Bockstaeletal-valueofenv-est2000

7/29/2019 Bockstaeletal-valueofenv-est2000

1/6

On Measuring Economic Values forNature

N AN C Y E . B O C KS T AE L ,

A. M YR I C K F R E E M AN , I I I ,

R A Y M O N D J . K O P P ,

P A U L R . P O R T N E Y , A N DV . K E R R Y S M I T H * , |

Dep art m en t of Agricu lt u ral an d Resou rce Econ om ics,

University of Maryland, College of Agricultu re,

College Park, M aryland 20742, Department of Economics,

Hu bb ard Ha ll, Bow do in Colle ge, Bru n sw ick , M ain e 04011 ,

Quality of th e Environm ent Division, Resources for the Future,

1616 P Street, NW, Washington, DC 20036, and

N ort h Carol in a Sta te Un iv ersit y, Cam pu s Box 8109,

Ral eigh , N ort h Carol in a 27695 -81 09

This paper describes how economists ascribe values tothe things people can choose. The economic value of an

ecosystem function or service relates to the contribution itmakes to human welfare, where human welfare ismeasured in terms of each individuals ow n assessmentof w ell-being. After developing how this definition is used,the paper describes problems and opportunities foradvancing the state-of-the-art in measuring ec onomicvalues for nature. These arguments are developed usingrecent studies that attempted to estimate economic valuesfor ec osystems on a global scale. One implication of thisevaluation is that there is a need for greater communicationbetween ecologists and ec onomists. Economic analysesmust reflect the intricate w eb of physical interrelationshipslinking activities that have harmful effects in one part ofan ecosystem to the potential effects on other parts. At the

same time, economic values for ecosystems acceptconsumer sovereignty and should be interpreted asdescriptions of the tradeoffs involved in evaluating w ell-defined c hanges to specific ec osystems.

In t roduct ion

Two things have happened in environmental policy andresearch over the last 25 years to focus attention on theeconomic significance of natural systems. First, there hasbeen increased interest in un derstanding the economicconsequen ces of environm ental regulation. Indeed, each ofthe last five U.S. presidentsst wo D emo crats an d t h reeRepublicanssissued executive orders directing federal agen-

cies to conduct benefit-

cost analyses of all major proposedregulations. This need to provide comparable and quantita-tive evidence of the benefits, not just the costs, of stricterenvironm ental safeguards instigated a m ajor new researcheffort within economics aimed at developing methods forvaluing nonmarketed public goods such as environmental

quality. [There was an already extensive literature on theapplication ofbenefit-cost an alysis for water p rojects datingto early contributions by Eckstein (1) an d Krutilla and Eckstein(2).However,th is research focused on the m arketed outputsfrom public investment projects as the tangibles an denumerated, but did not attempt to estimate, the economicvalue for the changes in environmental amenities or quality

attributed to these projects.For a history,see ref3.Discussionof the early work on en vironm ental valuation can be foundin refs 4 an d 5. For a presentation of valuation methodsdeveloped during this period, see refs 6 an d 7.] Second, thefocus of federal environmental protection gradually shiftedduring this period from an exclusive preoccupation withhuman health toward concerns about ecosystem integrity(see ref8). The result has been a kind of forced m arriagethat ha s regulators, ecologists,a nd oth er biologicalscientists,as well as economists, thinking about the costs and benefitsof protecting ecosystems and the services they provide.

From th e p erspective of m any na tural scientists as wellas policy makers, economists have been slow at developingeconom ic value estimates of ecological systems. A partialexplanation for this is that economists have difficultyunderstandingthe intricate web ofphysicalinterrelationshipsthat can link harm in one p art of the ecosystem to n egativeeffects in another. Another explanation, however, is thatvaluing ecosystem s in a way that is logically consistent withusual benefit-cost analysis is a very difficult business. Inanyevent,application ofconventionalvaluation methods tothe problem often p roduces an incomplete set of discon-nected values for a subset of ecosystem services, and thiscan lead to a dramatic underestimate of the benefits ofecosystem protection.

Given the absence of ecosystem values in the economicsliterature, a series of papers have emerged, authored largelyby natural scientists,that attem pt to fillth isgap. [See Costanzaet al. (9), Pimen tel et al. (10), and Ehrlich and Ehrlich (11).The first pa per is n ot exclusively authore d b y ecologists. Theautho rs included two econom istssR.C. dArge an d S. Farbe r.]

The problem that economists have with these approachesis not based on any fundamental difference in judgmentabout th e importan ce of ecosystems. Rather it is about thecorrect application of concepts ofeconom ic value and ab outwhat questions provide a meaningful basis for definingeconomic values for purposes of benefit-cost a nalysis orother policy-ma king n eeds.

