Ecosystem Service Framework

7/31/2019 Ecosystem Service Framework

1/23

Te Ecosystem ServiceFramework: A Critical

Assessment

Ecosystem Services Economics (ESE)

Working Paper Series

Environment for Development

R. David Simpson

January 2011

The United Nations Environment Programme

Division of Environmental Policy Implementation Paper N 5


2/23

Papers in this series are not ormal publications o UNEP. Tey are circulated to encourage thought and discussion. Te use and citation o this paper should take thisinto account. Te views expressed are those o the authors and should not be attributed to UNEP. Copies are available rom the Ecosystem Services Economics Unit,Division o Environmental Policy Implementation, UNEP, Nairobi, Kenya. Te views expressed in this Policy Series do not necessarily refect the views or policies o

UNEP or contributory organization(s). Te designations employed and the presentations do not imply the expressions o any opinion whatsoever on the part o UNEP orcontributory organization(s) concerning the legal status o any country, territory, city or area or its authority, or concerning the delimitation o its rontiers or boundaries.

Design and Layout: Catherine K. Kimeu/UNON

Printing: Publishing Services Section, UNON, Nairobi-Kenya, ISO 14001:2004 certiied

UNEP promotes

environmentally sound practices

globally and in its own activities.This publication is printed on 100 per cent

chlorine free paper from sustainably managed

forests. Our distribution policy aims to reduce

UNEPs carbon footprint.


3/23

Te Ecosystem Service

Framework: A Critical Assessment

R. David Simpson*National Center or Environmental EconomicsUnited States Environmental Protection Agency

January 2011

Ecosystem Services Economics

UNEP

* Te opinions expressed in this paper are the authors, and do not necessarily reect those o the United States Environmental Protection Agency. Teauthor would like to thank Emma Roach or research assistance and my colleagues at the National Center or Environmental Economics or manyhelpul comments.


4/23

2 T h e e c o s y s T e m s e r v i c e F r a m e w o r k : a c r i T i c a l a s s e s s m e n T

Table o contents

Preface .............................................................................................................................................................. 3

Abstract ............................................................................................................................................................. 4

1. Introduction............................................................................................................................................... 4

2. The value of ecosystem services: Some prominent examples ...................................................................... 6

3. Impediments to estimating ecosystem service values .................................................................................. 9

4. How does the ecosystem services framework change conservation policy? ............................................... 16

References ....................................................................................................................................................... 19


5/23

3T h e e c o s y s T e m s e r v i c e F r a m e w o r k : a c r i T i c a l a s s e s s m e n T

PreaceIn last ten years, the economic analysis o ecosystem services has caught the imaginations o the decision-makers.Economic values o ecosystem services at global level have been in the centre o debate. Te Millennium Ecosystem

Assessment (MA) and recently concluded, the Economics o Ecosystems and Biodiversity (EEB) have providedadequate rationale or economic approach to management o ecosystems. However, robust estimation o ecosys-tem services whether it is or payments or ecosystem services or or national level green accounting are scarce andinadequate. Te paper provides a critical angle to the whole debate. Te author questions both the justicationor adopting an ecosystem service ramework and the extent to which some prominent examples o valuation oregulating services can withstand detailed scrutiny. Te views expressed in the chapter may be seen as controversialby many readers, but they highlight some o the undamental theoretical and methodological considerations thatunderpin this eld o ecological economics.

Ibrahim Tiaw,Director, Division o Environmental Policy Implementation (DEPI)United Nations Environment Programme (UNEP)


6/23


AbstractNatural ecosystems provide services to society. Conservation advocates have shown considerable enthusiasm inrecent years or an ecosystem service ramework. Tey hope that an appeal to the value o ecosystem services will

result in greater unding. Yet it is not unusual to nd examples in which an author states at one point that the valueo ecosystem services is great, while shortly thereater stating that values have not yet been quantied.

Te values o ecosystem services have not yet been estimated with any generality or precision. While severalcelebrated studies have addressed valuation, most are problematic. Moreover, there are a number o reasons or

which valuing ecosystem services remains a daunting exercise. Foremost among these is a paradox o valuation: itwould be most useul to estimate the value o ecosystem services that are dispersed over a very broad public, but itis precisely under these circumstances that it is most dicult to use the tools o economic value estimation.

How, then, does an ecosystem service ramework afect conservation strategy when the value o ecosystem servicesremains uncertain? It certainly motivates more research. Unless and until such research yields more precise andcompelling evidence o values, however and there is little reason to be optimistic that it soon will the statusquo will likely prevail in conservation policy. Tat is, conservation will be largely motivated by the importanceo providing global public goods (largely carbon sequestration and biodiversity protection), and will requireinternational transers rom wealthier to poorer countries.

1. IntroductionIn recent years considerable enthusiasm has grown or an Ecosystem service ramework (Daily and urner 2008)to conservation policy. Gretchen Daily and Pamela Matson write that such an approach has sparked

a growing eeling o Renaissance in the conservation community. Tis ows rom the promisein reaching, together with a much more diverse and powerul set o leaders than in the past, ornew approaches that align economic orces with conservation, and that explicitly link human andenvironmental well-being. And this promise is owering thanks to substantial recent advances inkey areas o inquiry, such as ecology, economics, and institutions, and their integration. (Daily andMatson, 2006, reerences omitted)

Many hope that the ecosystem services ramework will provide a new and generous source o conservation unding.Heather allis and Peter Kareiva (2005) write that realization o the market worth o ecosystem services has thepotential to increase conservation unding by orders o magnitude.

Tis enthusiasm has sparked an impressive volume o workwithin the ecosystem service ramework (Daily andurner 2008). Fisher and coauthors (2009) document an exponential increase in the number o published papersemploying the terms ecosystem services or ecological services, beginning rom essentially none in the early1980s to more than 250 in 2007, the last year or which they have data. Some rough idea o the currency othe term can be gleaned rom the act that entering ecosystem services in the Google search engine returnsabout 4.7 million entries.1 Te Millennium Ecosystem Assessment, a multi-year, multi-million dollar internationalundertaking involving over 1,300 scientists rom around the world was conducted to assess the consequences oecosystem change, and consequent alterations in the ow o ecosystem services, or human well-being (MA 2005).Tis work may continue under a recently proposed Intergovernmental Platorm on Biodiversity and EcosystemServices (IPBES), modeled on the Nobel-prize-winning Intergovernmental Panel on Climate Change (IPCC). InNovember o 2008 representatives o 78 nations and 25 international NGOs met to consider establishment o an

1 Search conducted 17 April 2009. o get some idea o societal priorities and the zeitgeist, this is about hal the number returned when one searches orJulia Roberts, and slightly more than are returned or Matthew McConaughey.


7/23


IPBES (UNEP 2008a). At the meeting an approximately US$18.4 million Programme o Work and Budget waspresented (although not yet adopted; UNEP 2008b).

Organizations around the world are adopting an ecosystem services approach to ecological decision making. Yet theelements o that approach are not as settled as its widespread adoption might make it appear. One oten encounterspassages such as the ollowing: Although the societal benets o native ecosystems are clearly immense, they remainlargely unquantifedor all but a ew services (Ricketts, et al., 2004, emphases added; see also Kareiva and Rufo2009; Dailyet al. 2009). But i benets are largely unquantied what is the basis or concluding that they areclearly immense? Is there really much evidence supporting the contention that the services supplied by naturalecosystems are o great value and that they are being squandered by unwise land use decisions?

In this paper I suggest that evidence to that efect remains sparse. Tis is so or several reasons, and I will considerthem in turn. Te rst is that many o contributions to the literature on ecosystem service values would appear tobe intended to motivate research on ecosystem services than rather than to document the ndings o such research.Tere are certainly numerous, and in many instances eloquent, statements o the hypothesisthat natural ecosystemsprovide valuable services, but ewer careul tests o that hypothesis.

In some studies the interpretation o evidence concerning ecosystem service values is problematic. It is worthunderscoring that evidence o ecosystem service values will only motivate diferent conservation decisions i suchvalues outweigh costs. It is not sucient simply to note that there is somevalue to conserving what is already inplace without comparing that value to other possibilities.

Closely related to the above observation is the economic truism that value is determined on the margin. Terelevant concern is typically not that biodiversity or ecosystem services will perish in their entirety. Any monetaryestimate o such a calamity would necessarily be, to borrow Michael omans (1998) characterization, a seriousunderestimate o innity. Te relevant policy question, then, is whether preservingspecifc componentso ecosystemsprovides benets in excess o those that would arise rom their orgone uses.

