Public Disclosure Authorized Devolution of Resource Rights...

Devolution ofResource Rights,Poverty, and NaturalResource Management

A Review

Priya ShyamsundarEduardo AraralSuranjan Weeraratne

PAPER NO. 104

May 2005

E N V I R O N M E N T A L E C O N O M I C S S E R I E S

34114

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

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.

Devolution of ResourceRights, Poverty, andNatural ResourceManagement

A Review

Priya ShyamsundarEduardo AraralSuranjan Weeraratne

THE WORLD BANK ENVIRONMENT DEPARTMENT

May 2005

© 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 2005

iiiEnvironmental Economics Series

Contents

ABSTRACT vii

ACKNOWLEDGMENTS ix

ABBREVIATIONS AND ACRONYMS xi

Chapter 1Devolution of Resource Rights, Poverty, and Natural Resource Management — An Overview 1

Introduction 1Methods 2Community-based Natural Resource Management: A Global Phenomenon 2Different Forms of Decentralization 3Understanding Impacts 5

Poverty impacts 7Empowerment impacts 8Natural resource impacts 10Government impacts 10

From Devolution to Collective Action 11Challenges Ahead 14

Getting household incentives right 14Strengthening property rights 15Addressing heterogeneity and distributional issues 16Solving the scale problem 16Providing adequate support 17Enhancing sustainability 17

Conclusions 18

Chapter 2Success, Impacts, and Emerging Challenges in Community-Based Wildlife Management: WhatDoes the Evidence Show? 21

Introduction 21Universal Features, Local Differences 22

Income generation and distribution 23Devolution of responsibilities 24External agents 25

What Contributes to Effective Community Management? 25

iv Environment Department Papers

Environment in Poverty Reduction Strategies and Poverty Reduction Support Credits

Welfare and Resource Implications 29Benefits through village-level projects 29Household benefits 31Costs related to land loss and wildlife predation 32Reduced poaching and better resource conservation 34Improved perceptions, stronger rights and reduced conflicts 35

New Generation Concerns Resulting from Decentralization 36Communities versus households 36Heterogeneity within and between communities 37Competition between institutions 38Tenure over land and resources 39Mismatch between ecological and social scale of management 39Sustainability 41

Conclusions 41

Chapter 3Water User Associations and Irrigation Management Transfer: Understanding Impactsand Challenges 45

Introduction 45IMT Assumptions and Models 46Conditions For Self-Governing Irrigator Associations 48Assessing Impacts 51

Productivity and farm income 52Impact of IMT on government finance and personnel 54Impacts on O&M quality and resource conservation 56

Challenges Ahead 57Poverty reduction and IMT 59IMT and institutional development 59Maintenance 60IMT and water productivity 61Poor quality of IMT research 61

Summary and Conclusion 62

Chapter 4Assessing Decentralized Forest Management 65

Introduction 65Community Forestry Worldwide 66Community Management: Trends, Commonalities, and Differences 68

Three major institutional mechanisms 69Diverse benefit-sharing strategies 70From benefit sharing to power sharing 71From degraded to conservation forests 74

Enabling Environment 75Changes in the policy and legal framework 75Effective community demand 76Strong monitoring and enforcement mechanisms 76Complementary role of external organizations 77

Impacts of Community Forestry 79Household and community benefits 79

v

Contents

Environmental Economics Series

Short-term versus long-term benefits 80Gender inequities and distributional issues 82Positive natural resource impacts 83

Challenges to be Addressed 84Compensatory mechanisms for lost rights 84 Reconciliation of heterogeneous interests 84Greater community control over non-timber forest products 86Devolution of bureaucratic power and not just usufruct rights 87Tackling corruption 87

Conclusions 89

NOTES 91

REFERENCES 95

BOXES

1 Devolution and Decentralization — Some Definitions 22 Three Key Institutional Mechanisms and their Implications 53 Different Institutional Arrangements in CBNRM 244 Policy Reform — An Evolutionary Process 285 Evaluating Community Conservancies in Namibia—Community Investments

Improve Welfare 316 Transaction Costs Associated with Community Management in Kenya 347 Is Community Conservation Incentive Compatible? 378 Irrigation Management Transfer in Mexico 489 Features of Successful Irrigation Associations in the Philippines 51

10 IMT Impact on Cost of Irrigation Water 5211 IMT Impact on Agriculture 5312 IMT Impact Assessment in Mexico 5713 Recommended Improvements in IMT Research Methods 6214 Different Categories of Community Forestry Worldwide 6715 Authority and Rights: A Synopsis 7216 Features of Effective Monitoring and Enforcement 7717 Methods Used to Generate Short-Term Benefits in the Buldhana Forest

Division in Maharashtra India 8118 Compensatory Strategies to Decrease Opportunity Costs of Conservation 8519 Mechanisms to Overcome Heterogeneity: A Case Study from Nepal 8620 Communities as Managers and Not Just Users: The Case of Duru Haitemba,

a Successful Example from Tanzania 88

TABLES

1 Impacts of Decentralized Natural Resource Management 62 Examples of Benefits from CBNRM to Households, Communities, and the State 303 Increase in Wildlife Numbers in Kunene Region, Namibia, 1982–97 344 Impact on government subsidies and budget solvency of the irrigation agency 555 Summary of IMT Impacts 58

viiEnvironmental Economics Series

Abstract

Over the last several years, the rights of localcommunities over natural resources have beenstrengthened either through power-sharingagreements with the state, increased legalaccess to natural resources, or decentralizationwithin national agencies. Understanding theimpacts of these institutional changes isimportant both for governments and otherstakeholders. This paper focuses on fourquestions: (1) What is the scope and scale ofdecentralized natural resource management indifferent resource sectors? (2) What do we

understand about the impacts of devolution, interms of poverty reduction, resourceconservation, and financial implications forgovernments and local agencies? (3) What aresome conditions that contribute to success? (4)What does the future hold for decentralizedresource management; that is, what are someemerging challenges? These questions areaddressed in relation to three communalresource-management activities—community-based wildlife management, irrigationmanagement transfer, and community forestry.

ixEnvironmental Economics Series

Acknowledgments

We are grateful to the Trust Fund forEnvironmentally and Socially SustainableDevelopment for generous financial support.The financial support of the Government ofSweden is also gratefully acknowledged. Ourthanks also go to Kirk Hamilton, Team Leader,

Policy and Economics Team for his support andto peer-reviewers Jan Bojo and Robin Mearnsfor their excellent comments. We also thankRobert Livernash for his careful editing of thisdocument.

xiEnvironmental Economics Series

Abbreviations and Acronyms

ADB Asian Development Bank

AFR Africa

CAS Country Assistance Strategy

CEPOMS Committees on Environmental Policy and Management

CERSGIS Center for Remote Sensing and GIS

CIDA Canadian International Development Agency

CIEDP Committee for Integrating Environment into Development and Planning

DFID Department for International Development, United Kingdom

EAP East Asia and Pacific

EC European Commission

ECA Eastern Europe and Central Asia

EDP Environment Department Paper

EIA Environmental Impact Assessment

ENR Environment and Natural Resources

EPA Environmental Protection Agency

GPRS Ghana Poverty Reduction Strategy

GTZ German Technical Cooperation

HIPC Heavily Indebted Poor Countries

IDA International Development Association

IMF International Monetary Fund

IPRSP Interim Poverty Reduction Strategy Paper

JBIC Japan Bank for International Cooperation

JSA Joint Staff Assessment

LAC Latin America and Caribbean

MDAs Ministries, Departments, and Agencies

MDG Millennium Development Goals

MNA Middle East and North Africa

MoE Ministry of Environment

MTEF Medium Term Expenditure Framework

1Environmental Economics Series

Devolution of Resource Rights,Poverty, and Natural ResourceManagement — An Overview1

Introduction

Within the broad array of issues associated withpoverty and natural resources, thedecentralization and devolution of resourcerights is arguably one of the more importantpolicy issues confronting donors, practitioners,and governments.

In the natural resources area, profound changesare occurring in terms of who has access to andcontrol over resources. Over the last severalyears, the rights of local communities overnatural resources have been strengthened eitherthrough power-sharing agreements with thestate, increased legal access to natural resources,or decentralization within national agencies.Understanding the impacts of these institutionalchanges is important both for governments andother stakeholders.

Since decentralization efforts have been adevelopment trend for about two decades, it isappropriate to ask how and in what waydecentralized natural resource management hascontributed to improved livelihoods and betterresource management. There are also usefullessons to be learned from understandingdevolution experiences across different resourcesectors.

This study focuses on four questions:

• What is the scope and scale of decentralizednatural resource management in differentresource sectors?

• What do we understand about the impactsof devolution, in terms of povertyreduction, resource conservation, andfinancial implications for governments andlocal agencies?

• What are some conditions that contribute tosuccess?

• What does the future hold for decentralizedresource management; that is, what aresome emerging challenges?

In answering these questions, the study seeks tocraft solutions that build on both synergies andtrade-offs between poverty reduction goals andnatural resource management.

This study is based on a careful review ofliterature in three sectors: community-basedwildlife management, irrigation, and forestry.Chapters 2, 3, and 4 focus on each of thesesectors. The review was undertaken as part of alarger project that includes case studies oncommunity conservancies in Namibia,community forestry in India, and irrigationmanagement transfer in the Philippines. Resultsfrom the case studies are reported elsewhere.3

Priya Shyamsundar2

1

Environment Department Papers2

Devolution of Resource Rights, Poverty, and Natural Resource Management — A Review

Methods

The analysis in this paper is based on a reviewof over 175 articles and reports. We drewprimarily on peer-reviewed journal articles andworking papers written in the last five years.The articles are from established developmentjournals and the working papers and websitesreviewed are generally associated withinternational research organizations and donoragencies. While we have tried to identify muchof the recent work done on community-basednatural resource management in three sectors,our review is clearly not exhaustive.

Community-based Natural ResourceManagement: A Global Phenomenon

Decentralization and devolution (see Box 1)have enjoyed considerable momentum in thelast two decades. The most advanced forms ofdecentralization are arguably in the irrigationsector, but there is significant evolution in eachof the other fields.

Much of the experimentation in community-based wildlife management (CBWM) hasoccurred in Africa. CBWM gained ground in the1980s and 1990s, partly as a result of increasing

local and international resistance to strictprotected-area programs. It was also a responseto failures in state-run conservation.Decentralization in wildlife management ischaracterized by two overlapping phases. Thefirst phase is exemplified by IntegratedConservation and Development Programs(ICDPs) and the second phase by what iscurrently known as community-based naturalresource management (CBNRM). Bothprograms create economic incentives for localcommunities to conserve natural resources andboth involve the participation of localcommunities in decision-making. However,CBNRM programs further strengthen localcommunities in two ways: first, there is agreater focus on the sustainable use of naturalresources, whereas in ICDPs there is often aneffort to find substitutes for natural resources-based activities. Second, CBNRM seeks toempower local communities with greaterdecision-making power. Thus, while ICDPs maynot involve any form of decentralization,CBNRM programs are meant to lead todecentralization of managementresponsibilities. Chapter 2 discusses in detailsome of the challenges faced in attempting todevolve authority to local groups.

Box 1Devolution and Decentralization — Some Definitions

Decentralization has been used to characterize devolution of power within state bureaucracies, privatization,and increased political power to local authorities. Knox and Meinzen-Dick (2000) discuss decentralization aspart of a group of policies that are closely related to each other. These different policies include:

• Deconcentration — the transfer of decisionmaking authority to lower-level units of government• Decentralization — the transfer of decisionmaking and payment responsibility to lower levels of government• Privatization — the transfer of public sector functions to the private sector or private individuals• Devolution — the transfer of rights and responsibilities to user groups at the local level.

We understand devolution of resource rights to broadly mean a process by which state control over the use ofnatural resources is gradually and increasingly shared with local communities. This can happen with or with-out bureaucratic or political decentralization. It is generally accompanied by the creation or strengthening ofa subset of local institutions.

3

Devolution of Resource Rights, Poverty, and Natural Resource Management — An Overview


Irrigation management transfer (IMT)—theprocess of transferring irrigation managementfrom government to farmer organizations—wasfirst undertaken in the United States, France,Colombia, and Taiwan from the 1950s to the1970s; developing countries followed in the1980s and 1990s. To date, governments in atleast 25 countries in Asia, Latin America, andAfrica are reducing their roles in irrigationmanagement, while farmer groups or privateorganizations are taking them over (Vermillion1992). Since the mid-1980s, the centerpiece ofreforms has been the transfer of management(in rare cases, also ownership) of irrigationsystems—wholly or in part—to nongovernmen-tal agencies, combined with a reduced role forgovernment agencies in operation andmaintenance (O&M), fee collection, watermanagement, and conflict resolution. Chapter 3discusses the impacts of these changes on farmproductivity, water resources use, andgovernment finances.

Of the 3.9 billion hectares of global forests, it isestimated that some 77 percent are owned bygovernments, 11 percent are reserved for orowned by community and indigenous groups,and 12 percent are owned by individuals orfirms (White and Martin 2002). If onlydeveloping countries are considered, Martinand White estimate that communities groupshave access to and/or ownership over at least22 percent of forests. Thus, a sizable tract of theforest estate in poor countries is identified asbeing controlled by communities. Notsurprisingly, over the last two decades,community forestry as a forest managementstrategy has graduated from being a somewhatexperimental strategy to being more integratedinto conventional national forestry efforts(Arnold 2001). Decentralized forestmanagement is particularly prominent in South

Asia, Sub-Saharan Africa, parts of East Asia,and Central and Latin America. Chapter 4focuses largely on the different types ofexperiments under way and the lessons that canbe gleaned from these experiences.

Different Forms of Decentralization

Over the last decade, decentralization in naturalresource management has typically resulted inthe creation of a community user group tomanage a common pool resource. In some cases,the links between local government and thedecentralized NRM authority are indistin-guishable; however, often user groups functionwith only minimal supervision from the state.

Decentralization in wildlife management isgenerally characterized by either (a) devolutionof management powers to local governmentagencies; (b) the creation of new localconservation institutions; or (c) greaterauthority to traditional leaders. In Zimbabwe’s,well-known CAMPFIRE4 program, for example,authority is devolved to lower levels ofgovernment known as rural district councils.These councils retain rights over managementdecisions and revenue distribution. A differenttype of example is Zambia’s ADMADE5

program, where decentralization is at twolevels. First, each wildlife area is managedunder a wildlife management authority, whichis headed by the district governor. This area isthen divided into wildlife management sub-authorities, which are controlled by traditionalchiefs. In this model, traditional leaders holdwide-ranging powers. A third type of model isfound in Namibia, where communities canestablish a conservancy and gain exclusiverights to commercial tourism operations if theydefine a geographical area, define membership,develop operating rules, and so on (Jones



1999b). Each model comes with its ownadvantages and disadvantages (see Box 2);however, community management with littlereal authority to communities (as appears to bethe case in CAMPFIRE) is likely to result in lesslong-term change.

In the irrigation sector, a common form ofdecentralization involves a transfer ofresponsibilities from public irrigation agenciesto water users’ associations (WUA) or irrigators’associations. This is the case in many countriessuch as the Philippines, Indonesia, India,Senegal, and Colombia. Other types ofgovernance structures exist, but these are notthe norm. For instance, in Turkey, municipalgovernments became the new managementunit. In Vietnam, parastatal organizationsassumed responsibility, while in Sudan andNew Zealand, private / mutual companiesassumed responsibility. In the United States,Japan, South Korea, Mexico, and Taiwan—countries where irrigation management transferis deemed successful—post-transfer governanceentities tend to be farmer-elected boards ofdirectors, while management entities tend to becadres of professional staff appointed by theboard.

In forestry, devolution of authority to lowerlevels of government is evident in parts of LatinAmerica, especially Guatemala and Bolivia,where forest laws passed in the mid-1990sdelegate authority to municipal councils. In theBolivian case, decentralization to localgovernment has increased local power and theability to tax. More commonly, community-based forest management is characterized bythe creation of new village-level institutions.Village forest institutions in India, forest usergroups in Nepal, and community forestenterprises in Mexico are among the moreprominent examples. In many cases, especially

in Africa, traditional leaders have alsoestablished themselves as importantstakeholders. Quite frequently, there is evidenceof parallel local authority systems consisting oftraditional leaders on the one hand andgovernment-established structures on the other.

Box 2 summarizes some of the differentadvantages and disadvantages of the threetypes of commonly found decentralizedresource management structures. Differentinstitutional regimes may work better underdifferent political circumstances. The success ofa specific regime would also depend on pre-existing capacities and resource constraints.

• Decentralization to local governmentagencies has distinct advantages.Implementation is likely to be easierbecause of their ability to build on existingbureaucratic structures and the authorityoften vested in state organizations.However, collective action problems maynot be resolved unless local governmentsare strongly accountable to communitiesand there is mutual trust.

• Community-level organizations aretypically more likely to result in localsexercising actual control over naturalresources. These actions are likely to besustained over time as norms areinternalized. However, communityorganizations may be handicapped by alack of resources and capacity. Further,without active state intervention, inter- andintra-community conflicts may negatecommunity efforts. Local government that isaccountable to communities can overcomemany of the disadvantages faced by usergroups that are de-linked from localgovernment. However, local governmentcannot be assumed to be accountable to

5



rural communities or represent themadequately (Ribot 2002).

• Traditional leaders are often critical to thesuccess of community organizations.However, elite capture and favoritism arepotential outcomes that can have negativeimpacts on poverty reduction. Conflictsrelated to distribution of benefits andsharing of authority can also slow thebenefit stream from community action.

Understanding Impacts

There are few carefully undertaken studies thatidentify impacts of decentralized naturalresource management. Measuring impacts isfraught with difficulties associated withdifferentiating between on-going biophysicalchanges, the impacts of policy reforms, andbetween the effects of decentralization and theeffects of other overlapping policies (Robot2002). Further, lack of baseline information andcareful quantitative before and after studies

Box 2Devolution — Three Key Institutional Mechanisms and Their Implications

Advantages Disadvantages Lower levels of Government

� Opportunities for raising revenues for the state

� Implementation easier because of access to resources, incentives to exercise management control over resources, and reliance on pre-existing bureaucratic structures

� Possible opportunities to be more egalitarian in distribution of benefits

� Communities may not gain real authority in planning and decisionmaking; open access problems may not be resolved unless local government is strongly accountable to communities

� Capacity building at bureaucratic and community levels required

� Potential for resentment and conflicts between local government and communities and/or traditional leaders

Community-level Institutions

� Communities can exercise direct control over resources

� More likely to be sustained over time as norms and rules are internalized

� Potential for capture by traditional leadership and hierarchies

� Need for rapid capacity building if organizations are new

� Inter-community conflicts likely in the short run

� Potential for quick collapse in the absence of strong leadership and/or external support

� Growth of many issue-specific organizations with associated transaction costs

Traditional Leaders

� Implementation may be easier because of reliance on existing systems of communication & rules

� Communities can exercise direct control over resources through reciprocal social ties

� High potential for capture by traditional leadership and hierarchies

� Inter- and intra-community conflicts may arise in the short run

� Overlapping authority with local government may be a problem

Devolution to



makes it difficult to link specific outcomes todecentralization. The conclusions we draw thusrepresent impact observations and should betreated as a mapping of likely outcomes ratherthan as definite results. In general, this is anarea that needs more careful quantitativeanalyses.

Impacts of decentralized NRM can becategorized into poverty impacts, naturalresource impacts, and empowerment impacts.There are also various impacts on governmentfinances and personnel requirements. Table 1summarizes some of the impacts of CBNRM.

Table 1. Impacts of Decentralized Natural Resource ManagementIssue Wildlife Forestry Irrigation Poverty Impacts

Public (infrastructural) benefits not directly linked to resource use

Public infrastructural benefits that compliment private farm income and resource use

Private benefits (jobs, tourism ventures) mainly linked to resource conservation

Private benefits mainly linked to resource use (NTFPs, timber)

Private benefits from additional activities such as extension services and credit services that are not linked to resource use or conservation

Opportunity costs in terms of loss of land, possible increases in crop predation, and loss of subsistence meat/fodder uses

Opportunity costs mainly linked to current protection of forests and loss of subsistence fuel/fodder/NTFP uses

?

Transaction costs in terms of time in creating and sustaining new forms of group interaction

Significant monitoring costs Significant costs of infrastructure maintenance

? Production costs due to decrease in access to resource inputs used in small-scale enterprises

Production costs in terms of increase in water fees

Empo-werment Impacts

Community’s have re-gained limited traditional control over resources

Where control remains mostly in the hands of the state or is devolved to local government, community empowerment can be limited

Lack of secure tenure, limited rights to punish infractions, lack of bureaucratic support contribute to limited community empowerment

Heterogeneity in power, assets, and preferences result in winners and losers Natural Resource Impacts

Anecdotal and case study evidence of improvements in resource health and stock

Water use efficiency limited —observed only in cases where other conditions (such as volumetric pricing) are met

Scale issues remain unresolved Scale issues less problematic because of government system-wide control

Monitoring costs are reduced Some operational costs are reduced

Govern-ment Impacts

Staff costs may be reduced Reduction in staff a key component of reforms

7



Poverty impacts

CBNRM seeks to economically empower poor peo-ple by a) providing access to public goods throughdevelopment of local infrastructure; b) increas-ing opportunities to legally use natural resourc-es; and c) increasing private income opportunitiesthrough jobs, access to different services, and self-employment possibilities. Decentralization, how-ever, is not costless. It can result in significanttransaction and opportunity costs. In certain cas-es, productions costs also increase.

Community benefits are a trademark of naturalresource management programs. Some of thecommunity benefits are from directinternational assistance, while the rest comesfrom program earnings. Many of the wildlifemanagement programs in Africa, for instance,have resulted in local infrastructure such asroads and schools.6 Similarly, communityforestry has frequently resulted in village-levelassets. For example, Bray and others (2002)observe that community forestry in Mexico hasenabled communities to build “assets such aspotable water networks, schools, clinics, publicbuildings and social service safety nets …”Similarly, in the irrigation sector, successfultransfer of management to farmer groupsfrequently involves infrastructural investments.

While some studies do suggest that theinfrastructural benefits from CBNRM contributeto poverty reduction, this is more often acommon assumption — one that underliesmuch of development assistance forinfrastructure. A more pertinent concern iswhether community benefits necessarily createthe right incentives for sustainable resource use,particularly in forestry and wildlife.Infrastructural benefits (such as schools orcommunity halls) are rarely directly tied toprudent use of resources. They equally benefit

households who follow resource conservationrules as well as households that defect.

Household benefits are the most importantincentive mechanism for motivating successfulnatural resource management. Household-specific benefits from community-orientedwildlife management in Africa include wildlifedividends, guide and scouting jobs,employment in lodges and tour agencies,possibilities of selling handicrafts and tourism-related services, and availability of meat fromculling operations.7 Household benefits incommunity forestry accrue from increasedcontrol over timber and non-timber resourcesand revenue sharing with the government.Employment in forest-related enterprises isanother common source of revenue.

In irrigation, management transfer tocommunities is expected to increase farmproductivity and household profits in the longrun. However, the relationship betweenirrigation management transfer and agriculturaland economic productivity is indirect andambiguous. This ambiguity partly stems fromthe fact that impact evaluations of changes inproductivity rarely control for interveningvariables such as changes in rainfall or fertilizerapplication rates, as well as factors such asprices. Further, much of irrigation managementtransfer focuses on decreasing the burden ongovernment budgets rather than on increasingfarm productivity. Evidence of increased farmproductivity is found mainly in cases whererobust farmers’ organizations have earnedrevenues through diverse sources and byproviding better extension services to members.

The poverty impacts of CBNRM are as much aresult of the benefit stream that occurs as of thecosts that are incurred. The most importantcosts associated with CBNRM are opportunity



costs stemming from loss of access to land,forests, and wildlife. While estimates ofopportunity costs are limited, in specific cases,these costs can be large and result in negativeshort-term returns. For example, many state-supported community forestry programs resultin forest closures or loss of access to non-timberforest products in the short run. This can imposeimmediate costs on poor households who aredependent on forests for fuel-wood and othersubsistence products.8 However, short-termscosts can be minimized though careful planningas suggested by the examples in Chapter 4.Long-term costs of land loss are more difficult toaddress given the challenges involved inensuring judicious compensation.

Another set of costs associated withcommunity-based activities are transaction costsresulting from participation in meetings,monitoring, providing labor for maintenance ofinfrastructure, and membership fees. Very fewstudies actually document the burden placed bythese costs, but monitoring costs in particularcan be significant in wildlife and forestmanagement. In areas with wildlife, animalpredation-related costs are another majordifficulty.

Community management can increaseproduction costs as well. Participatory irrigationmanagement, for example, can lead to anincrease in the cost of irrigated water whensignificant subsidies existed before IMT andthese subsidies are simultaneously dismantled.Further, high-cost systems such as pumpirrigation may significantly increase the cost ofwater to farmers (Vermillion 1997).

The case studies reviewed point to variousbenefits and costs from community-basednatural resource management that can have a

huge impact on the poor. What is difficult toconclude is if the aggregate impacts are mainlypositive or not. Several factors make carefulassessment challenging:

a) Studies often do not distinguish betweengross and net benefits

b) Impact evaluations rarely consider beforeand after program results

c) Programs differ in their distributionalimplications

d) Trade-offs between short-term and long-term benefits are generally ignored.

Overall, our analysis suggests a need for morecareful quantitative impact analyses. We needstudies that build on baseline data andundertake before and after comparisons toassess the welfare impacts.

Empowerment impacts

Devolution of natural resources creates space forcommunities to have a “voice” in how forests,water, and wildlife are managed. Voice, however,depends on the contractual agreement betweenthe state and communities. Further, communi-ties are heterogeneous entities, and decentraliza-tion can have unequal impacts on differentcommunity members.

There is evidence to suggest that communitymanagement has empowered communities byallowing them to regain traditional control overnatural resources. This appears to be true inwildlife, forestry, and water management. Tothe extent that devolution increases access tolocal government officials, brings managementdecisions to the local domain, and improvesaccess to information about natural resourcechanges, CBNRM strengthens communityrights. Further, there are a number of non-

9



financial benefits as well. For example, Ashley(1998) suggests that in Namibia, communitymanagement of wildlife has resulted indevelopment of new skills, pride, a sense ofcontrol, and experience and confidence indealing with outsiders—outcomes that gobeyond initial development objectives.Examples that counter Ashley’s results arefound where a) local government agencies areperceived to promote natural resourcemanagement instead of economic development;and b) where decentralized structures re-createtraditional hierarchies, with some membershaving more and others less “voice.”9

An important result of many community-oriented wildlife management and forestryprograms is improved relations over timebetween authorities managing protected areasand local households, and a decrease inconflicts.10 Some of these changes are a productof strategies that try to empower bothcommunities and government officials to worktogether to meet common goals.

Devolution that delegates managementresponsibilities to local communities withoutcommensurate rights is unlikely to empowerhouseholds or provide the right incentives forcollective action (Meinzen-Dick and Gregoriao2004). This is a critical concern indecentralization. For instance, in forestmanagement, two approaches define the roleand contribution of communities. The “user-centered” approach, which is predominantlyseen in Asia, recognizes local communities asforest users and seeks to secure their co-operation by granting legal access to certainproducts or a share in forest-derived benefits.The “power-sharing” approach mainly looks toforest communities as potential managers or co-managers and devises arrangements to give

them varying degrees of managerial power. Thepower-sharing mechanism seems mostadvanced in Tanzania and the Gambia.11 Thereare some indications of a gradual trend incommunity forestry from a user-based to a moremanager-oriented system. Such a shift willcertainly give communities more authority tomake long-term investments in forestry.

Lack of secure tenure, limited rights to punishinfractions, and lack of bureaucratic support allcontribute to less than effective communitycontrol over natural resources. In many small-scale irrigation systems in West and CentralJava in Indonesia, for instance, water users’associations have a mandate for operations andmaintenance, but they do not have formal rightsover water and infrastructure (Samad andothers 2002). This makes them powerless tosettle disputes and enforce fee collection. Theyalso do not have the political and legal clout toenter into business contracts. This type ofsituation is not un-common in many cases ofcommunity management in all three sectors.

Ribot (2002) suggests that successfuldecentralization depends on a) how accountablelocal institutions are to communities; b) whetherthese institutions have adequate discretionarypower; and c) whether the transfer of this poweris secure. All of these factors appear to make thedifference between whether or notdecentralization empowers local communities.

Natural resource impacts

There is some evidence that community-manage-ment programs are meeting their conservationgoals. Indicators such as the number of animalsand signs of poaching are improving, either be-cause of improved monitoring and enforcementor because of changed preferences. Community



forestry also appears to contributing to better for-est health. Evidence on participatory irrigationimpacts on water- use efficiency or conservationis limited. This assessment, however, is somewhattentative since biophysical changes at any pointin time are a result of multiple policies, actions,and physical changes.

There is some anecdotal evidence ofimprovements in wildlife numbers associatedwith CBWM. For example, animal censusresults from 1991 and 1994 show that the SelousConservation Program in Tanzania may haveincreased animal populations (Songorwo 1999).Such evidence is also found in the Kuneneregion of Namibia, where the community-guardprogram has resulted in the recovery of flagshipspecies such as the desert elephant and theblack rhino (Jones 1999b). Many of the papersreviewed suggest that poaching has beenreduced as a result of these programs. However,the results are not always unambiguous.Community management changes theincentives the rural residents face, but does notstem hunting. Small-game hunting is generallythrough snares (and not firearms) and locals areless likely to be caught (Gibson 1999). Hence,areas with more community scouts may beseeing an increase in snares and small-gamehunting, and a decline in large-game hunting.

In quite a few cases, there is demonstrableevidence of community forestry resulting inhealthier forests. Tanzania is perhaps the bestexample of having positive natural resourceimpacts in Africa. Especially in the miombowoodlands of Duru Haitemba and Mgori, thereare visible signs of gain. Charcoal burning,rampant timber harvesting, and unregulated in-forest settlements have all disappeared. This hascoincided with a return of understory shrubberyand grasses and the return of bee swarms to

forests (Wily, 1999). In the Mwamishali village,“after only three seasons, the area exhibited analmost unbelievable ecological transformation,from desolate landscapes of bare ground andheavily browsed shrubs to impressive vistas ofgrass and vigorously sprouting shrubs” (Mlenge2002). There is also evidence from India andNepal that community forestry is contributingto improved tree cover.

There is little to show that irrigation user groupsinfluence water use or conservation. Forexample, in a well-known study of the Alto RioLerma Irrigation District in Mexico coveringover 100,000 hectares, little evidence was foundthat increasing farmer control over water hasled to changes in water allocation ordistribution (Kloezen and others 1997). In thecase of irrigation, conservation appears to bemore closely tied to infrastructuraldevelopments and water pricing reform ratherthan to increases in efficiency as a result ofinstitutional changes.

Government impacts

Devolution of natural resource rights to local usergroups invariably changes the roles and functionsof government agencies. Overall, for cash-strapped governments in developing countriesthere are some significant advantages to decen-tralizing resource management.

Much of the case study literature on wildlifemanagement reports some form of benefitsharing between the state and communities.12

Increased attention to wildlife management hasmeant that the state can tax rural communitiesfor a resource (wildlife) that may have beenpreviously untaxed. Some authors suggest thatgovernments have been willing to supportCBWM because they can now get communities

11



to protect wildlife and can share in consequentcommercial transactions. In popular parks andprotected areas, revenue sharing withcommunities may actually decrease governmentrevenues in the short run. However, to theextent that community participation results inconservation and better opportunities fortourism, government revenues are likely toincrease over time. The state also gains becauseof reductions in monitoring costs and decreasedconflicts between frontline guards andcommunities.

In community forestry, the impact on the state issomewhat similar. To the extent that communitycontrol increases forest cover (as appears to bethe case in several countries), devolution resultsin assets that are then shared between the stateand communities. Improved relations betweenforest departments and communities, apotential decrease in corruption, and decreasedmonitoring costs are other benefits.

In irrigation management, there is a more directand significant link between governmentfinances and devolution. As Table 1 in Chapter 3shows, devolution does seem to lead to adecrease in government subsidies and improvesthe budget solvency of irrigation agencies.Improvements in government finances resultmainly from decreased spending on operationsand maintenance, increased fee collection,decreased subsidies, and a gradual decline instaff in national irrigation agencies.

From Devolution to Collective Action

The natural resources under consideration—water, wildlife, and forests—are common poolresources; control over these resources is beingdevolved from the state to communities. Thus,collective action is at the core of CBNRM

programs. In fact, the effectiveness of CBNRMprograms is directly related to their ability tofoster cooperative behavior. Thus, what aresome factors that are most likely to lead tosuccessful cooperative action?

Scholarly discussions identify several group andresource characteristics as necessary forsuccessful cooperative use of natural resources(Ostrom 1990, Baland and Platteau 1996,Agrawal 2001).13 Included among theseconditions are (a) clear boundaries ofmembership to resource users groups and to theresource in question; (b) a small number ofstakeholders or group size; (c) transparency andeffective monitoring; (d) homogeneity instakeholder endowments (such as wealthassets), preferences (end choices), andinformation; and (e) positive net benefits fromcooperating. But do well-established theoreticaland empirical conditions identified by scholarsexplain community-based natural resourcemanagement? In other words, do CBNRMprograms fulfill the conditions required forsuccessful collective action?

