paying for biodiversity conservation services in ... · paying for biodiversity conservation...

Papers in this series are not formal publications of the World Bank. They are circulated to encourage thought and discussion. The useand citation of this paper should take this into account. The views expressed are those of the authors and should not be attributed tothe World Bank. Copies are available from the Environment Department of the World Bank by calling 202-473-3641.

Paying for BiodiversityConservation Services inAgricultural Landscapes

Stefano PagiolaPaola AgostiniJosé GobbiCees de HaanMuhammad IbrahimEnrique MurgueitioElías RamírezMauricio RosalesJuan Pablo Ruíz

THE WORLD BANK ENVIRONMENT DEPARTMENT

May 2004

ENVIRONMENT DEPARTMENT PAPER NO. 96Environmental Economics Series

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Administrator

29351

The International Bank for Reconstructionand Development/THE WORLD BANK1818 H Street, N.W.Washington, D.C. 20433, U.S.A.

Manufactured in the United States of AmericaFirst printing May 2004

iiiEnvironmental Economics Series

Contents

ACKNOWLEDGMENTS VABBREVIATIONS VII

EXECUTIVE SUMMARY IX

Chapter 1Introduction 1

Chapter 2Silvopastoral Practices 3

On-site benefits 4Biodiversity benefits 5Other benefits 5

Chapter 3Barriers to Adoption 7

Chapter 4Payments for Environmental Services 11

Chapter 5From Theory to Practice 13

What is being paid for? 13How should payments be made? 16Avoiding perverse incentives 19

Chapter 6Monitoring Results 21

Changes in land use 21Impact of land use change on environmental services 21Impact of the project on participating households 22

Chapter 7Conclusions 23

Assessing the success of the project 23Assessing the success of the approach 24Replicating the approach 24

iv Environment Department Papers

Paying for Biodiversity Conservation Services in Agricultural Landscapes

APPENDIXES

A — Land Use Change in Central America and Colombia, 1990–2000 27B — Degraded Pastures and Silvopastoral Practices 29

REFERENCES 33

BOX

1 World Bank Support for PES 12

FIGURES

1 Typical time profile of benefits of silvopastoral systems 82 Effects of PES on the profitability of silvopastoral systems 19

TABLES

1 Changes in pasture land and forest area in Colombia, Costa Rica, and Nicaragua 32 Average land use in farms in Quindío, Colombia 73 Initial investment costs for selected silvopastoral practices 84 Environmental service indices used in the RISEMP 155 Example of payment computation 18

vEnvironmental Economics Series

Acknowledgments

The authors are members of the team thatprepared and is implementing the RISEMPproject. Stefano Pagiola, Paola Agostini, Ceesde Haan, and Juan Pablo Ruíz are at the WorldBank; José Gobbi and Muhammad Ibrahim atCATIE (Costa Rica); Enrique Murgueitio atCIPAV (Colombia); Elías Ramírez at Nitlapan(Nicaragua); and Mauricio Rosales at LEAD-FAO.

The opinions expressed in this paper are theauthors’ own and do not necessarily reflect

those of the World Bank, CATIE, CIPAV,Nitlapan, or LEAD-FAO. We would like tothank Ken Chomitz, John Kellenberg, AgustinArcenas, and Benjamin Kiersch for helpfulcomments and suggestions. An earlier draft ofthis paper was presented at the Third BioEconWorkshop, Contract Mechanisms forBiodiversity Conservation, Montpellier, France,May 22–24, 2003; comments from participantsat that workshop are also gratefullyacknowledged.

viiEnvironmental Economics Series

Abbreviations

ABC American Bird Conservancy

CAPE Cape Action Plan for the Environment

CATIE Centro Agronómico Tropical de Investigación y Enseñanza

(Center for Teaching and Research on Tropical Agronomy)

CIPAV Centro para la Investigación en Sistemas Sostenibles de Producción Agropecuaria

(Center for Research on Sustainable Agricultural Production Systems)

CRP Conservation Reserve Program

EBI Environmental Benefits Index

FAO Food and Agriculture Organisation of the United Nations

GEF Global Environment Facility

LEAD Livestock, Environment and Development Initiative

NGO Non-governmental organization

PES Payments for environmental services

PSA Pago por Servicios Ambientales (Payment for Environmental Services, Costa Rica)

PSAH Pago por Servicios Ambientales Hidrológicos

(Payment for Hydrological Environmental Services, Mexico)

RISEMP Regional Integrated Silvopastoral Ecosystem Management Project

WBI World Bank Institute

ixEnvironmental Economics Series

Executive Summary

Adoption of improved silvopastoral practicesin degraded pasture areas is thought to providevaluable local and global environmentalbenefits, including biodiversity conservation,However, these practices are insufficientlyattractive to individual land users for them toadopt them spontaneously, particularly due totheir high initial costs. This paper describes thecontract mechanism developed for theRegional Integrated Silvopastoral EcosystemManagement Project, which is beingimplemented with financing from the GlobalEnvironment Facility (GEF). The project istesting the use of the payment-for-servicemechanism to encourage the adoption ofsilvopastoral practices in three countries ofCentral and South America: Colombia, CostaRica, and Nicaragua. The project has created amechanism that pays land users for the globalenvironmental services they are generating, sothat the additional income stream makes theproposed practices privately profitable.

Designing the mechanism required addressingissues such as (1) measuring the actual amountof environmental services being provided, sothat appropriate payments can be made; (2)providing payments in a way that resulted inthe desired change in land use; and (3)avoiding the creation of perverse incentives(for example, for land users to cut downexisting trees so as to qualify for additionalpayments for tree planting). Two variants ofthe proposed payment mechanism are beingtested, with participating land users assignedrandomly to one or the other. The project alsoincludes extensive monitoring of theeffectiveness of each mechanism in stimulatingadoption of the proposed measures and of theresulting impact on environmental services andon household welfare. These features, togetherwith the three-country approach, will providein the coming years a very rich dataset fortesting the use of contract mechanisms forbiodiversity conservation.

1Environmental Economics Series

Introduction

As natural habitats have come to beincreasingly restricted and degraded,increasing attention has been paid toconserving biodiversity in agriculturallandscapes. This can be both an end in itself,driven by the realization that agriculturallandscapes can have high levels of biodiversity,and a means of complementing conservation inprotected areas (Pagiola and others, 1997; Dailyand others, 2001). Classical approaches toconservation, attempting to preserve pristinehabitats within protected areas, are necessarybut insufficient in the face of growing pressureon land.

Efforts to enhance biodiversity in agriculturallandscapes need to consider the incentivesfaced by individual land users, who decidewhat practices to use on their land, generallywithout considering what biodiversity benefitsdifferent land use practices may have. Whenbiodiversity-friendly agricultural practices arethe most profitable, there is a happyconvergence of private and social interests. Thisis the case of jungle rubber in Indonesia, forexample (Thiollay, 1995; Tomich and others,1998). But biodiversity-friendly agriculturalpractices are not necessarily the most profitablefrom the perspective of individual land users.In some cases, the profitability of biodiversity-friendly practices can be boosted by inducingconsumers to pay a premium for their outputs,as in the case of shade-grown coffee (Pagiolaand Ruthenberg, 2002). But this approach

requires complex certification schemes and isnot always feasible.

A further approach, which has receivedincreasing attention in recent years, is toprovide direct payments for the provision ofbiodiversity services (Pagiola and Platais,forthcoming; Pagiola and others, 2002; Landell-Mills and Porras, 2002; Ferraro, 2001: Ferraroand Kiss, 2002). This approach internalizeswhat had been an externality, ensuring that it istaken into consideration in decisionmaking.

This is the approach taken by the RegionalIntegrated Silvopastoral EcosystemManagement Project (RISEMP), which is beingimplemented with financing from the GlobalEnvironment Facility (GEF). The project ispiloting the use of payments for environmentalservices as a means of generating biodiversityconservation and carbon sequestration servicesin watersheds at three sites in Colombia, CostaRica, and Nicaragua.

This paper examines the contract mechanismsdeveloped for the RISEMP. It begins bydescribing the specific context in which theproject is being implemented, that of degradedpastoral areas in Central and South America. Itthen describes the potential for silvopastoralpractices to address this problem, which wouldprovide both local and global benefits. But theon-site benefits of silvopastoral practices aloneare insufficient to justify their adoption by

1

Environment Department Papers2


farmers. Paying land users who adopt thesepractices for the biodiversity and carbonsequestration services they generate can tip thebalance towards adoption. The RISEMP ispiloting an effort to do so. The factors whichled to the design of the contract used in theRISEMP are described next. These include the

technical characteristics of the practices beingpromoted, the specific biodiversity and carbonsequestration being sought, and the economicsof silvopastoral practices from the land users’perspective. As this is a novel approach, theRISEMP includes extensive monitoring efforts.


Silvopastoral Practices

Cattle production has long been associated withdeforestation in Latin America (Barbier andothers, 1994; Binswanger, 1991; Browder, 1985;Downing and others, 1992; Kaimowitz, 1996;Kaimowitz and others, 2004; Mahar, 1988;Mertens and others, 2002; Myers, 1981; Repettoand Gillis, 1988; Schneider, 1994), and as suchhas been an important cause of the loss ofnatural habitat and biodiversity in the region. Inmost countries, the prevailing policy frameworkencouraged deforestation for timber extractionand conversion of forest areas to pastures andcrops, which were encouraged by subsidizedcredit, guaranteed prices, and other incentives.The extent of these policy distortions has beensubstantially reduced in recent years (Faminow,1998) but pressure from poor landholders and-insome areas-large scale ranches continues toresult in large-scale deforestation in many areas.In many countries, the legal framework

encourages this process, by granting titles toland that is deemed to be ‘improved’ (that is,cleared and used for agriculture).