I ecosystem services are not assigned their true value in land use decision-making, it is because such services arepublic goods; that is, they are benets which, when supplied by one person are necessarily accessible to many. Yetthe more compact the scale on which such public goods are provided the ewer the members o the public are

who benet rom their provision the less likely it is that ecosystem services will be underprovided. Consequently,we should be most concerned about the provision o ecosystem services whose benets are very widely dispersed.

Tis observation, in turn, leads to a couple o other issues. First, there is what I call below a paradox o valuation.Te things we would most like to be able to place an economic value on are those public goods whose benetsare the most widely dispersed. But these are precisely the goods whose value is most dicult to estimate. Second,the most compelling argument or conserving relatively pristine ecosystems may prove to be that they providethe global public goods o carbon sequestration and biodiversity protection. I this is, in act, the most important

argument or conservation, however, the ecosystem service ramework does not appear to be adding much newto the debate on conservation policy.

Tis last observation poses the main question motivating this paper. How doesadopting the ecosystem serviceramework alter the ways in which we think about conservation policy? I the point is simply that we ought toregard natural ecosystems as assets that provide value to society and should, thereore, compete with alternativeland uses as we make choices about how to allocate the earths surace among our wants and needs, the argumentis unexceptionable. It is also not novel, however. Underscoring such a undamental principle is useul, but it alonecannot account or current enthusiasm or the ecosystem service ramework.

Tat enthusiasm derives, rather, rom the sense that the ecosystem service ramework has already demonstrated,

or can soon be expected to demonstrate, the general economic superiority o conservation to alternative land uses.Most o the remainder o this paper is devoted to considering and, generally, rebutting these assertions. As theseare controversial points, I should hasten to point out a handul o caveats. I most certainly do not dispute that theservices o natural ecosystems are valuable to humanity. Nor do I dispute that the services o some such ecosystems


8/23


are morevaluable than are any alternative uses that might be made o the areas they occupy. Consequently, there aresurely instances in which land use could be made more rational and, generally, socially benecial by undertakingpublic policy to preserve natural habitats.

What has not been satisactorily established is the generality o such propositions. In the body o this paper I pursuetwo major themes. Te rst is that many well known tracts on ecosystem services do not, in act, make a general andcompelling case or their economic value. Te second is that there is a good reason or the rst observation: valuingecosystem services is an extraordinarily dicult undertaking. In the nal section o the paper I return to my mainquestion in light o these observations. What doesan ecosystem service ramework imply or conservation policy,and does adopting it move the debate orward?

2. The value o ecosystem services: Some prominentexamples

In this section I review some prominent examples o empirical work on the economic valuation o ecosystem services.Te point I wish to make with this review is that some o the works on which the perception that ecosystem servicevalues are important and signicant are controversial. Tey do not establish aprima faciepresumption that the benetso ecosystem preservation exceed those o alternative use. Nor, it should be pointed out, does the selection o a handulo studies whose methods can be aulted or the generality o whose conclusions can be disputed establish that ecosystemservices are notvaluable or important. It does, however, suggest that we should consider more careully the empiricalevidence or that assertion. Tis is the inquiry to which I turn ater discussing a handul o prominent examples.

The millennium ecosystem assessment

Publications discussing ecosystem services and their value oten reerence the recently concluded MillenniumEcosystem Assessment (MA; see, e.g., MA 2005). Despite the characterizations one sometimes reads o the MA,the reader should appreciate a ew things about its purposes and results. First, the Millennium EcosystemAssessmentdid not conduct and report new research. Rather, as the term I have emphasized suggests, it assessedwork already inthe literature. Second, the MA was, in the nal analysis, restrained in its assessment. It did not make as bold claimsor ecosystem service values as have appeared elsewhere in the literature.2 Finally, some commentators appear tohave conused the MAs suggested defnitionsand classifcation schemesor estimateso value. Te MA proposed adivision o ecosystem service values among provisioning, regulating, cultural, and supporting services. While thistaxonomy may be useul,3 denition and estimation remain very diferent undertakings. Te reader must, then,look elsewhere or empirical estimates o ecosystem values.

The Value o the Worlds Ecosystem Services and Natural Capital

Perhaps the best-known example o ecosystem service valuation remains a paper published in Naturein 1997 byRobert Costanza and twelve coauthors. In Te Value o the Worlds Ecosystem Services and Natural CapitalCostanzaet al. suggested that a minimum estimate o such values was US $33 trillion.4 I will consider in somewhat moredetail the benet transer approach taken by Costanza et al. below. For now, however, I will just note that the

2 A summary o the MAs ndings is oten reported as that 15 o 24 o the ecosystem services studied in the MA are in decline or are being degraded. While

perhaps cause or considerable alarm, this is not a statement concerning the value o ecosystem services.3 axonomic and denitional matters are not yet settled, however. No ewer than ve denitions o ecosystem services have been proposed in theliterature (see Costanza, et al., 1997; Daily 1997; MA 2005; Boyd and Banzha 2007; Fisher et al. 2007). Moreover, it is not clear to me what the lastdecades consideration o ecosystem services has added over and above earlier suggestions or the classication o total economic value o naturalsystems as proposed by the late David Pearce and R. Kerry urner (1992).


9/23


study set of a re storm o criticism, particularly, albeit not exclusively, rom economists (see Ayres, 1997; Smith,1997; oman, 1998; Freeman, 1998; and Pearce, 1998). I the point o Costanza, et al., were to demonstrate thevalue o natural ecosystems in sustaining lie, it delivered, as Michael oman (1998) noted, a serious underestimateo innity. I, on the other hand, it was intended to provide an economically meaningul measure, it reported alogical impossibility: awillingnessto pay that exceeded world income which would, in turn, limit the abilityto pay(Smith 1997). While it might reasonably be argued that this logical oversight was more semantic than substantive,5a more undamental criticism o the Costanzaet al. (1997) work is that it conused marginal and total values. Ineconomic theory the value o a good is determined by how much benet we receive rom a little more o that good,phrased in terms o the other goods we might give up to obtain it. Costanzaet al. (1997) have oten been criticized(see, recently, or example, Plummer 2008) or applying values obtained or ecosystems at one place and supposingthat such values could be applied to supercially similar but very diferently situated systems at other places. SinceCostanzaet al. (1997) continues to be widely cited and some subsequent work could be subjected to the samecriticisms (see, e.g., Mates 2007), readers o work on ecosystem services may be surprised to discover that mosteconomists who have reviewed this work have ound it seriously deective.

Economic Reasons or Conserving Wild Nature

Motivated in part by the criticisms ofered o the work by Costanzaet al. (1997), some o the same authors joinedwith Andrew Balmord and others to produce another piece that was published in Sciencein 2002. Balmord et al.(2002) attempted careully to identiy studies that reported the marginal value o land retained in natural habitatand compared it to the projected value o the same land i it were converted to more intensive use. Tis study hasalso been widely cited, but is also problematic in several respects.

Te rst concerns coverage and representativeness. Balmord and coauthors reported that they had surveyed over300 case studies, but ultimately identied only ve which compared the benets o ecological preservation withits opportunity costs and used broad and generally accepted measures o opportunity costs. O the ve studies

identied, two were not published in peer-reviewed journals (despite having been written, respectively, eight andour years beore Balmord et al. 2002). It is open to question whether such a limited sample is in any wayrepresentative o broader circumstances in conservation policy.

Moreover, two o the ve studies cited relied in part on the ecosystem service o carbon sequestration to generate valuesin excess o the opportunity cost o habitat preservation. Te inclusion o such values is entirely appropriate i theobjective is to demonstrate that theglobalbenets o conservation exceed the localcosts. However, the internationalcommunity has, by and large, still not developed a consistent set o rules or crediting developing countries orcarbon sequestered in orests and other ecosystems.6 Tis observation begs the question o what an emphasis onecosystem services is intended to accomplish or international conservation. I the objective is to convince localcommunities to preserve natural habitats because o the benets they will coner on themselves directly, work such asthat by Balmord et al. (2002) does not necessarily make the case (see also Kremen, et al. 2000, Naidoo and Ricketts2006). I the objective is to convince global donors that purchasing ecosystem services rom developing countrieswould be wise, work such as that by Balmord and his coauthors may be more compelling, but this argument hasbeen made in the conservation literature or many years (see, e.g., Pearce and Moran 1994).

4 Costanza et al. derived a range o value estimates rom $18 to 54 trillion, but, because o exclusions, suggested that the average estimate o $33 trillion

was a lower bound on actual values.5 Willingness to pay is, in economic theory, the amount someone would and can pay or a some good or collection o goods. It is necessarily limited by theresources at ones disposal, i. e., wealth. Tere is not anything necessarily logically inconsistent about saying that the benets we receive or ree romnature are o equal or greater subjective value to us than those we purchase in markets.