Because of the nature of the material availablefor this review, we were unable to preciselyidentify the conditions that underlie differentthriving CBNRM programs. Instead, we wereable to glean insights into the difficulties facedby CBNRM programs because of the absence ofsome of these conditions. Thus, we propose thatCBNRM programs are probably best startedwhere the governance conditions identifiedbelow are fulfilled. However, this is not alwayspolitically feasible. As a second best option,efforts to devolve responsibility to communitiesneed to carefully and systematically developmechanisms that can mimic the functionsenabled by these governance conditions.



Clear boundaries. Boundary identification is afirst step towards strengthening property rightsor control over natural resources. However, thewidespread presence of ambiguous andoverlapping rights over land, forests, and otherresources means that border classificationinevitably provokes serious conflicts betweenstakeholders. Once boundaries are established,management of resources becomes less difficult.

One promising strategy to resolve boundaryproblems in the beginning of the process is the“incentive” system used in Namibia (Jones1999). Community Conservancies in Namibiacannot register themselves and avail ofassociated rights unless they resolve theirboundary disputes. There are often seriousconflicts over land ownership and control. Forexample, in the case of the Torra, Khoadi/Hoas,and Dorr Nawas communities, conflicts arosewhen all three communities claimed certainareas of land, and some individuals in disputedareas registered under more than oneconservancy (Jones 1999b). These frictions canresult in long delays in conservancy creation.However, in this system, communities areforced to confront, negotiate, and resolve theirconflicts in the early stages of conservancycreation. The Namibian model may be a usefulone to emulate in other parts of the world. Theshort-term trade-off is delays in communityorganization, but this may result in longerlasting solutions.

Small stakeholder size. The broader literature onCPRs suggests that successful cooperativeaction is often characterized by small groupsize, which enables members to interact witheach other. While this concept is not fullyapplicable to wildlife management in Africabecause of dispersed populations, somesuccessful conservancy programs appear to belinked to low population densities.14 In such

cases, group size impacts community successthrough high per capita revenues rather thanthrough increased group interaction and trust.

In irrigation management, group size can be asignificant constraint. It is a lot easier for a smallnumber of large farmers to agree tomanagement rules than for a large number ofsmall farmers, as seen in the experience ofsmallholders in South Africa and elsewhere(Shah and others 2002). Irrigation systems madeup of many subsistence farmers are often thenorm in many developing countries. Thus, animportant policy question is whether user groupmanagement is feasible in such irrigationsystems. Given the current interest inparticipatory irrigation management, this is aquestion that needs careful consideration.

Resource scarcity and salience. Without someevidence of resource scarcity, there is littleincentive to conserve natural resources or toimprove the functioning of irrigation systems.For example, if irrigation water is abundantthere are few reasons for farmers to invest inorganizing themselves and bearing thetransaction costs of creating and sustaining usergroups. This is what happened with smallirrigation systems in Central and West Java,where river water is usually recycled bydownstream farmers. Attempts to organizefarmers did not succeed as the farmers did notperceive the marginal benefits of joining thegroup to be more than additional costs of timein meetings and labor contributions. Similarly,in forest communities, scarcity is likely to be animportant motivation for putting the time andresources into management activities(Bandyopadhyay and Shyamsundar 2004).

Salience of the resource is directly related toeconomic benefits from resource management.With irrigation and forestry, practically every

13



household has a stake in ensuring sustainableuse; irrigated water is vitally important to farmhouseholds, and firewood and non-timber forestproducts together serve energy, food, andmedicinal needs. Salience may be moreproblematic for wildlife management. Whilemost households are dependent on wildlife forfood, only a smaller number of householdsdirectly benefit from eco-tourism and associatedcommercial ventures that are the mainstay ofCBWM. Both scarcity and salience influence netbenefits from collective action. Thus, CBNRMprograms are sustainable only if they eitherbuild on existing perceptions of scarcity andsalience of the resource, or develop strategies toincrease the value of the resource to households.

Effective monitoring and enforcement. SuccessfulCBNRM programs are invariably associatedwith improved monitoring and enforcement. Inwildlife management, community programshave hired hundreds of local guards who havebetter information about local hunting, are heldresponsible for controlling poaching, and areoften answerable to local authorities and arethus accountable to users (Gibson 1999). Inirrigation management, Araral (2003) finds thatsuccessful irrigation associations conductinternal audits at least every two months andexternal audits regularly. They have a system ofchecks and balances to avoid nepotism betweenauditors and treasures. Further, the availabilityof indicators that farmers can clearly perceive isimportant. In China, volumetric pricing playsthis role and allows farmers to adapt quickly tochanges.

The most effective monitoring systems includepenalties for failure to cooperate. InBehroonguda’s community forestry program inIndia, patrolling duty is mandatory foreveryone in the village and failure to complyresults in a penalty and even loss of

membership. Monitoring of participation is alsoan effective way of building confidence in thecommunity. The following chapters providemany examples of effective monitoring andenforcement. This is a governance conditionthat CBNRM programs have mostlyinternalized.

Homogeneity in endowments, preferences, andinformation. There is a large literature on theimportance of homogeneity versusheterogeneity and inequality and how theyinfluence collective action (for example, seeBaland and Platteau 1996). Our review suggeststhat communities engaged in natural resourcemanagement are rarely homogenous entitiesthat harmoniously agree to undertake resourceconservation. Rather, as Leach and others (1999)state, “conflicting values and resourcepriorities—rather than shared beliefs andinterest—pervade social life.”

Differences in initial endowments, location, andpower relations within communities can lead tounequal costs from institutional change. Thiswas the case in the Maasai Kimana group ranch,where a combination of population pressure,immigration, and privatization of group rightsincreased the vulnerability of some socialgroups (Woodhouse 1997). The groupownership system resulted in householdswithout ranch membership. Women (andhouseholds) without access to irrigated landwere forced to travel increasingly long distancesto secure dry season grazing for their cattle.Similarly, joint forest management in India iscommonly perceived to exclude women fromdecisionmaking (Agarwal 2001). Decentralizedresource management can lead to winners andlosers. Heterogeneity clearly contributes toconflicts over a variety of decisions that needcollective agreement. For both these reasons, thestate will have to continue to play an important



role in CBNRM. Conflict resolution strategiesneed to be an intrinsic part of efforts to devolveresponsibility to communities.

Positive net benefits from cooperating. Over the lastfew years, user groups have mushroomedaround the world, which reinforces thepossibility of positive net benefits fromcooperating to manage natural resources. Inforestry and irrigation, in particular, there arealso many examples of endogenouslydeveloped self-governing institutions. Clearly,these groups exist because communities see anet benefit to working together. However, manyrecent user groups are subsidized by donoragencies or by the state, and there are very fewstudies that assess the “transaction costs” ofsuch collective action. Thus, the question ofwhether collective action in natural resourcemanagement is sustainable at its current scaleremains unresolved. Further, given donorinterest in decentralized approaches, and thehistorical failure of the state to manageresources well, CBNRM is likely to be scaled up.It is thus important to urgently assess options,including policies and household incentives, forincreasing net benefits.

Challenges Ahead

Community-based NRM is not a panacea formanaging natural resources in developingcountries. However, this is an institutionalchange that can significantly influence the livesof poor people and the resources on which theydepend. In order to make CBNRM moreeffective, donors and policymakers need torecognize and address five key challenges.

Getting household incentives right

Critical to the success of CBNRM programs isthat they are incentive-compatible at the

household level. CBNRM programs, for obviousreasons, focus on creating community usergroups, establishing community rules, andproviding community infrastructure. Whilethese are important mechanisms to put in place,they do not necessarily make CBNRM anattractive prospect for the individual household.

In wildlife management, for example, there areseveral reasons why community investmentsmay not be incentive-compatible at thehousehold level: 1) household benefits fromwildlife management are often small relative toagriculture;15 2) wildlife related ecotourismbenefits are often limited to a small number ofresidents, while a large number of householdsare affected by wildlife predation; 3) communitybenefits are publicly available—they are neithertied to conservation improvements, nor do theydeter illegal actions; and, 4) lack of tenurecreates limited “ownership” over programs andwildlife. A next step in the evolution of CBNRMhas to be to more carefully identify and invest inprivate incentives for wildlife management.

In irrigation, because of the close connectionbetween water and broad-based householddependence on agriculture, there is a greaterincentive to participate in collectivemanagement. Nonetheless, it appears thatsuccessful irrigation associations in thePhilippines provide a host of other incentives tomembers and officers such as credit,scholarships, health insurance, and transportallowances. These incentives, while un-relatedto irrigation, are specifically tied to householdwell-being.

In community forestry, given the dependence ofpoor households on forests, trade-offs betweenshort-term and long-term benefits can be amajor challenge. Our review of cases across theworld suggests that there are many examples of

15



strategies that increase short-term returns.Short-term returns occur mostly as a result ofimprovements in governance or because ofcareful compensatory mechanisms that decreasenegative impacts of forest closure. Attention tomechanisms that decrease household-level costsand increase household benefits will underscoresuccess in CBNRM.

Strengthening property rights

Rights extended by the state to communities tohelp manage natural resources cover a broadrange and can generally be categorized intorights of access, withdrawal, management,exclusion, and alienation of natural assets(Agrawal and Ostrom 2001). The literature onwildlife management, for example, suggeststhat communities invariably have access andwithdrawal rights, but there is considerablediversity in management rights. In almost allthe cases reviewed, the right of alienation or theright to buy and sell natural resources is held bythe state.

In many cases, CBNRM programs conferresponsibilities on communities withoutcommensurate rights (Meinzen-Dick and DiGregorio 2004). Clear rights enable communitiesto respond to local circumstances, creating theright conditions for success. This is the case withirrigation associations in the Philippines thathave the right to devise and change operationalrules within the ambit of their charters. Suchassociations are deemed to be among the moresuccessful associations in the country (Araral2003).

Issues of rights and control are closely tied tolegal recognition and land tenure security. Thecommunity forestry program in Nepal owesmuch of its success to the fact that local forestlegislation has created a legal basis for forest

user groups there. Tanzania and the Gambia aretwo other countries that have transferred legalownership of unreserved forest lands to localcommunities and provide ample testimony ofthe benefits of legal recognition.

Tenure is important for successful CPRinstitutions; however, there are also examplesworldwide that show that it is not always asufficient or even necessary condition forsuccess. Tenure needs to backed by certaintythat the state will enforce contracts. Thus,strengthening property rights at the local levelneeds to go hand-in-hand with strengtheningthe state’s ability to arbitrate and enforcecontracts.

Addressing heterogeneity and distributionalissues

Communities engaged in natural resourcemanagement are rarely homogenous entitiesthat harmoniously agree to undertake resourceconservation. As previously stated, collectiveaction programs have to deal with distributionalimpacts from institutional changes. Elite captureand gender discrimination are specific aspectsof this problem.

Koppen and others (2002), comparing thedifferential impacts of irrigation managementtransfer on poor and non-poor farmers, notethat it has often been assumed that the interestsof the poor sufficiently overlap with the generalinterest in the irrigation scheme, and that richand poor farmers alike have equal access tocanal water. However, the evidence fromAndhra Pradesh and Gujarat, India, suggeststhe strong possibility of elite capture of the IMTprocess, particularly in large-scale canalirrigation where land ownership and locationaladvantages are skewed in favor of biglandlords. Small farmers, who often participate



in repair and rehabilitation work, are oftenunaware of the existence of the water userassociation, while large farmers involvethemselves in committee work and makesdecisions. Thus, the challenge of makingparticipatory irrigation pro-poor still remains.Many such examples of unequal impacts ofcommunity management are found in forestsand wildlife management as well.

The presence of heterogeneity in communitiesand the possibility of elite capture do notnecessarily argue for scaling down the currententhusiasm for CBNRM. Rather, there may be acase for adopting a more flexible approach thatbuilds on existing conflicts and strategies. Thiswould require many steps, including a) earlyidentification of stakeholder conflicts and needs;b) linking benefits from institutional changedirectly to conflict resolution; and c)development of conflict resolution mechanisms,including a role for the state to mediateconflicts. Community “public” investments canalso counter the inherent inequalities that maybe reproduced in CBNRM projects.

Solving the scale problem

Management of fugitive resources, such aswildlife in Africa, suggests a need to focus on alarge ecological scale. However, the CPRliterature also suggests that social units that aresmall, in contact with each other, and have ahistorical connection are better suited to managecommons (Ostrom 1990). Thus, inevitably,decentralization can result in a mis-matchbetween what is required from an ecologicalperspective and what is known to work betterfrom a social management perspective. Themismatch between social and ecological scalecan mean that fugitive resources impose costson one community and benefits on another.Logan et al. (2002) discuss the case of the

Bulima CAMPFIRE district, where elephantforage can cause considerable crop damage inthe wet season. By the dry safari huntingseason, these elephants migrate to Tsholotshodistrict and benefit resident communities there.Not surprisingly, such inequitable impacts canlead to tension and overlapping claims.

In irrigation management, the issue of scaleplays out in the form of group size, and raisesthe question of whether user-groupmanagement is feasible in irrigation systemscharacterized by a large number of poorsubsistence farmers. In forestry too, it isimportant to consider issues of upstream anddownstream coordination, and what kinds ofinstitutions are required to ensure this within adecentralized context. Ostrom’s (1990) designprinciple of nested enterprises is beginning toemerge with the creation of federatedconservancies in Africa and elsewhere.16

In the irrigation sector, federations of irrigationassociations sometimes form a mid-level linkbetween the state and smaller user groups. Also,partly because of the crucial importance ofwater resources, and partially because ofpersonnel and infrastructure already invested inirrigation, the state plays a much moresignificant role as a meta institution that cancoordinate between smaller groups. Shouldintermediate institutions be crafted tocomplement decentralized user groups and playthe role of the “coordinator”? Or should“leaner” government bureaucracies play thisrole? This question needs further exploration ineach of the three sectors reviewed for this paper.

Providing adequate support

It is apparent from the case studies that anenabling policy and legal framework isimperative for more effective decentralized

17



forest management. Progress has been made,especially in Africa, where new forest laws havebeen passed in more than 30 countries since1990 (Wily 2002b). Common changes resultingfrom these laws include “changes in the characterof central forestry administrations, with wider civilsociety input in decision-making, sometimesrelocation of forestry departments into semi-autonomous institutions, and variant degrees ofdecentralization to local governments and policycommitment and new legal opportunity for forest-local populations to participate in forestmanagement” (Wily, 2002a). Tanzania and theGambia have led the way in creating legalprocesses support decentralization. Othercountries have a longer way to go.

Donors have played a critical role in facilitatingdecentralized natural resource management andbuilding the capacity of local organizations.Nongovernmental organizations are anotherimportant actor. In Namibia, for example, Jones(1999b) argues that the “‘light touch’ and high-quality facilitation” undertaken by IntegratedRural Development and Nature Conservation(IRDNC) has been intrinsic to the success ofcommunity conservancies. Further, the donorsinvolved in Namibia provided significant andsteady assistance, which was crucial forbuilding the program. Decentralized NRMneeds to be equipped with policy and legalchanges. Critical to this effort is consistentsupport and capacity building.

Enhancing sustainability

CBNRM programs are generally not self-sustained. Support for these programs usuallycome from the state, commercial activities, andfrom international donors. Communityconservation programs are dependent ontourism revenues, which vary depending oneconomic and political circumstances, while

development assistance is often a function ofagendas that are far beyond the control of localcommunities. Further, revenues earned onlypartly accrue to local communities because ofcost-sharing arrangements with the state.

Looking beyond the financial sustainability ofcurrent programs, we need to question whetherCBNRM promotes future investments in naturalresource stocks. In other words, does CBNRMprovide the right mix of information,empowerment, and financial incentives toenable households and communities to invest innatural resources? Wildlife investments mayonly become attractive to farm households if a)revenues from wildlife increase relative torevenues from agriculture; b) insurance schemescan reduce the costs of wildlife predation; and c)wildlife are seen more as private resources overwhich households have much more control.Until these conditions emerge, localcommunities may well treat wildlife as an assetthat is available to be run down and utilized forgreater “development.”

In irrigation, the sustainability issue surfaces intwo forms. First, without other measures suchas volumetric pricing, water-use efficiency mayremain unaffected even when irrigation systemsare under community management. Second,maintaining irrigation systems remains apervasive problem. The problem seems to beone of moral hazard and co-dependency—farmers need the irrigation agency for irrigationmaintenance, and they know the agency willbail them out because the agency depends, inpart, on irrigation fee collections and becausedilapidated systems are used as a justificationfor donor support. One solution is to make userassociations assume responsibilities for futurerehabilitation as a precondition for currentrehabilitation. However, given the financialfrailties of most small-scale water user



associations and their very limited capacity toprovide for capital expenditures, it is unlikelythat this issue will be resolved soon. A morepractical approach maybe to bundlemanagement transfer with a package of currentand future infrastructure improvements.

Our review suggests that decentralization mayreduce some operational and monitoring coststo the state, but long-term investments inrestoring capital assets (natural and physical)may remain a government responsibility. Thereare many examples of current benefit-sharingschemes between the state and resource-dependent communities. Contractualarrangements that outline and enforce long-term sharing of investment costs need furtherexploration.

Conclusions

Does devolution lead to better natural resourcemanagement? The answer appears to be yes.Increased local control motivates local interestin long-term investments, creates space for localdecision-making, and can increaseaccountability and management performance.Devolution also contributes to increasedinteraction between sector agencies andcommunities and helps decrease hostilitybetween government officials and localhouseholds. All of this enhances environmentaloutcomes. However, for these outcomes to befully realized, supra institutions that can play acoordinating role are required.

Does devolution contribute to povertyreduction? Devolution can empower localcommunities by a) providing access to publicgoods through development of localinfrastructure; b) increasing opportunities tolegally use and exercise control over natural

assets; and c) increasing private incomeopportunities. However, as this review shows,the ability of local communities to assert controlis circumscribed by various factors. In somecases, households also bear significant costs.Further, institutional changes have unequaleffects on different stakeholders. Increasinghousehold returns to CBNRM and resolvingproblems of unequal impacts of decentralizationare perhaps the most urgent practical challengesahead.

Is devolution good for the government? Sharingresponsibilities and control over naturalresources reduces some of the fiscal burden ongovernments. In the irrigation sector, this isclearly the case as national irrigation agenciespass on operations, maintenance, andmanagement responsibilities to water userassociations. Government staff reductions alsooccur. In forestry and wildlife management,financial costs decrease mainly because ofdecreased monitoring and enforcement costsand possible reduction in hostility betweencommunities and government officials.Decentralized resource management also affectsrevenue collection. In some cases, thegovernment may gain because of increases inresource stocks and commercial benefits. Inother cases, it will lose revenues as a result ofbenefit sharing. To the extent that localaccountability increases, corruption is also likelyto decline. Reductions in revenues, staff, andcorruption make implementation ofdevolutionary policies intrinsically challenging.Getting incentives right for bureaucracies is asimportant as getting them right for householdsand communities.

Community-based natural resourcemanagement has changed the relationshipbetween communities, the state, and natural

19



assets. In some parts of the world, this changehas been profound and communities enjoyextensive control over resources. In other partsof the world, the change has been more gradual,with the state and communities very slowlyrealizing impacts and implications. Devolution

results in the state handing over control ofcapital assets to local communities—a policywith both economic and political ramifications.This suggests a continued need for reform,experiments, and donor patience.


Success, Impacts, and EmergingChallenges in Community-Based Wildlife Management:What Does the Evidence Show?

Introduction

Community-oriented conservation gainedground in the 1980s and 90s in response toincreasing local and international resistance tostrict protected-area programs. These effortsalso emerged as a result of greater awareness ofthe difficulties of implementing state-runconservation. Engagement of communities innatural resource management has occurred inmany parts of the developing world. Suchchanges also permeate Western societies such asCanada, which has increasingly devolvedresponsibility over local resources to a numberof First Nations (Bradshaw 2003). In this paper,we take stock of community-oriented wildlifemanagement programs, and examine theirevolution, particularly in Africa. We seek tounderstand and identify new challenges theyface as they continue to meet community andconservation needs.

Decentralization in wildlife and biodiversitymanagement appears to be characterized bytwo overlapping phases. The first phase isexemplified by Integrated Conservation andDevelopment Programs (ICDPs), and the secondphase by what is currently known ascommunity-based natural resource management(CBNRM). Both programs create economicincentives for local communities to conservenatural resources and both involve theparticipation of local communities indecisionmaking. However, CBNRM programs

appear to further strengthen local communitiesin two ways. First, there is a greater focus onsustainable use of natural resources. This differsfrom what is often an effort in ICDPs to findsubstitutes for natural resources-basedactivities. Second, CBNRM seeks to providelocal communities with greater decisionmakingpower. Thus, while ICDPs may not involve anyform of decentralization, CBNRM programs aremeant to lead to decentralization ofmanagement responsibilities.

While there are many definitions of CBNRM, itis best understood by identifying its mainattributes. Kellert and others (2000, p. 706)suggest that CBNRM programs share fivedistinct characteristics: “a) a commitment toinvolve community members and local institutionsin the management and conservation of naturalresources; b) an interest in devolving authority fromcentral and/or state government to more local andoften indigenous institutions and peoples; b) a desireto link and reconcile the objectives of soci economicdevelopment and environmental conservation andprotection; c) a tendency to defend and legitimizelocal and/or indigenous resource and property rights;d) a belief in the desirability of including traditionalvalues and ecological knowledge in modern resourcemanagement.” These characteristics appear toform the basis of many CBNRM programsreviewed for this study.

Much of the literature on CBNRM is focused onwildlife in Africa. The continent of Africa has

Priya Shyamsundar

2



clearly been a fertile area for many successfuland not so successful attempts by states, donors,and international NGOs to establish CBNRMprograms. Projects such as CAMPFIRE inZimbabwe and ADMADE in Zambia are wellknown, and have motivated numerous othersimilar efforts (Newman and Webster 1993).17

Interestingly, a wide review of literatureresulted in little information on CBNRM inother parts of the world, even though there areample examples of ICDPs. This suggests thatthe existence of a rather profitable economicgood, wildlife hunting, may partially beresponsible for cooperative action to managenatural resources.

Community-based wildlife management(CBWM) programs have grown and evolvedover the last two decades. In this paper, we seekto understand the conditions that are mostconducive for successful programs. We reviewevidence of impacts and identify challenges thatare likely to stymie further development. Thecase-study literature, while broad and diverse,does allow us to draw some inferences aboutCBWM. Our review relies primarily on articlesin peer-reviewed journals over the last fiveyears. Where available, we have also reviewedreports from donor agencies and policy thinktanks.

We conclude that community-based wildlifemanagement, as a sub-project of community-based natural resource management, hasimproved perceptions among local communitiesabout wildlife and state representatives whomanage wildlife. However, it has yet to solveproblems of incentive compatibility. This islargely because opportunity costs to householdsof accepting conservation goals are high.Benefits from community conservation accrue inthe form of public and private goods. Public

benefits serve majority needs and are easier toensure; however, private benefits are required toguarantee individual household interest inwildlife and biodiversity. The case-studyliterature suggests that finding the balancebetween public and private benefits will becritical to the success of communityconservation.

While a wide array of governance mechanismsfall under CBWM, only some of these cedeactual control over wildlife to communities.Further, efforts to manage ecosystems in adecentralized fashion have yet to fullyrecognize “scale and coordination”-relatedissues and the need to link upwards fromdecentralized social units to broader geographicscales. The next stage of CBWM will have toinvolve institutional experiments that trulyempower communities to make managementdecisions, but within a framework of nestedinstitutions that serve the public-good nature ofwildlife and biodiversity.

Universal Features, Local Differences

In order to understand how community-basedwildlife management works, it is useful toexamine some regularities found acrossdifferent programs. The two most commonfeatures of community management of wildlifeare revenue generation and decentralization.Most CBNRM programs are built on thepossibility of being able to raise income fromconservation activities, and most attempt todevolve some management responsibilities fromthe state to a lower level organization. A thirdcommon feature is that community-orientedprograms rarely emerge without the support ofexternal agencies. Wildlife management inAfrica is littered with examples of programsthat share these features, yet differ from eachother in subtle but critical ways.

23

Success, Impacts, and Emerging Challenges in Community-Based Wildlife Management: What Does the Evidence Show?


Income generation and distribution

In nearly all the African cases reviewed, astrong external market and the potential forgenerating serious surpluses throughecotourism has been important in motivatingcommunity management of natural resources.The opportunity costs of conserving nature arehigh to local communities as well as thegovernment. Without adequate short-termreturns, there would be little motivation forbearing the transaction costs of communityorganization or the opportunity costs offorgoing agriculture. Thus, the existence ofmarket-oriented goods is central to CBNRMprograms. The classic example is a park, such asthe Mgahinga National Park in Uganda, wheregorilla tracking results in sizable profits. In1998, revenues from gorilla tracking wereestimated to be $190,000 per year (Adams andInfield 2003). The possibility of earning suchrevenues makes it feasible to considerintegrating conservation with economicdevelopment.

Wildlife is a revenue source for communitiesand local and national governments.Consequently, CBNRM programs almost alwaysentail benefit-sharing between the state andlocal communities. For example, in CAMPFIREin Zimbabwe, authorities retain 15 percent astaxes and up to 35 percent for management, anddistribute the remaining 50 percent to wards.This money is then reallocated betweendividends to households and administrativeexpenses (Bond 2001). Another example is theLuangwa Integrated Resource DevelopmentProject (LIRDP) in Zambia, where the early1990s saw approximately 40 percent of revenuesbeing targeted for community projects. Thispercentage was supposed to increase to 100percent as Game Management Area ownershipwas transferred to communities (Wainwright

and Wehrmeyer 1998). In general, the statefrequently retains a major share of the returnson commercial activities. This may be used as atax for budget support, or more likely foradministering other conservation areas. Thus,wildlife areas that enjoy commercial interestsoften cross-subsidize the state’s largerconservation efforts. This provides a strongmotivation for the state to protect its interests inand revenues from such areas.

CBNRM programs generally apportion localbenefits into household and community-levelprofits. Many CBNRM programs create localpublic goods in the form of schools, communityhalls, extension services to farmers, and roadand bridge improvements. Household benefitsinclude jobs, bush meat, tourism-relatedservices and some household enterprises.CAMPFIRE in Zimbabwe is one of the fewexamples where households receive a cashdividend for wildlife conservation. Thesedividends vary considerably across wildlifedistricts, with districts with low human andhigh animal populations being the mostprofitable (Bond 2001). Whatever the form ofhousehold benefits, community-level benefitsusually dominate household benefits.

An important consideration is that goodsmarketed through CBNRM frequently servemultiple needs. For instance, in many countrieswildlife is a source of crop predation and localbush meat, as well as tourism benefits. Thus,protection of wildlife can result in benefits andcosts. Local communities, with strong historicallinkages to nature, also have noninstrumentalvalues for nature and associated goods.Therefore, while CBNRM programs are closelytied to marketable economic goods, there arediverse motivations that lead communities toparticipate and benefit from these



arrangements. This makes program designinherently challenging.

Devolution of responsibilities

A basic tenet of CBNRM is devolution ofauthority to some local-level organization.Worldwide, there are three different ways inwhich devolution is undertaken: a) devolutionto local government agencies; b) the creation ofnew local conservation institutions; and c)greater authority to traditional leaders. Eachmodel comes with its own advantages anddisadvantages. Box 3 provides some examplesof these three types of institutional structures.

In Zimbabwe, CAMPFIRE conferred authorityto Rural District Councils (RDCs), which havethe authority to collect, retain, and distribute

revenues. While this resulted in somewhat rapidgrowth of CAMPFIRE programs, it appears tohave created little empowerment of localcommunities, since control remained in thehands of the state. Under LIRDP in Zambia,devolution of administration was donedifferently. Instead of vesting authority at theRDC or chiefdom level, a local leaderssubcommittee was set up with six local chiefs,one MP, and four ward chairmen of the rulingparty. This original design has since evolved.The new structure includes an elected executivecoordinating agency called the IntegratedResource Development Authority, areadevelopment committees, and village actiongroups that will manage 80 percent of allrevenues (Wainright and Wehrmeyer 1998).Thus, in the Zambian case, authority is morefully devolved to the local level.

Box 3Different Institutional Arrangements in CBNRM

Vesting control in the hands of local government agencies in Zimbabwe: In Zimbabwe, CAMPFIRE fol-lowed a pre-existing administrative government structure. A 1982 amendment to the Parks and Wild Life Actof 1975 designated Rural District Councils as appropriate authorities for wildlife in communal areas. Whilethe original intention of the policymakers in Zimbabwe was to devolve authority to local self-selected commu-nities, legal and administrative practicalities resulted in devolution of authority to the RDC level (Jones andMurphree 2001). RDCs have the authority to collect, retain, and distribute revenues obtained from CAMP-FIRE. Ward Development Councils (WADCOs) and Village Development Councils (VIDCOs) have some lim-ited authority under RDCs. In this system, authority and decisionmaking power is vested in the hands of thestate.

Authority to traditional chiefs in Zambia: Under the ADMADE program in Zambia, decentralization oc-curred at two levels. A Wildlife Management Authority, headed by the District governor, was created for eacharea with sufficient wildlife. Each such area was divided into chiefdoms. Traditional chiefs head a WildlifeManagement Sub-Authority, which is essentially a committee made up of teachers, unit leader, village head-man, ward chairmen, and district council representative (Gibson 1999). Under this structure, traditional chiefshave considerable power to allocate community funds and to hire local individuals as guards.

Crafting new institutions in Namibia: Under Namibian law, a group of communal area residents can createa communal conservancy if they are able to meet a set of conditions. They can establish a conservancy andgain exclusive rights to commercial tourism operations within the conservancy borders if they define a geo-graphical area, define membership, develop operating rules and plans for income distribution, elect a repre-sentative council, and become a legal entity (Jones 1999b). Under this system, communities are given the spaceand opportunity to come together and create their own conservancy. If they are able to resolve internal andexternal (boundary-related) conflicts, then they have considerable power and control over wildlife resourceuse.

25



An important caveat to devolution of controland authority is that CBNRN programsgenerally stop short of any actual transfer ofownership rights over either land or theresources that land supports. In Africanexamples such as CAMPFIRE, ADMADE, orLIRDP, there is no legal ownership enjoyedeither by the community or by householdswithin the community. Communities andhouseholds, however, do have a variety ofusufruct rights. A good example is Namibia.Once a community is able to define theboundaries of a conservancy and register itself(which can occur if there are no boundaryconflicts), it then has usufruct rights over thewildlife that is found on its land. Wildlife ismobile between conservancies. However, userrights belong to the conservancy on whichwildlife is found (Jones 1999b).

External agents

Community conservation programs that arepurely “community-driven” appear to besomewhat rare. In most of the examples wereviewed, we found that either an internationalor local NGO had an important role to play increating the community conservation program.Committed leadership from individuals (oftenassociated with an NGO or with the statebureaucracy) led to the initiation of programs.Frequently, successful local models ofcommunity conservation are picked up by thestate and replicated in other areas. In manycases, external donors influence the processwith resources and by putting pressure on thestate to devolve authority to local communities.This situation is true, for example, ofCAMPFIRE in Zimbabwe and ADMADEprograms in Zambia.

In Namibia, the pioneering work undertaken byan NGO—Integrated Rural Development and

Nature Conservation—on community gameguards and tourism projects influenced post-independence policies that created communityconservancies (Jones and Murphree 2001). Infact, Jones (1999b) argues that the “’light touch’and high-quality facilitation” undertaken byIRDNC in certain communities needs to bereplicated in other conservancies. Further, thedonors involved in Namibia providedsignificant and steady assistance, which wascrucial for building the program. USAID, forexample, contributed some $14 million toNamibia’s CBNRM efforts between 1992 and1999 (Jones 1999b). Similarly, in 1995, a $4million trust fund was established with GlobalEnvironment Facility resources to supportprojects in the Bwindi-Mgahinga Park parishesin Uganda (Adams and Infield 2003). Thisprogram is largely a product of goodinteractions among donors, park officials, andNGOs, particularly CARE International (Infieldand Adams, 1999).

What Contributes to Effective CommunityManagement?

Collective action is at the core of successfulCBNRM programs. It is therefore not surprisingthat some of the CBNRM programs in Africahave been informed and influenced by the vastliterature on collective action and naturalresource management (see for example, thework by Elinor Ostrom and colleagues at theWorkshop in Political Theory and PolicyAnalyses, Indiana University). But do the well-established theoretical and empirical conditionsidentified by scholars explain community-oriented wildlife management? The existingempirical evidence suggests that factors such aslow population-to-wildlife ratios, clearlydefined boundaries, improved monitoring, andan enabling policy environment can contributeto better results. Economic benefits and the



presence of supportive external agencies havealready been identified as important features ofCBNRM. Differences in property rights alsomatter.

Ostrom (1990) makes the case that successfulcooperative action is characterized by smallgroup size, which enables members to interactwith each other. While this concept does notfully hold for wildlife management in Africabecause of dispersed populations, some authorssuggest that the more successful conservancyprograms may be linked to low populationdensities (Viratnen 2003, Murombedzi 1999).Murombedzi (1999) argues that CAMPFIREbenefits are highest where human populationsare low and animal populations are high. Hediscusses the case of the Masoka ward in theZambezi valley, a poster-child for CAMPFIREbecause it has one of the highest per capitawildlife revenues. However, this ward also hasone of the highest rates of immigration, andward members encourage new settlers. Acommon perception among the communitieshere is that wards need to become less isolatedin order to attract development funds. Thus, agrowing concern is that CAMPFIRE areas withlow population densities may attract newresidents as a result of their economic growthand this would dampen wildlife benefits.