Table 1 summarizes changes in pasture andforest area in Colombia, Costa Rica, andNicaragua. Forest cover has been in retreatthroughout the region. The area under annualcrops has fallen in many countries (Nicaragua isan exception, as the end of unrest in the early1990s allowed a considerable expansion ofagricultural land into areas that had beenunsafe). There has been some expansion ofpermanent crops, although this trend hasreversed in recent years, due to low coffeeprices. Permanent pasture, on the other hand,has expanded steadily in all countries for whichdata are available, although at different rates.Appendix A provides more detailed data onland use changes in Central America and

Table 1. Changes in pasture land and forest area in Colombia, Costa Rica, and Nicaragua

Notes: a. data from 1990; b. data from 2000; c. 1990�2000 change.Source: World Bank World Development Indicators database.

Colombia Costa Rica Nicaragua

Area, 2000

(‘000 ha)

Change

1990-2000

(%)

Area, 2000

(‘000 ha)

Change

1990-2000

(%)

Area, 1995

(‘000 ha)

Change

1990-1995

(%)

Annual crops 2,818 –14.7 225 –13.5 2,457 25.2

Permanent crops 1,766 6.2 281 12.2 291 14.3

Permanent pasture 40,925 2.1 2,339 0.4 4,820 a ..

Natural forest area 49,650 –3.6 1,966 –7.5 3,278 b –26.4 c

2



Colombia in the last decade. These data showthat these patterns were common throughoutthe region.

In addition to the environmental problemscaused by the initial loss of forest, traditionalapproaches to pasture are often unsustainable.After an initial period of high yields, soilfertility is depleted and grass cover diminishes,resulting in soil erosion, contamination of watersupplies, air pollution, further loss ofbiodiversity, and degradation of landscapes.Lower income for producers results incontinuing poverty and in pressure to clearadditional areas.

Silvopastoral systems, which combine trees withpasture, offer an alternative to prevalent cattleproduction systems in Latin America. Theyprovide a deeply rooting, perennial vegetationwhich is persistently growing and has a densebut uneven canopy. These systems can begrouped in four major categories (Murgueitio,1999):

• Systems in which high densities of trees andshrubs are planted in pastures, providingshade and diet supplements whileprotecting the soil from packing anderosion.

• Cut and carry systems, which replacegrazing in open pasturelands with stables inwhich livestock is fed with the foliage ofdifferent trees and shrubs specificallyplanted in areas formerly used for otheragricultural practices. Cut and carrysystems have been particularly successful inCentral America and in Colombia(Benavides, 1994).

• Use of fast-growing trees and shrubs forfencing and wind screens. This system,

widely used in some countries of tropicalAmerica, provides an inexpensivealternative for fencing and supplementslivestock diets.

• Livestock grazing in forest plantations. Inthis system, grazing is used to control theinvasion by native and exotic grasses, thusreducing the management costs of theplantations.

Appendix B illustrates some of these systems inthe RISEMP project sites, as well as thedegraded pastures they are meant to replace.

On-site benefits

Silvopastoral systems can provide a range of on-site benefits (Dagang and Nair, 2003). Theintroduction of trees in pasture areas canimprove pasture productivity. Silvopastoralsystems tend to increase nutrient re-cyclingacross a deep portion of the soil profileoccupied by the root systems of a wide varietyof plants associated with silvopastoral systems.Depending on the species of trees being usedand on local climate characteristics, trees extractwater and nutrients from soil horizonsinaccessible to grasses, and deposit the nutrientson the ground with the natural fall of foliage,twigs, and fruits. The biomass and amount ofnutrients released by pruning the trees of theagroforestry systems varies depending on thekind of management in use. As much as 18tons/ha of dry matter can be deposited on theground annually, and the amount of nitrogenflowing through the system can reach values ofup to 380 kg/ha/year (Alpizar and others,1983). In addition, the trees can provide directbenefits in the form of products such as fruit,fuelwood, fodder, and timber. From the farmers’perspective, the benefits of silvopastoralsystems derive from (a) additional production


Silvopastoral Practices

from the tree component; (b) maintaining and/or improving pasture productivity; (c)diversification of production; and (d)contribution to the overall farming system (forexample, by providing fodder or income at atime when other sources do not) (Current andothers, 1995). The shade provided by trees mayalso enhance livestock productivity, especiallymilk production.

Biodiversity benefits

The increased complexity of silvopastoralsystems relative to traditional pastures meansthey often bring important biodiversity benefits(Dagang and Nair, 2003). These take two mainforms. First, they tend to support much higherspecies diversity than traditional pastures.Second, they help connect protected areas.

Silvopastoral systems have been shown to playa major role in the survival of wildlife speciesby providing scarce resources and refuge; tohave a higher propagation rate of native forestplants under these scattered trees; and toprovide shade for grazing animals, and shelterfor wild birds (Harvey and Haber, 1999). Foodavailability for wild birds is high insilvopastoral systems, and the complexstructure of the vegetation provides a moreadequate nesting substrate and better protectionagainst predators than other agroecosystems.Silvopastures and other agroforestry systemsalso harbor a larger and more complexassemblage of invertebrates than monoculturepastures (Dennis and others, 1996). Byproviding alternative sources of fuelwood andother wood products, silvopastoral systems canalso help reduce pressure on remaining naturalhabitats.

In agricultural landscapes characterized by thefragmentation of the natural habitats,

silvopastoral systems can serve as biologicalcorridors, helping to connect remaininghabitats. At the regional level, silvopastoralsystems may play an important role in theimplementation of the Mesoamerican BiologicalCorridor, given the vast area of pasturelands inCentral America and Colombia. It is expectedthat these corridors would provide adequatehabitat for wildlife while facilitating seeddispersal and the regeneration of the nativevegetation (Saunders and Hobbs, 1991).

Other benefits

Silvopastoral systems are capable of fixingsignificant amounts of carbon in the soil underthe improved pastures and in the standing treebiomass (Fisher and others, 1994). Research inColombia (Ramirez, 1997), Panama, and CostaRica (CATIE, 1999; Pfaff and others, 20000) hasshown that soils under silvopastoral systemshave higher carbon content. Additional carbonis sequestered by the trees found in suchsystems. Moreover, grass-based pastures tend tosequester most of the carbon in the deeper partof the soil profile (between 40 and 100 cmdepth), thus making it less prone to oxidation,and hence loss (Fisher and others, 1994;Beinroth and others, 1996).

Silvopastoral systems are also likely to affectwater services, though the specific impact islikely to be site specific. Infiltration generallyincreases with the presence of trees, reducingsuperficial runoff with its attendant soil erosion.Improved livestock management can helpreduce compaction, thus further reducingsurface runoff. The presence of trees also leadsto increased evapotranspiration, however, thustending to decrease water yield (Bosch andHewlett, 1982; Bruijnzeel, 1990).



In hilly areas, trees have an additionalprotective role in the ecosystem, that ofpreventing landslides (Bruijnzeel, 1990). Notonly is the presence of trees essential for soilprotection on slopes, but also the variety of

species is important. Trees of different rootdepths are required for effective soil anchorage,in particular during torrential rain eventsaccompanying tropical storms.


Barriers to Adoption

Despite their many benefits, silvopastoralsystems have only been adopted to a limitedextent (Dagang and Nair, 2003). National-leveldata typically do not distinguish land uses witha sufficient level of detail, but Table 2 showsaverage land use in the Quindío, Colombia,RISEMP project site. Pasture with no or low treedensity dominates all other land uses (Mejía,2004). Overall tree cover is low, although there isa significant amount of forest remnants, most ofwhich is riparian forest. Permanent crops,mostly coffee, account for about 10 percent ofthe area. Coffee was once the dominant land usein this area, but it has been replaced by pasturein the last decade due to low coffee prices.Fodder banks are practically non-existent: only 7in 110 farms surveyed had any, with an averageof less than 1 ha each.

An important constraint to the adoption ofsilvopastoral practices is their limitedprofitability from the perspective of individualland users. Establishing silvopastures can entailhigh initial costs, as shown in Table 3. Increasingthe livestock herd to take advantage of theincreased fodder production entails additionalcosts. In addition, there are opportunity costsresulting from the time lags before the systemsbecome productive (particularly important insystems with substantial tree components).

Figure 1 illustrates the typical time profile ofreturns to adoption of silvopastoral practices.The example shown is of a 20 ha farm inNicaragua raising livestock for milk and meat.About 15 ha are used for unimproved pasture,with 2 ha devoted to cultivating basic grains for

Table 2. Average land use in farms in Quindío, Colombia

Hectares Percentage

Annual crops 0.9 2.6

Permanent and semi-permanent crops 3.7 10.8

Pasture with no trees 21.9 64.1

Pasture with low tree density 0.6 1.8

Pasture with high tree density 0.0 0.1

Fodder banks 0.0 0.1

Plantations, riparian forest, and forest remnants 7.0 20.5

Total 34.2 100.0

Source: Mejía, 2004.