6 Developing such guidance is the mission o the REDD (Reduced Emissions rom Deorestation and Degradation) programme.


10/23


The Catskills Parable7

Te next example or consideration is one o the most oten-cited pieces o evidence or the importance oecosystem services. In the late 1990s New York City sought to secure the saety o water provided rom its reservoir

in the Catskills outside the city. In a widely praised (see, e.g., Levin 1999) and cited, albeit brie, piece in Nature,economists Geofrey Heal and Graciela Chichilnisky (1998) argued that the city had avoided the substantial capitaland operating costs o a new water treatment plant by devoting a raction o their expense to the preservation onatural habitat in the vicinity o the reservoir. While this example has oten been cited as evidence o the valueecosystem services provide, it is problematic on several counts. First, the city aced the requirement to construct atreatment plant based on aregulatoryrather than an actual demand. Te United States Environmental Protection

Agency required that municipalities drawing rom surace sources either treat their water or obtain a waiver romthe regulatory requirement. Tere was no evidence that New York Citys drinking water was unsae or requiredtreatment (NRC 2000; Sagof 2002). While one positive aspect o maintaining more natural habitat or wildliearound the reservoir was that it could then support more wildlie and biodiversity that some constituents valued,there was no evidence that citizens were willing to pay extra or water in order to support wildlie conservation, anobservation made at the time by some advocates o biodiversity protection (Salzman, et al. 2001).

Moreover, the assertion that what happened in the Catskills approximated a situation in which payments weremade or ecosystem services is not well supported by the evidence. Mark Sagof (2002) reports that while the 1997Memorandum o Agreement between the EPA and the City called on the latter to solicit purchase o 355,000 acreso land in the watershed, by February o 2002 only a little over 19,000 acres had been acquired, most o them inthe orm o conservation easements restricting use, rather than o ee simple purchase. Environmental organizationsexpressed dissatisaction with the pace o acquisitions (Sagof 2002). On the other hand, however, some eforts torestrict land use without what were regarded by owners and developers as adequate compensation were decried asthievery (Daily and Ellison 2002). However one characterizes what happens in the Catskills, it seems reasonableto conclude that it was not what it has oten been described as: a set o large-scale transactions entered intovoluntarily to realize the preservation o important ecosystem services.

Perhaps the most troubling aspect o the Catskills Parable is the lack o scrutiny to which it has been exposed bythose who cite it as evidence o the importance and value o ecosystem services. While the episode is widely citedin the literature on ecosystem services, tracing such citations back to their origin leads one to the 2-page 1998Chichilnisky and Heal Naturearticle, a piece that lists no reerences, but would appear to contain signicant andmaterial actual misstatements (Sagof 2002). It is troubling that so slight a piece that has been so inuential in thediscourse about ecosystem services. Tis provides a useul cautionary tale. One should be careul in the evidenceshe cites to support the importance o ecosystem services.

Vittel

Te next two examples are, I believe, more substantively valid, but each begs an important question. Te rstis o the Vittel water bottling acility near the Vosges Mountains in northeastern France. 8 Te company bottles

water rom a spring in the region, which it then markets worldwide.9 In recent decades the quality o water romthe underground aquier in the area has been threatened by more intense cattle operations in the region. Tepresence o the cattle could lead to increased nitrate in the groundwater and, eventually, a decline in the qualityo water rom the spring. As the process by which the nitrate would enter groundwater is well understood, as isthe relationship between nitrate and cattle ranching operations, Vittel readily perceived how it could maintain thevalue o its ranchise by compensating local ranchers or modiying their practices. Consequently, it entered intoan arrangement to do so.

7 I have borrowed this designation rom the title o Mark Sagofs (2002) paper, on which I have drawn heavily.8 I rely heavily on Perrot-Matre (2006) or this paragraph.9 One might also pause in passing to speculate as to the aggregate ecological consequences o maintaining the purity o a spring so that its water can

continue to be shipped to customers as many as 10,000 kilometers away


11/23


Te Vittel example has oten been cited as an example o the importance o ecosystem services (see, e.g., Ranganathan,et al., 2008; U. S. Forest Service 2007, Perrot-Matre 2006). Indeed, it does seem to be a very apposite case. Whileit does not speak to the diculties or methods o valuation per se, it is clearly an instance in which improvedenvironmental perormance would create large enough gains or the beneciary to induce it to compensate theparty that restricts its options in order to supply such benets. Te only question is whether this is an exception thatproves a rule. Reducingthe intensity o agricultural use is perhaps a step back toward a more natural ecosystem,but is a ar cry rom adopting measures to preserve relatively pristine landscapes. Te Vittel example is a good one othe benets o less intensive use o the landscape. Inasmuch as even this relatively straightorward efort to concludea transaction in relatively minimal ecosystem services was dicult to consummate, it begs the question o whatrealistically are the prospects or concluding deals or still more dicult, difuse, and diverse ecosystem services.

Pollination on the Finca Santa Fe

Another case study that briey (we will see momentarily why interest was transitory) attracted interest amongpractitioners and advocates occurred on a cofee plantation in Costa Rica. One ecosystem services that has attractedconsiderable attention has been pollination. Areas o retained natural habitat in the vicinity o agricultural eldsprovide reuge or insects and birds that may pollinate agricultural crops. In a careully conducted study aylorRicketts and his colleagues (2004) demonstrated that cofee yields were about 20% greater on plants within onekilometer o retained orest areas, presumably because they are better served by native pollinators.

Tis observation begs a question. All 480 hectares o habitat located within one kilometer o preserved oresthabitat in the study be Ricketts, et al., were owned by a single owner. I it were, in act, the case, that the valueo cofee production on the plantation would be higher i 157 hectares o adjoining orest were preserved, what

would prevent the owner o the cofee plantation rom acquiring the orest and preserving it or his own benet?In act, however, careul consideration o the numbers reported in the Ricketts, et al. (2004) study suggest thatthe cofee plantation owner might have been better of i, rather than acquiring the 157 hectares to assure that it

remained in orest, he had instead acquired the land, cut the trees, and planted more cofee on it. While yieldsmight have declined elsewhere on his land, the increase in production on the newly cleared land would have morethan compensated or it. As it turned out, events have revealed a related phenomenon. Douglas McCauley (2006)reports that the cofee plantation was subsequently uprooted and replaced by a pineapple arm. Pineapple does notrequire insect pollination at all.

Te point o this and the above examples is not to suggest that ecosystem services are not important. Tey surelyare in aggregate, and they may well be on the margin in a number o situations o policy interest. Te point,rather, is that cases in which such values are most easily demonstrated oten lead to unsurprising and not veryilluminating policy implications. Te value o ecosystem services can be most easily estimated when simple, otenprivate transactions can be structured to assure that they are maximized. Conversely, the cases o greatest policyinterest arise precisely when the benets o ecosystem services are difuse and most dicult to measure.

3. Impediments to estimating ecosystem service valuesIn this section I discuss why the value o ecosystem services is so dicult to measure, as well as why some o thepublished estimates o such values might be taken with a grain o salt.

The production unction approach

A number o recent papers and studies have suggested that researchers take a production unction approach to theestimation o ecosystem service values (see, e.g., Boyd and Banzha 2006; Daily, et al., 2009; SAB 2009). Underthis approach natural ecosystems would be regarded as an element o machinery that can, in combination with


12/23


manuactured capital, labor, and other such inputs, produce outputs o value to humanity. Such processes mightgenerate products or sale, such as agricultural commodities, lumber, or sh, or through the process o householdproduction they might create goods and services such as pleasant household settings or amily health that arevaluable even though they are not generally traded in markets.

Te economic logic behind the production unction approach is impeccable. Te value o an input to productionis, rom elementary economic theory, the price10 o the output in whose production it is employed times themarginal product o the input in the production o the output. Te problem arises, however, when we have toimpute such a price in the absence o market data. Empirical economists typically cannot observe the marginalproduct o an input in the production o an output. Te thought experiment implied What would happen tothe production o good Yi we varied the quantity o goodX

1by a small amount while holding the quantities o goods

X2,X

3, . . . ,X

Nconstant? can rarely be conducted in practice.

Te alternative is to speciy a production unction positing a unctional relationship between the quantities oinputs employed and the quantity o output11 produced and estimate it employing statistical techniques. Tisapproach has indeed been used, both in the general empirical economic literature and in work on ecosystem

services (see, e.g., Acharya and Barbier 2001). Several comments and caveats are in order, however. I begin withsome o the more basic.

Te rst is that estimating production unctions has long been regarded as problematic by empirical economists.Te reasons or their caution are instructive in the context o ecosystem services. When an economist estimates aproduction unction she must make certain assumptions about the variables that she assumes afect production.Naturally enough, i she wants accurately to estimate the efects o particular inputs on production, she mustsuppose that she has accounted or all relevant actors. I she has not, it is likely that some o the actors shepresumes to account or higher levels o production in one area relative to another are not actually contributing toproduction but, rather, capturing the efect o some other variables that has not been considered explicitly.