The definition and identification of the wildlifeconservancy borders in Africa is fraught withconflict between various stakeholder groups,head men, and so on. However, once theboundaries are established, management ofresources appears to become a somewhat lessdifficult task. Conservancies in Namibia have toclearly define boundaries in order to beregistered. This has often led to serious conflictsover land ownership and control. In the case ofthe Torra, Khoadi/Hoas, and Dorr Nawas

communities, conflicts arose when all threecommunities claimed certain areas of land andsome individuals in disputed areas registeredunder more than one conservancy (Jones 1999b).These frictions can sometimes becomeintractable and result in long delays inconservancy creation. However, in this system,communities have the “local” space andincentive to decide, negotiate, and resolve theirconflicts in the early stages of conservancycreation.

Two confounding factors that prevail even afterconservancy registration are the movement ofwildlife across boundaries and inability toexclude outsiders from moving livestock intoconservancies. In Namibia, by law, wildlifebelongs to the conservancies where they arefound (Jones 1999b). However, the lack of rightsto exclude outsiders (especially whentraditional user rights conflict with new rights)can create problems. Identification of stakes andconflict resolution is a continuing process. Thepractical implication is that communities needto be given considerable time to develop theirsystem of management. The larger policyquestion of security of tenure looms large in allboundary-related decisions.

The issue of property rights is perhaps the mostcontentious in community management ofnatural resources. CBNRM programs confer awhole spectrum of rights over natural resourcesto local communities. Thus, it is useful tounderstand that property rights come indifferent forms. Agrawal and Ostrom (2001)provide a useful framework for distinguishingbetween different types of rights. They suggestthat different bundles of rights can be identifiedin terms of rights of access, withdrawal,management, exclusion, and alienation ofnatural assets.

27



The case-study literature reviewed suggests thatcommunities invariably have access andwithdrawal rights to some degree. In cases suchas CAMPFIRE, most management decisions aremade by local government and benefits areconferred upon community members – thus,households have no management rights. InNamibian conservancies, on the other hand, agreat deal more management authority is heldby communities themselves. Yet, even in this“best practice” case, communities cannot reallyexclude others from using their lands forgrazing purposes (Jones 1999b). In almost all thecases reviewed, the right of alienation or theright to buy and sell natural resources is held bythe state. Arguably, the greater the number ofrights communities enjoy, the more likelycommunity management will succeed. While itis difficult to draw concrete conclusions in theabsence of careful empirical studies of theimpact of property rights on CBNRM, it can beargued that the lack of rights createdisincentives for long-term sustainablemanagement. In a later section, we discuss someadditional challenges associated with tenureand property rights.

Monitoring and enforcement is a problem thatfrequently comes up in discussions related toCBNRM. Programs that have been successful instrengthening conservation have largely beenable to do so because of improved monitoring.Hundreds of guards have been hired as part ofcommunity conservation programs. Programssuch as the Community Conservancies inNamibia, ADMADE in Zambia, or CAMPFIREin Zimbabwe hire guards from within thecommunity to control poaching. These guardstend to have better information about localharvesters of wildlife. They can also be heldaccountable for poaching activities, and areoften answerable to local committees. This

appears to have led to a decline in poaching,particularly of large mammals (Gibson 1999).

Improved monitoring is also expanding stateauthority into rural space. Communities arerarely owners of wildlife. Their ability to policethemselves thus enhances the conservation andrevenue goals of the state. The state gains in theshort term through benefit-sharing agreementsand if it can decrease its direct monitoringexpenses. In general, improved communitymonitoring increases the effectiveness of thestate’s ability to implement its conservationagenda.

Community management programs can besuccessful only if they are backed by either statepolicies or legislation. Thus, it is first useful toask why the state or bureaucrats within lineministries are motivated to provide thisenabling environment. Many authors argue thatCBNRM programs may be acceptable to thestate because they enhance the state’s ability totax rural communities for resources that werepreviously untaxed. As Kevin Hill (1996)suggests, “ CAMPFIRE .. not only is a wildlifeprogram; it is also a rural taxation program.” Whilenot true for all programs—for example,communal conservancies in Namibia haveexclusive rights over revenues from commercialtourism operations (Jones 1999b)—most oftenthe state does share some part of wildliferevenues. Donor pressure is another factor thathas helped motivate the state. In others cases,committed staff within state bureaucracies orNGOs have provided the leadership requiredfor CBNRM. Thus, changes have occurred as aresult of external pressure and new incentivesand information.

Policy changes leading to community-orientedprograms have been gradual (see Box 4). In



Zimbabwe and Namibia, successful and well-established commercial conservation efforts ledthe government to consider the possibility ofcommunal conservancies (Jones and Murphree2001). Zimbabwe gained independence in 1980,and a decade later, Namibia obtained itsindependence. This created a historicopportunity to make changes and equalizeopportunities. With independence, new lawswere written specifically to allow for theformation of communal conservancies (Jonesand Murphree 2001, Jones 1999b). In these

cases, the presence of successful private modelsand new policy space made CBNRM possible.

Barrow and others (2001) argue that policyevolution toward community conservation inEast African countries—such as Kenya,Tanzania and Uganda—has been relativelyslow. This is partly because of a lack of privatecommercial conservation models. Centristgovernments and a very strong conservationlobby may have also contributed to the slowermovement toward decentralization (Barrow and

Box 4Policy Reform — An Evolutionary Process

Because natural resource policies have historically treated communities as inimical to conservation, policyreforms are vital to CBNRM. Policies do not grow in a linear fashion toward a pre-determined goal. Rather, asour examples below show, they develop when new opportunities arise and are generally preceded by pilotprograms.

The Importance of Pilots

Policy rarely develops in the absence of concrete models to pave the way. In Tanzania, a pilot communityconservation program was launched in 1987 by the Tanzania National Parks (TANAPA) and the African Wild-life Foundation in northeastern Serengeti (Bergin 2001). By 1991, the program had been expanded to twonational parks. Simultaneously, a steering committee was established to consider expanding this program toall national parks. Between 1991 and 1995, TANAPA reformed itself and created a community conservationprogram that would affect all the parks under its authority. TANAPA was well supported in this effort byexternal agencies.

Indirect Policy Reform and Tenure Security

In designing communal conservancies, Namibian officials were fully aware of the problems created by the lackof tenure security over conservancies. Thus, even though they were unable to give communities full owner-ship of conservancies, they included conservancy- and ownership-related clauses in land policy legislation.As a result, a 1998 land policy approved at the Cabinet level made provision for “legally constituted bodiesand institutions to exercise joint ownership rights” over land (GRN 1998:3). This policy is seen as a first stepthat will potentially pave the way for ownership and security of tenure over land in conservancies (Jones1999b).

Policy Space as a Result of New Opportunities

In Namibia and Zimbabwe, independence enabled the creation of new CBNRM legislation. Pre-existing lawsallowed conservation and economic use of wildlife in private lands. Based on these models, decisionmakerswere able to create communal conservancies, which were viewed as a mechanism to equalize opportunities forusers of communal lands. However, this opportunity to change laws came during and after independencewhen a whole host of inequalities were addressed.

29



others 2001). Protected-area outreach programs,on the other hand, have been far moresuccessful. These programs, which generallyinvolve state authority within protected areasand community activities in buffer zones, maywell evolve into community conservationprograms.

Welfare and Resource Implications

As previously indicated, benefits in CBNRMprograms are generally of two types:community benefits and household benefits.Community benefits include improvedprotection from wildlife, social benefits such asschools and community halls, and in some casesdistribution of tourism revenues to households.Household benefits (which do not accrue to allhouseholds) include jobs as scouts or in tourismenterprises, privileged access to game meatfrom safari hunting or culling, and sale ofspecific goods and services to tourists or asother forms of development occur.

Some broad generalizations can be made aboutwho benefits most or least in Africa. Womenseem to have the least power in localorganizations and are often the most negativelyaffected by restrictions imposed on naturalresource use such as fuel-wood and foddercollection. Traditional leaders and older mentend to have decision-making authority, whilenew jobs created as a result of tourism tend tobe oriented toward the young (Virtanen 2003).Further, the individual household benefits tendto be concentrated among a few households(Emerton 2001, Bandyopadhyay and others2004). Table 2 below provides some examples ofbenefits from CBNRM to households,communities, and the state. As the table shows,CBNRM almost invariably results in benefit-sharing between local communities and thestate.

Benefits through village-level projects

Community benefits are a trademark ofconservation programs. In almost all theprograms scrutinized, some form ofinfrastructural benefits accrued as a result ofrevenue-sharing between the state and localcommunities. Some of the community benefitsare from direct international assistance, whilethe rest comes from program earnings. Gibson(1999) states that ADMADE in Zambia resultedin approximately 60 projects between 1989 and1992, including twenty-three teachers’ houses,nine maize grinding mills, and seven ruralhealth centers. Under Zambia’s LIRDP program,40 percent of revenues were targeted tocommunity projects, and this was to beincreased to 100 percent as ownership changesoccurred (Wainwright and Wehrmeyer 1998). Ingeneral, the case-study literature suggests thatcommunity benefits are more likely to occurrelative to household-level benefits.

In Namibia, there is clear evidence thatcommunities and households benefit fromconservation programs. Even as early as 1993,Kunene region communities gained fromhunting of surplus game, which resulted inmeat worth $25,000 and skins worth $3,040.Between 1993 and 1995, the Lianshulu Lodge inMudumbu National Park collected a levy of$1.25 per tourist per night, which wasdistributed to five communities. Similarly, theTorra conservancy received approximately$40,000 between 1996 and 1998 from a profit-sharing agreement (Jones 1999b).18 In a recentpaper, Bandyopadhyay and others (2004)evaluate the effect of community conservanciesin the Kunene and Caprive regions of Nambia(see Box 5). The study suggests that community-wide benefits may be the reason whyconservancies have an overall positive impacton the average household’s welfare, while



Table 2. Examples of Benefits from CBNRM to Households, Communities, and the State

Program Household benefits Community

benefits State benefits and revenue sharing

CAMPFIRE, Zimbabwe (Bond 01)

Wildlife dividends. Median dividend was approximately $4.50 per household in real terms in 1996. Some 100,000 beneficiary households in 1995. Jobs as guards and in tourism.

Numerous community projects

CAMPFIRE earned approximately $9.3 million between 1989 and 1996. 50 percent of revenues were meant for communities—but this is divided between administrative costs and community dividend.

ADMADE, Zambia (Gibson 1999)

Employment Wildlife scouts Cheap meat

60 projects (mostly infrastructure) between 1989-92

35 percent for communities and 65 percent to Parks and Wildlife Service. Gibson calculates that only 2 percent annual gross sport-hunting revenues reached communities.

LIRDP, Zamiba (Gibson and Marks 1995) (Gibson 1999) (Wainwright and Wehrmeyer 1998)

Employment, Extension and credit services to farmers Cheap meat

Road repairs, infrastructure projects.

40 percent of revenues to communities in early 1990s; to be increased to 100 percent.

Mgahinga Gorilla National Park, Uganda (Infield and Adams 1999 Archbald and Naughton-Treves 2001)

Employment $10,000 until 2001 used for schools in three parishes. Other livelihood activities and small infrastructure projects.

Park earned approximately $250,000 in 1998–99. New policy of 20 percent of park entry fees for community projects after 1996, not implemented until 2001.

Communal Conservancies, Namibia (Jones 1999a and 1999b)

Employment (between 1996-1998, wages to Torra community members from a tourism lodge amounted to $70,000) Bed Levies (eg. $8,000 distributed to 370 households in Caprivi in 1996) Community Guards Trophy Hunting ($30,000 to a San community from trophy hunting in 1998 —not clear what portion was redistributed to households)

Torra conservancy earned nearly $35,000 by August 1998 from an 18- month joint venture with the private sector.@

Community has exclusive rights to commercial tourism activities and revenues within conservancy boundaries.

Bwindi Impenetrable Forests (Archbald and Naughton-Treves 2001)

Donor programs included sustainable use of NTFPs and problem-animal control

$70,000 until 2001 for schools, clinics, roads in 19 parishes.

Park earned nearly $700,000 in 1998–99. New policy of 20 percent of park entry fees for community projects after 1996, not implemented until 2001.

31



conservancy participants themselves do notsignificantly gain.

Community programs create public goods, butdo not specifically create incentives for thosewho conserve, nor do they punish householdsthat engage in poaching and other illegalactivities. Thus, an emerging concern is thatcommunities do not necessarily link thesetangible benefits to the decision to conservenature. Another fear is whether there istransparent accountability of what happens tocommunity resources. Gibson (1999), forexample, estimates that only about 2 percent ofADMADE sport–hunting gross revenuesreached communities and that manycommunity projects were incomplete because oflack of resources. In Uganda, a new 1996 policyassigned 20 percent of park entry fees forcommunity projects, but this policy had yet tobe implemented as of 2001. Thus, whilecommunity benefits do occur, the on-the-ground

reality probably falls short of actualexpectations and intent. As in so manydevelopment programs, the money that actuallygets to the communities may be a lot less thanthe amount actually expected or realized inaggregate.

Household benefits

Household benefits include direct payments,guide and scouting jobs, employment in lodgesand tour agencies, possibilities of sellinghandicrafts and tourism-related services, andavailability of meat from culling operations.Under ADMADE and LIRDP in Zambia, forexample, hundreds of scout jobs were created.However, it appears many of these jobs wereacquired by friends of the chief or wardchairmen and projects grouped around thechief’s residence (Gibson 1999).Nonetheless,these programs have been a source of localeconomic development.

Box 5Evaluating Community Conservancies in Namibia—Community Investments Improve Welfare

In a recent paper, Bandyopadhyay and others (2004) evaluate the effect of community conservancies in theKunene and Caprive regions of Nambia. The study uses data from over 1,000 households to assess whetherconservancies are improving the welfare of households.

Source: Bandyopadhyay and others 2004.

Interestingly, the study finds that while the average household is better off as a result of conservancies, house-holds that report that they are “participants” are not that much better off. The analysis suggests that the wel-fare benefits from conservancy development may be somewhat evenly distributed between participant andnon-participant households. While cash benefits are limited, participants and non-participants also enjoy oth-er non-cash benefits such as meat and community infrastructure. These community-wide benefits may be thereason why the study finds that conservancies have a positive impact on the average household’s welfare, butconservancy participants themselves do not significantly gain.

Knowledge, Benefits and Costs Associated with Conservancies in Namibia Percent Households with some knowledge about conservancy plans and constitution 26 Households that report they are conservancy participants 34 Households with conservancy associated cash income 12 Households who perceived meat distribution to be a benefit from conservancies 21 Households who reported crop damage from wildlife 50



Wages are an important benefit fromconservation programs. The Torra conservancycommunity members in Namibia earned$70,000 in wages from 1996 to1998 through anagreement with a photographic safari companyto develop a tourism lodge. Residents receivedanother $6500 from providing other services tothe lodge. In Namibia, bed-levies are anothersource of household income. For example, in1996, the Etendeka Mountain Lodge in theKunene region distributed approximately $8,000to 370 households in communities neighboringthe lodge (Jones 1999b).

CAMPFIRE is one of the few programs thatoffers direct payments for wildlife conservation.In 1995, approximately 102,000 householdsgained wildlife dividend income. Bond (2001)estimates the median household wildlifedividend income earned to be $4.50 in realterms in 1996. This, however, amounted to lessthan 10 percent of the average gross incomeearned by households from agriculture; that is,it did not change the relative returns of wildlifemanagement in comparison to agriculture.Households also earn other income fromwildlife management, including salaries fromassociated jobs.

Legal access to meat is another very importantbenefit linked to conservation programs. Forexample, in a 1996 survey of participants in theSelous Conservation Program (SCP) inTanzania, over 23 percent of participantsperceived availability of game meat as aprogram benefit and said that nutrition hadimproved because of the occasional availabilityof game meat. This is an important finding,given that 50 percent of the respondents saidthat the program brought no benefit at all(Songorwa 1999). Usually meat is either madeavailable through hunting quotas or sold after

culling operations at a low cost. Both ADMADEand LIRDP make available less expensive meatfrom culling operations. However, it appearsthat this is accompanied by a litany ofcomplaints from households, either stating thatthe meat price is higher than they can afford orthat the quantity available is low (Gibson 1999,Songorwa 1999). Because access to meat is animportant historical and cultural benefit fromliving close to wildlife, it is often perceived as aright that needs to be recognized and respectedby CBNRM programs.

Costs related to land loss and wildlife predation

The Gwampa Valley case in Zimbabwe is aclassic example of resistance to CBNRMprograms as a result of the high opportunitycosts of land. In the Gwampa Valley, the colonialgovernment had in the past established a tsetseclearance programs that involved wildlifeslaughter. Because of this and other historicalfactors, communities associated “game with theprimitive and backward, .. and cattle andagricultural production were.. preferred to wildlifemanagement” (Alexander and McGreggor 2000).When CAMPFIRE was initiated in the Nkyayiand Lupane Districts of the valley, the localcouncil planned to stock wildlife to attractsafaris. This, however, meant dislocation of localcommunities and land, which was steadilyopposed. Revenue-starved councils saw theproject as a means for economic development.On the other hand, communities viewedCAMPFIRE as a program against modernizationand as promoting land losses. As a result, as of1995, CAMPFIRE in the Gwampa Valley was ina state of deadlock.

Similarly, Infield and Adams (1999) find thatcommunities around the Mgahinga Gorilla

33



National Park associate significant opportunitycosts with the park. Adams and Infield (2003)identify thirteen subsistence products (fromfarmland to fruits and seeds) that householdsmay have sacrificed for conservation. The mostimportant opportunity cost is a result ofapproximately 1000 hectares of productiveagricultural land that was incorporated into thePark. Adams and Infield (2003) estimate that theresulting yield losses cost communities some$850,000 per year. If these estimates areaccurate, this is a huge sum of money. It is notsurprising that local communities view the parkas a source of either “past or future”agricultural operations.

Wildlife predation-related costs are anothermajor problem for communities that live aroundconservation areas. To the extent thatconservation occurs, predation may alsoincrease. Sonogorwa (1999), studying villages inthe SCP in Tanzania, found that villagers andprogram personnel perceived an increase inwildlife predation. He describes howconservation programs try to make availableblank and flare cartridges to communities tocounter predation, but animals, particularlyelephants, learn and adapt to such measures.

In a CAMPFIRE ward in Binga District, villagersare estimated to lose 8.25 percent of their maize,sorghum, and millet annually, mainly fromhippo and elephant predation (Wunder 1997, inLogan and others 2002). Based partly on thesenumbers, Logan and others (2002) estimate thatthe ratio of household benefits to predation costin Binga is about 6:16 (Logan et al. 2002). Thepossibility of such large predation costs isreinforced by Mayaka (2002), who discussescommunities and hunting zones in the BenoueNational Park complex in Cameroon. Based on asurvey of 239 households, he estimates that

benefits, which are in the form of land royalty,only amount to 0.2 to 0.6 percent of predationplus guarding costs. Further, as of the time ofthe survey, the households had yet to receiveany royalties. In all the case studies reviewedthat estimated benefits and predation costs, thecosts appear to be far higher than benefits.

Even without conservation efforts, there wouldlikely be predation costs simply because ofgeographic proximity to wildlife. However,investments in conservation are perceived byhouseholds to increase these costs. The lack of acompensation policy or insurance againstwildlife predation invariably contributes tocommunity resistance to wildlife conservationschemes. Further, the loss of meat protein fromhunting restrictions further increases the costsof wildlife management programs. Logan andothers (2002) suggest that many CAMPFIREhouseholds forfeit some 25 percent of animalprotein due to hunting restrictions.

Community management of natural resourcesinvolves transaction costs in addition toopportunity costs. We understand little abouttransactions costs, even though they are likely tocontribute significantly to the success ofCBNRM. One study that does estimatetransaction costs is Mburu and others (2003)(see Box 6).

Reduced poaching and better resourceconservation

There is some evidence that communitymanagement programs are meeting theirconservation goals. Indicators such as numberof animals and signs of poaching are improving,either because of improved monitoring andenforcement or because of changed preferences.For instance, animal census results from 1991



and 1994 show that the Selous ConservationProgram in Tanzania may have increasedanimal populations (Songorwo 1999). Suchevidence is also found in the Kunene region ofNamibia (see Table 3), where the communityguard program has resulted in the recovery offlagship species such as thedesert elephant and the blackrhino (Jones 1999b).

Many of the papers reviewedsuggest that poaching has beenreduced as a result of CBNRMprograms. The LIRDP programin Zamibia has reportedreduced commercial poaching

Box 6Transaction Costs Associated with Community Management in Kenya

Community management of natural resources involves production and transaction costs. We understand littleabout transactions costs associated with community management, even though they are likely to contributesignificantly to the success of CBNRM. Transaction costs can be classified into search and information costs;bargaining and decision or contracting costs; and monitoring, enforcement, and compliance costs (Mburu andothers 2003).

Very few studies estimate transaction costs and compare them with “production costs,” such as opportunitycosts of land, costs of infrastructure, or costs associated with animal predation. One study that does is byMburu and others (2003), who estimate transaction costs associated with community management of theKimana Community Wildlife Sanctuary and the Golini-Mwaluganje Community Wildlife Sanctuary in Kenya.

Source: Mburu and others (2003).

Ex-ante costs are total costs per participating household, while the post-investment costs are costs per partic-ipating household per year. Thus, at least from this one study, we can gather that transaction costs are impor-tant but not as significant as other production costs. The study finds that transaction costs can be a significantpart of total costs during the stage when the community activity is being planned and investments are beingmade. Once the initial organization of community activities is undertaken, transaction costs as a proportion oftotal costs declines. Further, this study shows that a larger proportion of transaction costs are borne by locallandowners or stakeholders relative to production costs, which are often picked up by NGOs or the state.

Comparison of Production and Transaction Costs in two Community Projects in Kenya

Ex-ante stage costs

($ / household) Ex-post stage costs

($ / household / year) Production Transaction Kimana 19 24 44 20 GM 454 184 153 20

Table 3. Increase in Wildlife Numbers in Kunene Region,Namibia, 1982–97

Source: Durbin and others 1997. Based on ground and air sightings.

1982 1986 1990 1992 1995 1997 Springbok 650 2000 7500 Oryx 400 800 1800 Mountain Zebra 450 900 2200 Elephant 250 384 415 Black Rhino 65 93 114 130 Giraffe 220 300

through its game patrols and scout programs(IUCN 1989), and park records in Mgahingashow reduced illegal entry and resource use. Inthe Mgahinga National Park, data from the parkagency suggest that there was a decline in thenumber of people arrested, snares removed, and

35



encroachment for goat and cattle feedingbetween 1991 and 1997 (Infield and Adams1999). Infield and Adams note that these resultscould be as easily due to tightening of lawenforcement as due to actual communityconservation activities. Most communityconservation programs have an element of lawenforcement, and this aspect of the programgenerally seems to work.

However, the results are not alwaysunambiguous. For example, Gibson (1999)agrees that the LIRDP and ADMADE programsin Zambia have led to a decline in big gamepoaching, particularly by outsiders.Nevertheless, he argues that small-gamehunting with the use of traditional tools maywell have increased, both in Zambia and underother similar programs such as Zimbabwe’sCAMPFIRE program. Gibson argues thatcommunity management changes the incentivesthe rural residents face, but does not stemhunting. Small-game hunting is generallythrough snares instead of firearms, and localsare less likely to be caught. Hence, areas withmore community scouts are likely to see anincrease in snares and small-game hunting, anda decline in large-game hunting.

Improved perceptions, stronger rights andreduced conflicts

To the extent that devolution results incommunities having more access to localgovernment officials and brings some decisionsrelated to natural resources to the local level,community conservation presents localcommunities with stronger rights. To the extentthat these decentralized structures re-create“traditional hierarchies,” decentralizationresults in some members having more or less“voice.”

People’s values toward wildlife are constructedfrom historical experiences and are not the sameeverywhere. In the Gwampa Valley ofZimbabwe, views about wildlife were formedby colonial policies to eradicate tsetse by killingthousands of wild animals (Alexander andMcGregor 2000). Shaped by a culture of animalhunting, forced migration, and political strugglefor land, communities in the valley viewedCAMPFIRE with great resentment and had littledesire to use land for wildlife rather thanagriculture and livestock rearing. Thus,depending on historical circumstances,management of land for wildlife may be moreor less welcomed. This case suggests thatdevolution to local governments does notautomatically result in stronger rights overnatural resources. Arguably, communities havea better chance of expressing their objectionswhen schemes are decentralized.

To counter the example of Gwampa Valley, thereis evidence from the Kunene region in Namibiathat conservancies have empoweredcommunities. Ashley (1998) identifies a numberof non-financial benefits of CBNRM, includingdevelopment of new skills, pride and a sense ofcontrol, and experience and confidence indealing with outsiders. Ashley concludes thatcommunity management is resulting indevelopment and empowerment that go beyondinitial objectives. There is also evidence fromnational parks in western Uganda (Mgahinga,Bwindi, and Kibale) that community relationswith park authorities have improved as a resultof community outreach efforts (Infield andAdams 1999, Archbald and Naughton-Treves2001). Some of these changes are a result ofengaging with communities, even in a fairlycentralized system of park management.

An often-made contention is that that ruralAfricans have great intrinsic interest in



maintaining their traditional connections towildlife (Jones 1999a). Thus, CBNRM projects,which have allowed rural communities toregain control over their natural assets, havealso resulted in “social re-empowerment”(Owen-Smith and Jacobsohn 1991). Proponentsof CBNRM programs suggest that this kind ofempowerment leads to reduced conflict andimproves management of natural areas. TheCommunity Conservation Service (CCS) inTanzania is a case in point. A program thatgradually evolved and grew to cover all theparks under the Tanzania National Parks’authority, CCS has possibly empowered bothpark managers and local communities to engagein dialogue about park-people relationships. InTanzania, CCS profited from the fact that manycommunities viewed national parks as a sourceof development and income. There was thuslittle resistance to community conservation onceTANAPA was ready to come to the table (Bergin2001).

New Generation Concerns Resulting fromDecentralization

CBNRM, as a tool for sustainable development,is here to stay. However, it faces manychallenges that need resolution for effectivelymeeting its goals of conservation and economicdevelopment. Thus, in this section, we discusssome emerging concerns that need to beaddressed.

Communities versus households

In most of the examples we have reviewed,property rights are generally vested in someform of communal organization. Benefits alsoaccrue, often at the community level in the formof social infrastructure. These benefits are notinsignificant—for example, the Kenya Wildlife

Service distributed some $1. 25 million forcommunity activities in protected-area bufferzone areas between 1991 and 1995 (Barrow andothers 1996, in Emerton 2001). However, manyauthors argue that the focus on communitiesrelative to households is a problem forconservation (Emerton 2001, Gibson 1999).

The benefits-based approach to conservation asit currently exists is having a positive effect onhousehold preferences and attitudes towardnatural assets. However, there is reason tobelieve that this may not be sufficient to meetconservation goals. Individual householdbenefits tend to be small relative to totalhousehold income. Bond (2001) shows, forexample, that per-household dividends from theCAMPFIRE program have been decreasingsince 1989, and, amount to only about 10percent of income earned from agriculture.Further, except for the case of wildlifedividends, revenues are not broad-based andaccrue mostly to a small number of households.Recent data from seven conservancies inNamibia show that only 12 percent ofhouseholds obtained revenues that theyassociated with conservancies (Bandyopadhyayand others 2004). Further, each householdincurs significant costs. These costs includeopportunity costs from loss of access to land orforest resources, costs of animal predation, andtransaction costs associated with communitymanagement.

The potential negative returns to conservationat the household level lead to—and areprobably reinforced by—the community-levelinvestments. As previously stated, many of thecommunity services that are offered contributeto household investments in agriculture ratherthan wildlife conservation. Moreover,

37



community infrastructure projects do notchange incentives. This is partly becauseresidents do not necessarily associate thecreation of a local school with their efforts toconserve wildlife (see Box 7). Communityprojects also mimic public goods—benefitsaccrue equally to villagers who illegally huntand to villagers who truly try to conservewildlife.

Heterogeneity within and between communities

Communities engaged in natural resourcemanagement are rarely homogenous entitiesthat harmoniously agree to undertake resourceconservation. Rather, as Leach and others (1999)state, “conflicting values and resource priorities—rather than shared beliefs and interest—pervadesocial life.” Communities are characterized byheterogeneity of endowments and interests.Further, whether or not a group of householdscan be characterized as a community dependson the scale of analyses (Leach and others 1999).Even if there is some agreement about whatconstitutes a community, there are severalspecific problems that can arise.

Differences in initial endowments and powerrelations within communities can lead tounequal costs as a result of institutional change.This was the case in the Maasai Kimana groupranch, where a combination of populationpressure, immigration, and privatization ofgroup rights increased the vulnerability of somesocial groups. The group ownership systemcreated in 1971 resulted in households withoutranch membership. Women and householdswithout access to irrigated land had to travelincreasingly long distances to secure dry-seasongrazing for their cattle. Thus, Woodhouse (1997)argues that new markets and devolution ofpower to local authorities can allow the “practiceof customary hierarchy to be translated intodifferential advantage in the market,” leavingcertain stakeholders negatively affected.

Differing stakeholder needs can contribute toconflict and impact conservation efforts. In SCPareas in Tanzania, inter-community conflictsoccur because of the need to clarify demarcationof village boundaries and new rules. This hasmeant that farmers in one community cannotopen up lands for agriculture in other villageswithout seeking permission. Anti-poaching

Box 7Is Community Conservation Incentive Compatible?

There are various reasons why CBNRM may not create the right incentives for conservation. A few are listedbelow:

• Household benefits are small relative to agriculture.

• Household benefits are limited to a small number of residents, while a large number of households dependon wildlife.

• Community benefits are publicly available—they are neither tied to conservation improvements, nor dothey punish detrimental activities.

• Lack of tenure creates limited “ownership” over programs and wildlife.

• Scout programs provide incentives for improved monitoring of big-game hunting, but possibly not for small-game hunting.

• Agricultural crop production may be subsidized, an issue often beyond the control of communities.



units in participating communities that stopoutsiders from hunting wildlife in their areashave also added to inter-community conflicts(Songorwa 1999). In the Kunene conservancy inNamibia, there are two distinct groups—theyoung and old—who have differing needs andperspectives on the use of wildlife incomes.There are also examples of wildlife proponentsversus livestock-ranching interests (Jones1999b). Differences among these stakeholdershave contributed to delays in the creation ofconservancies.

Community investments can counter theinherent inequalities that can be reproduced inCBNRM projects. Batterby (1998) discusses howdifferential benefits accruing through the gestoindes terroirs villageois (GTV) program in thevillage of Toega in Burkina Faso are minimizedby investments in “social capital” andcommunity projects such as tree plantations,compost pits, and improvements in watersupply. Thus, the public investments thatCBNRM programs make are equalizing in waysin which private benefits are not. Theinvolvement of an NGO also makes a difference,as evidenced by the increasingly strong role ofwomen in Kunene region conservancies inNamibia in committee meetings and elections.Jones (1999) attributes the successful transitionof women from little involvement to a strongpresence to facilitation by IRDNC, an NGO, andto benefits perceived by women in participation.

Leach and others (1999) suggest that takinggroup heterogeneity into account evokes afundamentally different type of CBNRM. Theyargue for a more flexible approach that focuseson and takes pointers from ongoing strugglesand strategies. This approach recognizesinstitutions as “rules in use” and does not viewthe creation of a formal community-level

organization as a broad solution to resourcedegradation.

Competition between institutions

Community management of natural resourcesgenerally involves either the creation of a newinstitution or the assignment of new powers toan existing institution. In either case,competition among institutions is inevitable.Further, any changes made affect both formaland informal institutions (which are more likelyto be overlooked). To the extent that changes inrules affect power relationships, this willinvariably lead to conflict.

When community conservancies wereestablished in Namibia, it was decided thatregional councils would deliberately be leftwithout major control over wild resources. Thisdecision had significant implications when theprograms were implemented—regionalcouncilors and governors in Kunene andCaprivi provinces both refused to endorse theprogram initially, and, there are many instancesof bureaucrats slowing the process ofconservancy registration (Jones and Murphree2001, Jones 1999b). However, a variety ofstrategies seem to have contributed to greatercooperation by regional councilors. Detaileddiscussions with councilors; technical assistancefrom Ministry of Environment and Tourismofficials at council meetings when conservancyapplications are considered; and invitations tocouncilors to regional conferences to increaseawareness about Namibia’s successes relative toother countries appear to have helped (Jones1999b).

Competition among traditional leaders andelected conservancy committees has also beenproblematic in Namibia (Jones 1999b). There are

39



several examples of headmen attempting todivert authority from the elected committees. Afrequent solution has been to engage theheadmen in the committees and to providethem with a role to play. While this strategy hasclear benefits, it can also result in the headmenputting undue influence on the conservancyprocess. As Virtanen (2003) argues, traditionalauthority may be in-compatible with moderndemocratic notions of equality. Traditionalinstitutions are likely to treat outsiders orimmigrants unequally, and the possibility offavoritism and lack of accountability cannot beignored (Gibson and Marks 1995).

A lesson that emerges from the review is that anarrangement that involves all major localstakeholders, however fractious initially, islikely to spring returns in the long run. Localgovernment is perhaps the most importantstakeholder that needs to be engaged incommunity conservation efforts. Often, localgovernment is viewed as a source of corruptionand rent extraction and is distrusted; butwithout local government support, long-termsustainability is unlikely.