Two years ago, Costanza et al.(9) published a widely citedestimate of the aggregate annual monetary value of theservices provided bya number ofimportant ecosystems.[Seealso Ehrlich and Ehrlich (10), who provide an estimate of thetotal value of ecosystem services, an d Pimentel et al. (11),who provide an estimate of the total economic and envi-ronmental value of biodiversity.] This paper has stimulatedconsiderable debate. Pearce (12) recently reviewed its

meth ods an d con cluded, ... the article by Costanza and hisco-au thors is deep ly flawed. Yet its inte ntion wa s correct: toshow all of us just how valuable the natural world is. (page28). Pearce went on to ask an important question that isoften a part of the response to any criticisms raised with theCostanza et al. approach: Given the [desirable] purpose ofthat article, does it rea lly ma tter if the ana lysis is wrong? Inwhat follows, we attem pt to explain why it does. [Moreover,intended u ses ofth e pape r are not restricted to providing anindexof importance. In interviews about his paper, Costanzahas encou raged the use of these estimate s for policy. Pimm s(13) commentary on the Nature article reaffirms this use,suggesting: the real power of Table 2 (from Costan za et al.)

Part of the special issue on Economic Valuation.* Corresponding author e-mail: [email protected]; tele-

ph on e: (919)513-3761; fax (919)515-1824. University o f Maryland. Bowdoin College. Resources for the Future.| North Carolina State Un iversity.

Environ. Sci. Technol. 2000, 34, 1384-1389

1384 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 8, 2000 10.1021/es990673l CCC: $19.00 2000 American Chemical SocietyPublished on Web 02/25/2000


2/6

lies in its use for local decisions. The issue is whether andhow such num bers are to b e used in guiding pub lic policy.]

As clearly as is possible in a few pages, we spell out h oweconom ists ascribe value to th ingsswhether as pedestrianas a ball point pen or as complexas a saltwater marsh, notingthe special difficulties posed by the latter. We also discusshow m eaningful econom ic value questions are framed, andwhat qu estions do not lend them selves to this framing. Topreview our argument, it makes little sense to talk about theeconom ic value of ecosystems as if the choice were betweenhaving them as they are or n ot having them a t all, because

economic value is about tradeoffs and as such requiresdefining the alternatives clearly. Instead, the concept lendsitself to evaluating well-defined changes to ecosystems.Indeed, this is the only reasonable question to ask andcertainlyth e most relevant qu estion for most policyan alyses.Finally, we briefly describe some specific techn iques th athave been used to help in estimating the economic value forman y ofth e varied services that ecosystems provide,rangingfrom outdoor recreation to water purification. Specialattention is given to an approach that has frequently beenmisusedsthat of valuing a resource at th e cost of replacingits lost services. We conclud e by d iscussing why chan ges inthe condition of ecosystems are so d ifficult to value.

The Meaning of Economic Value

In common usage, the term value means importance ordesirability. The transfer between general usage and eachdisciplines perspective on how importance or desirabilityshould be gauged has created considerable debate amongecologists, econom ists, and philosophers. In econom ics,valuation concepts relate to hum an welfare. So the econom icvalue of an e cosystem function or service relates on ly to thecontribution it makes to human welfare, where hum anwelfare is m easured in terms of each individuals ownassessment ofhis or her well-being. Ofcourse, this is not theonly possible concept of value, nor is it always the mostrelevant. But for purposes of ben efit-cost analysis inassessing policy options and for purposes of determiningliability when natural resources have been harmed, thisco n cep t h as co n sid erab le p reced en ce as w ell as legal

standing.Getting environme ntal values to count a t all in the p olicyand legal domain should not be taken lightly. It is the resultofdecades ofargumen t that nonm arketed goods and servicesmust be given standing on a par with marketed goods. Inpart, the incorporation of nonm arket (and especially envi-ronmental) values into policy and liability considerationshas been possible because the econom ists concept of valuehas a long history of rigorous thought behind it.

For over 50years econom istsh ave ana lyzed the propertiesof the logically consistent con structs that support answersto two types of valuation qu estions: First, how does on econstruct a measure of how much better or worse off anindividual is with a (policy) change instead of without it?And second, how does one add up the gains and lossesexperienced by all ind ividuals to a ssess the result of a given(policy) chan ge for society as a whole? The an swers to bo thof these questions recognize that an absolute measure of apersons value for someth ing isu nach ievable. Moreover,evenif it were, aggregations of such measures over individualswould have little m eaning (see refs 14-16). Economistsvalue things only in comparative terms. When they saythey are valuing a change, econom ists are really defining atradeoff between two situations. The economic value of apolicy change is defined by the amount (either positive ornegative) of comp ensation that an individual would need inorder to be as well off (by his own reckoning of well-being)as he would have been without the policy-induced change(17, 18).

Thus, an economic value estimate is an answer to acarefully defined question in which two alternatives are beingcompared. For example, suppose a power plant is beingconsidered for a location that would eliminate a swimm ingbeach. Different people can have quite different values forthis change depending on whether they would use the beach,gain from th e lower cost of electricity, or both. Econom icvalue measures the amount a person would pay or be paid(in com pensation) to b e as well off with th e power p lant aswithout it. The compensation might be positive or negative.For a nonswimm er who benefits from lower electricityprices,

the com pensation will be different than for a frequent b eachuser. These answers do no t value the be ach per se. Insteadthey mea sure for each person the trade off that h e is willingto m ake.