Te canonical example o this phenomenon is the application o ertilizer in agriculture.12

I an econometricianwere to estimate the efects o capital, labor, and ertilizer applications on agricultural outputs, he might well ndthat the efects o ertilizer appear to be negative: armers who are observed to apply more ertilizer harvest lessoutput than those that use less. Why would this be? Are the armers making oolish choices, paying to buy more oan expensive input that is actually hurting, rather than helping, them?

Not necessarily. A more plausible explanation is that armers who apply more ertilizer are attempting to make upor the absence o another actor that has not been measured, and are not ully able to do so. Perhaps armers whohave inherently less ertile soil apply more ertilizer.

I suggested that this example is instructive in the context o ecosystem service value estimation. Te simplestsolution to this problem o unobserved variables would be to make the efort to observe them. Yet this is not the

solution typically prescribed by econometricians. Rather, it is more common to employ a bit o theoretical sleight-o-hand to obviate the problem. Te problem o unobserved variables arises because some o the explanatory actorsthat areobserved are correlated with some that are not, and consequently pick up the efect o the unobservablevariables. I it were possible to express the production relationship as a unction o variables that are uncorrelated

with any unobserved actors, the problem would be solved. Tis is exactly what econometricians try to do. Teytypically preer to estimate cost unctions or, better yet, prot unctions. A cost unction relates a producers costo production to thepriceso the inputs he purchases and the quantityo output he produces. A prot unction

10 More generally, we might say that the value o an input in utility terms is the marginal utilityaforded by the output times the marginal product o the

input in its production, but a still more undamental result o economic theory holds that the price o any product is determined by the ratio o themarginal utility it afords to that o some other good denoted as numeraire.11 More generally, we might say positing a relationship between the quantities o inputs employed and the quanties o outputs produced, but Ill keep the

example simple or now.12 I base this statement on the act that this was the example my instructor employed when I took my introductory econometrics course some thirty years ago.


13/23


relates prot to the prices o inputs andoutput. As such prices are typically not correlated with local productionconditions, estimation can proceed.13

Te point I want to emphasize here is that i including variables such as the efects o ecosystem services onproduction were easy, empirical economists would have done so a long time ago, as the need to develop moresophisticated methods o estimation would have been obviated. O course, virtually any worthwhile empiricalundertaking will be dicult. So, the relevant question is whether these diculties have been, or will be, overcomein empirical work.

A paradox o valuation

Even indiferent students o economics are likely to remember the paradox o value: why is water, which is soessential to lie, so cheap, while diamonds, which have such limited, and largely ornamental, uses so expensive?Te answer is that water is (generally) abundant with respect to the uses to which it is put, while diamonds are(generally) scarce relative to the demand or them. Te exceptions prove the rule: someone dying o thirst experiencing an extreme scarcity o water would surely trade all the diamonds he had or a drink.

One might propose a corollary, a paradox o valuation. I something is more or less valuable depending onwhether it is more or less scarce, then the only way to place a value on a good that does not have an evident marketprice is to identiy some circumstances under which it is more scarce and others under which it is less so. Now this

will be relatively easy to do i the public good whose value we are trying to establish is relatively local; that is, it hasa discernable efect on those who are close enough to its source as to enjoy its benets, but others who are moredistant receive only negligible benets.

I the benets aforded by such a public good are localized, however, it begs the question as to whether the good inquestion ought to be considered public at all. I the radius o dispersal is compact enough, why would it not make

sense or afected parties to merge their interests and internalize the externality by placing the source and thebeneciary o the good under common ownership? Tis is essentially what Vittel did in the example cited above;it internalized the externality o cattle ranching in its watershed by purchasing a portion o local ranchers landuse rights. For ecosystem services dispersed over somewhat broader areas a number o authors have documentedinstances in which local communities evolve rules or their management (see Ostrom 1991; Balland and Platteau2002; more generally, Coase 1960 discusses ways in which afected parties might reach agreement or the provisiono public goods or private goods with extensive externalities).

Tere are also instances in the literature in which it would appear that private parties may be behaving optimallyalready with respect to the provision o local, at least, ecosystem services. Tis would appear to have been the casein the pollination services example developed by Ricketts, et al., (2004) and discussed above. In a similar studyTorsnes (2002) nds that construction lots abutting retained natural areas command modestly higher prices thanthose at a greater distance away. It would appear, however, that, absent legal restrictions against doing so, residentialdevelopers would make more money by clearing more land or construction than they would lose by reduced salesprices on lots now rendered less attractive by the reduction in adjoining natural areas.

When there is enough spatial variability to acilitate easy measurement o values, then, it seems reasonable tosuppose that local people would develop an appreciation or such values and evolve institutions to preserve them.Te opposite situation is more problematic rom a policy perspective. What i the services provided by natural

13 Explaining the reasons or this gets complicated. Basically, principles o duality assure that production processes may be expressed equivalently by

relating the quantity o output to the quantities o inputs or by expressing the cost (or prot) associated with production to the prices o inputs and thequantity o output, in the case o the cost unction, or prices o inputs and output in the case o the prot unction. Tere is also a general preerenceor estimating prot rather than cost unctions, as the quantity o production may also be afected by unobserved actors. For this reason much o theearly empirical work on estimating cost unctions was conducted on regulated public utilities, on the argument that output was determined by regulatoryauthorities rather than by actors that might also afect costs.


14/23


ecosystems are provided to very large publics? Tis would be the case i the intensity o service provision does notvary greatly over the landscape. Tere would, then, be little variation, and little prospect or inerring value romobserving diferences in the volume o service ows received rom diferent areas. In extreme instances, such asthe sequestration o atmospheric carbon dioxide and the preservation o endangered species valued or moral andaesthetic reasons, the public goods are truly global and it would be impossible to value them by observation ocross-sectional variation.

Moreover, even i it were possible to identiy somevariation in the intensity with which ecosystem services areprovided across the landscape by diferent sources, it becomes increasingly dicult to disentangle the efect o anyparticular wetland, orest, meadow, etc., rom that o the multitude o others that would inuence production i thebenets o each do not dissipate relatively quickly in the distance rom source to receptor areas.

Tus, there is a paradox o valuation. We would like to know the value o services whose benets accrue oversuch large areas o the landscape that ecient arrangements or their private or communal provision are unlikelyto evolve. Yet these are exactly the circumstances under which the exercise o valuation will prove most dicult.

Spatial correlation and endogeneity

Te estimation o production, cost, or prot unctions is oten orgone or a related technique: hedonic estimation.A hedonic model is one in which the value o some good or asset such as a hectare o land is determined by isattributes.14 wo other undamental results rom economic theory are, rst, that the rental value o an asset, suchas a hectare o land, is determined by the prots that can be earned rom using that land in its highest and bestuse, and second, that the purchase price o such land is determined by the net present value o the stream o rentalearnings that could be realized rom its ownership. So, the value o ecosystem services should be reected in therental and purchase prices o land receiving them.

Tere is a long history o hedonic studies o land value based on the uses made o nearby lands []. Some o thestatistical issues discussed above reappear with a vengeance in this context, however. Since it is oten extremelydicult to measure the actual services provided to one parcel o land relative to those provided to another, a morecommon empirical procedure is to estimate the value o land as a unction o (among other actors) proximity toeatures o the landscape such as wetlands or orests. Tis raises a dicult issue. Reliable estimation o the empiricalrelationships between land values and the actors that explain them requires that allo the latter be included in theanalysis. I they are not, some o the other explanatory actors included may not themselves explain variations invalue so much as proxy or other actors that are not included. Tis becomes particularly problematic when landvalues are determined by similar actors in adjoining or nearby places in the landscape.

Te essential points to appreciate here are that 1) the goal o a hedonic pricing exercise is to understand how land

valueat one point in a landscape is determined by land

useat another; 2) it is more likely that more intensive

use

will be made o land when the eatures o its location imply that greater value will arise rom more intensive use; and3) the value-determining eatures o the landscape are likely to be highly correlated or properties that are relativelyclose to one another. Tis combination o considerations will likely imply that the benets o ecosystem services

will be underestimated in hedonic studies. Te reasoning encapsulates the points above. Consider a arm locatedadjacent to an area o retained natural orest. Te value o that arm may well be lower than that o a similar armsituated amidst other arms and ar rom remaining orest areas. Both the isolated arm and the orest adjoining itare likely situated in areas identied as having been less advantageous or arming.