Tenure over land and resources

Many CBNRM systems vest usufruct rights overwildlife with communities, while ownershiprights over wildlife and land remain with thestate. There are many reasons why this is done.There are also reasons to be concerned aboutwhether the lack of ownership will weakencommunity conservation efforts in the long run.

In Namibia, conservancies confer resourcerights and not land rights (Jones, 1999a). Thispractical decision was made to keep the processmoving. Further, the tradition of trying to

protect wildlife communally started in the early1980s in the Kunene region of Namibia, when itbecame obvious that wildlife numbers weredecreasing. Communities had little securityover rights (this was prior to independence), yetthey started a community guard program inKunene that was later expanded to the Capriviregion. Thus, as Jones (1999a) argues “peopledefined ‘ownership’ in terms of connection to wildlifebased on cultural values rather than property rightsderived from the state.”

Many authors (including Jones 1999a) speculatethat insecure tenure over land will result inmigration of human and livestock populationsinto successful conservancies and negativelyaffect the long-term sustainability of theseprograms. In Zimbabwe, as Murombedzi (1999)has argued, lack of tenure security in wildlife-rich areas may be responsible for attracting in-migrants (with potential negative effects onwildlife). It is, of course, not clear if security oftenure would change the migration pattern aslong as returns to agriculture are higher thanreturns to wildlife management.

Mismatch between ecological and social scale ofmanagement

Management of fugitive resources, such aswildlife in Africa, suggests a need to focus on alarge ecological scale. However, the CPRliterature also suggests that social units that aresmall, in contact with each other, and have ahistorical connection are better suited to managecommons (Ostrom 1990). Thus, inevitably,decentralization can result in a mismatchbetween what is required from an ecologicalperspective and what is known to work betterfrom a social management perspective.



Batterby (1998) discusses how the GTVapproach in West Africa, which transfers controlover local land use to community groups, ismost likely to succeed when disputes overoverlapping terrain does not exist; that is, when“action” space matches a community’s“geographic” space. However, while this may bethe case for agricultural areas, it is rarely seen inpastoral and conservation areas. In the case ofwildlife, one solution to resolve the mismatchbetween geographic and action space is toassign property rights over wildlife to the spacewhere they are found (as done in Namibia).Thus, fugitive resources can belong to differentcommunities during different periods of time.

Providing rights to communities over wildlifemay reduce conflict, but does not fully solve thescale problem. The mismatch between socialand ecological scale can mean that fugitiveresources impose costs on one community andbenefits on another. Logan and others (2002)discuss the case of the Bulima CAMPFIREdistrict, where elephant forage can causeconsiderable crop damage in the wet season. Bythe dry safari hunting season, these elephantsmigrate to Tsholotsho district and benefitresident communities there. Not surprisingly,such inequitable elephant activities can lead totension and overlapping claims. Asserting itsrights under CAMPFIRE rules, the TsholotoshoRDC in 1991 had ended a quota arrangementbetween the two districts (Logan and others2002).

Another confounding problem occurs becauselong-run fluctuations in wildlife resources arehard to predict, making management of wildlifestocks in decentralized settings ratherchallenging (Naughton-Treves and Sanderson1995). Even if the state sets hunting “quotas,”the fugitive nature of the wildlife and

uncertainty about optimal harvest rates can leadto quota revisions. This can contribute toconflicts between different communities whoshare this property. The scale issue is furtherexacerbated by the fact that it is entirelypossible to have too many wildlife at one scaleand too few at another scale. This problem is notuncommon and has led to the collapse of manyfisheries (Naughton-Treves and Sanderson1995).

A solution presented by Ostrom (1990) for thisproblem is the creation of nested institutions.This is beginning to emerge with the creation offederated conservancies in Africa. In hisexample of Cameroon, Mayaka (2002) indicatesthat a three-institutional setup is proposed forco-management of hunting areas around theBenoue National Park. The proposedframework includes: “Wildlife Village Committeein each village in the hunting area; Union of VillageCommittees, a legalized federation of villagecommittees; and, an Area Co-ManagementCommittee, a board formed by representatives of theUnion and government official” (Mayaka 2002).Whether such a structure (when it does getinstitutionalized) resolves scale issues is yet tobe seen. For example, in Namibia, someconservancies have chosen to stay small andopposed attempts toward the creation of superconservancies. At least one conservancy(Sesfontein) has broken up into two. Thus, Jones(1999b) suggests when communities can decidehow to manage common areas, new solutionsmay emerge that even “test” existing theories.

Sustainability

Community conservation programs aregenerally not self-sustained. Support for theseprograms usually comes from tourism and from

41



international donors. Tourism revenues varydepending on economic and politicalcircumstances, while development assistance isoften a function of agendas that are far beyondthe control of local communities. Further,revenues earned only partly accrue to localcommunities. This is to some degree because ofadministrative costs of park management, and,because high-revenue earning parks andprotected areas have to support otherconservation areas in any country. Thus,financial sustainability is an issue that eachconservation area (and communities whodepend on these areas) will need to confront.

Barnes and others (2002), in their study of fiveconservancies in Namibia, explicitly address theimpact of donor grants on communityconservation. They undertake a careful cost-benefit analysis of community investments inconservancies and show that these investmentsresult in a positive net present value. Whendonor support to these conservancy projects isremoved, the financial returns diminishconsiderably, but the projects are still viableinvestments. This study is one of the onlystudies identified in this review that appraisesdonor support in the context of communitymanagement of natural resources. It shows thatdonor support is essential. However, CBNRMcan thrive, at least in certain Namibianconservancies, with decreased externalresources—this is a very positive result for thefuture of CBNRM.

Looking beyond the financial sustainability ofcurrent programs, we need to question whetherCBNRM promotes future investments inconservation. In other words, does CBNRMprovide the right mix of awareness,empowerment, and financial incentives tosustain the wildlife industry into the future.

Wildlife investments may only becomeattractive to farm households if a) revenuesfrom wildlife increased relative to revenuesfrom agriculture; and b) wildlife are seen moreas private resources over which householdshave much more control. Until these conditionsemerge, local communities may well treatwildlife as an asset that is available to be rundown and utilized for greater “development.”

Conclusions

CBNRM programs are complex constructs thatseek to meet multiple objectives of economicdevelopment, empowerment, and conservation.They also involve some degree of devolution ofproperty rights (and power) from a bureaucraticstate organization to a community organization.As a result, a successful CBNRM program has tomeet a triple bottom-line, while simultaneouslyresolving stakeholder conflicts that are endemicto any shift in property rights regimes. This isparticularly difficult in developing countrieswhere rights are typically contested andambiguous.

Is decentralization sound environmental policy?Our review of wildlife management in Africasuggests that it is certainly a feasible and smartpolicy. However, CBNRM has a long way to goto achieve its conservation potential. Multiplechallenges are posed by the need to create theright incentives for conservation and by theopportunity costs of alternative land-uses.Broader concerns dealing with mismatchesbetween the scale required for ecosystemmanagement and social management, andsustainability remain.

Archbald and Naughton-Treves (2001)recommend a three-tiered strategy for ensuringthat stakeholders face the right incentives for



conserving biodiversity and wildlife. Theirapproach would include (a) public goodsdelivery programs that bring economicdevelopment to the community; (b)compensation or insurance schemes that reducespecific costs at the household level (such ascrop damage); and (c) some economic incentivesto households (dividends, employment) tochange behavior. A fourth arm to this is a strongmonitoring program that would both allow thestate and local communities to understandchanges in natural assets and createdisincentives to those who would bend agreed-upon rules. Our review suggests that the least-addressed of these four aspects is compensationfor crop predation. This is probably becausecompensation and insurance schemes areinherently susceptible to moral hazard andadverse selection problems. However, this is anarea that requires further exploration andempirical verification.

Is decentralization good social policy? The case-study literature suggests that decentralizationcan occur in many forms and with varyingdegrees of empowerment to local communities.In some cases, such as CAMPFIRE,decentralization to local officials does notappear to have facilitated increased communitycontrol over natural resources. In other cases,devolution has meant that local communitieshave some say over what kinds of ecotourismare undertaken and how the benefits of wildlifemanagement are distributed. While devolutionhas rarely resulted in full managementauthority over natural assets, it has led toimproved interactions between biodiversitymanagers and communities. As shown in theKimana case, decentralization undertaken forenvironmental management may not empowerthe most vulnerable social groups. In fact,decentralization can reinforce traditionalhierarchical structures, and this is the challenge

in using environmental decentralization as ameans to address social issues.

Does decentralization lead to economicdevelopment and poverty reduction? CBNRMprograms provide public goods to communitiesand offer opportunities for income andemployment. To the extent that they open upnew markets in ecotourism and offeralternatives to subsistence agriculture, they dohave a positive impact on rural livelihoods.There is also some evidence that the publicgoods provided through CBNRM improve theoverall welfare of local communities. However,opportunity costs remain and are arguablyrather high. Our review, of course, cannotanswer the question of whether CBNRM is abetter investment for poverty reduction relativeto other rural development schemes. While thismay be true in some areas that receive highnumbers of tourists, it may be less true inothers.

Finally, is devolution good for the government?The answer to this question appears to bepositive. Most CBNRM schemes involve somerevenue-sharing mechanism with thegovernment. Many authors contend, therefore,that CBNRM facilitates a rural wildlife tax. Thestate, as the “owner” of natural resources, issimply obtaining resource rents on its assets. Tothe extent that involving the communityincreases these rents, it is a win-win situationfor the state. The state also gains becauseCBNRM often leads to decreases in the state’smonitoring costs and reduces conflicts betweenforestry authorities and communities.

Community-based wildlife management as anatural resource policy has led to many changesand improvements in the lives of rural peoples.However, it is not clear that it has overcome theincentives problem, which is the basis of much

43



of current environmental degradation. As Box 7suggests, there are several reasons to beconcerned about whether communityinvestments are compatible with incentives.

The next step in the evolution of CBNRM is tomore carefully identify and invest in private

incentives for wildlife management. Long-termchallenges such as sustainability of CBNRM willdepend on the creation of nested institutionsand changes in tenure and other land-usepolicies.


Water User Associations andIrrigation Management Transfer:Understanding Impacts andChallenges

Introduction

Irrigation management transfer (IMT)—theprocess of transferring irrigation managementfrom government to farmer organizations—wasfirst undertaken in the United States, France,Colombia, and Taiwan from the 1950s to the1970s. Developing countries followed in the1980s and 1990s. To date, governments in atleast 25 countries in Asia, Latin America, andAfrica are reducing their roles in irrigationmanagement, while farmer groups or privateorganizations are taking them over (Vermillion1992). The countries adopting IMT as a policyinclude Chile, Peru, Brazil, Mexico, theDominican Republic, Colombia, Haiti, Senegal,Mauritania, Niger, Zimbabwe, Tanzania, Sudan,Somalia, Madagascar, Turkey, Pakistan, India,Sri Lanka, Bangladesh, Lao PDR, Vietnam,China, Indonesia, and the Philippines.

Since the mid-1980s, the centerpiece of reforminvariably has been the transfer of management(in rare cases, along with the ownership) ofirrigation systems—wholly or in part—to wateruser associations (WUAs), private corporations,or parastatal units, combined with thedownsizing or withdrawal of the government’srole in operation and maintenance (O&M), feecollection, water management, and conflictresolution. Early motivations behind IMTenvisaged that farmer management of publicirrigation systems would enhance theirperformance and bring about wide-ranging

socioeconomic changes that would enablefarmers to substantially improve farm incomes(Shah and others 2002). This was partly basedon the reported success with user-managedirrigation, which was documented—at least inAsia)—to be far more productive andfinancially viable compared to public irrigationsystems. These successes—including farmer-managed irrigation schemes (FMIS) in the hillsof South Asia, tubewell companies in NorthGujarat, lift irrigation schemes built andmanaged by sugar cooperatives in Maharasthra,and deep tubewell cooperatives inBangladesh—all showed that well-managedcollective irrigation by farmers could play animportant role in transforming their livelihoods(Tang 1992; Lam, Lee, and Ostrom 1997).

More recently, a number of researchers (Kloezenand others 1997; Vermillion 1997; Koppen andothers 2002; Samad and others 1999; Barker andMolle 2002; Shah and others 2002) havesuggested that the IMT discussion has shiftedmore toward getting irrigation off the backs ofgovernment than toward improving the lot ofthe farmers and the rural poor, the original goalof public irrigation investment over the past 50years. The driving force behind reforms—theneed to reduce the government’s recurrentexpenditures for irrigation. IMT—is nowconsidered beneficial even if it just saves thegovernment money. However, Shah and others(2002) argue that, at least in Africa, “nowhere isthere a significant body of positive experience to

Eduardo Araral19

3



suggest that straightforward IMT will work insmallholder irrigation as it has with large,commercial farmers in the United States,Mexico, South Africa, New Zealand, Columbia,and Turkey.”

While IMT is being undertaken worldwide andwith growing interest, there is surprisingly littleevidence about its results, particularly itsimpact on poverty reduction. The drivingassumption behind this growing interest is thatsuccessful IMT is, in principle, viable providedthe “process is right” and favorable socio-technical, legal, and political conditions arecreated. However, it is not well-established inthe literature whether IMT can simultaneouslysave money for the government, bring aboutmore cost-efficient management for the farmers,and achieve financial and infrastructuralsustainability. This paper, therefore, attempts toanswer the following questions:

• What do we understand about the impactsof IMT in terms of poverty reduction,financial, and personnel implications for thegovernment, O&M, and resourceconservation?

• What are the conditions most likely to resultin sustainable local resource governanceorganizations? Are there specific conditionsunder which irrigation user groups workbetter in order to address poverty issues?

• What are the emerging challenges in IMT?

To evaluate the evidence on the impacts ofirrigation decentralization, a careful review ofwritten literature was undertaken. Literatureand data sets came mainly from recent researchreports of the International Water ManagementInstitute (IWMI), particularly that of Vermillion

(1997), proceedings from the InternationalConference on Irrigation Management Transferheld in Wuhan, China in 1994, and discussionswith Avelino Mejia20 and his staff at theNational Irrigation Administration, Philippinesin 2003. Various sector papers from the WorldBank, as well as research reports and collectionsfrom the Workshop in Political Theory andPolicy Analysis at Indiana University, were alsoreviewed.

This chapter is organized into six sections. Thefirst section deals with the evolution of IMT andthe research questions and methodology. Thesecond section discusses various IMTassumptions and models. The third sectiondiscusses the impacts of IMT in terms ofproductivity and poverty, impacts ongovernment finance and personnel, and impactson O&M and resource use. The fourth sectiondeals with the conditions associated with theemergence and sustainability of self-governingirrigators’ associations. The fifth section dealswith emerging challenges in IMT, and the lastsection provides a summary and conclusion.

IMT Assumptions and Models

The underlying logic behind IMT is bestsummarized by Vermillion (1997). According tohim, there are three reasons why IMT is afeasible alternative to centralized irrigationsystems:

• While government bureaucracies lack theincentives to maximize managementperformance, farmers have a direct interestin enhancing irrigation management.

• IMT can increase the profitability ofirrigated agriculture sufficiently to offsetany increased cost of irrigation to farmers.

47

Water User Associations and Irrigation Management Transfer: Understanding Impacts and Challenges


• IMT will also save money for thegovernment in terms of decreasedresponsibilities for routine O&M.

Early models of IMT (1950s to 1970s) focusedmore on non-poor, market-oriented, large-scale,and business-like agriculture, including largefarms in the United States, Mexico, NewZealand, and Turkey. The objective of theseearly models included saving governmentmoney, improving O&M cost efficiency, andmaintaining or increasing the productivity ofirrigated agriculture. The national governmentusually initiated the process. The scopeincluded full transfer of O&M and financefunctions, but ownership of assets remainedwith the government (with the exception ofNew Zealand, which completely privatizedownership).

In contrast, most current models of IMT (1980sand 1990s)—in South and Southeast Asia, LatinAmerica, and Africa—are targeted at poor,small-scale, local-market-oriented agriculture.While the formal objectives remain the same—that is, saving government money, improvingO&M cost efficiency, and maintaining orincreasing productivity of irrigatedagriculture—current IMT is increasingly judgeda success even if it just saves governmentmoney and improves collection efficiency.Current IMT is also generally undertaken asdonor aid-funded projects (Shah and others2002; Groenfeldt and Svendsen 2000). Ingeneral, the scope of O&M and finance-transferred functions is partial, and the processis time bound. In particular, current models ofIMT differ from previous ones: there is greatervariability in terms of (a) transfer units andtheir size; (b) the new management unitresponsible after IMT; (c) the extent of functions

transferred; (d) ownership of assets; and (e)implementation and financing modalitiesadopted by donors.

For some models, such as those in thePhilippines (for communal irrigation), NewZealand, Colombia, Nepal (small systems), andChina, transfer units involve the entire scheme.In other models, transfer units may involve onlydistributary canals, as in the Philippines andNepal (for large irrigation systems), Sri Lanka,Nigeria, Egypt, and India. The size of transferunits varies, from a low of 150 hectares in thecase of communal irrigation in the Philippines,to as much as 14,000 ha in India, 25,000 ha inColombia, and 30,000 ha in Mexico.

In terms of the new management unit, mostIMT models transfer responsibilities to waterusers’ associations (WUA) or irrigators’associations (as in the case of the Philippines,Indonesia, Egypt, Bangladesh, Nepal, India,Senegal, and Colombia). In Turkey, municipalgovernments became the new managementunit. In Vietnam, parastatal organizationsassumed responsibility, while in Sudan andNew Zealand, private / mutual companiesassumed responsibility. In the United States,Japan, South Korea, Mexico, and Taiwan, whereIMT was deemed to have a generally positiveimpact, post-transfer governance entities tendedto be farmer-elected boards of directors, whilemanagement entities tended to be cadres ofprofessional staff appointed by the board.

In terms of the extent of O&M and financefunctions transferred, IMT models in thePhilippines, Indonesia, Nepal, Sri Lanka, Sudan,Nigeria, the Dominican Republic, and Colombiapractice partial transfer of responsibilities. Onthe other hand, models in Vietnam, China,Bangladesh, Egypt, Turkey, Senegal, Colombia,



New Zealand, and Mexico (see Box 8) practice afull transfer of O&M and financialresponsibilities. The powers and functionsdevolved to WUA after IMT includes authority,in varying degrees, to:

• Make rules and sanctions, with themaximum sanction of stopping wateravailable to the WUA

• Make O&M plans and budgets

• Set water charges

• Hire or release management staff

• Control intake

• Control the main canal system andsubsidiary canal systems

• Assume responsibility for futurerehabilitation

• Manage canal rights of way

• Contract and raise funds

• Make profits.

In terms of ownership of assets, the governmentretains ownership in most IMT models, with theexception of Senegal and New Zealand, whereassets were privatized. Experiments are alsounder way in the Philippines to transfer assetownership to WUAs in large systems (Eleazar2002).

Conditions For Self-Governing IrrigatorAssociations

What are the conditions most likely to result insustainable irrigation governanceorganizations? Are there specific conditionsunder which irrigation user groups work better?The creation and strengthening of water user

Box 8Irrigation Management Transfer in Mexico

The Mexican IMT Program is one of the most ambitious and successful of its kind worldwide, not only becauseof the large scale of its irrigated area and the speed of its implementation, but also because of the positiveimpacts claimed to be the result of the strategy followed. In less than a decade until the end of 1996, almost 2.9million hectares have been transferred to 373 WUAs representing 90 percent of the area served by the 80irrigation districts in Mexico. Kloezen and others (1997) described the characteristics of the Mexican programas follows:

• IMT did not come on its own, but followed and is part of a much wider set of liberal economic reformsarising from the economic crisis faced by Mexico in the 1980s.

• IMT was made workable as it met with a political commitment at the highest levels.

• IMT was accompanied by the introduction of a new water law that recognizes water rights to water userassociations, as well as the authority/responsibility of water users.

• IMT is a rapid top-down process that has met with relatively little resistance from farmers, as it was devel-oped on an already existing strong organizational base—the ejidos and the organization of private growers.

• The government, and later private organizations, provided training to the new WUA.

• WUAs agreed to jointly manage the system with the agency during a fixed /short time.

• The Mexican IMT program aims not to maximize direct user participation in O&M, but to involve farmers inrepresentative governance.

49



associations through institutional developmentis a standard component of most donor-fundedirrigation projects. Institutional development,however, is a poorly understood task. Usually, itis merely associated with providing technicalassistance consultancies, staff and beneficiarytraining, study tours, and financial andlogistical support. What is often left out is whatVincent Ostrom (1980) referred to as “theartisanship involved in the design, operation,appraisal, and modification of rule-orderedbehavior,” one which Uphoff (1986) alsoreferred to as “the ongoing nature of getting theprocess right.”

Ostrom (1992) suggested that “the rulesgoverning the supply and use of any particularphysical system must be devised, tried,modified, and tried again, and considerabletime and resources will be invested in learningmore about how various institutional rules,combined with particular physical, economic,and cultural environments, produce incentivesand outcomes. The choice of institutionstherefore is not a ‘one-shot’ decision in a knownenvironment, but rather an ongoing investmentin an uncertain environment. The time investedin this process is similar to the time invested inbuilding and operating a better physicalinfrastructure.” In a synthesis of 15 years oflessons learned on common pool resources,Ostrom, Dietz, Dolsak, Stern, Stonich, andWeber (2002) suggested the specific physicalcharacteristics of the irrigation systems and thecharacteristics of the irrigators’ associations areimportant factors affecting the emergence andsustainability of self-organized irrigators’associations.

McKean (1992) and Ostrom and others (1990,2001) suggested that self-governing associationsare more likely to form when the followingattributes of the physical resource are present:

• Resource scarcity and feasible improvement. Ifirrigation water is abundant, there are fewreasons for farmers to invest in organizing.This was the case of small irrigation systemsin Central and West Java, where river wateris usually recycled by downstream farmers.Attempts to organize farmers did notsucceed, as the farmers did not perceive themarginal benefits of joining the group to bemore than additional costs of time inmeetings and labor contributions. On theother hand, incentives for self-organizationmay be higher when water scarcity issubstantial, as in the case of irrigationsystems in the arid Ilocos region of thePhilippines (Ostrom 1990). In addition, ifthe irrigation system is substantiallydestroyed, organizing may not generatesubstantial benefits. Farmers may also havedifficulties adapting to exogenous shockssuch as prolonged drought and catastrophicflooding. This was the case with manyirrigators’ associations in the Bicol region inthe Philippines, which ceased to exist aftertyphoons led to the collapse ofinfrastructure (Araral 2003).

• Monitorable indicators. The presence offrequently available and reliable indicatorsabout the condition of the irrigation system,for example volumetric pricing practices inChina, affects the capacity of farmers toadapt relatively soon to changes that couldadversely affect their long-term benefitstream.

• Spatial and hydrologic boundaries. The largerthe boundaries, the higher the cost ofmonitoring and enforcement, particularly inthe case of small-scale farming, as shown inAfrica (Shah and others 2002).

The following characteristics of the farmersaffect the distribution of costs and benefits, thus



influencing the likelihood that a self-governingirrigators’ association will emerge:

• Salience. If farmers do not obtain a majorpart of their livelihood from irrigatedfarming, the high cost of organizing andmaintaining a self-governing irrigationsystem may not be worth their effort, assuggested by Koppen and others (2002) inthe case of big landlords in Andhra Pradeshand Gujarat in India.

• Heterogeneity of wealth. Bardhan and Dayton-Johnson (2002) suggest that heterogeneity ofwealth makes it more difficult to agree toallocation rules. This in turn makesenforcement and resource maintenancemore difficult, as appears to be the case inlarge-scale canals in Gujarat and AndhraPradesh in India (Koppen and others 2002).

• Group size. It is a lot easier for five largefarmers to come together to agree to therules of self-management than for 1,500smallholders, as shown by the experience ofSouth Africa and elsewhere (Shah 2002).

• Common understanding. If farmers do notshare a common understanding of how theirrigation system operates (topographic,hydrologic, climatic, socioeconomic factors,and the magnitude of problems), they willfind it extremely difficult to agree on futurejoint strategies (Ostrom 1990). This isparticularly true of aquifers, wheremonitoring the resource condition isdifficult and expensive.

• Trust and Reciprocity. Farmers who trust oneanother to keep agreements and usereciprocity in their relationships with oneanother face lower expected costs related tomonitoring and sanctioning one another

over time. Farmers who lack trust at thebeginning of a process of organizing may beable to build this form of social capital ifthey initially adopt small changes that mostfarmers follow before trying to make majorinstitutional changes, as experience inNepal has shown. Uphoff (1992, 2003)makes a strong case for the importance ofsocial capital for improved irrigationperformance.

• Autonomy. Farmers who are able todetermine rules without external authoritiescountermanding them have lower costs oforganizing. For example, in many small-scale irrigation systems in West and CentralJava in Indonesia (Samad and others 2002),while water users’ associations have themandate for O&M, they do not have formalrights to the water and infrastructure andare virtually powerless to settle disputesand enforce collection of irrigation servicefees in irrigation systems that cut acrossmultiple villages. They also generally lackthe legal and political clout to mobilizeloans and enter into contracts that wouldpermit them to enter into business ventures.As a result, IMT does not constitute adramatic change in management.

• Prior experience and local leadership. Farmerswith prior experience with other forms oflocal organization greatly enhance therepertoire of rules and strategies known bylocal participants as potentially useful toachieve various forms of regulation.Pradhan (2002) suggests, in the case ofNepal, that groups with more social capitalare better able to deal with collective actionproblems. Ternstrom (2003), on the otherhand, finds strong empirical support for therole of local leadership in the emergenceand sustainability of irrigation institutions.

51



Box 9 presents cases in the Philippines andidentifies how some of these factors have beenconducive to success. The growing theoreticalconsensus above does not, however, lead to aconclusion that most irrigation farmers willundertake self-governed regulation. The crucialfactor is not whether all attributes mentionedabove are favorable, but the relative size of theexpected benefits and costs as perceived by thefarmers. Farmers must perceive and actually

experience that the marginal benefits oforganizing will outweigh the costs involved. Inaddition, the macroeconomic policies of acountry are equally important in determiningthese costs and benefits.

Assessing Impacts

Irrigation management transfer has now beenundertaken in many parts of the world and

Box 9Features of Successful Irrigation Associations in the Philippines

Araral (2003) describes the typical characteristics of outstanding IMT irrigators associations in the Philippines;that is, those that have consistently won national awards. These characteristics are broadly consistent with andreinforce the above conditions for self-organized irrigators’ associations.

First, they have at least 10 years of collective experience as an organization. Second, the irrigation resource issalient enough to the users that they are willing to invest time and effort to create new institutions. A majorityof the members depend on rice farming for most of their livelihood. Third, the farmers have the autonomy todevise and change operational rules at least within the ambit of their charters. Fourth, at least a subset offarmers are able to engage in direct communication with each other, including the opportunity to bargain. Inaddition, they share the following characteristics:

In terms of organization, they have high membership rates (80-89 percent of farmers in their service areas), theboard of directors meets at least monthly or even more; the general assembly meets at least annually; they keeporganized records (membership, financial, minutes of meetings; articles of incorporation; water permits; gov-ernment reportorial requirements). In terms of organizational discipline, they regularly hold elections accord-ing to their by-laws; most conflicts are resolved within their IA without external assistance; enforcement mainlyrelies on hierarchically enforced norms, peer pressure, reputation, and altruism; agreed-upon sanctions areenforced 90 percent of the time; collective leadership is strong; at least 90 percent of members attend sectormeetings to discuss plan preparation and evaluation, and contribute to group work and donate a portion oftheir labor as equity fund to the association.

In terms of O&M, successful IAs are strong in the planning, coordination and execution of O&M policies andplans, particularly in the preparation of the cropping calendar, water distribution plan, maintenance and re-pair, and amortization schedule.

In terms of financial planning and implementation, the better IAs seem to have a clear and generally agreed-upon financial plan for the cropping calendar; exercise fiscal discipline (high viability index measured as aratio of income and expense); conduct internal audits at least every two months and external audits regularly;have a system of checks and balances to avoid nepotism between auditors and treasurers; and emphasizetransparency and accountability as a norm of behavior among officers.

Successful IAs have a cooperative or a tie-up with a cooperative for agricultural assistance in terms of inputs,marketing, and equipment. Their collective leaders are entrepreneurs in terms of fund raising and mobilizingresources; women are actively involved (half of IA officers are women); and they provide incentives to officersand members (such as credit for scholarships, mortuary, and hospitalization, as well as incentives for officerssuch as a transport allowance and health insurance).



continues to grow in interest among donors,governments, and farmers. What do weunderstand about the impacts of irrigationmanagement transfer? In this section, we reviewavailable case studies to identify impacts interms of poverty reduction, governmentfinance, operations and maintenance quality,and resource conservation.

Productivity and farm income

One of the underlying policy assumptionsbehind IMT is that it has a positive impact onpoverty reduction and agricultural growth. Thelogic behind this assumption is that IMT canreduce the cost of water and also diversifyrevenue sources for water users’ associations—both of which are expected to improveagricultural productivity and incomes.Reduction in the cost of water comes in the longterm from increased water-use efficiency.

Diversification of revenue sources, on the otherhand, results from the granting of corporatepowers and greater autonomy to the WUA.

The evidence on whether IMT leads to growthand increased agricultural productivity is rathermixed. Box 10 discusses a number of caseswhere IMT has led to a change in the cost ofirrigated water. As shown in Box 10, theavailable empirical evidence suggests that 1)where significant subsidies existed before IMT,the cost of irrigation to farmers may risesubstantially; 2) where there is little or nochange in subsidies, IMT may lead to a decreasein irrigation costs to farmers; and 3) high-costsystems such as pump irrigation are more likelyto significantly increase the cost of water tofarmers (Vermillion 1997). Thus, overall, IMTwill not necessarily reduce the costs of irrigatedwater to farmers. Instead, IMT may wellincrease the private cost of production tofarmers.

Box 10IMT Impact on Cost of Irrigation Water

The cost of irrigation to farmers can rise substantially after IMT, where significant subsidies existed beforeIMT. This is particularly true for high-cost systems such as pump irrigation. For example, Meizen-Dick andothers (1997) report that IMT in Senegal has led to a 200 to 400 percent increase in the cost of water, mainly dueto the significant loss of government subsidies. Privatization of irrigation in the Senegal River Valley, whichwas irrigated by lift pump schemes, led to a 78 percent increase in the cost of rice production for farmersbetween 1980 and 1993 (in constant 1980 prices) due primarily to discontinuance of subsidies for credit, inputprovision, and irrigation. The positive side, however, was that overpumping from lift irrigation was reducedas a result of improved supervision by farmer-hired staff. In Indonesia, Johnson and Reiss (1993) report thatwater charges to farmers increased five- to seven-fold as significant government subsidies were lowered. Inthe Dominican Republic, Yap-Salinas (year?) reports a 1,500 percent increase in the cost of irrigated water overan eight-year period when subsidies for pump irrigation were reduced.

On the other hand, where there is little or no change in subsidies, IMT may lead to a decrease in irrigation coststo farmers, particularly for surface irrigation. This was the case in the Philippines, where the cost of waterdeclined by 75 percent (Oorthuizen & Kloezen 1995); in Nepal, by at least 40 percent (Olin 1994); in Colombia(Vermillion & Garces-Restrepo 1996); in the United States, by 16 percent (Svendsen & Vermillion, 1994); and inNew Zealand, by at least 25 percent (Farley 1994). In these cases, farmers were granted financial autonomythat allowed them to determine irrigation tariffs, collect and keep greater water revenues, and take cost-cut-ting measures. These measures included reducing overhead costs and designing simpler repair and mainte-nance work, which improved operational efficiency and resulted in lower costs of water delivery.

53



The impact of IMT on agricultural productivityis measured in terms of an increase in cultivatedarea, cropping intensity, and yield, while impacton economic productivity is measured in termsof gross value of output, net farm income perhectare, and economic returns to irrigation. Box11 presents several studies that try tounderstand the linkages between managementtransfer and agricultural growth.

The relationship between IMT and agriculturaland economic productivity is indirect and oftenambiguous, especially when compared to therelationship between IMT and O&M

performance or financial viability. Theambiguity stems from the fact that most studiesdo not attempt to control for interveningvariables such as changes in rainfall or fertilizerapplication rates, as well as factors such asprices, subsidies, and markets. For instance, anassessment by Kloezen and others (1997) onagricultural and economic productivity of theIMT program in Mexico revealed thatfluctuations in productivity values cannot berelated directly to the program, but have to beviewed in the context of other economicchanges since the 1980s arising from neo-liberaleconomic reforms.

Box 11IMT Impact on Agriculture

In an assessment of small-scale IMT programs in West and Central Java, Indonesia, Samad and others (2000)found that the potential for improving productivity or profitability of irrigated agriculture through changes inirrigation system management was limited. The main reason is that water and land resources were alreadyintensively exploited. Cropping intensities are relatively high and water is generally recycled and reutilizedbetween systems along river courses. This was the key reason for the lack of substantial improvement inagricultural productivity after IMT.