Answers to this question are sensitive to the circumstancesofthe choices. Is there a similar beach a short distance awaythat is not affected? Would the elimination of the specificbeach ca use con gestion at others, resulting in inferior beachexperiences? Would loss of the beach also ha rm an endan -gered species that is specific to the area? [The inclusion ofprote ction of end angered sp ecies in the list ofe ffects reflectsthe fact that preserving species has value to hum ans, becau seof the value of a diverse gene pool,for exam ple.]The se issueshighlight a feature of economic values. Compensationmeasures cannot be d efined in isolation. They are entirely

dependent on the context and m aychange as there is a changein one or more elements of that context. This feature ofeconomic values requires analysts to be specific about whatis obtained with the change as well as the default situationthat exists without it.

Because compe nsation is a m easure of peop les tradeoffs,it can be mea sured in terms of any set of comm on un its onedesires. It is usually m easured in m oney. [While the logicun derlying these definitions leads to well-defined me asures,this does not m ean that am biguities are en tirely avoided. Arecent example of the potential for effects with differentchoices for the num eraire is given by Brekke (19). Anapplication that dem onstrates the consequences ofchoosingan alternative num eraire is found in ref20.]Given thiscon ceptof a tradeoff, two alternative compensation measures aredefinab le: on e that usesth e initialwell-being ofth e individualas the base or reference level and one that uses as the p ointof reference the level of well-being that would be obtainedafter the proposed change. The first concept answers theq u est io n : h o w mu ch (n egat ive o r p o sit ive) w ou ld t h eindividual need to receive, should the change take p lace, t omake him as well off as he would be in the absence of thechange? The other answers a parallel, but subtly different,q u est io n : h o w mu ch (n egat ive o r p o sit ive) w ou ld t h eindividual need to receive should the change not tak e place,to make him as well offas he would have been h ad the chan getaken place? Under some conditions, these two measurescan be expected to be ap proximately equal. In oth ers, theywill be quite different. One m easure is no m ore easily justifiedthan the other for all situations.

Additional caveats are necessary.Ind ividualsexpressions

of willingness to pay or willingness to accept compensationare cond itioned by their endowm ents of wealth. This doesn o t alt er t h e mean i n g o f t h e in d i vid u al co mp en sat io nmeasures, but i t does suggest that care must be taken ininterpreting any aggregation of th ese individual m easures.If som e groups ben efit from a n action while others lose, andthese grou ps ha ve significantly different wealth, the n p olicyactions based on the simple aggregation of individualcompensation measures can have serious distributionalimplications that need to be taken into account.

All of this argument assumes that individual behaviorresults from rational choices, an assumption that many findunsupportable. Beckers (21) Nobel address captures the

VOL. 34, NO. 8, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1385


3/6

broad generality of what is intended in this assump tion. Henotes that: the analysis assumes that individuals maximize[their] welfare as they conceive it, whether they be selfish,altruistic, loyal, spiteful or masochistic . .. Actions areconstrained by income, t ime, imperfect m emory and cal-culating capacities, and other limited resources, and also bythe opportunities available in the economy. These op -p o rtu n i ties are largely d et ermin ed b y t h e p rivat e an dcollective actions of other individuals and organizations(page 385, term in brackets was inserted).

His position is similar to most conventional economists.

It does not deny that in some circumstances people makemistakes, nor does it insist that people are always observedmaking seemingly rational choices. It does offer one way toexpressth e implications ofth ose choicessas an implied lowerbound on their economic value. The usefulness of thesemeasures depends only on the assumption that people dotheir best to realize the highest level of well-being given the ircircumstances. The other alternative strategies for definingindividua l values are inconsistent with such an interpreta tionbecause they replace each con sumers judgmen t of what isbest with some external prescription of how they shouldmakechoices.

Framing the Valuation QuestionMeasures of economic value are specific to the object of

choice (i.e., what has changed between the reference andthe policy induced states), the starting point or referencestate, and other aspects of the context of the decision (im plicitor explicit).[The impo rtan ce of conte xt in such exercises canbe illustrated by considering p urchases m ade at a youngchilds lemonade stand on a hot day. Using this youngentrepren eurs experience th at peop le pay $1 or more for ac old d r in k o n a h o t d a y a n d c on c lu d in g t h at m o d es tinvestmen ts will yield m illions would b e misleading. Thecircumstances of these choices influence the observeddecisions.Remove the child, other dimen sions of the contextbeing unchanged, and some (if not all) of the same peoplewould m ake different decisions. This is not evidence ofconsumer irrationality. Nobel Laureate Amartya Sen (22) inhis Frisch Memorial Lecture to the Econom etric Society

argued that choices tellus abou t how revealedp referencescan be influenced bythe specific features ofthe act ofchoice,such as the identity ofthe chooser, the m enu (or choice set),and the relation between the particular decision an d socialnorm s.] If the policy question is whether to im pose nationalregulations on agricultural nonpoint source pollution, thestatus quosas described by the current pattern of nutrientloadingssmu st be comp ared with the predicted alternativepatt ern ofre duce d load ings,wh ich will be different and havedifferent effects on aquatic ecosystems in different parts ofthe Un ited States. To value the com plete loss of a pa rticularecosystem function or service, one would need to com parethe state of the world as it exists with a well-defined pred ictionof what would exist should that specific function or servicebe eliminated. And in doing this, we would want to includesecondary effects on other ecosystem functions but hold

constant those things which are unaffected.When the stated objective is to measure econom ic value,

the con cepts and principles associated with performing thetask are not matters of opinion. Whether one likes or agreeswith the economic concept of value, the points listed abovemust be adequately addressed for the results to have thelogically consistent properties attributable to this type ofvaluation. The importance of this cannot be overstated,although it has be en ignored in several recent p ublicationsthat p urport to m easure the econom ic value of the worldsecosystems.

