585
Robert Rowe
RCG/Hagler. Bailly, Inc.
P.O. Drawer O
Boulder, CO 80306
Richard Ruppert
8236 La Jolla Shores Dr.
La Jolla,CA 92037
Clifford Russell
Vanderbilt Institute for Public Policy Study
1208 18th Ave., S
Nashville, TN 37212
Daniel Schruefer
National Oceanic and Atmospheric
Administration
6001 Executive Blvd.
Rockville.MD 20852
V. Kerry Smith
North Cai
Carolina State University
Department of Business and Economics
Box 8109
Raleigh, NC 27695-8109
JohnStoll
University of Wisconsin-Green Bay
Department of Urban and Public Affairs
2420NicoletDr.
Green Bay. WI 54311-2791
Robert Unsworth
Industrial Economics, Inc.
2067 Massachusetts Ave.
Cambridge, MA 02140
David Waddington
U.S. Fish and Wildlife Service
Federal Aid Division
4401 N. Fairfax Dr., MS322
Arlington, VA 22203
Steven Waters
Research Triangle Institute
3040 Cornwallis Rd.
Research Triangle Park, NC 27709
Thomas Wegge
Jones and Stokes Assoc.
2756 9th Ave.
Sacramento, CA 95818
Rodney Weicher
National Oceanic and Atmospheric
Administration
1825 Connecticut Ave.
Washington, DC 20235
Katherinc Wellman
National Oceanic and Atmospheric
Administration
1825 Connecticut Ave., Suite 625
Washington, DC 20235
Peter Wiley
National Oceanic and Atmospheric
Administration
6165 S. Steamboat Way
ML Airy, MD 21771
B-4
-------
r
APR 3
NOVJZ
Information Resources Center
US EPA (3404)
401 M Street, SW
Washington, DC 20460
-------
U.S. EPA Headquarters Library
Mail code 3201
1200 Pennsylvania Avenue NW
Washington DC 20460
------- |