In other countries, impact on cropping intensity is mixed. Cases in the Philippines, Sri Lanka, India, China,Nepal, and Vietnam report increased cropping intensities ranging from 80 to 250 percent, which were general-ly attributed to more responsive irrigation operations after IMT (Wijayaratna and Vermillion 1994; Nguyenand Luong 1994; Kloezen 1996; Kalro and Naik 1995; Uphoff 1992; Pradhan 2002). However, in the SenegalRiver Valley, privatization of irrigated agriculture support services was accompanied not only by a decline incropping intensities, but by an expansion in irrigated area. Farmers shifted to growing more of their crop onlyin the wet season, partly due to rising input prices and the greater complexity of dry-season irrigation afterIMT. In another region, privatization led to a near doubling of irrigated area between 1985 and 1993 and anincrease in cropping intensity from 86 to 93 percent (Wester, During and Oorthuizen 1995). Results for cropyields, however, were mixed and ambiguous and no information was generally available.

In cases where irrigation operations improved—that is, when more water was made available—there werereports of increased cultivated area, for instance 14 percent in Vietnam (Nguyen and Luong 1994), and 80percent in Nigeria during dry-cropping season (Musa 1994). The increase in cultivated area in Nigeria resultedfrom significant improvement in water distribution to tail-end areas. Crop diversification was reported inIndia, the United States, and Colombia as a result of more flexible water distribution practices, emphasis onvalue-added crops, and in response to market demands.

Scant information is available on IMT impacts in terms of increased income, gross value of output, and eco-nomic returns. In the United States, average farm incomes rose 15 percent due to reduction in water cost(Svendsen and Vermillion 1994). In Mexico, Johnson (1996) reports that annual economic returns remained thesame. In Colombia, the cost of water relative to cost of rice production increased, but net farm incomes andgross value of output increased substantially due to a shift to high-value crops.



While the impacts of management transfer onirrigation costs and on agricultural yield ingeneral are ambiguous, it is clear that IMT canresult in increased revenues to water userassociations as they grow into mature for-profitorganizations. When management transferenables WUAs to diversify revenue sources andbecome profit-oriented (and not just a nonprofitservice organization), this results in positiveimpacts. Diversified revenue sources come frommembership fees; seasonal and shareholderfees; trading of fertilizers and agrochemicals;rice marketing; tractor rental and interest fromsmall loans in Sri Lanka, from farm enterprises inChina and revenues from mini-hydro powerstations, and water-selling contracts and otherincome-generating projects in the United States.Diversified revenue sources enable the WUA toventure into agriculture business operations,such as wholesale procurement of farm suppliesthat help lower costs of farming for itsmembers, and to engage in cooperativemarketing that helps increase farm prices(Kloezen 1996). WUAs venturing asmultipurpose cooperative are also able toprovide patronage refunds in the form of cashand stock dividends to its members. In the BayiDistrict in Hebei Province in China, the wateruser association was able to develop nineenterprises between 1984 to 1992 after it becamefinancially autonomous. Sixty-five percent of its$60,000 profit was allocated for watermanagement costs; the rest was used as salariesand bonuses to workers, many of whom are alsomembers of the irrigation staff (Vermillion andothers 1994).

Impact of IMT on government finance andpersonnel

It is commonly assumed that IMT will savemoney for the government as it divests itself of

the responsibility to finance routine costs ofO&M of irrigation systems. It is also commonlyassumed that the savings can be used to reducegovernment expenditure in the irrigationsubsector or to reallocate funds to otherfunctions that cannot be handled or financeddirectly by the private sector (Shah and others2002). But much depends on political will,budgets, and financial policy (Vermillion 1997).Nonetheless, as Table 4 below shows, IMT doesseem to lead to a decrease in governmentsubsidies and improve the budget solvency ofirrigation agencies.

Most studies dealing with fiscal impacts,however, only document government spendingfor O&M and mainly at the scheme level(Vermillion 1997). There is little information onwhether savings from decreased governmentspending for irrigation O&M are being divertedto new construction, rehabilitation, or other useswithin the sector; to other sectors; or to permitan overall shrinking of governmentexpenditures.

The Philippine experience in IMT related togovernment personnel typifies IMT experienceelsewhere. IMT in the Philippines was intended,at least partially, to reduce recurrentgovernment O&M costs by reducing irrigationmanagement positions. Reports indicate thatoverall irrigation agency staff diminished (by asmuch as 50 percent) both at the system andadministrative levels as a result of a progressiveand long-standing IMT policy adopted since the1970s.

This decline, however, was gradual as theNational Irrigation Administration (NIA)waited for staff to retire. Vacated posts were notfilled as part of the personnel attrition policy ofthe Civil Service Commission. Other personnelmanagement practices at NIA included (a)

55



relocating staff members into systems that werenot being transferred (similar to the practice inIndonesia and Sri Lanka), or transfer of staffmembers into non-O&M activities such asconstruction of new systems (as in Turkey); (b)rehiring retired staff as project consultants; and(c) encouraging / requiring the irrigators’

association (IA) to hire former NIA staff as partof IMT contracts.

Where affected staff remain valuable to theIAs—for instance, water technicians who arealso local residents—IAs are more willing tohire them on a progressive salary-sharing

Impact of IMT on Government Subsidies Cases reporting, country, type of irrigation Reported Impacts Kim, Kwanbo, 2003, Korea, SI

Annual government savings of 20 percent due to efficiency improvements after IMT

Bagadion and Korten, 1991, Philippines, SI Wijayaratna and Vermillion, 1994, Philippines

Annual government savings of $12/ha due to increased cash and in-kind contributions from farmers. Revenues from irrigation service fees constituted 24 percent of the revenues of the government irrigation agency (NIA) in 1979 and 60 percent in 1990. Government spending for O&M fell from P25million in 1976 to zero in 1982, but limited subsidies were later introduced.

Kloezen, 1996, Sri Lanka, SI Government O&M spending decreased from $14.80/ha in 1985 to $6.50/ha in 1994.

Pant, 1994, Uttar Pradesh, India, LI Government subsidies for a typical tubewell decreased by 25 percent after IMT.

Mishra and Molden, 1996, Nepal, SI Government subsidies for O&M declined from $6.65/ha to $4.06 after IMT.

Vermillion and Johnson, 1990, Indonesia, SI

Annual government savings of $13.5million in O&M costs of small-scale irrigation.

Johnson, 1996, Mexico, Annual government subsidies for O&M fell from $40million in 1989 to zero in 1993, covering 2.4million hectares of service areas.

Impact of IMT on Budget Solvency of the Irrigation Agency Oorthuizen & Kloezen, 1995, Philippines, SI

The irrigation system’s annual budget deficit declined from an average of $19,178 in 1982–85 to $554 from 1986–89, the first four years after transfer.

Pant, 1994, Uttar Pradesh, India, LI Before IMT, annual losses of $876 and afterwards, consistent surpluses.

Bagadion, 1994, Philippines, SI Annual average loss of $42,218 for 1981–89 converted into annual surplus of $42,880 after IMT during 1990–92.

Johson, 1996; Gorriz, Subramanian & Simas, 1995; Mexico, SI

Deficits declined from $66million in 1989 to $44 million in 1993 after 80 percent IMT.

Vermillion & Garces-Restrepo, 1994, Colombia, SI

Budget deficits for 2–4 years before transfer and surpluses 2–4 years afterwards due to staff reduction and increases in revenue.

Table 4. Impact on government subsidies and budget solvency of the irrigation agency



scheme (i.e. the irrigation agency initiallyshoulders staff salary, but progressively sharesthe cost until the IA pays 100 percent). Officeand administrative personnel and irrigationsystem managers are less likely to be hired bythe IAs, either because of little value-added orbecause they are too expensive.

Impacts on O&M quality and resourceconservation

It is commonly assumed that IMT will bringabout improvements in the quality of irrigationO&M and will result in sustainableinfrastructure maintenance. Further underlyingthis assumption is the notion that farmer-usersof small-scale irrigation systems are potentiallycapable of operating and maintaining theirirrigation systems. It is also assumed thatinvolving local water users’ associations in thepre-construction and construction activities isan important means of developing the skills andstructures of the associations and for ensuringthat farmers are willing to contribute toward themaintenance of the investments (Bagadion andKorten 1991). Underpinning this assumption isthe notion that farmers have a direct interest inenhancing and sustaining the quality and costefficiency of irrigation management. Whengiven the authority and incentives to actcollectively, farmers will act to contain the costof water management while improvingoperational performance.

Evidence on the impacts of IMT on O&Mquality is mixed. In general, most studies reportpositive impacts in terms of improvedoperations—including increased water-useefficiency, reliability, adequacy and timeliness ofwater delivery, increase in service area,responsiveness of the IA to their members’needs, and more equitable water distribution.

The impact of IMT on maintenance, on the otherhand, is less positive. Farmers tend to under-invest in maintenance, and evidence suggests anoverall declining trend. In cases of lift irrigation,maintenance has generally worsened mainlybecause of increasing costs borne by farmers asa result of reduced government subsidies.

Not much evidence is available on the impact ofIMT on water resource sustainability. Some ofthe unanswered questions are as follows:

• Does the most progressive form of IMTcreate adequate incentives for farmers inwater-scarce areas to conserve water in asignificant way?

• What kinds of water conservationincentives are created by IMT in cases ofrun-of-the-river type irrigation and aquifer-dependent farms, where monitoring watertables is costly?

• What about in systems where monitoring ofgross consumption is easier, i.e. damirrigation?

• Would the assignment of water rights tofarmers create significant incentives?

• Are the incentives the same in the face ofhigh transaction costs (particularlymonitoring and sanctioning costs) in large-scale systems composed of thousands ofsmallholder farmers?

One can expect that incentives are more likely tobe created for farmers to conserve water underthe following conditions: (a) water is scarce butwater rights are clear and secure; (b) water ispriced as an economic good; (c) consumption isadequately monitored at the farm level (i.e. useof volumetric pricing) and at the basin level (i.e.

57



use of water accounting); (d)) irrigation systemsare sufficiently small to keep transaction costs ata minimum or when WUAs are appropriatelynested; (e) when water conserving technologiesare practiced; and (f) when IMT is mostprogressive.21 The empirical support for such aproposition can be suggested by indicators atthe farm level such as land and waterproductivity (measured in terms of output perunit command when land is scarce or outputper unit of water consumed when water isscarce), or at the basin level in terms of relativewater supply—that is, total water supply (asmeasured in terms of water accounting) relativeto crop demand. 22

Unfortunately, however, even for the mostprogressive IMT, most conditions that are more

likely to create conservation incentives areseldom met. Our review shows that IMT hashad a variety of positive impacts on farmhouseholds. However, these are conditioned bydifferent circumstances. Table 5 summarizes ourbest understanding of when IMT is likely tohave a positive impact and the conditions underwhich they are more likely to be negative.

Challenges Ahead

This review of IMT experience in a numbercountries indicates a number of emergingchallenges: (a) poverty reduction and IMT;(b) IMT and institutional development;(c) operations and maintenance; (d) waterconservation; and (e) the poor quality of IMTresearch.

Box 12IMT Impact Assessment in Mexico

The main impacts of the IMT Program, based on an evaluation of the Alto Rio Lerma Irrigation District (AL-RID) in Mexico covering 113,000 ha, were summarized by Kloezen and others (1997) as follows:

• Farmers’ increased control has not led to major improvements in operational performance, particularly inthe way water is allocated and distributed. There were also no changes in the area irrigated or croppingpatterns that can be attributed to IMT.

• Farmers’ increased involvement in decision-making and control has increased managerial accountability.Farmers are particularly positive about the improvement of services provided by the ditch tenders. They feelthat compared to the pre-IMT period under the government irrigation agency, the WUAs have more controlover the ditch tenders’ work and rent-seeking behavior.

• IMT has led to a better match between actual expenditures and farmers’ perceived needs, especially in main-tenance at lower system levels.

• Financial self-sufficiency has increased from 50 percent to 120 percent after IMT due to the ability of theWUA to achieve fee collection rates at over 100 percent. The hiring of professional administrative staff andthe use of good computer software to handle daily financial administration has resulted in better financialadministration.

• IMT has not resulted in an increase in the cost of water to farmers. Although the cost of irrigation to farmersremains low after IMT (less than 5 percent of gross value output), WUAs find it very difficult to convincefarmers that irrigations fees should be increased to keep up with inflation. None of the WUAs also created acontingency fund for future emergency repairs.

• No convincing evidence was found that IMT had a substantial effect on agricultural and economic produc-tivity. Fluctuations were attributed more to the dismantling of credit and subsidy systems, input price poli-cies, and price changes in world commodity prices.



Positive impact more likely if…. Negative impact more likely if…. Productivity and poverty

Cost of Irrigation water

There is little or no change in subsidies; the effect on any increases in water costs is overcome by improvements in efficiency or water availability.

Significant subsidies are phased out, particularly in high-cost systems such as pump irrigation.

Productivity for all income levels

Head-tail end-water distribution improves to increase cropped area. Opportunities for productivity improvement are high (i.e. when water and land resources are not being intensively exploited, cropping intensities are low), as well as opportunities for multiple cropping and diversification due to more reliable and flexible water supply.

Little impact if land and water resources are already intensively exploited and cropping intensities already high; productivity may not be due to IMT but may be due to dismantling of credit and subsidy systems, input price policies, and price changes in world commodity prices.

Revenue diversification

IMT enables water user associations to diversify revenue sources and become profit-oriented, not just service-oriented. Increased revenue for WUA helps lower costs of farming, increases farm-gate prices, provides benefits to members, generates local employment, and stimulates local economy.

Government Finance

Significant subsidies for recurrent O&M costs are phased out, including redundant staff. Collection efficiency increases; savings are reallocated toward supportive investments in the agriculture sector, or are used for support services such as credit, research and extension and farm-to-market roads, which are all vital for the success of IMT.

Future financial responsibility for rehabilitation is unclear or government commitment is not credible, such that farmers believe that government will always bail them out.

Sustainability O&M IMT is most progressive when WUAs are allowed to (a)

make rules and sanctions; (b) make O&M plan and budgets; (c) set water charges; (d) hire or release management staff; (e) control intake; 6) control main canal system; (f) control subsidiary canal system; (g) be responsible for future rehabilitation; (h) have canal rights of way; (i) have right to contract, raise, and disburse funds and make profit; (j) have well-specified management functions and delineation of authority; (k) have effective accountability and incentives; (l) have arrangements for viable and timely conflict resolution; and (m) have adequate resources that can be mobilized for irrigation management. In larger systems, governing boards are farmer-elected and managers are professional cadres; legal structures are able to handle increasing scales of complexity. Economic value of irrigated farming is high, as in high-value export crops IMT bundled with infrastructure is improvement, such as canal lining and use of automatic diversion weirs that substantially reduce O&M costs to farmers.

Few or none of the progressive requirements of IMT are present; high operating costs such as in lift irrigation and when economic value of crops are low, as in paddy farming

Water conservation

(a) water is scarce but water rights and water service are clear, secure, and sustainable; (b) water is priced as an economic good; (c) consumption is adequately monitored at the farm level (i.e. use of volumetric pricing) and at the basin level (i.e. use of water accounting); (d) when irrigation systems are sufficiently

(a) water rights are unclear / insecure and thus little incentive to conserve; (b) water is priced as a social good; (c) poor monitoring of consumption, and transaction costs are high and more difficult to

Table 5. Summary of IMT Impacts

59



Poverty reduction and IMT

IMT can contribute to poverty reduction if (a)head-tail distribution improves; (b) the effects ofany increases in water costs are overcome byimprovements in efficiency or wateravailability; (c) it leads to increased productionand productivity for farmers of all income levelsas a result of expansion of cropping area, anincrease in cropping intensities, and cropdiversification; and (d) farmer organizations areable to diversify their non-farm income sources.

However, many researchers suggest a recenttrend in the IMT discussion that focuses moreon getting irrigation off the back of thegovernment than improving the lot of thefarmers and the poor. For instance, IMT isconsidered to be beneficial even if it saves thegovernment money and improves cost-effectiveness of operation and maintenance,while improving or at least not weakening theproductivity of irrigated agriculture. Somescholars suggest a re-examination of thisincreasingly hands-off government policy toirrigation management.23 A key issue concernsthe appropriate responsibility-sharingarrangements between farmer groups and thegovernment. In developing countries, small-scale paddy farming, for example, is a high-risk,low-return investment—crop insurance andcredit markets are thin, infrastructure is poor,imported inputs are costly and the overall costof farming is high. Farmers view IMT as a wayin which governments reduce subsidies andpass on the burden of food security fully tothem, while farmers in industrialized economiescontinue to be heavily supported by hiddensubsidies (Araral 2003).

Koppen and others (2002), comparing thedifferential impacts of IMT on poor and non-poor farmers, note that it has often been

assumed that the interests of the poorsufficiently overlap with the general interest inthe irrigation scheme, and that rich and poorfarmers alike have equal access to canal water.They point out, however, that evidence fromAndhra Pradesh and Gujarat, India, suggeststhe strong possibility of elite capture of the IMTprocess, particularly in large-scale canalirrigation, where land ownership and locationaladvantages are skewed in favor of biglandlords. They note that small farmers areoften unaware of the existence of the water userassociation, but it is they who often participatein repair and rehabilitation work, while thelarge farmers involve themselves in meetingsand committee work and makes decisions. InAndhra Pradesh, in particular, a higherproportion of small farmers depend on canalwater, but the concentration of their plots,mainly in the tail-end, poses a disadvantageregarding water accessibility. So the challenge ofmaking IMT pro-poor still remains.

IMT and institutional development

Institutional development is a common featureof IMT projects, but is a poorly understood task.It is associated more with blueprints related totechnical assistance consultancies, staff andbeneficiary training, study tours, and, financialand logistical support. A particular challenge ofinstitutional design for IMT is how to copeeffectively with heterogeneity among resourceusers, especially with respect to theirpredisposition to cooperate in the absence ofclear sanctions. Conflicting values and interestsbecome most severe when groups areeconomically and culturally heterogeneous intheir relationship to the resource; for example,head-enders and tail-enders in an irrigationsystem (Lam 1998, Tang 1992), and whenmembers differ in their degree of dependenceon the resource (Berkes 1992).



Shah and others (2002), studying thesmallholder IMT experience in Africa, suggeststhat “IMT faces problems in smallholdercommunities not because they are less able or lesscooperative but because the management cost of agovernment built irrigation system—like mostservice institutions—increases more rapidly with thenumber of customers than with the volume ofbusiness.” They explain that a 1,500-hectaresystem that serves 1,500 irrigators costs muchmore to manage—in terms of logistics or servicedelivery, fee collection, and maintenance—andfaces more difficult collective action problemsthan a similar 1,500-ha system that serves fivelarge-scale farmers. Moreover, it is a lot easierfor five large-scale farmers to come together toagree to the rules of self-management than for1,500 smallholders. They conclude that IMTworks better in cases of non-poor, market-oriented, large-scale, and business-likeagriculture, as in the cases of large farms in theUnited States, Mexico, South Africa, NewZealand, and Turkey.

Vermillion (1997) notes that “for several years,WUAs have been promoted both as a governing anda management body for irrigation systems.Community organizers have helped WUAs todevelop a constitution and by-laws, select leaders,approve plans and budgets and apply sanctions.WUAs then directly manage O&M and finances.This model is probably not well-suited formanagement at higher levels of larger systems or inmore complex management environments.Accountability between farmers and leaders,especially in finances, is often weak and WUAsgenerally do not have a professional staff. As a result,many conclude that IMT is more successful only atsmall scales of management.” Thus, on the onehand IMT appears to work best in large-scalecommercial systems; on the other, in developing

countries, with a large number of small farmersthere are many challenges in scaling up.

Maintenance

Maintaining irrigation systems under IMTremains a pervasive problem. The problemseems to be one of moral hazard and co-dependency—farmers need the irrigationagency for irrigation maintenance and theyknow the agency will bail them out because theagency depends, in part, on irrigation feecollections and because dilapidated systems areused as a justification for donor support.Panella (1999), for instance, suggests that in thePhilippines, a pioneer in IMT, the keyachievement of 20 years of IMT has beenimproved fee collection performance. This ismainly because collection efficiency has been amajor criterion in the performance evaluation ofirrigation managers (NIA 2003). But despitethis, NIA has not become truly viable. Over halfof its operating budget is covered by hiddendonor subsidies. Its emphasis on viability hasled to extreme cost cutting, severe under-investment in O&M, and a staff orientationfavoring collection at the expense of O&M (seealso Hayami and Kikuchi 2000).

In Indonesia, Johnson and Reiss (1993) report anincrease in deterioration of pump-set equipmentfor lift irrigation after IMT. This was attributedto lack of local knowledge, skills, parts, andcash, even though farmer motivation was high.The same problem of accelerated deteriorationwas reported in Senegal (Wester, During, andOorthuizen 1995). In surface irrigation,however, attempts to attribute improvements inmaintenance to IMT can be problematic in caseswhere IMT includes physical rehabilitation orrepair of irrigation infrastructure.

61



Various authors have suggested that the moralhazard problem can be addressed by makingWUAs assume future responsibilities forrehabilitation as a precondition for currentrehabilitation. Given the financial frailties ofmost small-scale WUAs and their very limitedcapacity to provide for capital expenditures, it isunlikely that this issue will be resolved easily.The experience in small-scale farming has beenthat farmers are even hard up for theiroperating expenses and are less likely to be ableto save for capital expenditures. Improving thequality of irrigation infrastructure—such aslining of secondary canals and use ofproportional diversion weirs—would helpensure that maintenance costs to the farmers arerelatively small. IMT therefore would have to bebundled with infrastructure improvements if itis to have a positive impact on maintenance.

IMT and water productivity

One of the underlying assumptions behind IMTprojects is the potential for improved waterproductivity as a means to poverty alleviation.Yet, cost-effective alternatives to improve waterproductivity—particularly the development ofwater-saving technologies and managementpractices such as varietal improvements, zerotillage, raised beds, alternate wetting anddrying, aerobic rice, and systems of riceintensification—are often ignored. Asexperience in the Philippines has shown (Sagun2002), significant gains in water conservationcan be achieved through the use of thesefarming practices in conjunction with IMT.

Poor quality of IMT research

Vermillion (1997) suggests that the quality ofmost IMT studies is poor. The problem withpoor quality research is that there is now limited

understanding of what really works and whatdoes not. While there have been many studies ofIMT, they are riddled with problems of a lack ofinternal and external validity. For example:

• Most studies do not provide for a balancedcore set of performance indicators; most arebiased in support of system efficiency andgovernment finance, and biased againstpoverty impact, equity, and environmentalissues. In particular, little is known whetherprice effects, where they exist, are lowerthan the efficiency effects, particularly forsmall farmers and tail-enders.

• Most studies excessively rely on irrigationagency data and are seldom validatedindependently.

• Survey respondents are seldom selectedthrough a systematic random sampling;most often, surveys rely on conveniencesampling, thus weakening the validity ofthe results.

• Most studies are qualitative descriptionsand do not employ rigorous statisticaltechniques.

• There are no before-and-after and with-and-without comparisons of findings.

• Assessments of financial impacts are oftenundertaken from the government’sperspective.

• Potential alternative causes of increasedcrop yields are often ignored, such asincreased fertilizer application rates, rainfallpattern, cultivation technology, prices,subsidies, and markets.

• Documentation of changes in productivityis often not linked to IMT.



Box 13 provides some recommendations forimproving IMT-related research.

Summary and Conclusion

This paper reviews recent IMT literature toanswer a number of questions. First, what arethe conditions most likely to result in

Box 13Recommended Improvements in IMT Research Methods

• Most studies on IMT focus on efficiency criteria and impacts on government finance. Wherever possible,studies should include a balanced core set of performance indicators, including financial performance, qual-ity of O&M, equity, poverty impact (particularly on agricultural productivity), and incomes and impacts onthe environment.

• Most studies rely heavily on agency offices. Future studies should avoid excessive dependence on agencydata. Where such data are used, corresponding measures should be sampled independently (through farm-er interviews) to validate the data.

• Most studies are prone to problems of self-selection of respondents. To enable generalization, farmers shouldbe selected through systematic random sampling, normally stratified according to location of fields relativeto irrigation head-works.

• Most analytical approaches suffer from internal and external validity problems. More extensive use shouldbe made of before-and-after interrupted time series research designs that include data for at least 3 to 5 yearsbefore and 3 to 5 years after to firmly establish the timing of impacts of transfer.

• Most IMT studies are single case studies and have limited potential for generalization. Given the difficultyof conducting detailed time series analyses in a large number of schemes, case studies should be comple-mented by surveys of 20 to 30 randomly selected schemes where a smaller amount of data is collected oncore performance measures.

• Most IMT studies measure only financial impacts to the government. Costs to the government and farmersshould be measured, as well as changes in total cost of irrigation.

• Where transfer programs include improvements in irrigation infrastructure, research should include with-and-without comparisons, with controls for effects of physical improvements and management reform.

• Given the importance of documenting the physical sustainability of irrigation systems after IMT, impactstudies should include direct observations of physical conditions and functionality of irrigation infrastruc-ture.

• Assessments of agricultural impacts (such as changes in crop yields) or economic productivity changesshould control for changes in fertilizer application rates, rainfall patterns, technology or cultivation practic-es, prices, market conditions, and subsidies. Statistical methods should be used to assess the relative impor-tance of different causal factors.

• Assessments of economic productivity should compare changes in cost of irrigation to farmers with changesin the value of agricultural output.

• Changes in agricultural or economic productivity should be more clearly linked to transfer by documentingthe nature and timing of effects of IMT on changes in policies and procedures for O&M and financing, andthen relating O&M performance outcomes to agricultural productivity.

sustainable local resource governanceorganizations? Are there specific conditionsunder which irrigation user groups work better?What do we understand about the impacts ofIMT in terms of poverty reduction, quality ofO&M, financial and personnel implications forthe government, and impacts on resourceconservation? And finally, what are some

63



specific issues that are neglected as a result ofIMT?

As is often the case, there are no clear-cutanswers regarding decentralized managementof irrigation and its impacts. However, thecurrent review suggests that IMT can contributeto poverty reduction if (a) head-tail distributionimproves; (b) the effect of any increases in watercosts are overcome by improvements inefficiency of water availability; (c) productivityincreases as a result of expansion of croppingarea, an increase in cropping intensities, andcrop diversification; and (d) revenue sources forfarmers organizations are diversified. Whetherthese conditions are met in most cases isunknown because of the poor quality of IMTresearch. IMT does appear to decrease thefinancial burden on governments throughreduced subsidies and improved collectionefficiency.

Serious questions about the sustainability ofirrigated agriculture remain. IMT clearlycontributes to improving operations, butmaintenance remains a significant problem. Thecurrent trend is that of decreasing investment inmaintenance. IMT can also contribute to water

resource conservation when certain conditionsare met, such as secure water rights, volumetricpricing of water, farm- and basin-levelmonitoring and accounting, and availability ofwater conservation technologies. Conservationis aided when IMT results in considerablyincreased flexibility over water use.

The emergence and sustainability of self-organized irrigation institutions depends on therelative size of the expected benefits and costsas perceived by the farmers. The marginalbenefits of self-management in terms of costefficiency, responsiveness, service reliability,and productivity need to outweigh marginalcosts associated with cooperation for successfulirrigation institutions to prevail. There are manychallenges that remain in designing effectiveirrigation institutions. Perhaps the mostimportant of these is ensuring that managementtransfer considers the issue of heterogeneityamong farmers and ensures that poorer farmersare aided and not hurt by decentralization. Thesecond major challenge is addressing the moralhazard problem related to maintenance. Betterempirical research to understand impacts anddesign issues will go a long way in meetingthese challenges.


Assessing DecentralizedForest Management

Introduction

Decentralization of forest management is acrucial component of community-based naturalresource management. The process ofdecentralization and devolution of rights andresponsibilities over forestry has moved alongwith considerable impetus, especially in the lastdecade. It has graduated from being asomewhat experimental strategy to beingintegrated into more conventional nationalforestry efforts (Arnold 2001). This acceptanceof decentralized forestry reflects its appeal totwo important stakeholder groups: those whobelieve that it can contribute to rurallivelihoods, and those who view it as a durablemechanism for managing vulnerable forests.

Decentralized forest management typicallyincludes two analytically distinctcompartments: decentralization at the localgovernment level and devolution of authority tothe community level. However, a great deal ofdiversity exists at each level, and, there isconsiderable overlap between these two formsof decentralization. As Schafer and Bell (2002)note, a wide variety of institutionalarrangements have been put in place and alarge number of projects initiated around theworld, primarily designed at placing moreauthority and control in local hands.

Given this context, it is appropriate to take stockand evaluate the progress made in the greater

involvement of communities in themanagement of forests. A brief perusal of theliterature reveals a tremendous variation in thedifferent forms, degrees, and approaches ofdecentralized management. This paper, which isbased on a review of case studies, presents acritical appraisal of decentralized forestry witha view to identifying emerging patterns, trends,and issues that would be of importance toscholars, practitioners, and policymakers alike.The paper primarily draws on peer-reviewedjournal articles from established developmentand forestry journals in the last three to fouryears. Several working papers have also beenreviewed.

The chapter is organized in the followingmanner. The first section presents a brief globaloverview of community forestry. The nextsection provides a detailed discussion on somecommonalities and differences among differentdecentralized regimes. Conditions that seem tocontribute to successful decentralization inforest management—and how these conditionsrelate to those prescribed by proponents ofcommon property resource management—arealso addressed.25 The relative scarcity ofempirical evidence precludes a thoroughassessment of the poverty and natural resourceimpacts of community forestry. However, this isaddressed in the next section, based onavailable information. A final section focuses onemerging challenges, bottlenecks, and policyissues.

Suranjan Weeraratne24

4



Community Forestry Worldwide

Decentralized management strategies in forestryare particularly prominent in South Asia, Africa(mainly Sub-Saharan), parts of East Asia, andCentral and Latin America. As mentioned at theoutset, there is a vast array of institutionalmechanisms by which decentralization takesplace. A brief global synopsis follows.

Data on trends in forest-cover change andownership are limited, and reliable data areeven more difficult to obtain. Based primarilyon information from 24 of the top 30 forestedcountries in the world, White and Martin (2002)estimate that (globally) some 4 percent of forestsare reserved for communities, while 7 percentare owned by communities. The community-access numbers improve when forests in onlydeveloping countries are considered. Some 22percent of the forests in developing countriesare either under community ownership or areunder some form of community management(White and Martin 2002). Thus, a significantamount of forest land in poor countries isactually under the control of communitygroups.

In South Asia, India and Nepal in particular areacknowledged to have advanced forms ofcommunity management. The National ForestPolicy of 1988 in India marked a radical shift ingovernment policy and resulted in the initiationof the Joint Forest Management (JFM) program,which made it possible for the forestrydepartment to involve people in themanagement of certain forests.26 By 2001, 22states had adopted JFM and some 45,000 groupswere protecting 11.63 million hectares (mha), ornearly 17 percent of government forests(Agarwal 2001; GOI 2001, in Kumar 2002; Whiteand Martin 2002).27 Under the terms of JFM,

Village Forest Institutions (VFI) are givenconditional access to specified forest products inaccordance with the guidelines laid by stateforest departments. In addition, communityforest management is also undertaken by self-initiated groups.28

Nepal also has a long history of communityforestry. Its decentralization effort received afurther boost with the implementation of the1993 Forest Policy, which led to thedevelopment of Forestry User Groups (FUG). By2001, over 9,100 FUGs were protecting 0.66 mhaout of a total 5.8 mha of forested land area(Bahuguna 2000, in Agarwal 2001). In Nepal,the FUGs are largely state-initiated, withcommunities participating at different levels.However, as in India, there is ample evidence ofself-initiated community groups engaged in theprotection of forests outside of the FUGs.

In Africa, more than 20 countries haveimplemented new forest laws since 1990 (Wily2002a). Increasing prominence is given tocommunity forestry in these laws. It isestimated that in 2002, there were at least 4,500rural communities in Africa involved in someform of forest management (Wily 2002a).29

Africa provides a diverse mosaic ofdecentralization patterns. These range fromdevolution to the lower tiers of government, asin Zimbabwe (Shackleton and Campbell 2001),to significant community empowerment(characterized by close to fully decentralizeddecision-making processes), as in some villagesin Tanzania and Gambia (Wily 2002a). Uganda,Madagascar, Burkina Faso, Malawi, and theNiger among others provide cases where lower-level Village Management Committees areformed to lead community management. Suchcommittees have differing levels of authorityand state involvement. Management

67

Assessing Decentralized Forest Management


agreements with the government were signed inthe Niger, MOUs in Uganda, and contracts inMadagascar to implement the process ofcommunity forestry (Bojang 1999).

Western Ghana provides an interesting exampleof community forestry linked to the privatesector.30 Ghana Primewood Products Ltd. andDalhoff Larsen & Hornemann implemented aninnovative program in the Gwira-Banso region,where they developed a joint harvesting andreplanting venture with local communities.Community involvement in commercialplantations has also gained ground in Malawiand Tanzania (Wily 2002a).

Wily (2002a) organizes the different forms of co-management found in Africa into fourcategories: (1) collaborative, in whichcommunities are consulted with minor roles andpowers; (2) contractual, in which communityroles are more substantial, but still inequitablein terms of wood licensing and enforcement; (3)consigned, in which communities have alloperational powers, including licensing andenforcement; and (4) community based, inwhich jurisdiction is fully devolved tocommunities and sometimes includesownership. Box 14 extends Wily’s framework toinclude a “consultative management” category.

Box 14Different Categories of Community Forestry Worldwide (Based on Wily’s Typology)

Consultative CFM: Community is consulted but has only minor roles and powers.• The Government Forest Development Agency of Ivory Coast has established 69 Farmer-Forest Commis-

sions as routes for local consultation (Wily 2002b).• A Community Forestry Management Unit established in 1992 in Ghana promotes CF Committees as a con-

tact point for consultation in forest reserve planning (Wily 2002b).• Gokwe communal area in Zimbabwe.• Decentralized forest regimes in parts of Latin America.