3



-500

-250

0

250

500

750

1,000

1,250

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Year

Ne

t fa

rm in

com

e (

US

$)

Current practices

Silvopastoral practices

NPV (US$):

IRR (%):

440

12

No PES

the household’s own consumption and 3 ha inbrush. As shown in the figure, current land usesgenerate a net farm income of about US$800 ayear. The proposed investment involvesswitching 3 ha of pasture to improved pasturewith low tree density, and establishing a 0.75 hafodder bank. This would allow an increase inthe livestock herd from 14 to 15, but moreimportantly it would result in a substantialincrease in the productivity of the herd due tothe greater availability of higher quality fodderand the shade provided by the trees. Milk

production would increase from just under2,000 liters annually to over 4,000 liters. Oncethe silvopastoral systems have been established,net farm income would rise to about US$1,200 ayear-a 50 percent increase. In the first years,however, farm income would be substantiallylower because of the up-front investment costsand the time lag before the trees growsufficiently to provide benefits. Only in the fifthyear following the initial investment wouldfarm returns rise above those of the current landuse practices. As a result, these investments are

Table 3. Initial investment costs for selected silvopastoral practices

Quindío,

Colombia

Esparza,

Costa Rica

Matiguás-Río

Blanco,

Nicaragua

Improved pasture (US$/ha) 375 250

Planting 100 trees in improved pasture (US$/ha) 55 50 }265

Planting 1,000 leuceana trees (US$/ha) 1,000

Protein bank (US$/ha) 960 660 475

Live fencing (US$/km) 700 610 390

Source: Gobbi, 2002.

Figure 1. Typical time profile of benefits of silvopastoral systems

Note: 20ha farm in Matiguás, Nicaragua.


Barriers to Adoption

financially marginal: in this case, the rate ofreturn to adoption of silvopastoral practices isless than 12 percent, and the net present value isonly US$440 (over 50 years, at a 10 percentdiscount rate).

The low rates of return to the adoption ofsilvopastoral systems are typical. Estimatesprepared for the RISEMP show rates of return ofbetween 4 and 14 percent, depending on thecountry and type of farm (Gobbi, 2002). Otherstudies found similar results; White and others(2001), for example, found rates of return toadoption of improved pasture in Esparza, CostaRica, of 9 to 12 percent. These estimates, ofcourse, only consider the on-site benefits ofsilvopastoral practices. The biodiversityconservation and carbon sequestration benefitsare not considered in the farmers’decisionmaking.

This problem is compounded by a lack ofawareness by farmers of some of the on-sitebenefits offered by silvopastoral systems, suchas reduced dependency on chemical fertilizersand pesticides, savings in water for irrigation,soil protection and enhanced fertility, and thepotential for additional incomes fromharvesting fruit, fuelwood, and timber. Limited

knowledge of these on-site benefits furtherreduces the perceived benefits to land users.

Even if silvopastoral practices are financiallyviable, the high initial investment costs requiredpose problems for credit-constrained land users.In the Quindío project site, only 25 percent ofhouseholds had access to credit in the past fiveyears (Mejía, 2004). Access to credit is higher inthe Matiguás-Río Blanco area in Nicaragua,thanks to presence of several NGOs that offercredit. About 50-75 percent of households in thisarea report having used credit in the past fiveyears (Ramírez and others, 2004). However,credit is often only available for specificpurposes and with collateral requirements thatare difficult for farmers to meet.

The long-term nature of investments in mostsilvopastoral practices means that tenuresecurity is an important factor in their adoption(Deininger and others, 2003, Meinzen-Dick andothers, 2002). Tenure is not a constraint in thethree study sites, however. In the Costa Rica andColombia project sites, all farmers have formalownership of the land (though they may not allhave titles). In the Nicaragua project site, mostranchers occupy public land, but long-termoccupancy gives them secure tenure.


Payments forEnvironmental Services

From the land users’ perspective, thebiodiversity conservation and carbonsequestration benefits are externalities. As such,they do not take them into consideration inmaking their land use decisions, thus reducingthe likelihood that they will adopt practices thatgenerate such benefits, including silvopastoralsystems. Recognition of this problem and of thefailure of past approaches to dealing with it hasled to efforts to develop systems in which landusers are paid for the environmental servicesthey generate, thus aligning their incentiveswith those of society as a whole. The simplelogic of Payments for Environmental Services(PES) is that compensating land users for theenvironmental services a given land useprovides makes them more likely to choose thatland use rather than another.

There has been considerable experimentationwith PES and other market-based approaches inrecent years (Pagiola and Platais, forthcoming;Pagiola and others, 2002; Landell-Mills andPorras, 2002). Latin America has been aparticularly fertile ground for suchexperimentation (Pagiola and Platais, 2001).Costa Rica has developed an elaborate, nation-wide PES program, the Pago por ServiciosAmbientales (PSA) (FONAFIFO, 2000; Pagiola,2002). Under the 1997 Forestry Law, land userscan receive payments for specified land uses,including new plantations and conservation ofnatural forests. The PSA program is now being

supported by a World Bank loan and GEF grantunder the Ecomarkets Project (World Bank,2000) (Box 1). The town of Heredia hasestablished an ‘environmentally adjusted watertariff’, the proceeds of which are used to paylandholders to maintain and reforest watershedareas (Castro, 2001; Cordero, 2003). In aseparate initiative, hydropower producer LaManguera SA is paying the MonteverdeConservation League to maintain under forestcover the watershed from which its plant drawsits water (Rojas and Aylward, 2002). InColombia, irrigation water user groups andmunicipalities in the Cauca valley are paying toconserve the watersheds that supply them withwater (Echevarría, 2002b). In 2003, Mexicocreated the Payment for HydrologicalEnvironmental Services program (Pago porServicios Ambientales Hidrológicos, PSAH),which pays for the conservation of forests inhydrologically critical watersheds usingrevenue from water charges (Bulas, 2004). Insouthern Mexico, the Scolel Té project is payingfarmers to provide carbon sequestrationservices (Tipper, 2002). In Ecuador, the city ofQuito has created a water fund withcontributions from the water utility and theelectric power company to pay for conservationin the protected areas from which it draws itswater (Echevarría, 2002a).

The bulk of PES programs to date have focusedon water services, reflecting both the urgency of

4



addressing water issues in many developingcountries and the relative ease with which thebeneficiaries of water services can be identified(Pagiola and Platais, forthcoming). Theapproach has been used for biodiversitybenefits in a few cases, mostly with GEFsupport as in the case of the Costa RicaEcomarkets Project. Environmental NGO

Conservation International has also used theapproach (which it calls ‘conservationconcessions’ or ‘conservation incentiveagreements’) in several cases, including Guyanaand Peru (Hardner and Rice, 2002; Rice, 2003).Many of these efforts have focused on relativelyuntouched areas, however, rather than onagricultural areas.

Box 1World Bank Support for PES

The World Bank is working with several countries to develop PES programs, through loans, technical assis-tance, and capacity building. In addition to the RISEMP project, current operational activities involving PESinclude:

! Costa Rica. The Ecomarkets Project, which supports the country’s PES program, includes a US$32.6 millionloan from the World Bank to help the government ensure current levels of environmental service contractsand a US$8 million grant from the GEF to assist the program’s conservation of biodiversity (World Bank,2000).

! Guatemala. The Western Altiplano Natural Resources Management Project includes a component aimed attesting and piloting PES mechanisms at the local level and supporting the development of the requirednational policy framework and instruments (World Bank, 2003a).

! Venezuela. A GEF-financed project focusing on Canaima National Park is under preparation, including amechanism to channel watershed conservation payments made by hydropower producer CVG-EDELCA.

! Mexico. The World Bank provided technical support to the government’s efforts to establish the Paymentfor Hydrological Environmental Services program.

! Dominican Republic, Ecuador, and El Salvador. Pilot PES programs are under preparation in these coun-tries. El Salvador is at the most advanced stage.

! South Africa. The Cape Action Plan for the Environment (CAPE), under preparation, aims to use a PESapproach as one of the tools to encourage conservation in the Cape Floristic Region.

! BioCarbon Fund. The newly created BioCarbon Fund is examining the potential for buying carbon seques-tration services generated by land use change. For example, one proposal would pay for carbon sequesteredby improving shade-grown coffee systems in the Mexican uplands.

In addition, the World Bank’s training arm, the World Bank Institute (WBI), has provided training on PEStargeted to technical personnel in ministries, conservation agencies, and nongovernmental organizations (NGOs)involved in implementing PES programs.

It should be noted that the World Bank did not originate the PES concept. The World Bank has played animportant role in launching such projects primarily because its borrowing countries have requested its assis-tance in doing so. By virtue of its role in assisting many countries, it has been able to cross-fertilize efforts inindividual countries with the lessons learned in others (Pagiola and Platais 2003).


From Theory to Practice

Although the PES approach is intuitivelyappealing, putting it into practice is far fromsimple (Pagiola and Platais, 2003). Thetheoretical elegance of a blackboard conceptneeds to be translated into actualimplementation arrangements on the ground.The remainder of this paper describes theapproach adopted to do so by the RISEMPproject.

The RISEMP, which began implementation inJuly 2002, is seeking to pilot the use of thepayment for environmental services approach toencourage the adoption of silvopastoralpractices in degraded pastures areas in Centraland South America (World Bank, 2002). Theproject is being implemented in threemicrowatersheds: Quindío, in Colombia;Esparza, in Costa Rica; and Matiguás-RíoBlanco, in Nicaragua. Participating land usersenter into contracts under which they receive apayment for the environmental services thatthey generate. They receive annual paymentsover a two- or four-year period, based on theincrement in environmental services providedrelative to the baseline situation for thatparticular farm. Through this mechanism, theproject aims to establish silvopastoral systemson 3,500 ha, thus enhancing the environmentalbenefits generated in watersheds covering about12,000 ha.