Costanza et al. (9), Ehrlich and Ehrlich (10), and Pimen telet al. (11) claim to an swer the question: What is the total

value of the worlds ecosystems (or in Pimentel et al. ,biodiversity)? Costanza et al. (9) and Pimentel et al . (11)calculate total values by scaling up a variety of estimatesofvalues taken from other studies ofecosystem services andfunctions. The original studies valued small changes inspecific and localized components ofindividualecosystems,with each study implicitly holding constant other features ofthe global ecosystem and the economy. It is incorrect toextrapolate the value estimates obtained in any of thesestudies to a mu ch larger scale, let alone to suppose th at theextrapolated estimates could then be added together and

applied to the whole planet.Environmen talgoods are genericallyd ifferent from private

commodities. They are usually not divisible into discreteun its, like cups of coffee or ball point p ens. More impo rtan t,their services are available to many people simultaneously.[This is a key element in the definition of a public good ineconom ics; the good is not used up through one (or m ore)individua lsconsu m ption.] As a result, there are lim its to theanalogies we can draw between the economic values ofnonm arket environm ental goods and th ose associated withprivate goods. Nonetheless, at least one aspect of the errorsof scaling up values can be easily illustrated. Whatever theprivate good involved in consu m ption choices, be it ballpointpens or pa irs of socks, we can expect that it will be subjectto diminishing value as m ore of the good is purchased (and

consumed) within a given (conveniently defined) period oftime. In app lications, this prop erty imp licitly reflects the factthat even though not all pairs of socks are the same, in mostcases an individuals willingness to pay for another pair islikely to diminish the more he already has. In this example,the individua l has plenty of close substitutes for this potentialnew purchase. Likewise, an individuals willingness to payfor another pen will be cond itioned on how m any others healready has. Take away those pre viously purcha sed socks orpen s,a nd the willingness to payfor a new one isveryd ifferent.

Thus,simple mu ltiplication of a p hysicalqu antityby unitvalue (derived from a case studyth at estima ted the econom icvalue for a specific resource ) is a serious error. Small chan gesin an ecosystems services do n ot ad equately characterize,with simp le multipliers, the loss of a global ecosystem. Values

estimated at one scale cannot be expand ed by a convenientphysicalindexofarea,such as hectares,to another scale;norcan two separate value estimates, derived under differentcontexts, simply be added together. When we estimate acompe nsation measure for one elemen t ofan ecosystem, weassume that othe r aspects ofthe world that influence hu manwell-being are u nchan ged. For examp le, we m ight com putea compensation measure for the elimination of a specificwetland, holding others constant. In another analysis, acompensation measure for the elimination of a differentwetland might have b een estima ted, holding the first at itsinitial level. But th e two com pensation m easures cann ot beadded together to obtain the correct compensation in theevent that both wetlands are eliminated. [The relationshipbetween the sum of these two measures, on one hand, andthe value of the com posite change in the two ecosystems, onthe other, depends both on (i) the ecological relationshipsbetween wetlands and the functions they provide and (ii)individualpreferences for those functions and the size oftheincome effects their loss generates. The analytical featuresgoverning this relationship focus on whether the servicesare substitutes or com pleme nts for each other. Also,th e termincom e effects refers to the effect ofch an ges in incom e onthe values (measured in money) that individuals assign toa specific change. For small changes, where the monetarycommitments are modest,the income effects associated withdoing each in isolation versus the two together may bemo dest, but for large changes they will not be. See Diam ond(23) fo r an examp l e o f a case w it h s mall ch an g es an d

1386 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 8, 2000


4/6

H an eman n (24) for a detailed treatment of all cases.] If thetwo wetlan ds are substitutes in any way,sim ple addition willund erstate their joint value.

The failure of additivity in economics has an apparentpara llel in ecology. Ecosystem s respon d to chan ges througha varietyof physical, biological, and chem icalfeedback cycles.These feedback cycles are central to the processes that linkall species to each other and to their respective hab itats. Alinear aggregation rule treats each change as if it could bemade independent of the other constituent elements. Indoing so, it assumes independence within an d across the

ecosystems being considered, and it ignores the possibleeffects of feedback cycles. [One reviewer of an earlier draftof this paper suggests that som e ecologists are u ndertakingcomp utations of global gross p roduction for different eco-systems by m ultiplying point estimates of a rate by an areafor the ecosystem. Moreover it was suggested that theseapproaches displayed consistency with estimates fromsatellite imagery. We are not contending that economistsand ecologists do not approximate aggregate measures.Rather, our objective is to explain the concept of economicvalue and highlight the features that bear in important wayson how we select among the available ways to approximateit.The app roach m ust be consistent with the principles usedto define it. When it is not, or the assum ptions mad e in theapproximations become com pletely implausible;th ere is no

reason to believe the errors will decline with scale of theapproximation. In fact, where interdependencies occur,scaling u p b y sim ple add ition is likely to m agnify the errors.]