Collaborative CFM: Community is consulted to a higher degree and also possesses some management rights.But community involvement is largely restricted to user benefits; mainly involves joint forest managementand multi-stakeholder management.• Joint Forest Management in India.• In Zimbabwe, communities assist the Forestry Commission to manage Mafungabusi Forest Reserve and the

Pumula Forest Block. They protect mainly peripheral parts of the forests in return for agreed access to re-sources, mainly for fuel-wood and grazing.

• Zambia provides examples of multi-stakeholder co-management of local forests.• Community forestry program initiated in the Philippines.• Some forest regimes in Honduras, Guatemala, and Bolivia.

Community Based CFM: Jurisdiction is fully devolved and sometimes includes ownership.• Most visible in Tanzania and the Gambia. In Tanzania nearly 600 Village Forest Reserves have been declared

out of unreserved village lands, ranging from 5 to 10,000 ha, each actively managed and owned by thecommunity (Wily 2001). Prominent examples include the Duru Haitemba forest in the Arusha region andMgori forest in Singida region.

• In some cases, ownership is recognized but key management powers are retained by the state, as withcertain forests in Nigeria and South Africa. This is also often the case with regard to Community ForestryEnterprises in Mexico.

Source: Modified from L. Wily 2002a.



This is a somewhat simplified framework andmerely attempts to provide several broadclasses to fit the many worldwide examples ofcommunity forestry.

In Southeast and East Asia, the Philippinesstands out in particular in terms of community-based forestry, while reforms are taking place(albeit slowly) in Indonesia. The vast forestland of Indonesia is mainly under state control,with less than 1 percent of forests beingcommunity controlled (White and Martin 2002).However, there are many traditional“unofficial” community management systemsoverlapping with state regimes in Kalimantanand several other regions of the country. Recentlegislative reforms in the country have openednew possibilities for co-management and aretantamount to a tacit recognition of suchoverlapping (McCarthy 2000). Decentralizationis more advanced in the Philippines, though theDepartment of Environment and NaturalResources (DENR) has control and jurisdictionover the community forestry program (Gauld2000).

Mexico is well-noted for its decentralizedforestry and leads the way in Central and LatinAmerica. The level of devolution in Mexicanforests is almost unmatched anywhere else.According to some estimates, up to 80 percentof forests are in the hands of local communitieswith varying degrees of actual control (Bray andothers 2002). Mexico is a unique contrast to thesituation in most developing countries, wherecommunity forest management means themanagement of non-timber forest products(NTFP) or community woodlots for subsistenceuse. In Mexico, a significant number ofcommunities—organized as Community ForestEnterprises (CFE)—are engaged in the

management of common property forests for thecommercial production of timber. Currentestimates for the number of CFEs range from290 to 479 at various levels of consolidation andsustainability; most have been operating sincethe early 1980s (Alatorne 2000, in Bray andothers 2002). The 1986 Forestry Law formallyrecognized the rights of communities to formtheir own logging business (Klooster 2000a).Mexico’s progress in this respect is unrivalledand unprecedented, though commercial timberproduction based on the Mexican model (albeiton a smaller scale) is emerging in several otherLatin American countries, notably Peru,Guatemala, and Bolivia (Bray and others 2002).

Aside from Mexico, Bolivia, Honduras, andGuatemala have advanced decentralizedmechanisms in forestry and passed new forestlaws in 1990s. In these countries, authority isdelegated to local municipalities. As per the1996 Guatemalan and Bolivian Forestry Laws,the central governments of the respectivecountries devolved significant authority andfinancial incentives to municipalities toadminister public forests within their domains(Gibson and Lehoucq 2003; Andersson 2003).

Community Management: Trends,Commonalities, and Differences

A review of forestry case studies reveals afascinating diversity. This diversity is evident inthe choice of institutional mechanisms, benefit-sharing schemes, legal authority, and manyother aspects. The situational and contextualparticularities of most case studies makegeneralizations difficult. However, someconvergent themes and commonalities (withvarying degrees of differences) can be singledout. These themes are discussed in depth below.

69



Three major institutional mechanisms

The survey of the literature reveals that for themost part there are three primary institutionalmechanisms with regard to community forestry.These include devolution to local-levelgovernment, creation of new village-levelinstitutions, and allocation of new powers to anexisting traditional-leadership-based structure.It is hard to establish geographic trends andpatterns; all three mechanisms seem to bepresent in many different parts of the globe.

Devolution of authority to lower levels ofgovernment is evident in parts of LatinAmerica, especially Guatemala and Bolivia.Forest laws passed in the mid-1990s delegatedauthority to municipal councils in these twocountries. It is evident from the Bolivianexample that decentralization to localgovernment increased local power and theability to tax. The 1996 Forestry Law in Boliviagave municipal governments 25 percent controlover centrally collected royalties fromcommunity logging concessions (Andersson2003). In Lesotho, local government in the formof District Development Councils—subsequently named Rural Councils after theLocal Government Act of 1997— have beenwidely involved in issuing grazing permits,setting aside maboella or closed areas, andissuing fines for violations. They serve as a linkbetween the central government and localcommunities (Shackleton and Campbell 2001).Similar examples are seen in other parts ofAfrica, notably in Malawi, with respect to theDistrict Development Committees. AsShackleton and Campbell (2001) stress,decentralization to lower governmental levelsoften results in little community involvement inplanning or decision-making, and a largepercentage of the revenue is retained at thevarious regional/district/municipal levels.

In many countries, community-based forestmanagement is characterized by the creation ofnew village-level institutions. Village ForestInstitutions in India, Forest User Groups inNepal, and Community Forest Enterprises inMexico are among the more prominentexamples. It is rather difficult to make anassessment of the overall effectiveness of newcommunity-level organizations. Some, like theFarmer Forest Commissions in Ivory Coast,have resulted in limited managerial power andauthority for the community, while others suchas the CFEs in Mexico have yieldedconsiderably more autonomy to local people.

Another problem in assessing the impact of newcommunity-based institutions is that they co-exist with traditional collective actioninstitutions, sometimes building on existingtraditions and sometimes changing these tofavor new rules. The Van Panchayats in UttarPradesh (UP) and the Kangra Forest Co-operatives in Himachal Pradesh are two suchexamples, sometimes dating back to the earlypart of the 20th century. Sometimes JFM isimposed on existing self-initiated groups andthey are compelled to change their structuresand functioning to conform to JFM rules.31

In many cases, especially in Africa, traditionalleaders have established themselves asimportant stakeholders in community forestryand play crucial roles in their implementation.Quite frequently, there is evidence of parallel ordual local authority systems consisting oftraditional leaders on the one hand andgovernment-established structures on the other.The case studies on forestry cover a broadspectrum of involvement of chiefs and othertraditional leaders, from situations wheretraditional authority structures have beencompletely removed (parts of Botswana) tocases where these institutions exert considerable



control, as in Malawi (Shackleton and Campbell2001).

The Chimaliro Hills in Malawi present anexample of the often dominant role played bytraditional leadership. Malawi has always had along history of traditional structures and theeffectiveness of the village forest committeesdepended a lot on the influence of thetraditional leaders. The influence they carrymeans that the recognition of traditionalauthorities has become essential in most projectsand interventions in rural areas. This wasclearly in evidence in the case of the state-initiated Moribane Forestry Reserve inMozambique, where the local chief imposedseveral conditions on the Center for ForestryResearch and made use of every opportunity toassert his primacy in the area (Schafer and Bell2002). Another African example of traditionalhierarchies exerting strong power structures isthe Suledo Village Forest in Tanzania (Sjöholmand Luono 2002).

In Southeast Asia, case studies from Indonesiaand the Philippines demonstrate the existenceof traditional forms of community managementthat sometimes overlap with existing stateregimes. This was evident in a survey done inWest Kalimantan to assess people’sparticipation in forest management, whichunequivocally revealed the importance attachedby stakeholders to traditional forms ofmanagement (Colfer and Wadley 1996).

The choice of the institutional mechanism is avery important issue for community forestry.The optimal mechanism will depend on localpolitics; availability of resources, including localleadership; whether a new institution needs tobe created or not; and on whether there is anyprior experience with community forestry in thearea. A possible ideal solution is full devolution

to a local community organization with ongoingsupport from a local government orgovernment-supported agency.

Diverse benefit-sharing strategies

Classical forms of benefit-sharing strategies incommunity forestry include buffer zonedevelopments and employment opportunities.Buffer zones have the longest history and aredesigned to reduce local dependence upon theforest by providing alternate livelihoodopportunities and substitutes to forestresources. On-farm tree planting programsusually are included along with creditopportunities, and these often go hand-in-handwith environmental education programs (Wily2002b).

There is little uniformity to be found in aplethora of benefit/revenue-sharing schemes(sometimes even within countries). In India,most JFM state orders ensure participatingvillagers a 25- to 50-percent share of the netincome from timber on “final felling” of maturetrees. This usually entails a minimum of 5 to 10years of protection of the forests before thebenefits are reaped. For example, in the state ofWest Bengal, by 1998 poles worth Rs. 40 millionhad been harvested and 25 percent of the netproceeds shared with the members of theconcerned forest protection committees (Khareand others 2000). Recent amendments to theorders of two states, Andhra Pradesh (AP) andTamil Nadu, now entitle the partner villageinstitutions to 100 percent of the final produce.However, in AP’s case, at least 50 percent of theincome has to be reinvested in the JFM forest(Khare and others 2000). The benefits fromtimber are distributed both in cash and kind.

Malawi in East Africa presents more evidence ofdifferent village forest communities having

71



different benefit-sharing mechanisms. In theChimaliro Hills, for example, 70 percent of thebenefits go to the state, while the remaining 30percent goes to the community. In Mangweru,by contrast, the village council has full controlover revenue and none is claimed by the state.Evidence from case studies showed that manycommunities were plagued by inefficient andinequitable benefit distribution mechanisms.The San Martin Ocotlan community in Mexicois one such notable case (Klooster 2000b).

In addition to timber, fuel-wood, fodder, andnon-timber forest products (NTFP) are the othermain benefits from forests. In India, especiallyin areas where JFM facilitates the rapidregeneration of coppicing species like teakthrough community protection, it is believedthat increased availability of fuel-wood maybean early benefit prior to final harvesting.Availability of fodder grass is another earlybenefit, and is often the main incentive forparticipation in forest regeneration efforts inarid areas where livestock plays a crucial role inthe economy.

There is a common perception that villagershave free access to non-timber forest productsfrom JFM areas, which will provide them aregular flow of benefits in lieu of theopportunity costs they have to incur forprotection. However, the more valuable NTFPslike cashew nuts, bamboo, and fibrous grasseshave in the past been either excluded from freeaccess altogether or included under income-sharing agreements. The same also applies tocertain NTFPs that are nationalized (Khare andothers 2000). Recent policy reforms in Indiahave led to changes in NTFP revenue collection,and communities are likely to be benefitingmuch more than previously (Khare and others2000).

Many of the benefits from community forestryare derived mainly at the community level.Often funds derived are invested in communitydevelopment activities. In Nepal, income fromthe sale of timber, NTFPs, and other products isdirectly deposited in community funds andused for community development activitiessuch as the construction of school buildings andirrigation channels (Adhikari 2003). However,there are also several cases of benefits allocatedat a household level, yet in-depth informationon these cases remains elusive. The loggingcommunity of San Antonio in Mexico invested apart of its funds on new public buildings forthe village and also distributed 2,500 Pesos inprofits to each community member in 1994(Klooster 2000b).

From benefit sharing to power sharing

There are two main approaches to communityinvolvement in forest management (see Box 15).The first and still more prevalent strategyrecognizes local communities as forest users andseeks to secure their cooperation by grantinglegal access to certain products or a share inforest-derived benefits. The second approachmainly looks to forest communities as potentialmanagers or co-managers, and devisesarrangements to give them varying degrees ofmanagerial control and authority. Proponents ofdecentralization argue that the second approachwould encourage communities to govern moreeffectively and eventually lead to greatercommunity empowerment.

The user-centered paradigm predominates inthe Asian context. For example, managementpower mainly rests with the state in Indonesia.There is resistance to devolution of power fromthe forest bureaucracy and substantial influencefrom corporate interests wishing to retain



Box 15Authority and Rights: A Synopsis

control of forests and lands (McCarthy 2000).In India, substantial control over JFM resideswith the forest department. Under JFM, thelocal communities are given conditional accessto specified forest products. In most states, theVillage Forest Institutions (VFI) are registeredwith the forest department, with the latterreserving the right to dissolve VFIs orunilaterally cancel the JFM agreement in case

User-centric rights

Power-centric rights (includes user-centric)

� Right to access. In Zimbabwe, communities

assist the Forestry Commission to manage Mafungabusi Forest Reserve and the Pumula Forest Block. To the extent that they protect peripheral parts of the forests, they can access the same mainly for fuel-wood and grazing. This right seems to pertain more to protected forests.

� Right to share benefits with the State.

Benefit-sharing is with respect to timber and non-timber products. However, benefits are often restricted to specified products and contingent on conditionalities, as in the case of JFM in India, Mozambique (Tchuma Tchato), and Kenya (Arabuko Forest Reserve).

� Right to lease land. Used in the community

forestry program in the Philippines, this is a mechanism to transfer user rights for a period of time. Under the community forest management agreement, land is leased out for 25 years and is renewable for another 25 (Gauld 2000).

� Right to monitor use. Many community

programs have some form of a monitoring system. These range from community patrols to employing guards (India, Nepal, many African countries).

� Right of enforcement on violators. Often

these rights are also retained by the state. However, exceptions do exist, as demonstrated by some cases in Mexico and with regard to national forests in Tanzania and the Gambia.

� Right to form community logging business.

This is predominantly seen with the Community Forestry Enterprises in Mexico, with emerging trends elsewhere in the region, including Peru, Bolivia, and Brazil.

� Right to legal recognition and tenure

security. In Nepal, forest user groups are legally recognized. In Africa, most countries have pushed forward land reforms; legal recognition is best exemplified by Tanzania.

� Right to Ownership. Tanzania and the

Gambia are the standout examples. In some cases, ownership is recognized but key management powers are still with the state (Mexico, South Africa).

� Right to sell forests and convert to non-

forestry uses. In 1992, an amendment to article 27 of the Mexican constitution enabled Ejido farmers to sell forests or seek joint ventures with private investors.32 However, in many other cases, these rights are retained by the state.

the VFI is thought to have violated any of itsterms (Khare and others 2000). The exceptionsto this are the states of Gujarat and Haryana,where the VFIs have an autonomous status. InGujarat, for example, these institutions have anindependent existence as cooperativesregistered under the Co-operative Society Act(Sundar 2000).

73



In Nepal, the FUGs have varying degrees ofmanagement authority and seem to embraceelements of both the power-centric and user-centric paradigms. But, the right to sell forestsand the right to convert them to non-forestryuses resides with the state, i.e. some crucialelements of negotiating power remains with thestate (Chakraborty 2001). The last few yearshave seen several policy changes passed (orattempted) by the central government in Nepalthat are unfavorable to community management(Shreshtha 2001; Britt 2001). In 1997, seniorforest officials started a process to curtail themandate of community forestry. These includedraising taxes on community forest revenue, andrestricting the maximum forest area perhousehold to 1 ha in the hills and .25 ha in theTerai. This was followed by two amendments tothe Forest Act Bill in 1998 and 2002 that soughtto reduce user rights (Baginski and Blaikie2003).

The user-centric approach is less entrenched inMexico and in parts of Africa. In Mexico, itseems in many cases the ownership of forestsrests with the communities. They also influenceand implement logging plans. However, thegovernment is in charge of the overallmanagement framework (Klooster 2000b).

Wily observes that in Africa, more successfulinitiatives in community forest managementhave moved away from a user-centric approachto a more power-centric approach. The power-sharing mechanism seems most advanced inTanzania and the Gambia. They have led theway in providing programs and supportinglegal processes that directly encouragecommunities to bring currently unreservedforest areas under their own jurisdiction ascommunity forests and to apply to managenationally owned forests (Wily 2002a; Bojang

1999). After a series of user-centered projects inMozambique, new initiatives have beenlaunched to help local communities establishmanagement regimes that are community-driven and sustained, and where possible toroot these in recognition of local tenure (Anstey2000, in Wily 2001). In Uganda, the content ofjoint forest management agreements has movedsharply since 1998 from division of usufructrights and responsibilities into arrangementsthat designate local communities as managers.In Namibia, the demarcation of four new stateforest reserves gave way in 1998 to their re-designation as future community forests to beowned and managed at the local level (Wily2001). However, Wily (2002a) cautions that atpresent less than half of the community forestsin the continent enable the community to berecognized as owner-manager and to managethe forest in largely autonomous ways. Theremainder either limits local tenure orcircumscribes local jurisdictions in some criticalway. Often the state retains most or all control oflicensing and enforcement, meaning thatbringing offenders to book remains essentiallywith foresters.

Issues of legal recognition and land tenuresecurity are closely allied with authority andcontrol. The community forestry program in theTerai region in Nepal owes much of its successto the fact that local forest legislation hascreated a legal basis for the forest-user groupsthere. Tanzania and the Gambia are primarilyfocused on transferring legal ownership ofunreserved forest lands to local communitiesand provide ample testimony of the benefits oflegal recognition. In many ways, the Tanzaniancase is unique. The Village Forest Councils inthat country are considered autonomous legalentities able to sue and to be sued. The rulesagreed upon by the village committees are



passed as village by-laws (Wily and Dewees2001). Land reforms have also taken place inmany other African countries—includingUganda, Mozambique, Swaziland, and SouthAfrica in the last decade—with an emphasis onimproving land tenure of the local communities.These new provisions have resulted indecreased powers of governments in forestreserves, upgrading of customary land tenurerights, and limited the powers of states toappropriate more customary lands (Wily 2002a).

While tenure security is very important forsuccessful CPR institutions, there are severalexamples worldwide that show that it is notalways a sufficient or even necessary conditionfor success. The complexity of forestrychallenges—even with land tenure security—isdemonstrated in the case of the YaneshaForestry Cooperative (COFYAL) in the PalcazuValley in Peru. The Yanesha had legalrecognition and had received titles to their landunder the 1974 Law of Native Communities inPeru (Morrow and Hull 1996). Yet, in spite ofthe legal recognition, the Yanesha did not feelsecure about their land as they were subject toencroachment from constant migration.33 Underthese circumstances, the Yanesha expressed aprimary interest in the short-term profitability(immediate cash returns from the resource) ofthe forest. Morrow and Hull conclude that thisemphasis on short-term profits may have beenmotivated by feelings of insecurity about theirability to protect their lands from outsideencroachment and the lack of adequateincentives for long-term protection of theforests.

At the same time, there is evidence from thePhilippines and parts of South Asia and Africawhere indigenous groups have had their ownnatural resource management systems on state

land. In other words, there are also exampleswhere the lack of tenurial security did notcoincide with a lack of interest in environmentmanagement. The villagers, even in the absenceof any kind of formal recognition, stillundertook to protect the forests as they realizedthat as primary stakeholders hugely dependenton forests, it was in their best interests to engagein conservation.

From degraded to conservation forests

A distinguishing pattern of early communityforestry was the restriction of local roles to areasof lesser biodiversity or tourist interest, limitedcommercial importance, and also to forestsclassified as “degraded.” However, as Wilyobserves from her study of Africa, mostgovernments seem to show an increasingproclivity to place well-stocked forests andprotected areas under community control (Willy2002a). This is particularly evident in Sub-Saharan Africa, though progress has also beenmade in several other parts of the world.

Uganda, South Africa, and Ethiopia have beguncommunity forestry in official forest reserves,including those of the highest conservationpriority. In Tanzania, community forestrycurrently operates in both protection andproduction reserves and unreserved forestareas, and covers all forest types from moistmontane, dry woodland, and coastal mangrovesto commercial plantations. The country isbelieved to possess the most expansivecommunity forestry program in Africa (Wilyand others 2000). In fact, Wily notes that inTanzania “in principle no forest is considered toolarge or too small, too valuable or too degraded tocome under community based management.”However, it has to be noted that though mostAfrican states decreasingly draw lines as to

75



where communities can be involved, restrictionsstill exist in some countries.34

A singular feature of the initial JFM program inIndia was its restriction to degraded forests.Approximately 40 percent of the country’s totalforest area is considered to be degraded (havinga crown cover of less than 40 percent andassessed by satellite imagery),35 and thuseffectively 60 percent of the total forest area inthe country was precluded from JFM inaccordance with the original guidelines.However, a Government of India resolution in2000 recommended that JFM be extended towell-stocked forest areas (Kumar 2002). Nepalfollows a similar pattern, with communityforestry mainly restricted to degraded forests,though there have been several instances ofwell-stocked forests handed over to Forest UserGroups (Agarwal 2001).

Enabling Environment

There is a copious amount of literature thataddresses conditions needed for successfulmanagement of common property resources(Wade 1988; Ostrom 1990; Bromley 1992;Keohane and Ostrom 1995).36 However, asurvey of the literature on forest managementreveals that in many cases, the communities onthe ground rarely display these qualities.Communities are highly differentiated withmultiple interests and actors, each of whom caninfluence decision-making in different ways. Onthe basis of the evidence gathered, a few moreapparent conditions for the success ofdecentralized forestry are discussed below.

Changes in the policy and legal framework

It is apparent from the case studies that anenabling policy and legal framework is

imperative for more effective decentralizedforest management. Progress has been made,especially in Africa, in promulgating new forestlaws in more than 30 countries since 1990 (Wily2002b). Common changes in policy in these newlaws include “changes in the character of centralforestry administrations, with wider civil societyinput in decision-making, sometimes relocation offorestry departments into semi-autonomousinstitutions, and variant degrees of decentralizationto local governments and policy commitment andnew legal opportunity for forest-local populations toparticipate in forest management” (Wily 2002a).Tanzania and the Gambia have led the way insupporting legal processes that allowcommunities to bring currently unreservedforest areas under their own jurisdiction and tomanage nationally owned forests (Wily 2002b;Bojang 1999).

Changes in legal and policy frameworks havehad an impact on other parts of the world too.For instance, the 1986 Forestry Law in Mexicoofficially recognized community rights tologging and led to the establishment of theCommunity Forestry Enterprises (CFE). TheForest Policy of 1988 in India was distinct frompreceding policies in prioritizing conservationand local subsistence needs over revenuegeneration for the state. And, as noted earlier,changes in forest legislation created a moresecure legal basis for the Forest User Groups(FUG) in the Terai in Nepal.

Land tenure reforms are also important increating space for community forestry. Again,Africa has made considerable progress in thisrespect. Reforms in many countries have led tomore rigorous constraints on the use of theroutine right by governments to appropriateland for public purposes, including the creationof Government Forest Reserves (Wily 2002b).37



However, as noted earlier, the existence oftenure is not always a good indicator of success.In general, policy change and policyimplementation tend to be an incrementalprocess, with some reversals (for example, inNepal).

Effective community demand

To reap maximum benefit, community forestmanagement has to be driven primarily bycommunity preferences. Communities have tobe convinced of the utility of involvedparticipation. This often implies that thebenefits to be derived from devolution ofpowers have to be in excess of the estimatedcosts. Examples of high returns to communityforestry abound, and this possibly explains therapid spread of this mechanism in the 1990s (seesection on Impacts). However, collective actionis not a given, and transaction costs can be asignificant burden. The failure of the YaneshaForestry Cooperative in Peru was due toincreasingly high transaction costs and thefailure of the cooperative to yield expectedbenefits (Morrow and Hull 1996). Hence, a highbenefit/cost ratio would facilitate greatercommitment and involvement from thecommunities.

Strong monitoring and enforcementmechanisms

Effective mechanisms in monitoring andenforcement are vital for efficient communityforest management. Patrolling is a widely usedmechanism of monitoring. For example, inBehroonguda in India, patrolling duty ismandatory for everyone in the village; failure tocomply results in a penalty of Rs. 100. A personwho does not show up for patrolling duty threetimes is removed from membership of the VSS

(D’Silva and Nagnath 1999). This kind ofeffective night patrolling by villagers inBehroonguda accounted for the seizure of illegaltimber worth Rs. 234,311, and imposition ofpenalties amounting to Rs. 63,343 by 1998(D’Silva and Nagnath 1999). Monitoring ofparticipation is also an effective way of buildingconfidence among the community. This wasevident in the San Antonio community inMexico, where regular, well-attendedcommunity assemblies were a common feature.The community’s 71 adult men share anobligation to participate; those who do notattend are fined with a day’s communal work(Klooster 2000b).

Effective reporting is also a crucial ingredient inmonitoring. This was present in the Suledo case,where the patrollers made sure informationreached the right people. Accounting andreporting practices that provided members withhealthy flows of information were also evidentin the San Antonio and UZACHI communitiesin Mexico (Klooster 2000b).

There is considerable variation in terms ofenforcement strategies used. These includemostly fines and in some cases jail sentences (aswas seen in Isagala village in Tanzania).Graduated sanctions (as enforced in the Terairegion in Nepal and parts of Africa, India, andMexico), where the frequency and severity ofthe offense is taken into account, seem to bequite effective (see Box 16). Enforcementmechanisms are found under both community-and local-level jurisdiction. In Bolivia andGuatemala, such mechanisms are devolved tothe local municipalities. Sometimes the absenceof adequate legal provisions for the communitycan be problematic for meting out suitablepunishment for transgressors, as referringoffenders to the relevant authorities can oftencause delays and bureaucratic wrangling.

77



However, there are also instances whereenforcement mechanisms were primarilydevolved to the community level. This isespecially evident in Tanzania, both in respectto the management of community-ownedVillage Forest Reserves and national forests. Inthe Tigray woodlots in Ethiopia, violators arepunished by community-controlled local courts(Gebremedhin and others 2002).

It is crucial that monitoring and enforcement iseffectively implemented. The adverse effects ofinadequate implementation are illustrated in thecase of Bolivia. There, the central governmentdisbursed funds to municipalities for theprovision of forestry services.38 However, therewas also no provision for the withdrawal ofcentral government funds in the event of non-compliance by municipal authorities.

Consequently, the mayors in thosemunicipalities where government monitoringand supervision was not stringent did not havesufficient incentives for implementation.Andersson notes: “If benefits can be obtained by anactor without a contribution, a temptation alwaysexists to free ride on the efforts of others. The actor isnot always motivated to contribute to the collectivegood. If municipal government officials notice thatone of their principals, the SuperintendenciaForestal, keeps sending them checks withoutmonitoring the quality of the municipal forestryservices, municipal officials have an incentive toshirk or even produce no effort at all.”

Complementary role of external organizations

In spite of the presence of a fairly large numberof self-initiated groups, community forestry for

Box 16Features of Effective Monitoring and Enforcement

Communal Participation and Patrolling—a carrot-and-stick approach to ensure participation.• Patrolling in Suledo village forest in Tanzania (Sjoholm and Luono 2002) is mostly undertaken by young

men. Though salaries are not paid, they are exempted from other village development initiatives and some-times received a percentage of the fines collected.

• In the Terai region in Nepal, some user groups employed chowkidars (watchmen) with salaries ranging fromRs. 200–1,200 per month. In general, community forestry members paid these salaries. In some cases theforest department paid salaries (Chakraborty 2001).

• Assemblies in the forests of Ixtlan, Mexico, vigorously debate decisions that affect every aspect of communi-ty forestry. Those who do not participate regularly receive fines, which are deducted from their share of theforestry profits (Klooster 2000b).

Fines and sanctions—effective when the frequency and severity of offenses are taken into account.• In the Mahila Upakar user group in the Banke district of Nepal, grazing fines for goats were Rs. 5 for the first

time; Rs.10 for the second; and Rs. 15 for the third time (Chakraborty 2001).These fines were substantial,considering that the average agricultural daily wage levels in the Banke district were Rs.20-30 for womenand Rs.50-60 for men.

• In the village of Isagala in the Shinyanga region of Tanzania, a traditional law known as Mchenya is appliedby punishing everyone who does not abide by the management plan for the village forest. The fine startswith one bull, which is valued at (in 2002) between $50 and $100, but the fine increases to more bullsdepending on the kind of infringement and the frequency of infringement (Mlenge 2002).

• In the Lohgarh Hill Resource Management Society (HRMS) in Haryana, India, strict rules were designed toensure that bhabbar was not cut with fodder grass by the villagers. Household fines were imposed for first-time offenders; these were upgraded to a withdrawal of harvesting rights of grass for those who trans-gressed a minimum of three times (Hobley and Shah 1996).



the most part is not initiated by the communityitself. In Africa and in South Asia, the initialimpetus for community forest management hasmost recently come from the government,though this may well be in response tocommunity and NGO action. NGOs, donors,and other social and community groups play avital role in facilitation, mediation, and advisorycapacities.

Donors have played a critical role in facilitatingcommunity forestry and building the capacity oflocal organizations. For example, donor supporthas been instrumental to the expansion ofcommunity forestry in India. The World Bankhas supported JFM programs in the states ofMadhya Pradesh, Andhra Pradesh, UttarPradesh, Kerala, West Bengal, and Maharashtra.The Department for International Developmentsupports JFM programs in Himachal Pradeshand Karnataka, and the Overseas EconomicCooperation Fund of Japan supports a large-scale JFM endeavor in Rajasthan (Khare andothers 2000). Donors have also put pressure onthe Government of Nepal to reverse some recentpolicies thought to be unfavorable tocommunity forestry.

Donor projects often contain conditionalities onwhich the aid is contingent. This was evident inthe Philippines, where the United States Agencyfor International Development (USAID) and theAsian Development Bank imposed stipulationsupon the government. The release of donorfunds for community forestry was contingentupon the government introducing policyreforms promoting market efficiencies andcompetitiveness (Gauld 2000). This resulted in aforestry program that emphasized technicalproduction rather than one that highlightedcommunity empowerment and equitabledistribution of forest resources (Gauld 2000).Thus, outsider-driven initiatives can be fraught

with problems, stemming from multiplestakeholder agendas. Failure to take intoaccount contextual, institutional, andtechnological issues can hinder successfulimplementation. This was starkly illustrated inthe case of the USAID- and WWF-fundedYanesha Forestry Cooperative in Peru, wherethe imposition of complex rules and the lack ofattention paid to the social customs of theYanesha stymied the smooth functioning of theoperation and eventually led to its termination(Morrow and Hull 1996).

Nongovernmental organizations are anotherimportant actor in decentralized forestmanagement. In the Gokwe communal area inZimbabwe, state-controlled ResourceManagement Committees were viewed withsuspicion by some local communities and evenconsidered spies of the state. In this climate ofdistrust and suspicion, NGOs helped todissipate tension and facilitated talks betweencommunities and the state (Shackleton andCampbell 2001). In Chimaliro Hills in Malawi,the Wildlife Society of Malawi was instrumentalin mediating access to forests, and providedtraining, marketing, and back-up support tomake beekeeping in forests a success(Kayambazinthu 2000). Similarly, in Mexico(notably San Antonio), NGOs assistedcommunities in improving accounting,reporting, and auditing practices (Klooster2000b). In India, there are examples ofcommunity forestry being initiated by NGOactivity. In other parts of the world, such asBolivia and Guatemala, NGOs have exertedpressure on the local municipalities to investmore resources into the provision of communityforestry services. Thus, NGOs provide crucialassistance in three areas: (1) mediating betweenthe state and distrustful communities; (2)initiating community forestry programs and

79



building community capacity to undertakethese programs; and (3) acting as watchdogsand ensuring that the state assists and enablescommunity participation.

Impacts of Community Forestry

In spite of an abundance of academic literatureon community management of forests, in-depthempirical evidence that addresses impacts (onnatural resources, poverty, and empowerment)is infrequent. Most of the scholarly material hasfocused on the evolution and the modusoperandi of community forestry and not on itsaftermath. This relative paucity of informationon the impacts of community forestry makes athorough assessment difficult. In order todeepen our understanding of impacts, weconsider four sub-issues: (1) household andcommunity benefits; (2) short-term versus long-term benefits; (3) gender inequalities anddistributional effects; and (4) natural resourceimpacts.

Household and community benefits

Examples abound of benefits being acquiredboth at household and community levels.Household benefits include the distribution ofmonetary rewards from income generated bycommunity forestry and the creation ofemployment opportunities. Furthermore, inaddition to timber, fuel-wood, fodder, and non-timber forest products (NTFP) are the othermain household benefits from forests.

Evidence in the Behroongudan case study inAndhra Pradesh shows that income generatedfrom NTFPs and forest-related employmentschemes (such as coppicing shoots and singlingwork) accounted for nearly 43 percent of theaverage family income in 1998 (D’Silva and

Nagnath 1999). In 1998, villagers received Rs.359,500 from the sale of 3,198 teak poles thinnedfrom 100 ha as part of silvicultural operations.Further, wages offered by the forest departmentfor these works (averaging Rs. 40-50 a day)were better than the agricultural wage of Rs. 25for a female and Rs. 30 for a male. In 1998 alone,3,656 days of employment were generated,equivalent to over two months of full-timeemployment for the 68 people who wereinterested in and available for forest work(D’Silva and Nagnath 1999). Since the advent ofJFM, the villagers of Behroonguda have stoppedmigrating to nearby towns and villages insearch of work, as employment is now locallyavailable.