The project was prepared with support of themulti-donor Livestock, Environment and

Development Initiative (LEAD), hosted by theFood and Agriculture Organisation (FAO). It isfinanced by a US$4.5 million GEF grant, withthe World Bank acting as implementing agency.In each country, field activities are beingundertaken by local non-governmentalorganizations (NGOs): the Centre for Researchon Sustainable Agricultural Production Systems(Centro para la Investigación en SistemasSostenibles de Producción Agropecuaria,CIPAV) in Colombia, the Center for Teachingand Research on Tropical Agronomy (CentroAgronómico Tropical de Investigación yEnseñanza, CATIE) in Costa Rica, and Nitlapanin Nicaragua, with CATIE coordinating thework. The American Bird Conservancy (ABC) isproviding technical assistance for thedevelopment of a common and consistentmethodology for the monitoring of biodiversityat the three project sites.

What is being paid for?

Contracting for land users to providebiodiversity benefits is all very well in theory,but in practice it is clearly unrealistic to askthem to deliver biodiversity. A way is needed tocommunicate what is desired to potentialparticipants in ways that they can understand.The typical solution has been to offer to pay notfor biodiversity itself, but for land uses that arehospitable to biodiversity (Pagiola and others,2002).

5



But land use alone can be a relatively bluntinstrument. In Costa Rica’s PSA program, forexample, most contracts call for conservation ofexisting forest, and pay all participants the sameamount (FONAFIFO, 2000; Pagiola, 2002).While this approach has the virtue of simplicity,it fails to recognize the very different levels ofservices that different land uses can provide.The biodiversity-friendliness of agriculturalpractices is not a binary, yes/no proposition. Onthe contrary, there is a spectrum of effects,ranging from relatively inhospitable systemssuch as monocultures with heavy agrochemicaluse to relatively hospitable systems such asorganic coffee grown under a diverse shadecanopy of native species. Location also matters:biodiversity-friendly practices in proximity toprotected areas, for example, might be morevaluable by helping to buffer and protect them.Failing to take these differences into accountrisks either under-paying for desirable landuses, or over-paying for relatively less desirableones (Pagiola and Platais, forthcoming).

The solution adopted in the RISEMP was toprepare a list of land uses and associate eachwith a point system upon which payments arebased. This approach is similar to that of theEnvironmental Benefits Index (EBI) used in USConservation Reserve Program (CRP) (NCEE,2001). Separate indices were developed for thebiodiversity conservation and carbonsequestration benefits of each land use. Thesetwo indices were then aggregated to form anenvironmental service index to be employed asthe basis for calculating payments toparticipants. A similar index for water benefitswas not included, partly because of the lack ofdata needed to develop it, and partly becauseimproved water flows would be nationalbenefits, and thus are not eligible for GEFfunding. The biodiversity conservation and

carbon sequestration indices are presented inTable 4.

The biodiversity conservation index was scaledwith the most biodiversity-poor land use(annual crops) set at 0.0 and the mostbiodiversity-rich land use (primary forest) set at1.0. Within this spectrum, the points given toeach specific land use were set by a panel ofexperts, taking into consideration factors suchas the number of species (of plants, birds, smallmammals, and insects), their spatialarrangement, stratification, plot size, and fruitproduction. Higher scores were given to landuses that have greater potential to maintain theoriginal biodiversity of the region. Note that theindex estimates the environmental benefits of allland uses, and not only silvopastoral practices.

This approach can take into consideration thedifferent impact that different land uses arelikely to have on biodiversity. There are, ofcourse, limitations. The biodiversity impactdepends not only on the characteristics of theland use, but also on its location, its extent, andits relationship to other land uses. At the pilotscale of the RISEMP, issues of location are notsignificant, as all three pilot areas werespecifically chosen for their proximity toprotected areas or to corridors between them.All three project areas were selected in partbecause of their location in ecologically-sensitive areas. The Quindío project site is inone of the most severely degraded regions ofColombia, with few, mostly unconnectedremnants of natural habitats. Restoring a degreeof habitat heterogeneity and connectivity wouldincrease the chances of survival of speciesrequiring large home ranges in an areaconsidered as a priority for bird conservation.The Esparza area in Costa Rica is in the vicinityof conservation areas such as La Fortuna, the



Monteverde Reserve complex, and the AlbertoBrenes Biological Reserve. More biodiversity-friendly land use practices would help thechances of survival of several species occurringin these protected areas. The Matiguás-RíoBlanco watershed in Nicaragua is part of thebuffer zone of the Cerro Musún Natural

Reserve, and is very close to one of the priorityareas for bird conservation in the country. If thisapproach were to be scaled up and applied on abroader scale, location effects could beincorporated either by varying the points foractivities in different locations or by varying thepayment per incremental point. Issues of scale

Note: The environmental service index is the sum of the biodiversity and carbon sequestration indices.

Table 4. Environmental service indices used in the RISEMP(Points per hectare, unless otherwise specified)

Land use

Biodiversity

index

Carbon

sequestration

index

Environmental

service

index

Annual crops (annual, grains, and tubers) 0.0 0.0 0.0

Degraded pasture 0.0 0.0 0.0

Natural pasture without trees 0.1 0.1 0.2

Improved pasture without trees 0.4 0.1 0.5

Semi-permanent crops (plantain, sun coffee) 0.3 0.2 0.5

Natural pasture with low tree density (< 30/ha) 0.3 0.3 0.6

Natural pasture with recently-planted trees (> 200/ha) 0.3 0.3 0.6

Improved pasture with recently-planted trees (> 200/ha) 0.3 0.4 0.7

Monoculture fruit crops 0.3 0.4 0.7

Fodder bank 0.3 0.5 0.8

Improved pasture with low tree density (< 30/ha) 0.3 0.6 0.9

Fodder bank with woody species 0.4 0.5 0.9

Natural pasture with high tree density (> 30/ha) 0.5 0.5 1.0

Diversified fruit crops 0.6 0.5 1.1

Diversified fodder bank 0.6 0.6 1.2

Monoculture timber plantation 0.4 0.8 1.2

Shade-grown coffee 0.6 0.7 1.3

Improved pasture with high tree density (> 30/ha) 0.6 0.7 1.3

Bamboo (guadua) forest 0.5 0.8 1.3

Diversified timber plantation 0.7 0.7 1.4

Scrub habitats (tacotales) 0.6 0.8 1.4

Riparian forest 0.8 0.7 1.5

Intensive silvopastoral system (>5,000 trees/ha) 0.6 1.0 1.6

Disturbed secondary forest (> 10 m2 basal area) 0.8 0.9 1.7

Secondary forest (> 10 m2 basal area) 0.9 1.0 1.9

Primary forest 1.0 1.0 2.0

New live fence or established live fence with frequent pruning

(per km)

0.3 0.3 0.6

Multi-story live fence or wind break (per km)

0.6 0.5 1.1



and contiguity are harder to address. Somebiodiversity benefits may be obtained only afterappropriate land uses cover a minimum area, orif the areas covered are contiguous rather thanscattered. To an extent, these effects might beaddressed by adding bonus points if the areacovered by a given land use passes a threshold.Such an approach could quickly result in anexcessively complex point system, however.Another approach would be to set a minimumparticipation threshold for the PES program totake effect; this approach was followed by NewYork City, for example (A. Appleton, pers.comm.)

A similar procedure was used to establish thecarbon sequestration index, with different landuses given points according to their capacity tosequester stable carbon in the soil and in hardwood through the years. Recent studies indicatethat secondary forest can fix an average of 10tonnes of carbon per year in wood and in thesoil. As secondary forest has a value of 1.0 in theindex, 0.1 points correspond to an estimatedsequestration of 1 tonne of carbon. Data fromstudies conducted by CATIE were used tocalibrate the carbon sequestration index.

As data were insufficient to derive country-specific indices, the same index is being used inall three countries. Data from the monitoringefforts will be used to improve the indices, andit is expected that these will differ from countryto country.

Should downstream water users be willing topay for hydrological services, the approachcould also be extended by adding an indexdenoting the contribution of each land use typeto the desired water services, thoughdeveloping such an index would certainly provedifficult.

Note that under RISEMP, biodiversity andcarbon sequestration benefits are given equalweight in calculating payments. The two indicescould easily be de-coupled, however, withseparate payment levels for each kind ofenvironmental service. Alternatively, differentweighting schemes could be used to giveproportionally more weight to one or the other,depending on the interest of those making thepayments.

This index approach was tested with potentialparticipants, and is proving quite intelligible tothem in practice. Dissemination materials suchas posters and handbooks have been preparedshowing precisely what the payments would befor specific land uses.

How should payments be made?

The second challenge in developing anappropriate contract is the need to understandthe economics of the farming system, so that theappropriate amount and form of payment canbe determined. Payments for environmentalservices will have the desired effect only if theyreach the land users in ways that influence theirdecisions on how to use the land.

Analysis of the time path of benefits generatedby silvopastoral systems showed that they areunattractive to land users primarily because oftheir substantial initial investment, and becauseof the time lag between investment and returns,as shown in Figure 1 above. This leads to thehypothesis that a relatively small paymentprovided in the early period of adoption wouldbe sufficient to ‘tip the balance’ between currentand silvopastoral systems. This effect works byincreasing the net present value of investmentsin silvopastoral practices, but also by reducingthe initial period in which adoption of these



systems imposes net costs on land users. By thetime payments end, the silvopastoral practicesthemselves are ready to begin generatingincome for land users. The payments alsoalleviate the liquidity problems faced by manyland users and help them finance the requiredinvestments.