Clearly, extrapolating from small-scale studies cann ot helpus to estimate the econo mic value for the worldsecosystems,but is there an y reason able answer to this question?We knowthat Costanza et al.s estimate of an ann ual econom ic valueof 33 trillion U .S. dollars is a logically incon sistent m easureofwhat individuals would be willing to pay to avoid the loss,if only because this estimate exceeds their total ability topay.Simplyp ut, if, as Costanza et al.(9) estimate, the worldsGNP is 18 trillion dollars, the worlds population does nothave 33 trillion d ollars to spe nd ann ually. [We do not inten dthe argument to suggest that if all the goods and servicesimportant to people were traded on ideal private markets

the GNP would then correctlymea sure these goodsecon omicvalues. The issue is simp ly that a p erson cann ot offer to giveup m ore than he or she has. Some m ight argue that incom edoes not really define spending limits b ecause individualscan b o rro w. H o wever, t h i s p resu mes t h e exis ten ce o f someone willing to lend. There are no creditors if everyoneisb orrowing, and n o one would lend to those with no prospectof ever paying back the loan. Thus, the offer to exchange anamount greater than ones income to obtain ecosystemprotection is infeasible.] Costanza et al. ( 25) attempted toanswer t his criticism to th e Pearce re view cited earlier. Theyobserved that using the GNP as an app roximate m easure ofincome is inappropriate because ecosystems provide realincome (contributions to hum an welfare) that is neverreflected in any ma rket.We do not deny ecosystems provideservices that peop le do not have to pay for. But by definition,willingness to pay to protect the worlds ecosystems fromdestruction reflects how much individuals would be willingto give up in other things to obtain this outcome. It is anexplicit tradeoff that defines willingness to pay. Our criticism,as well as that of Pearce, simply suggests a person cannotgive up m ore tha n he h as. Costanza et al.s estimate of worldGNP is an approximation of what the world h as, translatedinto money terms, before the proposed change; it is fromthis amou nt of m oney that the aggregate willingness to p ayfor the ecosystem m ust come. [It is true tha t GNP is not a nideal measure of total available income. There are activitiesthat generate income that are not reflected in GNP (e.g., theso-called informal econ omy). It is also conceivable that

people would re-allocate their t ime in response to some

exogenous ch ange, giving u p more leisure in exchange for

work, to realize a higher level of monetary compensation.

This would entail a modified definition of willingess to pay

but would b e generally consistent with our a rguments (see

App end ix D in ref26). However, none of these ad justments

to GNP would alter our basic point. The discrepan cy between

the Costanza et al. (9) estimates for the ann ual willingness

to pay and any measure of income available is too great. No

one would expect that these adjustments would increase the

measure of aggregate income to make their estimate plau-

sible. As a result, this dram atic overstatemen t rem ains one

clear signal of their mistake in understand ing the und erlying

concept.]

Suppose instead that one asks what com pensation the

people ofth e world would require in order to volun tarilygive

up the worlds ecosystems. This willingness-to-acceptform

of the economic valuation question uses the alternative

baseline an d highlights the need to define clearly the

alternative state that applies when the compensation is paid.

This is required to m easure the com pensation n ecessary to

equate each individuals well-being, given the current level

of the worlds ecosystems, with that in an alternative state.

But what is this alterna tive state: a different, newly em erging

set of ecosystems or a complete void?

If the answer is a different set of ecosystems, then wew ou ld n eed a g oo d d escrip t io n o f t h e s ervi ces t h ese

ecosystems would provide before we could define exactly

what is lost an d, in turn , estima te the econom ic value of

this loss using the comp ensat ion principle. It is possible that

ecologistsalready know someth ing about the ecologicalstates

that would exist were certain current states destroyed through

mans mismanagement. However, i t seems unlikely that

anyone cou ld reliablyp redict how the world would look were

man to destroy simultaneously all currently existing eco-

systems. The feedback effects an d interactions implicit in

ecological systems p ose a daun ting mod eling task.

If the alternative state is nothingness, then the answer

to the willingness to accept question is trivial. There is n o

fin it e com p en sat ionthat individuals would accept to agreeto the loss of the worlds ecosystems, and they would pay

everything they had to avoid it. To an economist, this is the

definition of an essential good, a good for which there is no

finite compensation for its complete elimination. In this

sense, ecosystems are essential. At this level of generality,

there m ay of course be oth er essential inputs to the process

ofsustaininghuman well-being.For example,an institutional

system that defines norms of behavior and provides some

rule oflaw may also be deem ed essential, as well as the social

system, conditioning interpersonal interactions. Evaluating

any one of them relative to the default of their com plete

absence would yield similar answers.

There is a final reason why valuing the worlds ecosystem s

in monetaryterms makes no sense.For manyenvironmental

valuation problems, mon etary valuations work because theyprovide a com m on m etric in which to express trad eoffs. But

for the extensive changes proposed by the above authors,

the econ omic system would chan ge comp letely, generating

quite different sets of prices and incomes. Since m oney has

no m eaning independ ent of the system in which prices and

incomes are determined, it makes little sense to ask how

mu ch individuals would payto retain the worldsecosystems.