Income acquired from the collection andprocessing of palm kernels in northeasternBrazil was shown to account for 39 percent ofcash income and 34 percent of total householdincome during the seasonal slack period inagriculture. Many of the poorer farmers weredependent on this cash for purchasing seed andother inputs required for planting in the newseason (May and others 1995, in Arnold 2001).In western Niger, it was found that “income fromforest products from the commons rose as a share ofhousehold income from 2 percent in the harvestseason to 9 percent in the hot and rainy seasons and11 percent in the cold season” (Hopkins and others1994, in Arnold 2001). The authors alsoobserved that cash income from these sourceswas sufficient to purchase between 9 and 28percent of the households’ annual caloric needs.

Logging and employment are key benefits fromcommunity forestry in Mexico. The loggingcommunity of San Antonio in Mexicodistributed 2,500 pesos in profits to eachcommunity member in 1994 (Klooster 2000b).The Community Forest Enterprise (CFE) of El



Balcon on the Pacific Coast north of Acapulcogenerates around 250 full-time and part-timejobs and has fixed capital assets of over $4million, not counting the natural capital of theforest (Bray and others 2002). The Oaxacacommunity operates a CFE that offers part-timeemployment to nearly a quarter of thecommunity. The El Rosario de Xico communityin Veracruz, Mexico, offers year-roundemployment for all of the 24 communitymembers (Merino 1997b, in Bray and others2002). In the San Antonio community, work is“easily available for those who seek it, with rotationwhen necessary” (Klooster 2000b). As Kloosterfurther notes, there is a clear attempt here toincrease local employment opportunities bypaying for work in reforestation, roadmaintenance, and clearing the debris left bylogging.

Examples from the Gwira-Banso region ofGhana, Chimaliro Hills in Malawi, and theSuledo village forest in Tanzania showcase thepotential of beekeeping as an additional incomegeneration opportunity for the villagers(Appiah 2002; Shackleton and Campbell 2001;Sjöholm and Luono 2002). The richness of theseforests in a variety of flowering plants andreliable water sources make them conducive tobeekeeping. It offers a source of livelihood forthe poor who can use honey for making andselling local brew and for medicinal purposes(Sjöholm and Luono 2002).

Often, funds derived through communityforestry are invested in communitydevelopment activities. As Bray and othersobserve, the post-CFE period in Mexico hasenabled “the communities to use profits toinvest in the enterprise and to build communityassets such as potable water networks, schools,clinics, public buildings, and social servicesafety nets in the form of free medical care andold-age pensions (almost unheard of in rural

Mexico) and fulfill functions left unattended bythe government” (Bray and others 2002). InUZACHI communities in Mexico, profits havebeen spent on collective goods such as roadimprovements, an auditorium for communityassemblies, and the construction of churches(Klooster 2000b). Nuevo San JuanParangaricutiro has established two saw mills,kilns for drying lumber, a furniture factory, aresin distillery, and also started a program ofagricultural diversification (Lemus 1995, inKlooster 2000b). In Ixtlan, forestry proceedshave enabled diversification into transport,agricultural promotion, and ecotourism(Klooster 2000b).

In Nepal, income from the sale of timber,NTFPs, and other products is directly depositedin community funds and used for communitydevelopment activities such as the constructionof school buildings and irrigation channels(Adhikari 2003). Similarly, in Gambia, incomegenerated by the community forest goes into alocal fund. At least 40 percent of this fund isused for forest rehabilitation activities, while theremaining 60 percent is used for villageinfrastructure development (Sonko andChamara 1999). In the Mwamishali village inthe Shinyanga region of Tanzania, the sale ofproducts from the community forest reserve hasbeen utilized to build four classrooms, two pitlatrines, and a head teacher’s office at a totalcost of 4,200,000 Tsh (Mlenge 2002). Theseexamples show that the returns from commonproperty management are often invested insupporting local public goods.

Short-term versus long-term benefits

A major challenge in community forestry isensuring that there are short-term benefits tocommunity management. Often, communityforestry can result in closures of forests for

81



conservation and long-term sustainability. Thiscan impose significant immediate costs oncommunities. Our review of cases across theworld suggests that there are examples ofreturns in the short term.

In Mexico, the transition from concessionlogging to community logging translated intoimmediate profit benefits for the localcommunities. The forest management systemunder the concessionaires was dysfunctional,with widespread timber smuggling and loggingwastes. In the village of San Martin Ocotlan, forinstance, independence from concessionairesresulted in immediate profit increases of up to600 percent, even after doubling the wages ofcommunity-member loggers (Klooster 2000b).Thus, Mexico presents a good example of howcommunity forestry can result in immediateprofits because of improvements in governanceand consequent changes in the distribution ofbenefits.

In India, especially in areas where JFMfacilitates the rapid regeneration of coppicingspecies like teak through community protection,it is believed that increased availability of fuel-wood may be an early benefit prior to finalharvesting. In Gamtalao Khurd village inGujarat, nearly 12 tons of fuel-wood washarvested during forest cleaning operationsafter just one year of protection (Arul andPoffenberger1990, in Khare and others 2000).Availability of fodder grass is another earlybenefit and is often the main incentive forparticipation in forest regeneration efforts inarid areas where livestock plays a crucial role inthe economy. In Kaliakua village in Gujarat,where a JFM program has been implemented,villagers harvested Rs. 494,000 worth of grass in1997–98 (Prasad 1998, in Khare and others2000). In some Van Panchayats in Naini TalDistrict of Uttar Pradesh, the grass yielddoubled after a year of protection (Mansingh1991, in Khare and others 2000).

Box 17Methods Used to Generate Short-Term Benefits in the

Buldhana Forest Division in Maharashtra India

The Buldhana forest division is situated in the north-central part of Maharashtra. The JFM program was offi-cially implemented here in 12 villages in the summer of 1996. Encouraged by the success of these 12 villages,within one year almost 100 more villages in the Buldhana division chose to adopt the JFM strategy, many ofthem as self-initiated groups. Part of this success was due to three main avenues of immediate income gener-ation that were implemented in the Buldhana forest division (Ghate 2000):

§ The Forest department undertook forestry work like a plantation, transporting dead wood from inside for-ests, and carrying out activities of coppicing, singling, and dressing of tree stumps in the forest area adjoin-ing these villages, thus resulting in the generation of wage income.

§ Forest Protection Committees took up the collection and marketing of Anjan (good fodder) on a large-scalebasis. These are found copiously in the surrounding forest area and are often in demand in the Buldhanaand Khamgaon markets. The persons involved in collection were given fixed wages (Rs. 60 per day), whilethe profits were deposited in the committee’s account. This activity fetched immediate income for the house-holds. The amount that went into the committee’s account was used for executing several developmentalschemes.

§ The FPCs encouraged self-employment, small business ventures, and cooperative dairy development. Com-mittees sought loans from cooperative banks to buy their own trucks to transport forest produce to themarket. Some bought cattle and engaged in dairy business privately, while some committees initiated coop-erative dairies with their own funds.



Evidence from the Ethiopian woodlots in theTigray region shows few significant short-termbenefits, due to the temporary closure of foreststo protection (Gebremedhin and others 2001).This seems indicative of a larger pattern inseveral other places. In some cases, the lack ofimmediate benefits makes it difficult to ensurethe participation of the villagers. Yet, as wasportrayed in the Tigray region in Ethiopia, if thevillagers are convinced about the long-termviability and profitability of the exercise, theywill still participate in forest protection. Theaverage estimated value of current benefits wasEB 2783 per woodlot in villages where benefitswere received (Gebremedhin and others 2001).However, it was estimated that a typicalwoodlot of eucalyptus trees would be worthmore than EB 80,000 per ha on average, ifallowed to grow, and that the eventual wealthof communities in the Tigray region would bemore than EB 5 million per community onaverage (Gebremedhin and others 2001). Thus,the villagers were convinced about the long-term wealth of community forests and that theywould benefit eventually, and only a smallfraction expressed uncertainty about futurebenefits as a problem.

Gender inequities and distributional issues

The often peripheral role played by women is acommon feature found in much of the literatureon community forestry. State-initiated groups inSouth Asia—Village Forest Institutions in Indiaand Forest User Groups in Nepal—broadly havea two-tier organizational structure consisting ofa general body and an executive committee. It isestimated that women constitute less than 10percent of JFM general bodies (Agarwal 2001).The situation is similar in self-initiated groups,where the customary exclusion of women fromvillage decisionmaking bodies is replicated. In

the FUGs in Nepal, the unit of membership isthe household, and often it is only the man’sname that is included on the membership list(Seely 1996, in Agarwal 2001). In general,women’s representation in the executivecommittees is also very low. The Government ofIndia passed a guideline in 2000 stipulating that50 percent of the membership in the generalbody in all JFM endeavors should be women(Khare and others 2000), but it remains unclearhow much of an impact this has really had.

Despite the general trend of low femaleparticipation in formal groups, there areexceptions in South Asia, including user groupsmade up exclusively of women or mixed groupswith a high female presence. All women usergroups are found primarily in the hills of UPand parts of Nepal. In Nepal, all women FUGsare said to constitute 3.8 percent of all FUGs(Government of Nepal 2000, in Agarwal 2001).User groups with high levels of women are alsofound in parts of UP, AP, and Nepal.Bheroonguda village in Andhra Pradesh isparticularly noted for high levels of femaleparticipation and empowerment and boasted afemale president for five years (D’Silva andNagnath 1999). Africa followed a similarpattern, with high female participation beingthe exception rather than the norm. The Suledovillage forest in Tanzania was a prominentexception, with half of the membership of theenvironmental membership committee requiredto be women (Sjöholm and Luono 2002).

In addition to gender inequalities, discrepanciesin benefits distribution appear to be a commonfeature, though this is an area that needs furtherresearch. Generalizations are particularly hardto make on poverty impacts due to the minutenumber of specialized studies done on thesubject. Adhikari’s (2003) work in Nepal revealsthat poorer households in forest-dependent

83



communities earn less from community foreststhan middle-income and rich households.Adhikari’s study shows that the average “poor”household obtains NRs. 7,756 from communityforestry, while the more “rich” householdsobtain an average of NRs. 24,466 per year. Thepreviously cited example of Western Nigershowed that with regard to cash income fromforest products, the poorest third of householdsin the study were more dependent on thissource of income than the richest third, andwomen (for whom it represented 27 percent oftheir income) were more dependent than men(for whom it represented 10 percent) (Hopkinsand Others 1994, in Arnold 2001).

Distributional discrepancies are also evident inSan Martin Ocotlan in Mexico. In 1995, 28percent of the workers earned just 3 percent ofthe income at the lowest end of the scale; 34percent of workers were in the next categoryand accounted for 15 percent of the totalearnings. In other words, in San Martin Ocotlanin 1995, more than 60 percent of the work forceearned approximately 18 percent of the totalwages. This was in stark contrast with the 7percent of the workforce at the top end of thescale (earning between 5000-10,600 pesos), whoearned more than one third of the total wages(Klooster 2000b).

Positive natural resource impacts

There is demonstrable evidence in quite a fewcases of community forestry resulting inpositive natural resource impacts. Based on theevidence in the literature, it seems conceivableto visualize a positive trend in general.

Tanzania is perhaps the best example of havingpositive natural resource impacts in Africa.Especially in the Miombo woodlands of DuruHaitemba and Mgori, there are visible signs of

gain. Charcoal burning, rampant timberharvesting, and unregulated in-forestsettlements have all disappeared. This hascoincided with a return of understory shrubberyand grasses and the return of bee swarms toforests (Wily 1999). Other positive impacts inDuru Haitemba and Mgori include fewer areasof erosion, steadier stream flow from the hills,and a complete absence of ring-barked trees orevidence of overgrazing (Berglund 1997, inWily2001). In the Mwamishali village, “after onlythree seasons, the area exhibited an almostunbelievable ecological transformation, from desolatelandscapes of bare ground and heavily browsedshrubs to impressive vistas of grass and vigorouslysprouting shrubs” (Mlenge 2002).

In the village of Behroonguda in India, 250 ha ofdegraded forest was subjected to silviculturaltreatment by 1997, in the process achieving theirtarget 7 years ahead of schedule. Improvedsilvicultural treatment had positive effects ongrowth in Behroonguda and led to a 67 percentincrease in the regeneration of seedlings(D’Silva and Nagnath 1999). Further, bybuilding bunds across streams and excavatingpercolation tanks, soil and moistureconservation appears to have occurred.

Existing studies from Mexico suggest that theestablishment of CFEs may have led to astabilization of forest cover in the community.Satellite images of El Rosario de Xico show asignificant expansion of forest canopy from 1982to 1993, and the community has declared 4 haunder strict protection for the conservation ofAbies Hickelii, a species of endangered oyamel fir(Bray and others 2002). Satellite images ofUZACHI communities in Oaxaca show that theforest area has increased by 500 ha in the last 18years as a result of community reforestation andlimits on agriculture in forest areas (Alatorre2000, in Bray and others 2002).Some forest



communities in the Sierra Jurez of Oaxaca haveconsistently logged below the authorizedvolumes in their management plans in anexplicit endeavor to conserve the resource.Moreover, as noted by Bray and others,communities have regularly expressed a desireto reduce their volume of extraction wheninventories indicate that they may be extractingat an unsustainable level. This was clearlyevident in the Quintana Roo communities ofNoh Bec and Laguna Kana, as they reducedtheir logging volume by 29 percent and 37percent respectively (Bray and others 2002).Some communities also show a hearteningcommitment toward broader biodiversityprotection. In La Trinidad, in an ancient forestarea of 365 ha, the community declared 29percent of it as a biodiversity protection area(Bray and others 2002).

In his study of community forestry in parts ofNepal, Edmonds (2002) discovered thatresource extraction was approximately 14percent less in forest user groups. Khare andothers report that satellite images from southernWest Bengal, where 90 percent of the forest isunder JFM, showed that within 2 years, 4,100 haof forest moved from degraded scrub (with lessthan 10 percent crown cover) to the open forestcategory (10-40 percent crown cover). However,in many communities the continued adoption ofsilvicultural methods designed towardmaximum timber extraction is worrying andneeds to be focused more toward local needsand sustainable management (Khare and others2000).

Challenges to be Addressed

Some of the more crucial emerging problems,challenges, and issues in community forestrythat confront the development andpolicymaking community are addressed below.

Compensatory mechanisms for lost rights

Evidence shows that poor, landless, forest-dependent people can be adversely affectedfrom the temporary closure of forests forprotection. They are denied access to productsthat are essential for day-to-day subsistence.The opportunity costs of closure are higher forthe poor, as compared with the less forest-dependent, wealthier people. Several scholars(Khare and others 2000; Chakraborty 2001;Klooster 2000a) have referred to the importanceof finding alternative fuel sources andalternative employment mechanisms as a wayof providing compensation for lost rights. Box18 provides some examples of howcompensatory mechanisms have beendeveloped in successful cases of communityforestry.

Reconciliation of heterogeneous interests

In all community-based natural resourcemanagement initiatives, there needs to be arecognized group of people who are eligible toparticipate in natural resource managementactivities and receive benefits from use.However, defining the group—who to includeand who to exclude—is a complex problem andremains one of the most difficult aspects in theimplementation of community-based programs.Assessment of the multitude of case studiesshow that there is no uniform pattern to befound; examples abound of communities thatare created on their own (by communityagreement, in accordance with customary landrights, traditional mechanisms, etc.) and arecreated externally by outside actors.

There is a common propensity of policymakersto view communities from a reductionistperspective and treat them as homogenousentities. In reality, however, this is seldom the

85



case. There is often tremendous internalstratification within the community, based onethnicity, race, gender, wealth, and a myriadother divisions. As Gauld observed in his studyof community forestry in the Philippines, itseems that “communities are typically, notuniversally defined on the basis of their geographicalfoundations as occupying a particular geographicalspace” (Gauld 2000). The tendency is also toequate spatial proximity with politicalhomogeneity.

Whether heterogeneity within the communityleads to more or less collective action is anempirical question addressed by many, yet stillremains a subject of much debate andcontention. Often in forest management, thisheterogeneity manifests itself in the form of richversus poor, distant users versus proximateusers, and in terms of stakeholders with widelydiffering needs and priorities. For example, inthe timber-company initiated project in Gwira-Banso in Ghana, the local communities were

Box 18Compensatory Strategies to Decrease Opportunity Costs of Conservation

Wage income through forest-related activitiesThe village of Behroonguda in Andhra Pradesh, India, was very successful in creating alternative wage em-ployment schemes related to forest protection. Employment from forestry work in Behroonguda generated14,180 days of work in 1998 and contributed Rs. 2,360 in wages out of a total average family income of 9, 665 inthat year (D’Silva and Nagnath 1999). These activities included soil and moisture conservation, silviculturaloperations, and other support activities.

Decreasing dependence on fuel-woodThe use of appropriate alternative fuel technologies in Behroonguda village reduced fuel-wood consumptionand decreased pressure on the local forest. The distribution of subsidized smokeless chullas (stoves) to half thehouseholds in the village cut down firewood needs by 25 percent. Eight families that received bio-gas genera-tors operating on animal dung ceased collecting firewood altogether. These measures have resulted in a de-cline in the village’s annual firewood consumption by 20 percent from 110 tons a year to 88 tons (D’Silva andNagnath 1999).

Alternative resource collection/use mechanismsThe community of San Antonio in Mexico imposed many restrictions on individual uses, but for the most partalso succeeded in redirecting traditional rights. For example, firewood gathering was coordinated with log-ging activities in order to decrease fire hazards and favor pine regeneration. Community members also havean agreement to forgo cutting pine trees for traditional roofing shingles, but as compensation for this, theforestry business supplies free tin roofing materials and lumber when needed. Some other communities main-tain designated areas for cutting fuel-wood (Klooster 2000b). An example from the Terai region in Nepalshows how local people satisfied their needs for fuel-wood and other essential forest products through illegalextraction from a nearby government-owned forest (Chakraborty 2001).

Compensation for lost rightsThe Nuevo San Juan community in Mexico integrated individual tenure rights with community interests. Allforest lands in the area were parceled out to community members for resin tapping. Leaders establishing thecommunity logging business choose to respect individual usufruct rights to forest plots; resin collection con-tinues on these plots in accord with individual interests. When logging plans slate an area for cutting, commu-nal interests take over, but possessors of resin-tapping plots get a stumpage payment as an incentive forprotecting trees (Klooster 2000b).



primarily concerned with immediate economicgains, while the timber companies were morefocused on tree planting and long-term timberproduction (Appiah 2002). Similar conflictsarose in Moribane in Mozambique and inseveral cases in India. It is crucial that theconcerns of all parties be taken into account andthe different interests reconciled. There is oftena role for NGOs in situations like this, as wasseen in the Gokwe case in Zimbabwe. Yet, theliterature also provides cases whereheterogeneous communities designedinnovative mechanisms, as epitomized in thecase below.

Given the diversified nature of mostcommunities, it is indeed crucial that moreattention and care be given to tackle intra-community inequities. The literature includesmany cases of elite dominance and contrivancewith state apparatus to control decision-making.The San Martin Ocotlan community in Mexicois one prominent case where the elites of thecentral village controlled decision-making, aswell as a disproportionate share of the benefitsat the expense of people in the outlying villages(Klooster 2000a; Klooster 2000b). It is necessarythat community meetings be assembledregularly, encourage participation of all

involved stakeholders (especially women as theprimary gatherers of fuel-wood and NTFP), anddraw up proper accounting and reportingmechanisms to ensure greater accountability. Asmentioned before, forest management has to bereoriented to meet current needs and long-termlivelihood security of the poorest, facilitate localinstitutions to identify existing users, and layclear guidelines for ensuring everyone’sparticipation (Khare and others 2001).

Greater community control over non-timberforest products

NTFPs are particularly significant in SouthAsian countries. For example, on average 40percent of the state forest revenue and 75percent of forest exports are derived fromNTFPs in India (World Bank 1993, in Khare andothers 2000). The commercial importance ofthese products means that some of the mostcommercially valuable NTFPs (like cashew,bamboo, or fibrous grasses) are eithernationalized or monopoly collection rights aregranted to certain government and privateorganizations. Thus, they do not fall under thepurview of 100 percent usufructs or underrevenue sharing as a JFM benefit (Agarwal andSaigal 1996, in Khare and others 2000).

Box 19Mechanisms to Overcome Heterogeneity: A Case Study from Nepal

Varughese and Ostrom’s (2001) study of heterogeneity in forest communities in the middle hills of Nepalsuggests that the more successful forestry communities devise innovative mechanisms to surmount problemscreated by heterogeneity. The authors hypothesize that when the benefits of community forestry are sufficient-ly high enough, the local communities will endeavor to find mechanisms to overcome heterogeneity. Forexample, in Raniswara in the Gorkha District and Bandipur in Tanahaun District, conflicts between proximateusers and distant users of the same forest area were overcome by devising a mechanism whereby the moredistant users agreed to pay an additional fee to the community as compensation for forsaking their share ofguard duties (Varughese and Ostrom 2001). In addition, those who cannot participate in joint maintenance,harvesting, or monitoring can pay special membership fees so as to avail themselves of forest products atbelow-market rates. In Raniswara, special membership is noted after payment of a fee. In Bandipur, membershave to purchase tokens of different colors; each color is specific to their membership type.

87



Recent Government of India legislation has beendirected toward removing some of the controlson NTFPs. This legislation—adopted by statesin 1998—gives ownership rights of all NTFPs totribals in certain tribal-dominated areas (Khareand others 2000). Yet, implementation has beenslow and progress has been made mainly inMadhya Pradesh (Khare and others 2000). InNepal, the marketing of many of the NTFPs isoften heavily controlled by the state or throughchains of “middle men.” Detailed informationon NTFPs in other parts of the developingworld is hard to find, yet, it seems that they aremore widely devolved to communitydecisionmaking levels, especially in Africa (Wily2002a). Thus, a survey of the literature clearlydemonstrates the need for greater communitycontrol over NTFPs and more equitabledistribution of NTFP benefits.

Devolution of bureaucratic power and not justusufruct rights

A lack of commitment by the state, particularlythe forestry bureaucracy, to truly release powerto a local level is considered a major factorlimiting the success of community forestry(Shackleton and Campbell 2001; Agrawal andGibson 1999). Ideally, the delegation of extraresponsibility toward communities should beaccompanied by commensurate power andauthority. However, this often does not happen,and can lead to delays in implementation. InIndia, for instance, it is argued that legal andadministrative arrangements for communityforestry tend toward centralization and thatlocal organizations under JFM are little morethan a proxy for the forest department, with thelatter still controlling key aspects (Jodha andBhatia 1998, in Shackleton and Campbell 2001).Similar evidence is found in the Gokwecommunal area in Zimbabwe and community

forestry in parts of Indonesia and thePhilippines.

Cases where devolution of authority goes hand-in-hand with recognition of local ownership ofthe forest are still not widespread. Wilyobserves that Tanzania and the Gambia havegone the furthest in developing a “workingregime of authority over the forest. Typically thisworking regime of authority is laid out in steps thatassist the community to define and demarcate theforest area and to develop a rational and sustainablemanagement plan for its protection and use, to beoperated at their cost and through regimes that theythemselves devise” (Wily 2002b). Crucially, theend result in both country processes isrecognition of the community as owner-manager of the community forest, rather thanonly licensee, user, or even manager. More than200 community forests have been created in theGambia through this process thus far, and morethan 500 in Tanzania (Wily 2002b). A positiveexample from the Duru Haitemba forest inTanzania and the process by whichdecentralization occurred is presented in Box20.

Tackling corruption

Corruption, lack of accountability, absence ofaccounting and recording mechanisms arefrequent problems encountered by forestmanagement communities. Concerns aboutcorruption arise not only because it is ethicallywrong, but because of its adverse effects on thedevelopment process and in particular itsimpact on the disadvantaged, who bear thebrunt of corruption and are least able to fightcorrupt bureaucracies (Hill 2000). In anassessment of a World Bank-assisted forestryproject in Madhya Pradesh, India, Hill notesthat opportunities for corruption can be reduced



with improved decentralized planninginvolving beneficiaries and better monitoring ofperformance and impact. In addition, the factthat communities and not individuals wereinvolved in management and benefit sharingtends to minimize corrupt practices.

Corruption was especially severe in thepreviously discussed example of San MartinOcotlan in Mexico, and was furthercompounded by elite dominance in the central

Box 20Communities as Managers and not just Users:

The Case of Duru Haitemba, a Successful Example from Tanzania

The Duru Haitemba is a hilly Miombo woodland of 9,000 hectares in the Arusha region of northern Tanzania.It was surveyed for gazettement as a national forest reserve in 1991 and forest guards posted to protect theforest (Wily and Dewees 2001). There was resistance from villagers to the reservation. Feeling that their cus-tomary tenure had been ignored, villagers made excessive use of forests. By eliminating a local sense of propri-etorship, the government eliminated local guardianship or recognition (Wily 1999).

The government then mooted a benefit-sharing concept, which was rejected by villagers. Thus, a new ap-proach was needed and it was decided that villagers should have full control and authority over forests for atrial period. Informal support for this endeavor was provided by the local district council. It was decided thatthe gazettement process would be suspended pending demonstration by villagers that they could halt forestdegradation (Wily 1997).

Eight villages around Duru Haitemba participated in this process. Each village formed a Forest ManagementCommittee in late 1994/95, appointed volunteer forest guards, and drew management action plans for thepart of the forest that fell within their village sphere. Access rules were central to the plans. Outsiders’ accessto the forest stopped. Village members agreed to limit use to certain zoned areas. Around half of each villageforest was closed off to all use, given its degraded condition. Damaging uses like charcoal production werebanned. Grazing was restricted to specified months and zones.

The government, impressed with progress, then began to withdraw forest guards and suspended gazette-ment. There was visible evidence of recovery and forest gain within three years. Charcoal burning, rampanttimber harvesting, and unregulated in-forest settlement have since disappeared. The return of understoryshrubbery, grass cover, and bee swarms to forests illustrates the positive natural resource impacts.

The composition of the forest management committees has shifted gradually from village elders to ordinaryvillagers (Wily 1997). There is also an increase in the number of villagers practically involved in the commit-ment toward forest conservation. Once villagers began actively managing forests, it became evident that “theyneeded not just the administrative support they had received from the local council, but legal backing as well” (Wily1997). To facilitate this process, each village was provided assistance to rephrase their management plans andrules as village by-laws. Formal approval for these by-laws was granted by mid-1995 under the District Au-thorities Act by the Full District Council. Thus, as Wily notes, “each village is by law, the legal authority andmanager” of that part of Duru-Haitemba forest that is adjacent to its own settled village area and specified inthe relevant village by-law as falling under their jurisdiction (Wily 1997).

village, and a skewed profit distributionmechanism that disadvantaged outlyingvillages. After many years without profits,villagers in the outlying areas demanded auditreports and were staggered to find the amountof corruption within the central committee.Similar problems of corruption were avoided inother communities in Mexico (notably SanAntonio) due to the presence of greatercommunity involvement and better accountingand reporting procedures. Thus, an important

89



side benefit of decentralized management islowered corruption. However, to be fullyrealized, this will require additional capacitybuilding at the community level. There is a rolefor training and NGO support to developcommunity-level accounting and reportingmechanisms.

Conclusions

This paper, which is based on a review of casestudies from around the world, sought a criticalappraisal of decentralized forest managementwith a view to identifying emerging patterns,trends, and issues that are of critical importanceto scholars, practitioners, and policymakersalike. In this concluding part of the paper, werecapitulate some of the main conclusions andfindings drawn from the literature.

It is evident that community forestry isprevalent worldwide from India to LatinAmerica and from Sub-Saharan Africa toSoutheast Asia. The phenomenon ofdecentralized forest management has especiallygained momentum in the last decade. There isalso a steady, if somewhat hesitant, movementof community forestry being extended fromdegraded to more rich forests.

The case studies reveal that decentralizationtakes places primarily through threeinstitutional mechanisms: (1) devolution tolower levels of government; (2) creation ofvillage-level institutions; and (3) the allocationof more power to existing traditionalinstitutions. Successful examples abound in allthree different mechanisms. It is hard toestablish geographic trends and patterns, as allthree seem to be present in many different partsof the globe.

A critical issue that dominates the discourse oncommunity forestry is the debate on powersharing versus benefit sharing. Even thoughthere is a gradual movement toward morepower sharing (especially pronounced inAfrica), the user-centric concept stillpredominates. The power-centric approach, asthe case studies have demonstrated, isimportant to success and mainly looks to forestcommunities as potential managers or co-managers and devises arrangements to givethem varying degrees of managerial control andauthority. Power sharing can occur throughmany processes. Reforms in land laws that givemore tenure security to forest communities,general changes in legal and policy frameworks,and the willingness to devolve bureaucraticpower are all crucial elements that facilitatepower sharing.

Benefits from community forestry are basicallyderived both at the community level and at thehousehold level. Benefits differ at each level,and can be both long term and short term. Ingeneral, it seems rather difficult to secureparticipation of the communities in the absenceof fairly quick benefits, yet as the example fromEthiopia demonstrated, if the communities areconvinced about the long-term viability of theproject, they will still take part. However, it iscrucial that effective compensatory mechanismsbe drawn up in the short run to make up forloss rights due to the temporary closure offorests. These mechanisms range from thecreation of employment opportunities to theprovision of alternative fuel sources. Genderinequalities and distributional discrepancies arealso frequently witnessed in community forestryprograms, and, the mechanisms through whichthese inequalities can be minimized needfurther scrutiny.



Another crucial aspect of community forestry iseffective monitoring and enforcement.

Communities use a variety of mechanisms toensure collective participation in managementof resources. These range from employment ofguards to protect forests to fines for notattending committee meetings. Though a vastspectrum of strategies are employed to ensurerule compliance, it seems graduated sanctions—where the penalty imposed is contingent on theseverity and frequency of infringement—is aparticularly effective mechanism.

It is evident from the review that there is a clearneed for more in-depth and carefully conductedresearch on the impacts of decentralizedforestry. Assessment of impacts is indisputablythe best way of measuring the success ofdecentralized forest management; thus, more

structured and focused emphasis needs to bepaid to impacts. In particular, more work needsto be done on distributional inequities, effectson poverty, and empowerment of communities.This is a clear research agenda for the future,and it is hoped that more specialized work iscarried out that would be helpful in thoroughlyanalyzing the effectiveness of community forestmanagement.

The paper also identified some emergingchallenges and policy issues. These include theneed to have greater community control overnon-timber forest products, provision ofcompensatory strategies, effective mechanismsto overcome a host of problems related toheterogeneity and tackling corruption, andproviding greater accountability of communityforest practices.


Notes

1. This study is supported by the Trust Fundfor Environmental and Socially SustainableDevelopment. It is part of a larger piece ofanalytical work undertaken by the Policyand Economics Team, EnvironmentDeparment, on “Understanding the Impactsof Devolution on Natural ResourceManagement.” This work includes threecase studies on communal conservancies inNamibia, community-based forestmanagement in India, and irrigationmanagement transfer in the Philippines.

2. Policy and Economics Team, EnvironmentDepartment.

3. Please see Bandyopadhyay andShyamsundar (2004), and Bandyopadhyay,Humavindu, Shyamsundar, and Wang(2004), which are available as World BankPolicy Research Working Papers atwww.econ.worldbank.org.

4. Communal Areas Management Program forIndigenous Resources

5. Administrative Management Design forGame Management Areas.

6. For instance, between 1989 and 1992,ADMADE in Zambia resulted inapproximately 60 projects, including 23houses for teachers, 9 maize grinding mills,and 7 rural health centers (Gibson 1999).

7. A high-end example is the TorraConservancy in Namibia. Communitymembers are documented to have earned

$70,000 in wages from 1996 to 1998 throughan agreement with a photographic safaricompany (Jones 1999b).

8. Such opportunity costs can result frommany different forms of natural resourcemanagement and are not specific tocommunity based management.

9. For instance, evidence from joint forestmanagement in India suggests that theseinstitutions re-create traditional patterns ofdiscrimination against women (Agarwal2001, Sarin 2001).

10. Among many such examples is evidencefrom national parks in western Uganda(Mgahinga, Bwindi and Kibale), whichsuggests that community relations withpark authorities have improved as result ofcommunity outreach efforts (Infield andAdams 1999, Archbald and Naughton-Treves 2001).

11. These countries have supported legalprocesses that encourage communities tobring unreserved forest areas under theirown jurisdiction and to apply to managenationally owned forests (Wily, 2002a;Bojang, 1999).

12. In ADMADE in Zambia, 35 percent ofrevenues were allocated to communityactivities and the rest retained by thegovernment. In CAMPFIRE, 50 percent wasretained by the state and 50 percent dividedbetween administrative costs andcommunity dividends.



13. There is considerable theoretical andempirical literature on conditions fordurable commons institutions and whycooperation in managing natural resourcesoccurs in some cases and fails in others.Agrawal (2001), for example, identifies 24conditions (see Table 1) based on threesynthesis books by Wade (1988), Ostrom(1990), and Baland and Platteau (1996).

14. Murombedzi (1999), for example, arguesthat CAMPFIRE benefits are highest wherehuman populations are low and animalpopulations are high.

15. Agricultural crop production may besubsidized, an issue often beyond thecontrol of communities.

16. Nested enterprises refer to multi layeredlinked institutions.

17. CAMPFIRE and ADMADE stand forCommunal Areas Management Program forIndigenous Resources and AdministrativeManagement Design for Game ManagementAreas

18. It is not clear whether these monies werethen distributed within the community orwhether they were used for communityprojects.