Based on this analysis, it was decided toprovide a relatively small, up-front payment toparticipating land users. This payment is ofUS$75 per incremental point, per year over afour-year period, up to a maximum of US$4,500per farm (US$6,000 in Colombia, where inputprices are higher). Both of these aspects deservefurther discussion.

In principle, the amount should be no less thanthe land users’ opportunity cost (or they willnot participate), and no more than the value ofthe benefit provided (or it would not beworthwhile to provide the service). In practice,the actual value of the benefit provided isextremely difficult to estimate, and particularlyso for benefits such as biodiversityconservation. In contrast, the farmers’opportunity cost can usually be estimatedrelatively easily. For this reason, as well as tolimit the budgetary requirements of thepayment, payment levels are usually set atslightly more than the opportunity cost of themain alternative land uses. All the existingsystems of payments for environmental servicesimplicitly or explicitly use this approach. CostaRica’s PSA program, for example, currentlypays US$45/ha/year for forest conservation.This payment has proven to be quite attractive,with far more applications for this contract thanthe program has been able to finance. (Incontrast, a payment of US$538/ha over 5 yearsfor reforestation has proven to be less popular,as many landowners consider the payment

offered insufficient to justify the investment.) InMexico, a specific study was commissioned ofthe opportunity cost of land (Jaramillo, 2003) toprovide a basis for payments levels under thePSAH; no study was made of the magnitude ofbenefits. Zelek and Shively (2003) propose ascheme to pay the opportunity costs of Philipinofarmers who adopt practices that sequestercarbon. Paying the opportunity cost of adoptingthe desired practice also accords well withGEF’s policy of paying for the incremental costsof generating global environmental benefits.

In terms of payments for carbon emissionsreductions, the US$75/point/year paymentlevel is equivalent to paying US$7.5 per tonne ofcarbon sequestered, or US$2 per tonne of CO2equivalent. This compares favorably to currentworld prices for carbon emissions reduction ofUS$3–5 per tonne of CO2 equivalent (WorldBank, 2003b)—although these paymentstypically require a higher degree of assurance ofthe permanence of the emissions reduction anda more intensive monitoring regime than theRISEMP offers. No similar comparison ispossible for payments for biodiversityconservation.

In general, emerging guidelines for paymentsfor environmental services indicate thatpayments should be on-going rather than finite(Pagiola and Platais, forthcoming). In CostaRica’s PSA program, for example, payments forforest conservation contracts are for 5 years, butthey are renewable indefinitely by mutualagreement. The logic for this is simple: ifenvironmental services are to be generated overa long period of time (presumably, indefinitely),then payments for these services should also bemade over a similarly long period. Endingpayments sooner creates the risk that land userswill revert to their previous land use practices.



Years from contract signing

0 1 2 3 4

Land use

Crops (annual, grains, tubers) (ha) 2.5 2.5 2.5 2.5 2.5

Natural pasture with low tree density (ha) 14.5 13.0 12.0 11.0 11.0

Natural pasture with high tree density (ha) 0.0 1.0 2.0 3.0 3.0

Fodder bank (ha) 0.0 0.5 0.5 0.5 0.5

Scrub habitat (ha) 3.0 0.0 0.0 0.0 0.0

Secondary forest (ha) 0.0 3.0 3.0 3.0 3.0

Total area (ha) 20.0 20.0 20.0 20.0 20.0

Wire fences with trees (km) 0.0 0.5 1.0 1.5 1.5

Environmental service score (points)

Crops (annual, grains, tubers) 0.0 0.0 0.0 0.0 0.0

Natural pasture with low tree density 8.7 7.8 7.2 6.6 6.6

Natural pasture with high tree density 0.0 0.9 1.8 2.7 2.7

Fodder bank (monocrop) 0.0 0.5 0.5 0.5 0.5

Scrub habitat 3.9 0.0 0.0 0.0 0.0

Secondary forest 0.0 5.7 5.7 5.7 5.7

Wire fences with trees 0.0 0.5 1.0 1.5 1.5

Total points for the farm 12.6 15.4 16.2 17.0 17.0

Baseline points 12.6

Incremental points 2.8 3.6 4.4 4.4

Income from environmental services (US$) 126 210 270 330 330

This is a risk that has been observed time andtime again in projects that attempted to changeland use practices, such as soil conservation orreforestation projects (Lutz and others, 1994).This risk was thought to be relatively low in thisinstance, as the silvopastoral practices, onceestablished, are privately more profitable (seeFigure 1). Moreover, the payments representonly a small portion of the necessary investmentcosts, thus making it unlikely that land userswould adopt practices they intend to abandonsolely to receive the payments. In an effort todetermine the long-term sustainability of themechanism, a sub-group of participants is beinggiven a slightly modified contract, in which thepayments are frontloaded: rather than receivingthem over a four-year period, farmers with this

alternative contract will receive a similaramount over a two-year period. Farmers wereassigned randomly to one or the other contract.

Table 5 illustrates the application of this contractfor the 20 ha farm in Matiguás, Nicaragua, usedin the previous example. In the baseline year,the farm has 2.5 ha under annual crops, 14.5 haunder natural pasture without trees, and 3 haunder brush (tacotal). Motivated by the project,it converts 3 ha of its pasture to higher treedensities: 1 ha a year for the first three years. Italso plants an 0.5 ha fodder bank, and fences offthe scrub areas so that secondary forest canregenerate. Finally, it plants trees along 1.5 kmof its fence lines. Using the environmentalservice index in Table 4, the resulting scores can


Table 5. Example of payment computation



be calculated for the baseline and for eachsubsequent year. These scores are then used tocompute the payments due to the farmer,including the initial baseline payment forexisting services (see below) and the mainpayment for incremental services providedunder the project. Figure 2 shows the impact ofthese payments on the time profile of benefits toadopting silvopastoral practices, and theresulting impact on the profitability of theinvestment. What had been a marginally viableinvestment now becomes more attractive.

Avoiding perverse incentives

The initial plan involved paying land users onlyfor incremental improvements in land usepractices. The extent to which land users hadalready adopted practices that conservedbiodiversity or sequestered carbon prior to theproject was to have been reflected in theirbaseline environmental service index, and onlyincrements to this index were to becompensated. It soon became clear that this

approach entailed a substantial risk of creatingperverse incentives. “Bueno, corto todo,” was acommon reaction by land users when told theywould not be compensated for pre-existingtrees: “fine, I’ll cut them all.” It might have beenpossible to avoid this risk among projectparticipants by imposing contractual restrictionson such actions, though this would certainlyhave required an increased monitoring effort,and thus increased costs. But there was also abroader risk that non-participants insurrounding areas would postpone adoptingsilvopastoral practices that they might havebeen tempted to adopt, so that they might waitfor a project to come and compensate them fordoing so. As a result, the initial plan wasmodified to allow for a payment to be made forpre-existing environmental services. A one-timepayment of US$10/point will be made for thebaseline points, up to a maximum of US$500 perfarm. This payment has the further benefit ofhelping to alleviate financing constraints toimplementing silvopastoral practices.

Figure 2. Effects of PES on the profitability of silvopastoral systems


-500

-250

0

250

500

750

1,000

1,250

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Year

Ne

t fa

rm in

com

e (

US

$)

Current practices

Silvopastoral practices

Silvopastoral practices with PES

NPV (US$):

IRR (%):

440

12

No PES

1,450

19

With PES



As part of the effort to avoid perverseincentives, the contract also specifies thatburning in pastures is banned (except in areasdevoted to food security, where burning isallowed in the first two years), and that thecontract will be terminated if the participantscut down primary or secondary forest in theirfarms.

A related problem is that of minimizing‘leakage’-that is, avoiding environmentally-

damaging activities being simply displaced, sothat there is little net benefit. The RISEMPminimizes this problem by computing thepoints on which payments are made over theentire farm and basing payments on the netpoints over the entire farm. If land users cutdown trees in one plot even as they plant themin another, the negative points earned from theadverse change will offset those gained from thepositive change.


Monitoring Results

Monitoring is always important, as it allows theeffectiveness of the project to be determined. Itcan also allow mid-course corrections to bemade if they should prove necessary. The needfor monitoring is particularly high in pilotprojects, which are intended to serve as guidesfor future projects.

Changes in land use

For the purpose of monitoring compliance withthe contract and computing payments owed,observation of participants would be sufficient.However, the RISEMP has the broader objectiveof piloting the application of PES systems tosilvopastoral practices. This requires not onlymonitoring the behavior of participants, butdetermining that the project has beeninstrumental in affecting this behavior.Monitoring the degree to which the project isencouraging participants to undertake thedesired changes in land use entails monitoringthe changes in land use of the participantsthemselves, and of a control group (so that theimpact of the project itself can be distinguishedfrom other trends that might affect land use).

To address these issues, monitoring will beundertaken on three groups. In addition to thetarget group of participants (80 households inColombia, 100 in Costa Rica, and 100 inNicaragua), a control group of 30 householdswill be monitored in each country. This control

group will be selected so as to have similarcharacteristics as the households in the targetgroup, but will not receive any payments forenvironmental services. The target group itselfwill be partitioned into two groups, so that theimpact of technical assistance provided by theproject can also be tested. The main part of thegroup will receive both the payments andtechnical assistance, while a subgroup of 30households will only receive the payments forenvironmental services (to the extent that theyadopt the recommended practices), without thetechnical assistance. Each household in each ofthese groups will be monitored every other year.Their land use will be monitored, and a socio-economic survey will be conducted.