In recent years econometricians have derived techniques or dealing with the combined efects o correlation in valuesacross space and omitted variables. Such techniques presume the existence o adequate instrumental variables

14 Hedonic is derived rom the same Greek root as hedonism, and, like the latter term, relates to pleasures. A hedonic valuation exercise attemptsto estimate the price o a good by the pleasures or, more generally, attributes its ownership ofers.


15/23


actors that may explain the decision not to devote nearby land to more intensive use while not afecting the valueo land that has been placed into more intensive use. Such variables are dicult to identiy, however. While worksuch as that o Irwin and Bockstael (2002) illustrates the approach, results cannot yet be considered denitive.

Taking the Con out o Econometrics15

Despite the conclusion o the last paragraph, most hedonic studies relating land value to the retention o relativelypristine ecosystems show that proximity to ecosystem services afords positive benets. Inasmuch as the actorsexplicated above would seem to suggest that estimates o value would be biased downwardsby the actors considered,one might suppose that existing studies make the case that ecosystem services are valuable a ortiori.

Tere are, however, other concerns. It is dicult to do controlled experiments in economics on a temporal, spatial,social, and nancial scale relevant to the resolution o important policy questions. Tus, rather than having theresearcher control the circumstances under which experiments are conducted, as would be the case in laboratorysciences, economists must typically look or natural experiments in which circumstances have ortuitously cratedsituations in which variables o policy interest in our case, areas preserved or the provision o ecosystem services difer while other relevant circumstances remain relatively constant. Perect natural experiments are vanishinglyrare, however, and researchers must instead attempt to correct or the inuence o variables they cannot controldirectly.

Tis is typically done in practice by supposing that there exists a unctional relationship between all o the explanatoryvariables o interest and the dependent variable (in the case o a hedonic model, the price o a property). Teempirical procedure works then by seeing how well the unction ts the data.

Tis procedure is problematic, however, as it does not comport well with the statistical theory on which empiricalprocedure is supposedly based. Tat theory presumes hypothesis testing. Te researcher orms a hypothesis or

example, that property values increase with the proximity o a property to a wetland and can test this conjectureby seeing whether or not the estimated efects o proximity are greater than those that might arise simply romrandom variation in the data.

In testing such a hypothesis, however, researchers inevitably must posit certain maintained hypotheses: thatthe set o other explanatory variables they have included is appropriate and complete, that the unctional ormthey have chosen is correct, and that the spatial and, in some instances, temporal coverage o the data is adequateand representative. It is, however, very unusual or a researcher to collect a set o data, perorm a single statisticalprocedure on it (in the parlance o the discipline, a single regression), and report the results o that procedure.Rather, it is common or researchers to experiment with a number o diferent specications typically, diferentunctional orms, combinations or variables, and/or special techniques and report only a subset o the resultsobtained rom doing so. o give an example, Robert Costanza and his coauthors (2008) in their study o the valueo coastal wetlands in protecting inland areas against hurricane damage, report that they experimented with ninediferent models, but report results or the one that, in their judgment, best t the data.16

Tere is widespread agreement that this type o pre-test bias is prevalent and problematic (see, e.g., Leamer 1978;Kennedy 2002; Glaeser 2007), although there is probably less agreement as to the severity o the problem and theappropriate solution. Tere are certainly examples rom other areas o economics in which researchers have exposedthe sensitivity o results to specic assumptions (see, e.g., Easterly, et al., 2004).

15 Te title o this subsection is borrowed rom Edward Leamers 1978 paper (and plea) Lets ake the Con out o Econometrics.16 My point in citing this work is not to criticize the authors so much as to credit them or having been more orthright than are many o their peers in

describing their procedures.


16/23


The battle o the Bayesians

Even i researchers do violate the principles o statistics in conducting specication searches, is there a reason tosuppose their results will be biased in predictable ways? Tere may be. Edward Glaeser writes:

Economists are quick to assume opportunistic behavior in almost every walk o lie other thanour own. Our empirical methods are based on assumptions o human behavior that would notpass muster in any o our models . . . While economists assiduously apply incentive theory to theoutside world, we use research methods that rely on the assumption that social scientists are saintlyautomatons. (Glaeser 2006)

What are the incentives that might motivate economic researchers? One that has oten been suggested (see,e.g., Pearce 2005) is publication bias.17 Academic economists obtain recognition and advancement in their eldby publishing in scholarly journals. It is seen as almost a de acto requirement or publication that one nd asignicant in the statistical sense o that term18 result. It may then, be that journals disproportionately acceptpapers which report statistically signicant results (suggesting that they may then reject large numbers o papers

that do not report such results). O course, i we do apply economic theory to economists, it seems unlikely thatlarge numbers o economists would spend the time and efort required to write up results they knew would likelybe rejected when submitted or journal publication. Tey might instead prescreen research projects, bringing toruition only those that yield signicant results. Tey might also be tempted to engage in specication searchesuntil they derive results that are (ostensibly, ignoring the pre-testing that may have occurred) statistically signicant.

It is too cynical to suppose that researchers slant their results simply to get them published. Another reason orwhich researchers might nd and report strong positive results is that the researchers rmly believe that such resultsare correct. Te process o specication search can be given a Bayesian19 interpretation. Suppose that I believe thatthe services perormed by a particular orest ecosystem are o considerable value. Suppose also that I amass a set odata, choose what seems to be a plausible unctional orm, and perorm a statistical analysis to test the hypothesis

that the value o ecosystem services provided by this area o orest are no greater than the opportunity cost oits preservation. Finally, suppose that I cannot reject the hypothesis. I can react to this inormation in one or acombination o two ways. First, I can say Tats surprising! I must have been wrong, and change my mind. Or,I can say Im pretty sure Im right about the value o ecosystem services; since I didnt get the right result, someother aspect o my analysis must be wrong. Im going to try a diferent specication and see i I get the right result

with it.20

Te reader o such work is then aced with her own problem o inerence, inormed by her own belies. I conrontedwith a claim that ecosystem services are o signicantly greater value than the opportunity cost o their preservation,but dogged by the suspicion that such a claim is based on extensive specication searching, does she believe theclaim or discount it? It seems that the general answer will be that she will ollow her own prior belies on the matter.

Such a scenario could lead to what I suggest might be a battle o Bayesians. A researcher believing that ecosystemservices are important and valuable would credit those ormulations o the empirical model that led to theconrmation o his prior belies. A skeptical reader could dismiss the researchers ndings on the presumption thatthe researcher had continued to test specications until he ound the one that comported best with his prior belies.

17 Publication bias is also perceived as a problem in the pharmaceutical research industry, where some allege that results avorable to a research undersinterests tend to be published, while those that are not are stuck in the le drawer.

18 A result is said to be statistically signicant at the x% level i the probability that it would have been ound to be as large as it is i it did not, in act,have any consistent efect, is less than x%.

19 Te Reverend Tomas Bayes (c. 1702 1761) was a British mathematician best known or ormulating Bayes Law, a theorem describing how probabilities

are updated in the presence o new inormation. Bayesian analysis concerns the ways in which prior distributions are updated on the basis o newinormation so as to orm posterior distributions.20 I do not believe that this is unusual. I have certainly proceeded in this ashion mysel at times. It is not unusual to overhear one colleague ask another

Did that new specication you tried work? where I presume to work would be to produce statistically signicant results o the sign and magnitudeexpected.


17/23


Which is right? In a sense and despite the contradiction between their positions both are. Each is equallyentitled to her prior belies concerning ecosystem services, and there is nothing necessarily irrational or inconsistentin each updating her belies as she has. Such divergent positions cannot be reconciled absent agreement on anexperiment a set o data on which a hypothesis can be tested that each agrees will resolve the matter.21

An extreme example o a Bayesian stando: Stated preerence studies

Nowhere is the problem o dening agreed-upon procedures or the estimation o value more vexing than withstated-preerence studies. Economists oten distinguish their discipline rom other social science approaches bynoting that they base their analyses on what people actuallydo rather than on the subjective thought processes thatpeople saymotivate their actions. It is, then, not surprising that sharp disputes have broken out between economistsciting the above distinction and those who abandon it byaskingpeople what theywoulddo rather than observing

what people havedone.22

Adherents o the latter procedure conduct stated preerence (as contrasted to revealed preerence studies).Examples include contingent valuation studies, in which survey respondents are asked how much they would payor a particular good or service,23 and conjoint analysis and choice experiments, in which respondents are askedto choose among diferent combinations o public and private goods to determine their preerence or the ormerand the consistency o their preerences between them.