The proposed change is so great as to completely realign

both the global ecosystem a nd the global economic system

as well.Such a chan ge tran sforms the units one is attem pting

to use to measure value.Comparingthem usingcurrent prices

and incomes becomes meaningless.



5/6

Methods of Reveal ing Peoples Values

Even if we knew how to frame the questionsWhat is theeconom ic value of the worlds ecosystems?sthe answer isunlikely to inform any real-world decision. While theconsequences ofhum an actions can be enormou s,the actualpolicych oices faced are rarely ifever at the global ecosystemscale. For such profound and all-encompassing questions,conventional econom ic valuation is inappropriate for all thereasons given above.

When the trad eoffs we want to know abou t are associated

with specific changes in specific resources or ecosystemservices, there are a variety of valuation methods available.(For more detailed descriptions, see refs 7or 6). Where serviceflows involve ma rketed commod ities such as food and timber,information derived from dem and functions for these ma rketgoods can provide a basis for estimating the values of theseservice flows.Informa tion isalso needed on sup plycon ditionsso that the cost of the labor, capital, and other resourcesdevoted to the harvesting of these comm odities can bededucted to determ ine the n et value of the surplus due tothe ecosystem contributions. As a counter example, valuingthe oceans and estuariesfood production at market priceswithout d educting the costs (in term s of huma n labor, fuel,etc.) of harvesting the fish would be incorrect.

Economic values for the nonmarket services of natural

ecosystems, services that contribute to hum an hea lth and toquality of l ife, can sometimes be derived from peoplesdecisions to use related resources or to substitute otherresources when the natural ecosystems qu ality is impaired(e.g., refs 27-29). It is also possible to use stated preferencechoices or contingent valuation to estimate these values byproposing choices with well-defined, potentialconsequencesor costs (e.g., refs 30 an d 31).

In some specific circumstances, the cost of replacing afunction of an ecological system with a hum an-engineeredsystem can be used as a measure of the economic value ofthe function itself(e.g., ref32).In a classic examp le,Gosselinket al. (33) used an estimate of the cost of a tertiary sewagetreatment system as the economic value of the nutrientremoval function of a wetland. However, replacemen t costcan be a valid measure of economic value only if threeconditions are m et. The conditions are th e following: (i)that the human-engineered system provide functions thatare equivalent in quality and magnitude to the naturalfunction; (ii) that the hum an-en gineered system is the leastcost alternative way of performing this function; and (iii)that individuals in aggregate would in fact be willing to incurthese costs if the natural function were no longer available.Full compliance with these conditions is rarely achieved,and as a result, simplistic transfers of replacement costestimates are usually m isleading.

Where the change we wish to evaluate involves thedestruction or impairmen t of a localized ecosystem, such asthe elimination of a wetland or the conversion of largeportions of a watershed from forest to developed uses,economic valuation admittedly faces a more serious chal-lenge. [For a discussion of these difficulties, see ref 34.] Achange in the condition of a local ecosystem may alter itsability to provide service flows of both consumptive (e.g.,fish an d timb er) and noncon sump tive uses (e.g., aestheticsand wildlife). [See Kahn and Kemp (35) an d St ran d an dBockstael (36) for attempts to trace ecosystem impairmentto end p roducts.]It ma y also diminish its capacity to providefunctions that are indirectly valued by humans, such asclimate stabilization or nutrient recycling. [Because anecosystem p rovides these services over time, it is also possibleto con sider an ecosystem as a natural a sset. Accordingly itmay seem reasonable to ask whether treating ecosystems asassets would change the interpretation of their economic

values. The short answer is no. Using the con cept of an asset

in valuation requires that the total value of the asset reflect

the value ofth e time profile of services it provides. Converting

this time profile of values generated by allocation choices

and their implied tradeoffs to a current value becomes

important. This process leads to consideration of the role of

discoun ting in defining an equivalent curren t value. All of

these details seem more complex than the static choice

between two a lternatives we used to define a lower boun d

for an economic value. They are not. They are simply one

way to account for the differential properties of some types

o f g oo d s. Weit zman (37) h as recen t ly es tab lis h ed t h erelevance of the simple, static choice analogy to comparing

two optimal p rofiles. He p roves it can be e valuated as if it

were a static choice, paralleling what we used to define the

concept ofan economic value.]It mayeven diminish features

of the ecosystem that are important to hum ans at a m uch

more abstract level, features that play an imp ortant role in

our heritage and culture. Because econom ic values depend

on individuals own assessmen ts of their well-being, an

und erstanding of the ways in which specific impairmen ts of

an ecosystem diminish their lives is essential.

Our ignorance does not preclude the need for these

answers, nor ha s it prevente d us from giving partial answers

when complete ones were unavailable. Thirty years ago,

environm ental econom ists argued that dam ming a wild andpristine river in Idaho to provide electricity was a bad idea.