Impact of land use change onenvironmental services

To verify that the silvopastoral systemspromoted under the project actually generatethe expected environmental benefits,biodiversity and carbon sequestration will bemonitored in all land use types in the three pilotareas. For biodiversity, counts of bird specieswill be the main indicator of biodiversity used,but they will be complemented by studies ofbutterflies, ants, and mollusks. Factors such asendemicity and rarity in the species observedwill be taken into consideration. Water qualitywill be monitored only in the Colombian site, asfunds did not permit a more general assessment

6



of the contribution of these systems to improvedwater quality.

The changes will be compared to baselinemeasurements made at the start of the project. Astudy of the Matiguás-Río Blanco sites inNicaragua, for example, found 131 bird species(Pérez and others, 2004). The key test for theproject will be whether biodiversity increasessignificantly compared to the baseline.

The results of the monitoring will also be usedto revise and refine the biodiversity and carbonsequestration indices. These revised indicescould be used to determine payments under anyfuture project. They will not, however, affectpayments under the current project, which willbe made based on an ex ante estimates of theglobal environmental benefits of each land use,as expressed in the environmental service index.

Impact of the project on participatinghouseholds

In addition to the project’s impact on the globalenvironment, it is also important to understandits impact on household welfare: does welfareincrease, and if so how much and in what way,and are there differences in how welfareincreases across income groups? Payments forenvironmental services have been hypothesizedas having the potential for improving thewelfare of the poor in target areas (Pagiola andothers, 2003), but there has been little empiricalwork on this topic to date. Data collectedthrough the socioeconomic survey will help toaddress these questions. It will allow low-income households to be identified, and willprovide a variety of measures of householdwelfare, including income-related measures(total income and income variability) and otherindicators (such as health status).


Conclusions

The RISEMP project demonstrates that directcontracts for biodiversity conservation inagricultural landscapes are not just a theoreticalcuriosity. Such contracts are possible, if the linkbetween land use and biodiversity is known.Specific details depend on the economics of thesystem being promoted.

It will be some time before the effectiveness ofthe mechanisms discussed in this paper can bedetermined. The intensive monitoring beingundertaken will allow a very detailed analysisof this effectiveness, including consideration ofnumerous exogenous factors that might affect it.This project will thus allow both an overallconclusion on the effectiveness of the approachand provide data for its refinement. Already,however, a number of key questions can beidentified, some specific to the particularapproach used in the RISEMP, some that applymore broadly to PES approaches in general, andsome that bear to the potential for replicatingthe approach on a wider scale.

Assessing the success of the project

Will the desired land use changes be induced? Theextent to which the desired land use changes areinduced depends in part on the payment levelbeing sufficient to ‘tip the balance’ betweencurrent and improved practices. Theappropriate level of the payment was a subjectof intense debate. The higher the payment

offered, the greater adoption is likely to be.However, higher payments also militate againstthe cost-effectiveness of the approach.Moreover, a higher payment per point alsoincreases the risk that participants will adoptotherwise un-profitable practices onlytemporarily, so as to receive the payment, withthe intention of abandoning them later. Thepayment level was initially set at US$50 perenvironmental service index point, but wasraised to US$75 after field staff reported thatparticipants considered US$50/pointinsufficiently attractive to justify widespreadadoption of silvopastoral practices. Given thenovelty of the approach, there is also a potentialcredibility problem of the project’s promise topay for environmental services. The baselinepayments probably played an important role inthis sense, over and above their benefits inavoiding perverse incentives.

Will changes in land use be sustainable? TheRISEMP project is based on the hypothesis thatsilvopastoral practices, once adopted, are moreprofitable to land users than current practices. Ifthis hypothesis is correct, then adoption shouldbe sustainable with no further assistance. To testthis hypothesis some participants are being paidover a two-year rather than a four-year period.

Will improvements in biodiversity conservation besignificant? Baseline studies show that there issome pre-existing biodiversity in each of the

7



project areas. The challenge for the project willbe to demonstrate whether it has improved thisbiodiversity significantly. The project willmonitor changes in biodiversity levels closely.

Will the poor benefit? If the farm models preparedfor the project (illustrated in Figures 1 and 2) arecorrect, the project has the potential ofincreasing the net income of participatinghouseholds quite substantially. Whether thiswill occur is being closely monitored.Constraints that might prevent poorerhouseholds from participating, and hence fromobtaining these income increases, are also beingexamined closely.

Assessing the success of the approach

How cost-effective is the PES approach in terms ofbiodiversity conservation? Both the cost and theeffectiveness of this approach remain to bedetermined. The two are to some extentinversely related. Payments will be high if landusers adopt practices with high point values,and low if they adopt practices with low pointvalues. The transaction costs of implementingthe project must also be considered. Some haveargued that an incremental conservation dollarwould be most effectively spent on otherapproaches, such as protected areas. Certainlyprotected areas are likely to have much lowertransaction costs than the approach discussedhere. But their implementation costs may wellbe higher. Establishing a protected area wouldrequire buying the land from its current owners.That is, it would require compensating them forthe loss of the entire flow of benefits it mightgenerate in its most profitable use. The PESapproach only requires compensating them forthe difference between the net benefits theyobtain under the conservation use and the mostprofitable use. Moreover, buying the land

outright requires paying for the entire presentvalue of the future flow of benefits up-front. Incontrast, a PES approach makes payments overtime. PES will also attract lower-opportunitycost lands, while a protected area approachusually privileges conservation benefits andthus may include higher-opportunity cost land.The PES approach is likely to be particularlyadvantageous if, as may be the case in theRISEMP, a short-term payment is sufficient toresult in sustainable adoption of the desiredland uses. Finally, buying land outright maysimply not be politically feasible, or may entailundesirable social consequences because of theneed to relocate the landowners. In the case ofthe Costa Rica Ecomarkets Project, the PESapproach was found to be much more cost-effective than establishing a protected area ofthe same size (World Bank, 2000).

PES may be the cheaper way to conserve agiven area, but the level of conservation is likelyto be lower. If the land were bought outrightand placed in a protected area, it could bemanaged optimally from the conservationperspective. Under a PES approach, land use isdetermined by the combination of conservationbenefits (as reflected in the payment) and landuser preferences; in many cases, this will lead toa compromise result. Moreover, some of thereasons that make PES cheaper may also lead tolower conservation benefits: lower opportunitycost land is not necessarily the most desirablefrom a conservation perspective, for example.

Replicating the approach

How can transaction costs be reduced? Thetransaction costs involved in implementing aPES approach are a key determinant of its cost-effectiveness, its sustainability, and itsreplicability. They also play a critical role in the


Conclusions

extent to which poorer land users canparticipate (Pagiola and others, 2003). Becauseof its pilot nature, the RISEMP has relativelyhigh costs for detailed monitoring and otheractivities that would not necessarily be neededin a scaled-up project. The environmentalservice index used in the RISEMP allows a veryfine-tuned targeting of payments to expectedbenefits, but it also imposes relatively highmonitoring costs. There is a need to find proxyindicators that are highly correlated withbiodiversity conservation but are easy andcheap to monitor, ideally using remote sensing.If forest cover provides an adequate proxy, forexample, it would be relatively cheap toobserve. The current environmental servicesindex cannot be monitored solely from remotesensing, as it includes elements of the type andquality of vegetative cover. A crucial questionwhich needs to be explored is that of thetradeoff between the precision of the index andthe transaction costs involved in implementingit.

How can the approach be made sustainable? Most ofthe discussion in this paper would be broadlyapplicable to PES approaches that addressbenefits other than biodiversity. Wherebiodiversity services differ, however, is in thelong-term sustainability of payments. Emerginglessons indicate that payments under PESprograms usually have to be made on a long-term basis if the desired services are to begenerated sustainably (Pagiola and Platais,2003, forthcoming). The specific practicespromoted by the RISEMP project may notrequire a long-term payment, but this is likely tobe the exception rather than the rule. In themore general case in which the systems with thehighest external benefits are not the mostprofitable to land users, the short-term paymentapproach used in the RISEMP is unlikely to

result in sustainable adoption of the desiredland uses. Rather, longer-term, probablyindefinite payments will be needed. In turn, thismeans that sustainable long-term financingsources will be required. Even if only short-termpayments are sufficient, substantial additionalfunding flows will be required if the approach isto be extended beyond the pilot areas. For waterservices, potential sources for such financingcan readily be identified-although capturingthem can be difficult. Moreover, fundingstreams for water services can in principle bevery sustainable, as they are tied to services thatwill continue to be used indefinitely (Pagiolaand Platais, forthcoming). All availablefinancing sources for biodiversity conservation,however, including the GEF, tend to focus onrelatively short-term projects. Placing funds intoa trust fund so as to generate a stream of futurerevenues is one option (GEF, 1999), but itrequires substantial up-front financing. Becauseof the greater ease of generating long-termpayment streams for water services, basingpayments on water service provision mayappear to be an attractive option. Themunicipality of Matiguás, for example, isinterested in using this approach to protect itswater supply. This approach should certainly beexploited as much as possible, but twoconstraints need to be borne in mind: first,water services are very site-specific, and somany areas would not be eligible for payments.The project areas near Matiguás, for example,are downstream of the water intakes for themunicipal water supply system and so wouldnot be included in a water service-based PESprogram. Second, the most desirable activitiesfrom the perspective of generating waterservices are not necessarily the same as thosethat generate the biodiversity and carbonsequestration services sought in the RISEMPproject. Basing payments on water services,



therefore, will require additional scientificresearch to improve the understanding of howland use affects water services. In general,therefore, water-based payments will not

generate all the desired biodiversity and carbonsequestration benefits, and the need for separatefinancing for this purpose will remain.