Stated preerence methods have oten been used in the valuation o ecosystem services. Richardson and Loomis(2009; see also Loomis and White 1996) survey a number o stated preerence studies that have been conductedon the value o threatened and endangered species. Naidoo and Adamowitz (2005) consider the value o birds toprotected areas in Uganda. Many such studies exhibit a troubling eature, however. As Brown and Shogren (1998)commented in reviewing 18 earlier stated preerence studies, the average person was willing to pay about $1000to protect 18 diferent species . . . . Many will nd these gures suspiciously high. Tey are suspiciously high

because there are so many public goods to whose maintenance a survey respondent might be asked to contribute.More generally, critics have long alleged that respondents to stated preerence surveys regard them as opportunitiesto purchase moral satisaction (Kahneman and Knestch 1991) albeit without having to actually pay or it rather than as considered statements o budget-constrained choices. Peter Diamond (1996) has suggested thatrelatively simple adding-up tests can be conducted to determine the consistency o stated-preerence studies,although other authors suggest that Diamonds conditions are more stringent than need be implied by receivedtheory (Smith and Osborne 1996).

Tis is precisely the problem, however: diferent commentators disagree as to what received theory requires, and hence,as to whether stated preerence studies provide valid estimates o the social values o things like imperiled ecosystemservices. While V. Kerry Smith was writing over a decade ago, it his likely that his assessment remains accurate:

Indeed, there is a curious dichotomy in the research using CV or nonmarket valuation.Environmental economists actively engaged in nonmarket valuation continue to pursue very technicalimplementation or estimation issues, while the economics proession as a whole seems to regard the

21 Tis standard echoes Karl Poppers (1959) notion distinguishing science as a realm o inquiry whose propositions must be alsiable i they are to bemeaningul. I we are to credit statements such as the services provided by a natural ecosystem are more valuable than the opportunity cost o itsmaintenance, or the opposite, we must have some agreed-upon set o techniques or measuring values and opportunity costs. My concern in much othe empirical literature on ecosystem services is that one side or the other in the debate can too easily say you have reported the wrong experiment.

22 See, or example, Nordhaus and Kokkelenberg (1999), who wrote:Some techniques such as hedonic-price or travel-cost studies rely on behavioral or market-based estimates; while these estimates are subject to signicantmeasurement problems, they are conceptually appropriate in economic accounts. Other techniques, such as contingent valuation, are not based on actual

behavior, are highly controversial, and are subject to potential measurement errors. [emphases added]23 I simpliy greatly in the interests o brevity. Open-ended questions o the orm How much would you pay or x? are no longer asked by most state-o-the-art practitioners o stated preerence studies. A more likely question would be a variant o Would you pay $y or x? possibly ollowed by I youwould pay $y, would you also pay $y + z or x? Te diference between the answers given to willingness to pay and willingness to accept questionshas also generated great controversy in the literature.


18/23


method as seriously awed when compared with indirect methods. Tey would no doubt regardthis urther technical research as oolish in light o what they judge to be serious problems with themethod. (Smith, 1997; p. 42)

Te proession has, essentially, agreed to disagree as to the validity o the approach.24 Obviously, this state o afairsmakes any estimates o ecosystem values based on stated preerence approaches problematic.

Beneft transer

Another approach has been employed in a number o ecosystem service valuation studies. Probably the mostprominent o these was the study published in Naturein 1997 by Robert Costanza and his colleagues. In that piecethe authors combined data rom scores o earlier studies reporting the value o services generated by particularecosystems. Tey then extrapolated these service values on a per-hectare basis to other ecosystems o similar typesaround the world. As has already been noted, this procedure is problematic. I, as may oten be the case, ecosystemsthat are unusually valuable because o their proximity to centers o population or sources o pollution are thesubjects o study, extrapolation o estimated values to more remote areas will be inaccurate.

Te general technique o applying estimates derived in one setting to the valuation o services generated in anotheris known as benet transer, and has been the subject o considerable research (see, e.g., the survey in Navrud andReady 2007). Best practice now gravitates to unction rather than estimate transer, wherein researchers transera set o parameters describing monetary benets as a unction o underlying site attributes rather than simplya numerical measure o monetary benets themselves. Other approaches attempt to calibrate preerences byimposing consistency across diferent methods o estimation and categories o value (see, e.g., Smith, et al. 2006).

One troubling aspect o existing benet transer studies and related meta-analyses25 however, is that theyoten nd that the methodsemployed in individual studies may explain the resultsthey derive (see, e.g., Richardson

and Loomis 2009). At the very least this suggests that some approaches may be more reliable than others. Moregenerally, any approach that aggregates other studies can be no more reliable than the studies on which it is based.

One might counter the above observation by saying that aggregation o individual studies should, by the law o largenumbers, smooth the errors ound in each. Tis would be true i the underlying studies themselves were unbiased.I, however, the potential sources o bias discussed above are matters o concern, the biases o the individual studies

would be transerred to the study that aggregates them.

4. How does the ecosystem services ramework change

conservation policy?Its more enthusiastic advocates have hailed the adoption o an ecosystem services ramework as a watershed eventin the evolution o conservation policy. Is it likely to prove to be one?

Te underlying logic o the ecosystem services ramework is impeccable. Natural ecosystems do, o course, providea host o valuable services to society. Tere is no denying that the water purication, pollinator habitat, recreationalopportunities, erosion protection, soil regeneration, etc. provided to humanity are valuable. O course, there are

24

We agree to disagree is the phrase employed by a presenter at a recent workshop at which I discussed his and other papers using stated preerencemethods. Much the same message was conveyed by a prominent representative o the other side o the stated preerence debate, Peter Diamond, in apersonal communication in 2005.

25 A meta-analysis is a study o studies. It typically takes the estimates o earlier studies as a set o dependent variables which it then seeks to explain asunctions o the location, methods, or other attributes o the studies.


19/23


also questions as to how such values compare to alternative uses, how valuable they are on the margin, and whatdisamenities natural ecosystems also generate. Some contributors to the literature on ecosystem services note, orexample, that there may be rapidly diminishing returns in certain services with respect to the size o the ecosystemproviding them (see, e.g., allis, 2005),26 and that completelyunnatural systems might provide equal or superiorservices in some respects (see, e.g., urpie, et al., 2003)27 With respect to disamenities, it is instructive to note thatmany o Europes wetlands were drained in 18 th and 19th century eforts to control malaria (Bate 2002; McCauley2006).

It should also be remembered that many o the values ascribed to ecosystem services arise precisely because o thebenets they provide to decidedlyunnatural systems: large areas o monoculture crops benet rom pollination,vulnerable urban structures benet rom ood control, etc. I the benets o maintaining areas in natural ecosystemsdepend on the size and location o unnatural areas, there is obviously a balance to be struck between preservation anddevelopment. Tis observation would not come as any revelation to advocates o an ecosystem services approach,but it does underscore the importance o evaluating marginal tradeofs.

It may be problematic i the general conclusion drawn rom an analysis o conservation vs. development tradeofs is

that they are currently drawn with too much emphasis on development. Tere is an important result in economicsknown as the general theory o the 2nd best (Lipsey and Lancaster 1956). Put in very colloquial terms, it says thattwo wrongs may make a right in some circumstances, or, perhaps put more accurately, correcting one problemmay not improve welare when other problems are not addressed. While the situation is certainly very complicated,it is not clear, or example, that taking steps to provide greater ecosystem services or one community would alwaysand necessarily be the best measures to combat other environmental problems, such as global warming. Kareivaand Rufo (2009), or example, note that Massive inrastructure projects, monocultures o plantation trees,seawalls and levees, [and] biouels that ultimately accelerate land conversion are among the strategies suggestedor combating climate change. While noting again that the science is complicated and appears to remain in somedispute, one cannot simply assume that the maintenance o natural ecosystems is necessarily consonant with otherecological or social objectives.

Tese observations may be particularly trenchant as it would appear that concern or ecosystem services arises largelyrom conservation advocates concern or the preservation o biodiversity. Tere is some discussion in the ecosystemservice literature concerning the degree to which biodiversity preservation and ecosystem service provision are, inact, consistent objectives (see, e.g. Balvanera, et al., 2001; allis and Kareiva 2005; urner, et al., 2007; Nelson2009).

Tese concerns are likely to be particularly acute when considering particular strategies or realizing the purportedvalue o ecosystem services. In the absence o clear evidence o the demand or ecosystem services, conservation(and, in some instances, development) donors have oten underwritten integrated conservation and developmentprojects (ICDPs) in which local communities are subsidized to undertake projects such as the collection onontimber orest products, ecotourism, and pharmaceutical research and development based on natural products.

Such ICDPs have proved to be very controversial, with both their ecological (see, e.g., erborgh 1999; erborgh,et al. 2002) and economic (see, e.g. Wells and Brandon 1991; Ferraro 2000; Ferraro and Simpson 2002) benetsin considerable dispute. Many commentators are likely to be extremely disappointed i the current interest inecosystem services turns out to be only a recycled argument or a conservation strategy that has already been triedand, in the opinion o many, ailed dismally, in developing countries.