They reached th is conclusion by comp aring the value, over

t ime, o f p o w er g en erated t o in d i rect meas u res o f t h e

economic value of the wild river as a unique recreational

experience. Even without taking account of the more subtle

ecosystem intera ctions, it was possible to recognize that th e

nonm arket recreational ben efits provided over time by the

wild river exceeded the benefits of the dam. [See the Hells

Canyon controversy discussed in Fisher et al. (38). At the

time, these authors were unable to measure allth e recreation

ben efits from th e wild river. They were ab le to describe ho w

the sam e econom ic factors contributing to the time profile

of power benefits would be likely to affect the recreation

ben efits provided by the river.This description allowed them

to ask a very imp ortant rh etorical questionsh o w mu ch d othe initial preservation benefits for the river need to be to

favor preservation?This question recognized the essence of

the link between economic value and a choice. Each choice

reveals a lo wer b o u n d fo r t h e econ o mi c valu e. As a

consequence, public choices for preservation need only

demonstrate that the economic value of what is protected

exceed that of the alternative state defining the choice.] Not

allsuch an alyses ofp olicych ange will favor the en vironm ent

however.For example,in a developingcountrywhere malaria

and other insect-carrying diseases are still a major threat,

what little we know of the ecological damages associated

with pesticide use will un likelyou tweigh their hum an hea lth

benefits when trad eoffs use econom ic value meth odology.

The very nature of ecosystemsstheir comp lexity in

structure and functionsmakes understanding the conse-quences of ecosystem damage a daunting task even for

ecologists. The larger the scale of the chan ge,t he more subt le

and com plexthe functions impaired,and the farther removed

from the en d prod uct of value to hum ans, the m ore difficult

willbe th e task ofestimating econom ic values.But this is not

s u rp ris in g s in ce if w e d o n o t yet , as a s ociet y, fu lly

com preh end the role different ecosystems playin our world,

we cannot give meaningful values by any definition. It does

not m ean we shou ld abando n the effort of measuring well-

defined economic values for those consequences we do

un derstan d. They are often sufficient to answer the questions

posed by p olicy decisions associated with ecosystems.

1388 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 8, 2000


6/6

Discussion

No one would suggest that economic values should rule theday. Econom ic valuation measures are on ly one com ponen tof the criteria available for evaluating p olicy. Correctlyinterpreting what econom ic value measures m ean does notrequire exclusive reliance on the results from such calculations(see ref39).

Th is p a p er is n o t i n t en d e d a s a ju d gm e n t o n t h eapprop riate definition of value. It is about furthering goodscience. Without cooperation, ecologists and economists

cannot serve the interests ofsociety. Moreover, cooperationis imp ossible without a better app reciation and respect foreach oth ers discipline. The belief by some econom ists thatecologicalmodels require a biocentric perspective on societalobjectives, with all species given equ al weight in determ iningresource allocations, is not correct. Likewise, the belief bysome ecologists that economists place values on ecologicalservices only if they contribute to marketed commodities isequally incorrect.

Neither discipline will actually p rovide th e exclusiveorganizing principles for how people will in fact intera ct withecosystems in practice. In free societies such out com es arisefrom a collection of un coordinated (or at best looselycoordinated) activities that are con strained by the rules wecollectively adopt. In this context, improving the com-

mu nication between disciplines is the only hope for betterinforming these processes. People will ultimately makechoices with or without the insights from a genuine collectiveeffort.

Acknowledgments

Thanks are due Robert Solow, Scott Farrow, an d threeanonymous referees for comments on an earlier draft.

Literature Cited(1) Eckstein, O. Water Resource Developm ent: The Econom ics of

Project Evalua tion ; Harvard University Press: Cambridge, MA,1958.

(2) Krutilla, J. V.; Eckstein, O. Mu ltiple Purpose River Developm ent:Studies in Applied Econom ic Analysis; Johns Hopkins Press:Baltimore, 1958.

(3) Hanemann, W.M. Preface.In Pricingth e European Environm ent;Navrud , S., Ed.; Oxford Un iversity Press: New York, 1992.

(4) Maler , K.-G. Environm ental Econom ics: A Theoretical Inqu iry;Johns Hopkins University Press for Resources for the Future:Baltimore, 1974.

(5) Freeman , A. M., III. The Benefits of Environm ental Im prove-m ent: Theory and Practice; Johns Hopkins University Press forResources for the Futu re: Baltimore, 1979.

(6) Braden , J., Kolstad, C. D., Eds.; Measuring the Demand forEnv iron m en ta l Qu alit y; North-Holland : Amsterda m, 1991.

(7) Freeman , A. M., III. The Measurement of Environm ental andResou rce Valu es: Th eory an d Met h od s; Resour ces for t heFuture: Washington, DC, 1993.

(8) U.S. Environm ental Protection Agency. Reducing Risk: SettingPriorities and Strategies for Environm ental Protection; U.S.Governm ent Printing Office: Washington, DC, 1990.

(9) Costanza, R.; dArge, R.; de Groot, R.; Farber, S.; Grasso, M.;Han no n, B.; Limb urg, K.; Naeem , S.; ONeill, R. V.; Paru elo, J.;Raskin, R. G.; Sutton, P.; van den Belt, M. Nature 1997, 387(May), 253-260.

(10) Ehrlich, P.; Ehrlich, A. Betrayal of Science and Reason ; IslandPress: Washington, DC, 1996.