Appendix A —Land Use Change in Central Americaand Colombia, 1990–2000

Notes: Annual crops category includes temporary pasture... indicates data not available.

Source: World Bank World Development Indicators database.

Area (‘000 ha) Change in area (%)

1990 1995 2000 1990-1995 1995-2000 1990-2000

Annual crops Colombia 3,305.0 2,399.0 2,818.0 -27.4 17.5 -14.7 Costa Rica 260.0 225.0 225.0 -13.5 0.0 -13.5 El Salvador 550.0 582.0 560.0 5.8 -3.8 1.8 Guatemala 1,300.0 1,355.0 1,360.0 4.2 0.4 4.6 Honduras 1,462.0 1,600.0 1,068.0 9.4 -33.3 -26.9 Nicaragua 1,963.0 2,457.0 .. 25.2 .. .. Panamá 499.0 500.0 500.0 0.2 0.0 0.2 Permanent crops Colombia 1,661.9 2,077.4 1,765.8 25.0 -15.0 6.2 Costa Rica 250.2 291.0 280.8 16.3 -3.5 12.2 El Salvador 259.0 273.5 250.7 5.6 -8.3 -3.2 Guatemala 487.9 553.0 542.2 13.3 -2.0 11.1 Honduras 358.0 346.9 358.0 -3.1 3.2 0.0 Nicaragua 254.9 291.4 .. 14.3 .. .. Panamá 156.3 .. .. .. .. .. Permanent pasture Colombia 40,093.8 40,093.8 40,924.8 0.0 2.1 2.1 Costa Rica 2,328.3 2,338.5 2,338.5 0.4 0.0 0.4 El Salvador 640.2 750.1 793.6 17.2 5.8 23.9 Guatemala 2,504.7 2,602.3 2,602.3 3.9 0.0 3.9 Honduras 1,499.3 1,532.9 1,510.5 2.2 -1.5 0.7 Nicaragua 4,819.6 .. .. .. .. .. Panamá 1,473.7 .. .. .. .. .. Forest area Colombia 51,519.5 49,649.9 -3.6 Costa Rica 2,124.1 1,965.8 -7.5 El Salvador 192.7 120.2 -37.6 Guatemala 3,383.0 2,851.7 -15.7 Honduras 5,974.9 5,381.9 -9.9 Nicaragua 4,455.4 3,277.8 -26.4 Panamá 3,394.0 2,873.0 -15.4


Appendix B —Degraded Pastures andSilvopastoral Practices

Photo 1 — Silvopastoral systems, in which pastures include significant tree cover, tend to be more productiveand to provide a much better habitat for biodiversity. Here, cattle graze in pastures with high tree cover inEsparza, Costa Rica.

All images in Appendix B are by Stefano Pagiola


Paying for Biodiversity Conservatioon Services in Agricultural Landscapes

Photo 2 —Degraded pasture in

Matiguás, Nicaragua

Photo 3 —Improved pastures with

high tree cover inMatiguás, Nicaragua


Appendix B — Degraded Pastures and Silvopastoral Practices

Photo 5 —Fodder banks inQuindío, Colombia

Photo 4 —Degraded pastures inQuindío, Colombia


References

Alpizar, L., H.W. Fassbender, and H.Heuveldop. 1983. “Estudio de SistemasAgroforestales en el Experimento Centraldel Catie.” Turrialba: CATIE (processed).

Barbier, E.B., J.C. Burgess, J. Bishop, and B.Aylward. 1994. The Economics of the TropicalTimber Trade. London: Earthscan.

Beinroth, F.H., M.A. Vázquez, V.A. Snyder, P.F.Reich, and L.R. Pérez Alegría. 1996. “FactorsControlling Carbon Sequestration inTropical Soils.” Washington: US Departmentof Agriculture, Natural ResourcesConservation Service.

Benavides, J. E. (ed.). 1994. Árboles y ArbustosForrajeros en América Central. Serie Técnica.Informe Técnico No.236. Turrialba: CATIE.

Binswanger, H. 1991. “Brazilian Policies thatEncourage Deforestation in the Amazon.”World Development, 19, pp.821-829.

Bosch, J.M., and J.D. Hewlett. 1982. “A Reviewof Catchment Experiments to Determine theEffect of Vegetation Changes on Water Yieldand Evapotranspiration.” Journal ofHydrology, 55, pp.3-23.

Bruijnzeel, L.A. 1990. Hydrology of Moist TropicalForests and Effects of Conversion: A State ofKnowledge Review. Paris: UNESCOInternational Hydrological Programme.

Browder, J. 1985. Subsidies, Deforestation, and theForest Sector of the Brazilian Amazon.Washington: World Resources Institute.

Bulas, J.M. 2004. “Implementing Cost Recoveryfor Environmental Services in Mexico.”Paper presented at World Bank Water Week,Washington, 24-26 February 2004.

Castro, E. 2001. “Costarrican Experience in theCharge for Hydro Environmental Servicesof the Biodiversity to Finance Conservationand Recuperation of Hillside Ecosystems.”Paper presented at the InternationalWorkshop on Market Creation forBiodiversity Products and Services, OECD,Paris, 25-26 January 2001 (processed).

Centro Agronómico Tropical de Investigación yEnseñanza (CATIE). 1999. “Capacidad yRiesgos de Actividades Forestales en elAlmacenamiento de Carbono yConservación de Biodiversidad en FincasPrivadas del Área Central de Costa Rica.”Paper presented at the IV Semana Científicadel CATIE, Turrialba, Costa Rica. 6-9 April,1999 (processed).

Cordero, D. 2003 “Tarifa de Agua Basada enCosto de Protección de Cuencas, enHeredia.” Heredia: Empresa de ServiciosPúblicos de Heredia (ESPH) (processed).

Current, D., E. Lutz, and S.J. Scherr. 1995. “TheCosts and Benefits of Agroforestry to



Farmers.” World Bank Research Observer,10:2, pp.151-180.

Dagang, A.B.K., and P.K.R. Nair. 2003.“Silvopastoral Research and Adoption inCentral America: Recent Findings andRecommendations for Future Directions.”Agroforestry Systems, 59, pp.149-155.

Daily, G.C., P.R. Ehrlich, and G.A. Sánchez-Azofeifa. 2001. “Countryside Biogeography:Utilization of Human Dominated Habitatsby the Avifauna of Southern Costa Rica.”Ecological Applications, 11, pp.1-13.

Deininger, K., S. Jin, B. Adenew, S. Gebre-Selassie, and B. Nega. 2003. “TenureSecurity and Land-Related Investment:Evidence from Ethiopia.” Washington:World Bank (processed).

Dennis, P., L. Shellard, and R. Agnew. 1996.“Shifts in Arthropod Species Assemblagesin Relation to Silvopastoral Establishment inUpland Pastures.” Agroforestry Forum, 7:3,pp.14-21.

Downing, T.E., H.A. Pearson, and C. Garcia-Downing (eds). 1992. Development orDestruction: The Conversion of Tropical Forestto Pasture in Latin America. Boulder:Westview Press.

Echevarría, M. 2002a. “Financing WatershedConservation: The FONAG Water Fund inQuito, Ecuador.” In S. Pagiola, J. Bishop,and N. Landell-Mills (eds.), Selling ForestEnvironmental Services: Market-basedMechanisms for Conservation andDevelopment. London: Earthscan.

Echevarría, M. 2002b. “Water User Associationsin the Cauca Valley: A VoluntaryMechanism to Promote Upstream-

Downstream Cooperation in the Protectionof Rural Watersheds.” Land-Water Linkagesin Rural Watersheds Case Study Series.Rome: FAO (processed).

Faminow, M.D. 1998. Cattle, Deforestation, andDevelopment in the Amazon: An Economic,Agronomic and Environmental Perspective.Oxford: Oxford University Press.

Ferraro, P.J. 2001. “Global Habitat Protection:Limitations of Development Interventionsand a Role for Conservation PerformancePayments.” Conservation Biology, 15:4, pp.1-12.

Ferraro, P.J., and A. Kiss. 2002. “DirectPayments for Biodiversity Conservation.”Science, 298, pp.1718-1719.

Fisher, M.J., I.M. Rao. M.A. Ayarza, C.E.Lascano, J.I. Sanz, R.J. Thomas, and R.R.Vera. 1994. “Carbon Storage by IntroducedDeep-rooted Grasses in the South AmericanSavannas.” Nature, 371, pp.236-238.

Fondo Nacional de Financiamiento Forestal(FONAFIFO). 2000. El Desarrollo del Sistemade Pago de Servicios Ambientales en Costa Rica.San José: FONAFIFO.

Global Environment Facility (GEF). 1999.“Experience With Conservation TrustFunds.” GEF Evaluation Summary ReportNo.1-99. Washington: GEF.

Gobbi, J. 2002. “Enfoques SilvopastorilesIntegrados Para el Manejo de Ecosistemasen Colombia, Costa Rica y Nicaragua:Análisis Económico-Financiero Ex-Ante dela Inversión en los SSP Propuestos ParaCada País.” Turrialba: CATIE (processed).


References

Hardner, J., and R. Rice. 2002. “RethinkingGreen Consumerism.” Scientific American,May, pp.88-95.