Yet some authors who have extolled the promise o the ecosystem service ramework write hopeully o thepossibility o ecosystems services . . . aligning conservation values and poverty alleviation (allis and Kareiva

26 Te author notes that increases in orest cover above 70% o a land area have no demonstrable efect on water purication. See also Kareiva and Rufo2009.

27 Te authors note that South Arican mountains might provide better water supply to the cities and armers below them i the mountains were coveredin concrete than i they native vegetation were restored.


20/23


2005), largely in hopes o tapping into international development unding.28 o those who have ollowed theevolution o conservation strategy over the past two decades, such suggestions elicit a disheartening sense odj vu.

While the new emphasis on ecosystem services and the operational mechanisms o payments or their provision difer in some important ways rom the ICDP strategies that have oten prevailed in recent decades, they arealso similar in one undamental way. o the extent that the ecosystem services emphasized in recent conservationliterature call or the preservation o natural ecosystems in order to provide localpublic goods services such as

water purication, erosion control, pollinator habitat, etc. they presume that local communities have provedincapable o recognizing and saeguarding these services on their own.

It is almost tautological that poor communities are poor because they ail in some respects to take the actionsthat would be required to improve their collective welare. Having stipulated as to this act, however, it seemsdisingenuous to suppose that poor communities will be better of i they conserve ecosystems whose preservation

would also benet an international conservation community whose underlying interest is in the preservation obiological diversity, not the welare o local communitiesper se.

How then, should the ecosystem service ramework afect conservation policy? It seems reasonable to suppose

that research should continue into the valuation o ecosystem services. I clear and compelling evidence can beassembled to suggest that the local value o preserved ecosystems will compensate the opportunity costs o theirpreservation, perhaps local development and international biodiversity interests can be aligned.

For the reasons cited above, however, research has not yet established that case. Given the circumstances, we mightconsider two paradigms in deciding on policy. Te rst is that provided by the theory o options. When aced

with uncertainty and irreversible consequences, it is prudent to exercise an abundance o caution. Te second,competing, view is provided by Ockhams razor. While there may be other possibilities, the most parsimoniousexplanation or why local people do not preserve their ecosystems is that they do not believe it is in their interestto do so.

While some poor communities may be suciently orward-looking as to wait or the resolution o uncertaintybeore making irrevocable choices concerning their ecological assets, it would be surprising i this were the norm.So, whether or not one thinks local communities are making wise choices in liquidating their natural capital,29they are likely to continue to do so in the absence o international payments or conservation. In the nal analysis,then, it seems that the consideration o ecosystem services does not leave conservation planners in a much diferentposition than they were beore with respect to practical conservation strategy. Payments will likely be requiredrom the wealthier to the poorer nations o the world on the basis o carbon sequestration in standing orests orintangible values o biodiversity i natural ecosystems are to be preserved. We may choose to call the goods or

which such payments are made ecosystem services, but the designation makes little diference with regard to thepractical need to structure international payments or conservation.

28 Te passage quoted here ollow immediately ater that quoted on p. 2 above, in which allis and Kareiva suggest that the realization o the market wortho ecosystem services has the potential to increase conservation unding by orders o magnitude.

29 It is worth remembering in this context that some authors who are sympathetic with the ecosystem services perspective believe such choices are, in act,wise, in the absence o international payments; see Kremen et al. 2000, Naidoo and Ricketts 2006.


21/23


Reerences

Acharya, Gayatri and Edward Barbier. 2000. Valuing groundwater recharge through agricultural production in the Hadejia-Nguru wetlands innorthern Nigeria.Agricultural Economics23 (3). 299 310.

Acharya, G., Bennett, L.L., 2001. Valuing open space and land-use patterns in urban watersheds. Journal o Real Estate Finance and Economics 22(23), 221237.

Ayres, R. U., On the Economic Valuation o Ecosystem Services, (Fontainebleu: INSEAD), 1997.Baland, Jean-Marie, and Jean-Philippe Platteau. 1996. Halting resource degradation: Is there a role or communities?Rome: FAO.Balvanera, Patricia, Gretchen C. Daily, Paul R. Ehrlich, aylor H. Ricketts, Sallie-Anne Bailey, Salit Kark, Claire Kremen, and Henrique Pereira.

2001. Conserving biodiversity and ecosystem services. Science291 (5511). 2047.Balmord, Andrew,Aaron Bruner, Philip Cooper, Robert Costanza, Stephen Farber, Rhys E. Green, Martin Jenkins, Paul Jeferiss, Valma Jessamy,

Joah Madden, Kat Munro, Norman Myers, Shahid Naeem, Jouni Paavola, Matthew Rayment, Sergio Rosendo, Joan Roughgarden, Katerumper, R. Kerry urner. 2002. Economic reasons or conserving wild nature.Science297, 5583. 950 953.

Benson, E.D., Hansen, J.L., Arthur, J., Schwartz, L., Smersh, G.., 1998. Pricing residential amenities: the value o a view. Journal o Real EstateFinance and Economics 16, 5573.

Bin, O., Polasky, S., 2005. Evidence on the amenity value o wetlands in a rural setting. Journal o Agricultural and Applied Economics 37 (3),589602.

Boyd, James, and Spencer Banzha. 2007. What are ecosystem services? Te need or standardized accounting units. Ecological Economics. 63 (2 3).

616 626.Brown, Jr., Gardner, and Jason Shogren. 1998. Economics o the Endangered Species Act.Journal o Economic Perspectives12. 3 20.Chao, P.., Floyd, J.L., Holliday, W., 1998. Empirical Studies o the Efect o Flood Risk on Housing Prices. Institute or Water Resources Report

98-PS-2.Cho, S.-H., Bowker, J.M., Park, W.M., 2006. Measuring the contribution o water and green space amenities to housing values: an application and

comparison o spatially weighted hedonic models. Journal o Agricultural and Resource Economics 31 (3), 485507.Coase, Ronald. 1960. Te problem o social cost.Journal o Law and Economics1. 1 44.Colby, B.G., Wishart, S., 2002. Quantiying the inuence o desert riparian areas on residential property values. Te Appraisal Journal LXX (3),

304308.Costanza, R., R. dArge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R. ONeill, J. Paruelo, R. Raskin, P. Sutton, and M.

van den Belt. 1997. Te Value o the Worlds Ecosystem Services and Natural Capital, Nature387: 253-260.Costanza, Robert, Octavio Prez-Maqueo, M. Luisa Martinez, Paul Sutton, Sharolyn J. Anderson and Kenneth Mulder. 2008. Te Value o Coastal

Wetlands or Hurricane Protection.Ambio 47 (4). 241 248.Daily, Gretchen C., and Katherine Ellison. 2002. Te new economy o nature: Te quest to make conservation protable.Daily, Gretchen C., Stephen Polasky, Joshua Goldstein, Peter M Kareiva, Harold A Mooney, Liba Pejchar, aylor H Ricketts, James Salzman, Robert

Shallenberger. 2009. Ecosystem services in decision-making: time to deliver. Frontiers in Ecology 7 (1). 21 28.Daniel, V., Florax, R., and Rietveld, P., 2007. Flooding risk and housing values: an economic assessment o environmental hazard. Working Paper:

07-02, pp. 26. Purdue University, College o Agriculture, Department o Agricultural Economics.Earnhart, D., 2006. Using contingent-pricing analysis to value open space and its duration at residential locations. Land Economics 82 (1), 1735.Easterly, William, Ross Levine, and David Roodman. 2004. New data, new doubts: A Comment on Burnside and Dollars Aid, Policies, and

Growth.American Economic Review94(3). 774-780.Ferraro, Paul J. 2001. Global Habitat Protection: Limitations o development interventions and a role or conservation perormance payments.

Conservation Biology15(4): 990-1000.Ferraro, Paul J., and R. David Simpson. 2002. Te Cost-efectiveness o Conservation Perormance Payments. Land Economics78(3): 339-353.Freeman III, A. Myrick, Notes or Valuing Nature Workshop: Commentary on Methods and Data Used by Costanza, et al., Paper prepared or the

Bowdoin College Environmental Studies Program Workshop on Valuing Nature, August 1998.Geoghegan, J., 2002. Te value o open spaces in residential land use. Land Use Policy 19, 9198.Geoghegan, J., Wainger, L.A., Bockstael, N.E., 1997. Analysis spatial landscape indices in a hedonic ramework: an ecological economics analysis

using GIS. Ecological Economics23, 251264.

Glaeser, Edward. 2006. Researcher Incentives and Empirical Methods. National Bureau or Economic Research. Working Paper 0329.Irwin, E.G., Bockstael, N.E., 2001. Te problem o identiying land use spillovers: measuring the efects o open space on residential property values.American Journal o Agricultural Economics 83 (3), 698704.