(11) Pimentel, D.; Wilson, C.; McCullum, C.; Huan g, R.; Dwen, P.;Flack, J.; Tran, Q.; Saltman, T.; Cliff, B. BioScience 1997, 47(December), 747-757.

(12) Pearce, D. Environm ent 1998, 40 (2), 23-28.(13) Pimm, S. L. Nature 1998, 387 (May), 231-232.(14) Baum ol, W. J. Welfare Econom ics and th e Theory of the State;

Harvard U niversity Press: Cambridge, MA, 1952.(15) Samuelson, P. A. Q. J. Econ. 1956, 70 (February), 1-22.(16) Sen , A. Choice, Welfare an d Measurement; MIT Press: Cam-

bridge, MA, 1982.(17) Hicks, J. R. Econ. J. 1939, 49 (196), 696-712.(18) Kaldor, N. Econ. J. 1939, 49, 549-552.(19) Brekke, K. A. J. Pu blic Econ. 1997, 64 (April), 117-124.

(20) Bockstael, N. E.; Strand, I. E. West. J. Agric. Econ. 1985, 10(December), 162-169.

(21) Becker, G. S. J. Political Econ. 1993, 101 (June), 385-409.(22) Sen, A. Econom etrica 1997, 65 (July), 745-779.(23) Diamon d, P. A. J. Environ. Econ. Ma nage. 1996, 30 (May), 337-

347.(24) Hane man n, W. M. The Econom ic Theory of WTP and WTA. In

Valuin g Environ m ental Preferences: Theory and Practice of theContingent Valuation in the U.S.,ECan d Developing Countries;Bateman , I.J.,Willis, K. G.,Eds.;Oxford UniversityPress: Oxford,1999.

(25) Costanza, R.; dArge, R.; de Groot, R.; Farber, S.; Grasso, M.;Han no n, B.; Limb urg, K.; Naeem , S.; ONeill, R. V.; Paru elo, J.;Raskin, R. G.; Sutton, P.; van den Belt, M. Environm ent1998,40 (2), 26-27.

(26) Just, R.;H ueth, D.; Schmitz, A.;App lied Wel fare Econ om ics an dPublic Policy; Prentice-H all: Englewood Cliffs, NJ, 1982.

(27) Bockstael, N. E.; McConnell, K. E.; Strand, I. E. Mar. Resour.

Econ . 1990, 6 (3), 1-

18.(28) Smith, V. K.; Hua ng, J. C. J. Political Econ . 1995, 103 (February),

209-227.(29) Bartik, T. J. J. Environ. Econ. Manage. 1988, 15 (March), 111-

127.(30) Mitchell, R. C.; Carson, R. T. Using Surveys to Value Public

Goods: The Contingent Valuation Method; Resources for theFuture: Washington, DC, 1989.

(31) Carson, R. T.; Han ema nn , W. M.; Kopp, R. J.; Krosnick, J. A.;Mitchell, R. C.; Presser, S.; Ruu d, P. A.; Smith , V. K. The NOAAPanels Scope Test for Contingent Valuation ; Center for Envi-ronm ental and Resource Econom ics,Duke University: July1997;Working Pap er.

(32) Shabm an , L. A.; Batie, S. S. Coastal Zone Ma nage. J. 1978, 4 (3),231-237.

(33) Gosselink, J. G.; Odium, E. P.; Pope, R. M. The Value of TidalMa rsh ; Center for Wetlands Research, Louisiana State Univer-sity: Baton Rouge, LA, 1974; LSU-SG-70-03.

(34) Toman , M. Ecosystem Valuation: An Overview of Issues an dUncertainties. In Ecosystem Function and Hum an Activities:

Recon cilin g Econ om ics a n d Ecology; Simp son, D., Christensen,N. L., Eds.; Chap man and Hall: New York, 1997.

(35) Kahn, J.; Kem p, W. M. J. Environ. Econ. Manage. 1985, 12 (3),246-263.

(36) Strand, I.; Bockstael, N. Interaction Between Agriculture andFisheries: Empirical Evidence and Policy Implications. In

Agricu ltu ral Ma n agem en t an d Econ om ics: Com m od ity an dResou rce Poli cies in Agricu ltu ral System s; Just, R., Bockstael N.,Eds.; Sprin ger-Verlag: New York, 1991.

(37) Weitzman, M. L. A Contribution to the Theory of WelfareAccou n tin g; Working Paper No. 6988; National Bureau ofEconomic Research: July 1999.

(38) Fishe r, A. C.; Kru tilla, J. V.; Cicchet ti, C. J. Am . Econ. Rev. 1972,62 (December), 605-619.

(39) Arrow, K. J.; Cropper , M. L.; Eads, G. C.; Hah n, R. W.; Lave, L.B.;Noll,R.G.;Portney, P.R.;Russell,M.;Schmalensee, R.;Smith,

V. K.; Stavins, R. N. Science 1996, 272 (April), 221-

222.

Receiv ed for rev iew Ju n e 16, 199 9. Rev ised m an u scrip t re-

ceived Nov em ber 17, 1999. Accepted December 1, 1999.

ES990673L


Documents

Bockstaeletal-valueofenv-est2000