Harvey, C., and W. Haber, 1999. “Remnant Treesand the Conservation of Biodiversity inCosta Rican Pastures.” Agroforestry Systems,44, pp.37-68.

Jaramillo, L. 2003. “Estimations of theOpportunity Costs of Forested Land inMexico.” México: Instituto Nacional deEcología (processed).

Kaimowitz, D. 1996. Livestock and Deforestationin Central America in the 1980s and 1990s: APolicy Perspective. Bogor: CIFOR.

Kaimowitz, D., B. Mertens, S. Wunder, and P.Pacheco. 2004. Hamburger Connection FuelsAmazon Destruction: Cattle Ranching andDeforestation in Brazil’s Amazon. Bogor:CIFOR.

Landell-Mills, N., and I. Porras. 2002. SilverBullet or Fools’ Gold? A Global Review ofMarkets for Forest Environmental Services andTheir Impact on the Poor. London: IIED.

Lutz, E., S. Pagiola, and C. Reiche. 1994. “Cost-Benefit Analysis of Soil Conservation: TheFarmers’ Viewpoint.” The World BankResearch Observer, 9:2, pp.273-295.

Mahar, D. 1988. “Government Policies andDeforestation in Brazil’s Amazon Region.”Environment Department Working PaperNo.7. Washington: World Bank.

Mejía, C. 2004. “Línea de Base de Colombia.” InJ.A. Gobbi (ed), “Enfoques SilvopastorilesIntegrados para el Manejo de Ecosistemas:Reporte sobre la Línea de Base.” InternalPaper No.2. Turrialba: CATIE.

Meinzen-Dick, R., A. Knox, F. Place, and B.Swallow (eds). 2002. Innovation in NaturalResource Management: The Role of PropertyRights and Collective Action in DevelopingCountries. Baltimore: Johns HopkinsUniversity Press for IFPRI.

Mertens, B., R. Poccard-Chapuis, M.-G. Piketty,A.-E. Lacques, and A. Venturieri. 2002.“Crossing Spatial Analyses and LivestockEconomics to Understand DeforestationProcesses in the Brazilian Amazon: the Caseof Sao Felix do Xingu in South Para.”Agricultural Economics, 27, pp.269-294.

Murgueitio, R.E., 1999. “Sistemas Agroforestalespara la Producción Ganadera en Colombia.”Paper presented at the SeminarioIntensificación de la ganadería enCentroamérica: Beneficios Económicos yAmbientales, Turrialba, Costa Rica, May 24-26.

Myers, N. 1981. “The Hamburger Connection:How Central America’s Forests BecameNorth America’s Hamburgers.” Ambio, 10:1,pp.3-8.

National Center for Environmental Economics(NCEE). 2001. The United States Experiencewith Economic Incentives for Protecting theEnvironment. Report No.EPA-240-R-01-001.Washington; United States EnvironmentalProtection Agency.

Pagiola, S. 2002. “Paying for Water Services inCentral America: Learning from CostaRica.” In S. Pagiola, J. Bishop, and N.Landell-Mills (eds.), Selling ForestEnvironmental Services: Market-basedMechanisms for Conservation andDevelopment. London: Earthscan.



Pagiola, S., J. Kellenberg, L. Vidaeus, and J.Srivastava. 1997. Mainstreaming Biodiversityin Agricultural Development: Toward GoodPractice. Environment Paper No.15.Washington: World Bank.

Pagiola, S., N. Landell-Mills, and J. Bishop.2002. “Making Market-based MechanismsWork for Forests and People.” In S. Pagiola,J. Bishop, and N. Landell-Mills (eds.), SellingForest Environmental Services: Market-basedMechanisms for Conservation andDevelopment. London: Earthscan.

Pagiola, S., and G. Platais. 2001. “SellingBiodiversity in Central America.” Paperpresented at the International Workshop onMarket Creation for Biodiversity Productsand Services, Paris, 25-26 January 2001(processed).

Pagiola, S., and G. Platais 2003. “ImplementingSystems of Payments for EnvironmentalServices: Initial Lessons of Experience.”Paper presented at the Workshop onEcosystem Services in the Tropics:Challenges to Marketing Forest Function,Spring 2003 Meeting of the InternationalSociety of Tropical Foresters, YaleUniversity, April 5-6, 2003.

Pagiola, S., and G. Platais. Forthcoming.Payments for Environmental Services: FromTheory to Practice. Washington: World Bank.

Pagiola, S., G. Platais, and A. Arcenas. 2003.“Ensuring the Poor Benefit from Systems ofPayments for Environmental Services.”Paper presented at the Workshop onReconciling Rural Poverty Reduction andResource Conservation: IdentifyingRelationships and Remedies, CornellUniversity, Ithaca, NY, May 2-3, 2003.

Pagiola, S., and I.-M. Ruthenberg. 2002. “SellingBiodiversity in a Coffee Cup: Shade-grownCoffee and Conservation in Mesoamerica.”In S. Pagiola, J. Bishop, and N. Landell-Mills(eds.), Selling Forest Environmental Services:Market-based Mechanisms for Conservation andDevelopment. London: Earthscan.

Pérez, A.M., I. Siria, M. Sotelo, I. Arana, F.Ramírez, E. Ramírez, and M. Ibrahim. 2004.“Resultados Preliminares del Monitoreo dela Biodiversidad en Sistemas Silvopastorilesde Matiguás y Río Blanco, Dpto deMatagalpa, Nicaragua.” Gaia, No.5, pp.1-26.

Pfaff, A., S. Kerr, F. Hughes, S. Liu, G. Sanchez,D. Schimel, J. Tosi, and V. Watson. 2000.“The Kyoto Protocol and Payments forTropical Forest: An InterdisciplinaryMethod for Estimating Carbon-offsetSupply and Increasing the Feasibility of aCarbon Market under the CDM.” EcologicalEconomics, 35:2, pp.203-221.

Ramírez, G.H. 1997. “Evaluación de DosSistemas Silvopastoriles Integrados porCynodon Plectostachyus, LeucaenaLeucocephala y Prosopis Juliflora.”Proceedings V Seminario-TallerInternacional Sistemas Sostenibles deProducción Agropecuaria y PrimerSeminario Internacional Palmas en Sistemasde Producción Agropecuaria para elTrópico, Cali, July 31-August 3, 1997. Cali:Fundación CIPAV.

Ramírez, E., AY. Cárdenas, and O. Dávila. 2004.“Línea de Base de Nicaragua.” In J.A. Gobbi(ed), “Enfoques Silvopastoriles Integradospara el Manejo de Ecosistemas: Reportesobre la Línea de Base.” Internal PaperNo.2. Turrialba: CATIE.


References

Repetto, R., and M. Gillis (eds.). 1988.Government Policies and the Misuse of ForestResources. Cambridge: CambridgeUniversity Press.

Rice, R. 2003. “Conservation Concessions -Concept Description.” Washington:Conservation International (processed).

Rojas, M., and B. Aylward. 2002. “The Case ofLa Esperanza: A Small, Private,Hydropower Producer and a ConservationNGO in Costa Rica.” Land-Water Linkagesin Rural Watersheds Case Study Series.Rome: FAO.

Saunders, D.A., and R.J. Hobbs. 1991. “The RoleOf Corridors in Conservation: What do weKnow and Where do we go?” In D.ASaunders and R.J. Hobbs (eds.), The Role ofCorridors. Surrey: Beaty & Sons.

Schneider, R. 1994. “Government and theEconomy on the Amazon Frontier.” LACRegional Studies Program Report No.34.Washington: World Bank.

Thiollay, J.-M. 1995. “The Role of TraditionalAgroforests in the Conservation of RainForest Bird Biodiversity in Sumatra.”Conservation Biology, 9, pp.335-353.

Tipper, R. 2002. “Helping Indigenous FarmersParticipate in the International Market forCarbon Services: The Case of Scolel Té.” InS. Pagiola, J. Bishop, and N. Landell-Mills(eds.), Selling Forest Environmental Services:Market-based Mechanisms for Conservation andDevelopment. London: Earthscan.

Tomich, T.P., M. Van Noordwijk, A.Budidarsono, A. Gillison, T. Kusumanto, D.

Murdiyarso, F. Stolle, and A.M. Fagi (eds).1998. Alternatives to Slash-and-Burn inIndonesia: Summary Report and Synthesis ofPhase II. Bogor: ICRAF.

White, D., F. Holmann, S. Fuijsaka, K. Reategui,and C. Lascano. 2001. “Will IntensifyingPasture Management in Latin AmericaProtect Forests? - Or is it the Other WayRound?” In A. Angelsen and D. Kaimowitz(eds), Agricultural Technologies and TropicalDeforestation. Oxford: CAB International.

World Bank. 2000. “Ecomarkets Project: ProjectAppraisal Document.” Report No.20434-CR.Washington: World Bank.

World Bank. 2002. “Colombia, Costa Rica, andNicaragua Regional IntegratedSilvopastoral Approaches to EcosystemManagement Project: Project AppraisalDocument.” Report No.21869-LAC.Washington: World Bank.

World Bank. 2003a. “Guatemala WesternAltiplano Natural Resources ManagementProject: Project Appraisal Document.”Report No.25660-GUA. Washington: WorldBank.

World Bank, 2003b. “State and Trends of theCarbon Market 2003.” Washington: WorldBank.

Zelek, C.A., and G.E. Shively. 2003. “Measuringthe Opportunity Cost of CarbonSequestration in Tropical Agriculture.” LandEconomics, 79:3, pp.342-354.

paying for biodiversity conservation services in ... · paying for biodiversity conservation...

Documents