Irwin, Elena G. 2002. Te efects o open space on residential property values. Land Economics78(4). 465-481.Kahneman, Daniel, and Jack Knetsch. 1992. Valuing Public Goods: Te Purchase o Moral Satisaction, Journal o Environmental Economics and

Management22, 1. 57-70.Kareiva, Peter, and Susan Rufo. 2009. Using science to assign values to nature. Frontiers in Ecology 7 (1).Kennedy, Peter. 2002. Sinning in the basement: What are the rules? Te ten commandments o applied economics. Journal o Economic Surveys 16 (4).

569 589.Kremen, C. Niles, J., Dalton, M., Daily, G., Ehrlich, P., Fay, P., Grewal, D. and R. P. Guillery.2000. Economic incentives or rain orest conservation across scales. Science288:1828-1832.Leamer, Edward. 1983. Lets take the con out o econometrics.American Economic Review73 (1). 31 43.Leggett, C.G., Bockstael, N.E., 2000. Evidence o the efects o water quality on residential land prices. Journal o Environmental Economics and

Management 39 (2), 121144.Levin, Simon. 1999. Fragile dominion: Complexity and the commons. Cambridge, MA: Perseus.

Mahan, B.L., Polasky, S., Adams, R.M., 2000. Valuing urban wetlands: a property price approach. Land Economics 76 (1), 100113.McCauley, Douglas. 2006. Selling out nature. Nature443. 27 28.Millennium Ecosystem Assessment (MA). 2005. Overview o the Milliennium Ecosystem Assessment http://www.millenniumassessment.org/en/

About.aspx#. Accessed 17 April 2009.


22/23


Naidoo, Robin, and Wiktor L. Adamowicz. 2005. Economic benets o biodiversity exceed costs o conservation at an Arican rainorest reserve.Proceedings o the National Academies o Science102(46): 16712-16716.

Naidoo, R., and .H. Ricketts. 2006. Mapping the economic costs and benets o conservation. PLoS Biology4(11): e360. DOI: 10.1371/journal.pbio.0040360.

National Research Council (NRC) o the U. S. National Academies o Science. 1999. Watershed management or potable water supply: Assessing theNew York City Strategy. Washington: National Academies Press.

Netusil, N.R., 2005. Te efect o environmental zoning and amenities on property values: Portland, Oregon. Land Economics 81 (2), 227246.Nordhaus, William, and Edward Kokkelenberg. 1999. Natures Numbers: Expanding the National Economic Accounts to Include the Environment.

Washington: National Academies Press.Pearce, David. 1998. Auditing the Earth. Environment40, 2 (1998), 23-28.Pearce, David, and Dominic Moran. 1994. Te Economic Value o Biodiversity. London: Earthscan.Pearce, David, and R. Kerry urner. 1990. Economics o Natural Resources and the Environment. Baltimore: Johns Hopkins University Press.Plummer, Mark. 2009. Assessing Benet ranser or the Valuation o Ecosystem Services. Frontiers in Ecology 7 (1). 38 45.Sengupta, S., Osgood, D.E., 2003. Te value o remoteness: a hedonic estimation o Ranchette prices. Ecological Economics 44 (1), 91103.Shultz, S.D., King, D.A., 2001. Te use o census data or hedonic price estimates o open-space amenities and land use. Journal o Real Estate

Finance and Economics 22 (2/3), 239252.Fisher, Brendan, R. Kerry urner, and Paul Morling. 2009. Dening and Classiying ecosystem Services or Decision Making. Ecological Economics

68, 3 (January 2009). 643 653.Loomis, J.B., and D. S. White. 1996. Economic benets o rare and endangered species: summary and meta-analysis. Ecological Economics18.

197206.

Mates, William. 2007. Valuing New Jerseys natural capital: An assessment o the economic value o the states natural resources. State o New JerseyDepartment o Environmental Protection. Available online at http://www.state.nj.us/dep/dsr/naturalcap/nat-cap-3.pd . Accessed 23 May 2009.Navrud, Stle, and Richard Ready (editors). 2007. Environmental Value ranser: Issues and Methods. Dordrecht, Netherlands: Springer.Pearce, David. 1998. Auditing the earth. Environment40, 2. 23-28.Perrot-Matre, Danielle. 2006. Te Vittel payments or ecosystem services: A perect PES case? London: International Institute or Environment

and Development.Popper, Karl. 1959. Te Logic o Scientifc Discovery. Basic Books, New York.Ranganathan, Janet, Ciara Raudseppe-Hearne, Nicolas Lucas, Frances Irwin, Monika Zurek, Karen Bennett, Neville Ash, and Paul West. 2008.

Ecosystem Services: A Guide or Decision-Makers. Washington, DC. World Resources Institute.Richardson, Leslie and John Loomis. 2009. Te total economic value o threatened, endangered and rare species: An updated meta-analysis. Ecological

Economics68. 1535 48.Sagof, Mark. 2002. On the value o natural ecosystems: Te Catskills parable. Politics and the Lie Sciences21 (1). 19 25.Salzman, James, Barton H.Tompson, Jr., and Gretchen C. Daily, Protecting Ecosystem Services: Science, Economics, and Law, Stanord

Environmental Law Journal20, 2 (May 2001), 309-332.Science Advisory Board (SAB) to the U. S. Environmental Protection Agency. 2009. Valuing the Protection o Ecological Systems and Services.

Washington: U. S. Environmental Protection Agency.Smith, V. Kerry. 1997. Mispriced planet. Regulation 20 (3), 16-17.Smith, V. Kerry. 1998. Pricing what is useless: A status report on non-market valuation o environmental resources in Fomer, H, ietenberg, (eds)

International year book o Environmental and Resource Economics, Edward Elgar, Cheltenham, p. 156- 208.Smith, V. Kerry, Subhrendu K. Pattanayak, and George L. vanHoutven. 2006. Structural Benets ranser: An Example Using VSL Estimates.

Ecological Economics60: 361-371.Smith, V. Kerry, and Laura L. Osborne. 1996. Do contingent valuation estimates pass a scope test? A meta-analysis. Journal o Environmental

Economics and Management31(3). 287-301.Smith, V. Kerry, Christine Poulos, and Hyun Kim. 2002. reating Open Space as anUrban Amenity. Resource and Energy Economics24: 10729.allis, Heather. 2006. Introduction to ecosystem services. Science Chronicles 4 (3). Online at http://www.naturalcapitalproject.org/pubs/tallis_

science_chronicles_2006 accessed 14 May 2009.allis, Heather, and Peter Kareiva. 2005. Ecosystem Services. Current Biology15, 18 (September).erborgh, John. 1999. Requiem or Nature. Washington: Island Press.erborgh, John, Carel van Schaik, Lisa Davenport, and Madhu Rao. 2002.Making parks work: Strategies or conserving tropical nature. Washington:

Island Press.Torsnes, Paul. 2002. Te Value o a Suburban Forest Preserve: Estimates rom Sales o Vacant Residential Lots, Land Economics78. 426-441.oman, Michael. 1998. Why not to calculate the value o the worlds ecosystem services and natural capital. Ecological Economics35. 57 60.urner, Will R., Katrina Brandon, Tomas M. Brooks, Robert Costanza, Gustavo A. B. da Fonseca, and Rosimeiry Portela. 2007. Global Conservation

o Biodiversity and Ecosystem Services BioScience. Online at http://www.bioone.org/doi/abs/10.1641/B571009. Accessed April 28 2009.urpie J.K., Heydenrych B.J., Lamberth S.J. 2003. Economic value o terrestrial and marine biodiversity in the Cape Floristic Region: implications

or dening efective and socially optimal conservation strategies. Biological Conservation 112, 233-251United Nations Environment Programme (UNEP). 2008a. Report o the ad hoc intergovernmental and multi stakeholder meeting on an

intergovernmental science policy platorm on biodiversity and ecosystem services. Available online at http://ipbes.net/en/adHoc_intgvmntlMulti_stakeholderMeeting.aspx . Accessed 17 April 2009.

United Nations Environment Programme (UNEP). 2008b. Programme o work and budget o an Intergovernmental Science-Policy Platorm onBiodiversity and Ecosystem Services. Available online at http://ipbes.net/Documents/IPBES_meeting_Report_UNEP_IPBES_1_6%20_en.pd.

Accessed 17 April 2009.United States Forest Service. 2007. Valuing Ecosystem Services. Capturing the rue Value o Natures Capital. Available online at http://www.s.ed.

us/ecosystemservices/pd/ecosystem-services.pd . Accessed 3 May 2009.Wells, Michael and Katrina Brandon. 1991. People and Parks: Linking protected areas with local communities. Washington: World Bank.


23/23