19. Workshop in Political Theory and PolicyAnalyses, Indiana University.

20. Manager, Institutional DevelopmentDivision.

21. An IMT can be considered progressive if , ata minimum, it provides for the followingauthorities and responsibilities to the WUA:(a) make rules and sanctions, with themaximum sanction of stopping wateravailable to the WUA; (b) make O&M planand budgets; (c) set water charges; (d) hireor release management staff; (e) control overintake; (f) control over main canal system;(g) control over subsidiary canal system; (h)responsibility for future rehabilitation; (i)

canal rights of way; (j) right to contract,raise and disburse funds and make profit.

22. See— http://www.iwmi.cgiar.org/tools/perform.htm#ani —for methodologicalapproach.

23. Conversation with Ganesh Shivakoti,Associate Professor in Resource Economics,Asian Institute of Technology, when he peerreviewed a draft version of this paper

24. Department of Political Science, McGillUniversity

25. Common property resource theorists haveidentified a broad set of requirements forthe successful implementation of naturalresource management. Elinor Ostrom’seight design principles for successful CPRare prominent among these (Ostrom 1990).

26. The Forest Policy of 1988 was distinct frompreceding policies. Environmentalconservation and meeting local people’ssubsistence needs were articulated as themain priorities instead of revenuegeneration through timber extraction (Khareand others 2000).

27. The total land area under communitymanagement is likely to be much higher, asthe above figures do not include many self-initiated groups.

28. Self-initiated groups are particularlyevident in the states of Orissa, Bihar, UttarPradesh (UP), and Jharkhand. Orissa isestimated to have approximately 5,000community forest management groups, ofwhich around 1,200 are said to participatein JFM (Singh 2002).

29. Wily notes that over 1,000 communityforests operate and at least 100 significantnational forest reserves are under (orcoming under) state-people co-managementregimes.

30. Under Ghana’s timber utilization contract,timber companies are required to allocate a

93

Notes


certain percentage of royalties for theprovision of social amenities (Appiah 2002).

31. For instance in UP, the JFM order of 1997required the existing Van Panchayats toaccept in writing that the Panchayat rulesno longer applied under JFM (Khare andothers 2000).

32. For more information, see “World Bankloans $ 15 million for Mexican forests”,Environment News Service, 2/24/97— http://forests.org/archive/samerica/15mexfor.htm.

33. The Palcazu Yanesha have a long history ofinsecure land tenure, encompassing severalhundred years of exploitation andalienation from lands they legally held, butwere not able to defend from FranciscanMissions or European or Mestizo colonists(Smith 1974, in Morrow and Hull 1996). Inaddition, at the time COFYAL wasdeveloped, the Palcazu communities werefacing increasing population pressures dueto constant in-migration from neighboringcommunities. The opening of a new road inthe 1980s leading to the valley made thesituation worse. In the words of oneCOFYAL officer, “there was no proposedmechanism to guarantee that the nativecommunities’ lands could be defended againstinvasions of loggers, traders, and colonists oncethe road came in” (Morrow and Hull 1996).

34. As Wily (2002a) observes in Zambia, onlylocal forests are the target of communityforestry rather than national forests. InCameroon, community forests may beestablished only in unclassified areas andare restricted to 5,000 ha and 10-yearagreement periods.

35. In the absence of precise parameters fordefining degraded forests, the term isinterpreted differently by different states.While some state JFM orders have left it to

the discretion of local forest officers todifferentiate degraded from non-degradedforests, others have defined it so preciselythat only a small amount of the total forestarea can be brought under JFM (Khare andothers 2000).

36. Ostrom’s design principles for commonproperty management are best known(Ostrom 1990). They include (a) the resourceand the users of the resource should beclearly defined and appropriators should beable to sustain legal claims; (b) the presenceof a small number of stakeholders; (c)homogeneity in stakeholders andhomogeneity in their endowments andpreferences; (d) a set of small, simple rulesrelated to access and user patterns, agreedto by the stakeholders and congruent withlocal conditions and provision rules; (e)stakeholders affected by operational rulesshould be able to participate in modifyingthem; (f) transparency and efficientmonitoring of information between stakeholders; and (g) access to rapid, low-costconflict resolution mechanisms.

37. As Wily (2002b) notes, “Procedures are beingmade more publicly accountable and almosteverywhere require fuller consultation withthose affected. The new land laws of Uganda(1998), Tanzania (1999), Mozambique (1997),and South Africa (1991, 1996, and 1997) areexceptional examples, but likely to be followedby proposed new legislation in Lesotho,Swaziland, Namibia, and Malawi. Comparabledevelopments are under way in FrancophoneWest Africa, versions of which are beingdelivered in the Ivory Coast (1998), Mali(2000), and Niger (1993).”

38. The 1996 Forestry Law indicates thatmunicipalities that receive forestry royaltiesmust within 6 months of the receipt of thesefunds create a Municipal Forestry Unit. In



addition, municipal governments areresponsible for identifying and demarcatingpublic forested lands in the municipalterritory that should be used exclusively forlocal communities. The Forestry Law also

asks municipal governments to providetechnical assistance to local forest users todevelop forest management plans and tohelp local users to acquire formal forestproperty rights (Andersson 2003).


References

Adams, W., and M. Infield. 2003. “Who is onthe Gorilla’s Payroll? Claims on TouristRevenue from a Ugandan National Park.”World Development 31 (1): 177-190.

Adhikari, B. 2003. “Property Rights and NaturalResources: Socio-Economic Heterogeneityand Distributional Implications of CommonProperty Resource Management.” WorkingPaper 1-03. South Asian Network forDevelopment and Environment Economics(SANDEE): Katmandu, Nepal.

Agarwal, B. 2001. “Participatory Exclusions,Community Forestry, and Gender: AnAnalysis for South Asia and a ConceptualFramework.” World Development 29 (10):1623-1648.

Agrawal, A. 2001. “Common PropertyInstitutions and Sustainable Governance ofResources.” World Development 29 (10):1649-1672

Agrawal, A. and C. Gibson. 1999. “Enchantmentand Disenchantment: The Role ofCommunity in Natural ResourceConservation.” World Development 27 (4):629-649

Agrawal, A., and E. Ostrom. 2001. “CollectiveAction, Property Rights, and Devolution ofForest and Protected Area Management,” in

R. Meinzen-Dick, ed., “Collective Action,Property Rights and Devolution of NaturalResource Management: Exchange ofKnowledge and Implications for Policy -Proceedings of the InternationalConference.”

Alexander, J., and J. McGregor. 2000. “Wildlifeand Politics: CAMPFIRE in Zimbabwe.”Development and Change 31: 605-627.

Andersson, K. 2003. “What MotivatesMunicipal Governments? Uncovering theInstitutional Incentives for MunicipalGovernance of Forest Resources in Bolivia.”Journal of Environment & Development 12 (1):5-27.

Appiah, M. 2002. “Co-Partnership in ForestManagement :The Gwira-Banso Joint ForestManagement in Ghana.” Environment,Development and Sustainability 3 (4): 343-360.

Araral, E. 2003. “Capacity Assessment ofIrrigator’s Associations.” ProjectPreparation notes prepared for the WorldBank-funded Philippines: Riverbasin andWatershed Management Project.

Archbald, K., and L. Naughton-Treves. 2001.“Tourism Revenue-Sharing around NationalParks in Western Uganda: Early Efforts toIdentify and Reward Local Communities.”Environmental Conservation 28 (2): 135-149



Arnold, J. 2001. Forests and People: 25 years ofcommunity forestry. Food and AgriculturalOrganization of the United Nations. Rome:FAO.

Ashley, C. 1998. “Intangibles matter: Non-financial Dividends of Community-BasedNatural Resource Management inNamibia.” WWF LIFE Program Windhoek.

Azziz, Y. A. 1994. “Irrigation managementtransfer: Development and turnover toprivate water user associations in Egypt.”Paper presented at the InternationalConference on Irrigation ManagementTransfer, September 20-24, Wuhan, China.

Bagadion, B. U., and F. F. Korten. 1991.“Developing irrigators’ organizations: Alearning process approach,” in M. Cernea,ed., Putting people first: Sociological variablesin rural development. (2nd ed.) Washington,D.C.: World Bank.

Bagadion, Benjamin U. 1994. “Jointmanagement of the Libmanan-CabusaoPump Irrigation System between farmersand the National Irrigation Administrationin the Philippines.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20-24,Wuhan, China.

Baland J-M, and J-P. Platteu. 1996. HaltingDegradation of Natural Resources. Is there aRole for Rural Communities? Oxford, U.K:FAO and Oxford University Press.

Bandyopadhyay, S., M. Humavindu, P.Shyamsundar, and L. Wang. 2004. “DoHouseholds Gain from Community Based

Natural Resource Management. AnEvaluation of Community Conservancies inNamibia.” Policy Research Working PaperNo. 3337. Washington, D.C.: World Bank.

Bandyopadhyay, S., and P. Shyamsundar. 2004.“Fuelwood Consumption and Participationin Community Forestry in India.” PolicyResearch Working Paper No. 3331.Washington, D.C.: World Bank.

Bardhan, P. and J.D Johnson 2003. UnequalIrrigators: Heterogeneity and Commons

Management in Large Scale MultivariateResearch. In Ostrom, Dietz, Dolsak, Stern,Stonich and Weber, eds., 2002. The Dramaof the Commons. Washington DC: NationalAcademy Press.

Barker, R., and F. Molle. 2002. “Perspectives onAsian Irrigation.” Paper presented at theConference of Asian Irrigation in Transition– Responding to the Challenges Ahead.Asian Institute of Technology, Bangkok,Thailand.

Barnes, J.I., J. MacGregor, and L.C. Weaver.2003. “Economic Efficiency and Incentivesfor Change within Namibia’s CommunityWildlife Use Initiatives.” World Development30 (4): 667-681.

Barrow, E., H. G., and M. Infield, 2001. “TheEvolution of Community ConservationPolicy and Practice in East Africa.” InHulme and Murphree, eds., African Wildlifeand Livelihoods. The Promise and Performanceof Community Conservation. Oxford, U.K.:James Currey Ltd.

97

References


Batterby, S. 1998. “Local EnvironmentalManagement, Land Degradation and the‘Gestion des Terroirs’ Approach in WestAfrica: Policies and Pitfalls.” Journal ofInternational Development 10: 871-898.

Bergin, P. 2001. “Accommodating NewNarratives in a Conservation Bureaucracy.TANAPA and Community Conservation.”In Hulme and Murphree, eds., AfricanWildlife and Livelihoods. The Promise andPerformance of Community Conservation.Oxford, U.K.: James Currey Ltd.

Berkes, F. 1989. “Cooperation from theperspective of human ecology.” In F.Berkes, ed., Common Property Resources,London, UK: Belhaven.

Bichara, A. 2002. “Conversation during theproject preparation mission for the Riverbasin and Watershed Management Project.”Legazpi City, Albay, Philippines.

Bojang, F. 1999. “Overview of the InternationalWorkshop on Community Forestry inAfrica.” In Proceedings of the InternationalWorkshop on Community Forestry inAfrica: Participatory forest management; astrategy for sustainable forest managementin Africa. April 26-30, Banjul, the Gambia.

Bond, I. 2001. “CAMPFIRE and the Incentivesfor Institutional Change.” In Hulme andMurphree, eds., African Wildlife andLivelihoods. The Promise and Performance ofCommunity Conservation. Oxford, U.K.:James Currey Ltd.

Bradshaw, B. 2003. “Questioning the Credibilityand Capacity of Community-based

Resource Management.” CanadianGeographer 47 (2):137-150.

Bray, D., L. Merino-Perez, P. Negreros-Castillo,G. Segura-Warnholtz, J. Torres-Rojo, and H.Vester. 2003. “Mexico’s Community-Managed Forests as a Global Model forSustainable Landscapes.” ConservationBiology 17 (3): 672-677.

Britt, C. 2001. “Mixed signals and governmentorders: The problem of on-again off-againcommunity forestry policy.” Forest, Trees andPeople Newsletter 45: 29-33.

Bromley, D. 1992. Making the commons work:Theory, Practice and Policy. San Francisco,CA: Institute for Contemporary StudiesPress.

Buchy, M., and D. Race. 2001 “The Twists andTurns of Community Participation inNatural Resource Management in Australia:What is Missing?” Journal of EnvironmentalPlanning and Management 44 (3): 293-308.

Campbell, B., A. Mandondo, N. Nemarundwe,B. Sithole, W. De Jong, M. Luckert, and F.Matose. 2001. “Challenges to proponents ofCommon Property Resource Systems:Despairing Voices from the Social Forests ofZimbabwe.” World Development 29 (4): 589-600.

Cernea, M., ed. 1985. Putting People First:Sociological Variables in Rural Development.New York: Oxford University Press.

_____. 1987. “Farmer Organization andInstitution Building.” SustainableDevelopment. Regional Development Dialogue



8(2): 1-24. Nagoya, Japan: United NationsCentre for Regional Development

Chakraborty, R. 2001. “Stability and outcomesof common property institutions in forestry:evidence from the Terai region of Nepal.”Ecological Economics 36: 341-353.

Colfer, C., and R. Wadley. 1996. “Assessing‘Participation’ in Forest Management:Workable Methods and UnworkableAssumptions.” Working Paper No. 12.Bogor, Indonesia: Center for InternationalForestry Research.

D’Silva, E., and B. Nagnath. 1999. “Local PeopleManaging Local Forests: BehroongudaShows the Way in Andhra Pradesh, India.”Working Paper Series No. 3. Asia ForestNetwork.

DSI (Deviet Su Isleri), EDI (EconomicDevelopment Institute, World Bank), andIIMI (International Irrigation ManagementInstitute). 1996. “Assessment of IMT inTurkey.” Paper prepared for theInternational Seminar on ParticipatoryIrrigation Management, April 10-17,Antalya, Turkey.

Edmonds, E. 2002. “Government-initiatedcommunity resource management and localresource extraction from Nepal’s forests.”Journal of Development Economics 68: 89-115.

Eleazar, L. 2002. Participatory Approaches toIrrigation Development and Management inthe Philippines: The Next Generation.Paper presented at the Workshop on AsianIrrigation in Transition-Responding to theChallenges Ahead. April 22-23, 2002. AsianInstitute of Technology, Bangkok, Thailand.

Emerton, L. 2001. “The Nature of Benefits andthe Benefits of Nature. Why WildlifeConservation has not EconomicallyBenefited Communities in Africa.” InHulme and Murphree, eds., African Wildlifeand Livelihoods. The Promise and Performanceof Community Conservation. Oxford, U.K.:James Currey Ltd.

Farley, P. J. 1994. “Privatization of irrigationschemes in New Zealand.” Short ReportSeries on Locally Managed Irrigation, no. 2.Colombo, Sri Lanka: International IrrigationManagement Institute.

Fomete, T., and J. Vermaat. 2001. CommunityForestry and Poverty Alleviation in Cameroon.Network Paper 25h. Rural DevelopmentForestry Network.

Garcés-Restrepo, C., and D. Vermillion. 1994.“Irrigation management transfer inColombia: An assessment of seventransferred districts.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20-24,Wuhan, China.

Gauld, R. 2000. “Maintaining CentralizingControl in Community-based Forestry:Policy Construction in the Philippines.”Development and Change 31: 229-254.

Gazmuri, R. 1994. “Chilean water policy.” ShortReport Series on Locally ManagedIrrigation, no. 3. Colombo, Sri Lanka:International Irrigation ManagementInstitute.

Gebremedhin, B. J. P., and G. Tesfay. 2001.Nature and determinants of collective action for

99

References


woodlot management in Northern Ethiopia.Socio-economics and Policy ResearchWorking Paper 40. Nairobi, Kenya:International Livestock Research Institute.

Ghate, R. 2000. “Joint Forest Management:Constituting New Commons—A case studyfrom Maharashtra, India.” Paper submittedfor the eighth biennial conference of theIASCP, entitled Constituting the Commons:Crafting Sustainable Commons in the NewMillennium, May 31-June 4, IndianaUniversity, Bloomington, Indiana.

Gibson, C., and F. Lehoucq. 2003. “The LocalPolitics of Decentralized EnvironmentalPolicy in Guatemala.” Journal of Environment& Development 12 (1): 28-49.

Gibson, C., and S. Marks.1995. “TransformingRural Hunters into Conservationists: AnAssessment of Community-Based WildlifeManagement Programs in Africa.” WorldDevelopment 23 (6): 941-957.

Gibson, M. 1999. Politicians and Poachers. ThePolitical Economy of Wildlife Policy in Africa.Political Economy of Institutions and Decisions.Cambridge, U.K.: Cambridge UniversityPress.

Gitomer, C. S. 1994. “Price regulation in thereform of irrigation management in ruralChina.” Paper presented at the PoliticalEconomy Workshop of the Joint Committeeon Near and Middle Eastern Studies of theSocial Sciences Research Council, February10-13, Berkeley, California.

Gorriz, C., A Subramanian, and J. Simas. 1995.“Irrigation management transfer in Mexico:Process and progress.” Paper presented at

the International Seminar on ParticipatoryIrrigation Management, February 3-15,Mexico.

Gosselink, P., and P. Strosser. 1996.“Participatory rural appraisal for irrigationmanagement research: Lessons from IIMI’sexperience.” Working Paper 38. Colombo,Sri Lanka: International IrrigationManagement Institute.

GRN. 1998. National Land Policy. Governmentof the Republic of Namibia. Windhoek.

Groenfeldt, D., and M. Svendsen, eds. 2000. CaseStudies in Participatory IrrigationManagement. Washington: World BankInstitute.

Hanrahan, M. 1990. “Evaluation of the On-FarmWater Management Project in theDominican Republic.” Prepared for the U.S.Agency for International Development.Duplicated.

Hobley, M., and K. Shah. 1996 “What makeslocal organizations robust? Evidence fromIndia and Nepal.” Natural ResourcesPerspectives No.11. July 1996. OverseasDevelopment Institute.

http://dlc.dlib.indiana.edu/documents/dir0/00/00/02/64/dlc-00000264-00/ghater041500.pdf

http://www.cbnrm.net/pdf/aldenwily_l_002_cfm.pdf

http://www.york.ac.uk/inst/sei/prp/pdfdocs/Policy_Process_Analysis_Report_1_Nepal_Final.doc.pdf



IIMI (International Irrigation ManagementInstitute) and BAU (Bureau of Socio-Economic Research and Training of theBangladesh Agricultural University). 1996.“Study on privatization of minor irrigationin Bangladesh: Final report.” TA: BAN 182.Dhaka: Ministry of Agriculture and AsianDevelopment Bank.

Infield, M., and W. M. Adams. 1999.“Institutional Sustainability andCommunity Conservation: A Case Studyfrom Uganda.” Journal of InternationalDevelopment 11: 305-315.

Johnson, C. 2002 “In the Eyes of the State:Negotiating a ‘Rights-Base Approach’ toForest Conservation in Thailand.” WorldDevelopment 30 (9): 1591-1605.

Johnson, S. H. III. 1996. “Irrigation managementtransfer in Mexico: Moving towardsustainability.” Paper prepared for theInternal Program Review, April 1–2.Colombo, Sri Lanka: International IrrigationManagement Institute.

Johnson, S.H., III. 1995. “Selected experienceswith irrigation management transfer:Economic implications.” Water ResourcesDevelopment 11 (1): 61–72.

Johnson, S. H., III, D. Vermillion, M. Svendsen,W. Xinyuan, Z. Xiying, and M. Xuesen.1995. “Institutional management andperformance changes in two irrigationdistricts: Case study from Hebei Province.”In Selected papers from the InternationalConference on Irrigation ManagementTransfer, Wuhan, China, September 20–24.Rome: International Irrigation ManagementInstitute and FAO.

Johnson, S.H. III, M. Svendsen, and Z. Xiying.1994. “Performance impacts of transfer.”Paper presented at the InternationalConference on Irrigation ManagementTransfer, September 20–24, Wuhan, China.

Johnson, S. H., III, and P. Reiss. 1993. “Canfarmers afford to use the wells afterturnover? A study of pump irrigationturnover in Indonesia.” Short Report Serieson Irrigation Management Transfer, no. 1.Colombo, Sri Lanka: InternationalIrrigation Management Institute.

Jones, B.T. 1999a. “Policy Lessons from theEvolution of a Community –BasedApproach to Wildlife Management, KuneneRegion, Namibia.” Policy Arena. Journal ofInternational Development 11: 295-304.

Jones, B. T. 1999b. “Rights, Revenues andResources: The Problems of PotentialConservancies as Wildlife ManagementInstitutions in Namibia.” Evaluating EdenSeries, Discussion Paper No. 2. IIED.

Jones, B., and M. Murphree. 2001. “TheEvolution of Policy on CommunityConservation In Hulme and Murphree, eds.African Wildlife and Livelihoods. The Promiseand Performance of Community Conservation.Oxford, U.K.: James Currey Ltd.

Joshi, A.1998/1999. “Progressive Bureaucracy:An Oxymoron? The Case of Joint ForestManagement in India.” Net Work Paper24a. Rural Development Forestry Network.

Kalro, A. H., and G. Naik. 1995. “Outcomes ofirrigation management transfer andfinancial performance of water users’associations in India: Some experiences.”

101

References


Paper presented at the Workshop onIrrigation Management Transfer in India,December 11–13, Indian Institute ofManagement, Ahmedabad, India.

Kayambazinthu, D. 2000. “Empowering ruralcommunities to manage natural resources:Where the does the power lie? The case ofMalawi.” In Shackleton, S.E., and Campbell,B., eds. 2000. Empowering communities tomanage natural resources: case studies fromSouthern Africa. SADC Wildlife SectorNatural Resource.

Keating, M. 1993. The Earth Summit’s agenda forchange. Geneva: The Center for OurCommon Future.

Kellert, S., J. Mehta, S. Ebbin, and S. Lichtenfeld.2000. “Community Natural ResourceManagement: Promise, Rhetoric andReality.” Society and Natural Resources 13:705-715.

Keohane, R., and Ostrom, E. 1995. LocalCommons and Global Interdependence.London: Sage.

Khare, A., M. Sarin, N. Saxena, S. Palit, S.Bathla, F. Vania, and M. Satyanarayana.2000. Joint Forest Management: policy, practiceand prospects. World Wide Fund for Nature-India and International Institute forEnvironment and Development (IIED),United Kingdom.

Kloezen, W. H. 1996. “Going beyond theboundaries of water user groups: FinancingO&M in Sri Lanka.” Paper prepared for theInternal Program Review, April 1–2.Colombo, Sri Lanka: International IrrigationManagement Institute.

Kloezen, W., C. Garces-Restrepo, and S.Johnson. 1997. “Impact Assessment of IMTin the Alto Rio Lerma Irrigation District,Mexico.” Research Report 15. Colombo, SriLanka: IIMI.

Klooster, D. 2000a. “Community Forestry andTree Theft in Mexico: Resistance orComplicity in Conservation?” Developmentand Change 31: 281-305.

Klooster, D. 2000b. “Institutional Choice,Community, and Struggle: A case Study ofForest Co-Management in Mexico.” WorldDevelopment 28 (1): 1-20.

Knox, A. and R. Meinzen-Dick. 2000.“Collective Action, Property Rights, andDevolution of Natural ResourceManagement: Exchange of Knowledge andImplications for Policy - A WorkshopSummary Paper.” CAPRi Working Paper 11.Washington DC: International Food PolicyResearch Institute.

Koehane, R., and E. Ostrom. 1995. LocalCommons and Global Interdependence.London: Sage.

Koppen, B., R. Parthasarathy, and C. Safiliou.2002. “Poverty Dimensions of IMT in largescale canal irrigation in Andrah Pradeshand Gujarat, India.” Research Report 61.Colombo, Sri Lanka: IWMI.

Korten, D. 1980. “Community Organization andRural Development: A Learning ProcessApproach.” Public Administration Review 40(5): 480-511

Kumar, S. 2002. “Does ‘Participation’ inCommon Pool Resource Management Help



the Poor? A Social Cost-Benefit Analysis ofJoint Forest Management in Jharkhand,India.” World Development 30 (5): 763-782.

Kwanbo, K. 2003. “Governing the KoreanIrrigation Institutions: Evaluation of theFarmer’s Self Governing Model DistrictProgram.” Paper Presented at the Workshopin Political Theory and Policy Analysis,Indiana University.

Lam, W. F. 1998. “Governing Irrigation Systemsin Nepal: Institutions, Infrastructure andCollective Action.” Institute forContemporary Studies.

Lauraya, F. M., and A. Lea R. Sala. 1994.“Alternative support systems to strengthenirrigators’ associations in Bicol, thePhilippines after irrigation managementturnover.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

Leach, M. R. Mearns, and I. Scoones. 1999.“Environmental Entitlements: Dynamicsand Institutions in Community-BasedNatural Resource Management.” WorldDevelopment 27(2): 225-247.

Levine, G. 1982. Relative water supply: Anexplanatory variable for irrigation systems.The Determinants of Developing CountryIrrigation Project Problems Technical Report6. Washington, D.C.: U.S. Agency forInternational Development.

Lise, W. 2000. “Factors influencing people’sparticipation in forest management inIndia”. Ecological Economics 34: 379-392

Logan B. I., and W. G. Mosely. 2002. “Thepolitical ecology of poverty alleviation inZimbabwe’s Communal Areas ManagementProgramme for Indigenous Resources(CAMPFIRE).” Geoforum 33: 1-14

Logo, P. 1999. “Dialectics for the promotion ofcommunity forestry in Africa from the topand from the grassroots: actual situationand prospects.” In Proceedings of theInternational Workshop on CommunityForestry in Africa: Participatory forestmanagement; a strategy for sustainableforest management in Africa. April 26–30,Banjul, the Gambia.

Maurya, P. R. 1993. “Partial turnover ofmanagement of Nigerian large scaleirrigation project to farmers: Constraintsand solution.” In Fifteenth InternationalCongress on Irrigation and Drainage, TheHague, The Netherlands. Transactions 1-E :51–65. New Delhi: InternationalCommission on Irrigation and Drainage.

Mayaka, T. 2002. “Wildlife Co-Management inthe Benoue National Park Complex,Cameroon: A Bumpy Road to InstitutionalDevelopment.” World Development 30 (11):2001-2016.

Mburu J., R. Birner, and M. Zeller. 2003.“Relative Importance and Determinants ofLandowners’ Transaction Costs inCollaborative Wildlife Management inKenya: An Empirical Analysis.” EcologicalEconomics 45: 59-73

McCarthy, J. 2000. “The Changing Regime:Forest property and Reformasi in Indonesia.”Development and Change 31: 91-129.

103

References


McKean, M. 1992. Management of TraditionalCommon Lands in Japan. In Bromley, ed.,Making the Commons Work: Theory,Practice and Policy. San Francisco, CA:ICS Press.

Meinzen-Dick, R. and Di Gregorio. 2004.“Collective Action and Property Rights forSustainable Development.” Overview. 2020Vision for Food, Agriculture, and theEnvironment., Brief No. 1. Washington, DC:IFPRI.

Meinzen-Dick, R., M. Mendoza, L. Sadoulet, G.Abiad-Shields, and Subramanian. 1997.“Sustainable water users’ associations:Lessons from a literature review.” In AshokSubramanian, N. Vijay Jagannathan, andRuth Meinzen-Dick, eds. “Userorganizations for sustainable waterservices.” World Bank Technical Paper 354.Washington, D.C.: World Bank.

Meinzen-Dick, R., A. Manzardo, and R.Reidinger. 1995. “Participation inirrigation.” Environment DepartmentParticipation Series Paper no. 3.Washington, D.C.: World Bank.

Merrey, D. J. 1996. “Institutional designprinciples for accountability in largeirrigation systems.” Research Report 8.Colombo: International IrrigationManagement Institute.

Mishra, V. S., and D. J. Molden. 1996.“Management turnover of West GandakIrrigation System, Nepal.” Short ReportSeries on Locally Managed Irrigation, no.14. Colombo, Sri Lanka: InternationalIrrigation Management Institute.

Mlenge, W. 2002. “Revival of customarylandcare in Shinyanga Region, Tanzania.”Forests, Trees and People Newsletter 46: 21-28.

Moench, Marcus. 1996. “Groundwater policy:Issues and alternatives in India.” IndiaCountry Paper, no. 2. Colombo, Sri Lanka:International Irrigation ManagementInstitute.

Morrow, C., and R. Hull. 1996. “Donor-InitiatedCommon Poll Resource Institutions: TheCase of the Yanesha Forestry Cooperative.”World Development 24 (10): 1641-1657.

Murombedzi, J. 2001. Committees, Rights,Costs and Benefits. Natural ResourceStewardship and Community Benefits inZimbabwe’s CAMPFIRE Programme. In,“African Wildlife and Livelihoods. ThePromise and Performance of CommunityConservation,” edited by Hulme andMurphree. James Currey Ltd, Oxford.

Musa, I. K. 1994. “Irrigation managementtransfer in Nigeria: A case of financialsustainability for operation, maintenanceand management.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

National Irrigation Administration (Philippines)203. “The Study of the Irrigator’sAssociation Strengthening Project inNational Irrigation Systems.”

Naughton-Treves L. and S. Sanderson. 1995.“Property, Politics and WildlifeConservation.” World Development, 23 (8):1265-1275



Ndibi, B., and E. Kay. 1999. “Measuring thelocal community’s participation in themanagement of community forests inCameroon.” Biodiversity and Conservation 8:255-271.

NESPAK (National Engineering ServicesPakistan). 1994. Report on field trip toDominican Republic and Mexico. Lahore,Pakistan: Water and Power DevelopmentAuthority and International Bank forReconstruction and Development.

Newman, K. and J. Webster, eds. 1993. AfricanBiodiversity: Foundation for the Future. Aframework for integrating biodiversityconservation and sustainable development.Biodiversity Support Program. Gland,Switzerland: World Wildlife Fund.

Nguyen M. Ta, and L. Thuan Ha. 1994.“Irrigation management transfer inVietnam.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

Olin, M. 1994. “Transfer of management ofwater users in stages I and II of theBhairawa-Lumbini Groundwater IrrigationProject in Nepal.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

Oorthuizen, J., and W. H. Kloezen. 1995. “Theother side of the coin: A case study on theimpact of financial autonomy on irrigationmanagement performance in thePhilippines.” Irrigation and Drainage Systems9:15–37.

Ostrom, E. 1992. Crafting Institutions for Self –Governing Irrigation Systems. San Francisco,CA: International Center for Self-Governance.

Ostrom, E. 1990. Governing the Commons: TheEvolution of Institutions for Collective Action.Cambridge, U.K.: Cambridge UniversityPress.

Ostrom, E., D. Thomas, N. Dolsak, P. Stern, S.Stonich, and E. Weber, eds. 2002. The Dramaof the Commons. Washington, D.C.: NationalAcademy Press.

Ostrom, E. 2001. Decentralization andDevelopment: The New Panacea in Keith

Dowding, Hughes and Margetts, eds. Challengesto Democracy: Ideas, Involvement andInstitutions. 2001. New York: Palgrave.

Owen-Smith and Jacobson, M. 1991.“Pastoralism in Arid and Semi-Arid North-West Namibia.” Paper presented at theNordic Man and Biosphere Meeting.

Pant, N. 1994. “The turnover of public tubewells in Uttar Pradesh: A case study of asuccessful cooperative society.” Paperpresented at the International Conferenceon Irrigation Management Transfer,September 20–24, Wuhan, China.

Pretty, J. N. 1995. “Participatory learning forsustainable agriculture.” World Development23 (8): 1247–1263.

Quisumbing, A., E. Payongayong, J. Aidoo, andK. Otsuka, 2001. “Women’s Land Rights inthe Transition to Individualized Ownership:Implications for Tree-Resource Management

105

References


in Western Ghana.” Economic Developmentand Cultural Change 50: 157-181.

Rana, J., R. P. Satyal, S. P. Rajbhandari, K. R.Sharma, and D. J. Molden. 1994.“Experience of management transfer tousers in Nepal.” Paper presented at theInternational Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

Rangan, H., and M. Lane. 2001. “IndigenousPeoples and Forest Management:Comparative Analysis of InstitutionalApproaches in Australia and India.” Societyand Natural Resources 14: 145-160.

Rao, C. S. 1994. “Farmers groups and theirviability in irrigation management transfer:A case study in Sreeramsagar Project,Andhra Pradesh, India.” Paper presented atthe International Conference on IrrigationManagement Transfer, September 20–24,Wuhan, China.

Ribot, J. 2002. Democratic Decentralization ofNatural Resources. Institutionalizing PopularParticipation. Washington, D.C.: WorldResources Institute.

Richards, M. 1997. “Common Property ResourceInstitutions and Forest Management inLatin America.” Development and Change 28:95-117.

Sagardoy, J. A. 1994. “Lessons learned fromirrigation management transfer programs.”Paper presented at the InternationalConference on Irrigation ManagementTransfer, September 20–24, Wuhan, China.

Sagun, M. 2002. “Conversation during field visitat the Sta. Maria River Irrigation System,Laguna.” Philippines July 2002.

Samad, M., D. Vermillion, S. Pusposotardjo, S.Arif, and S. Rochdyanto. 2000. “AnAssessment of the Small Scale IMT Programin Indonesia.” IIMI Indonesia CountryPaper. Colombo, Sri Lanka: IIMI.

Samad, M., and D. Vermillion, 1999.“Assessment of Participatory Managementof Irrigation Schemes in Sri Lanka. Partialreforms, partial benefits.” Research Report34. Colombo, Sri Lanka: International WaterManagement Institute.

Samad, M., and M. A. Dingle. 1995.“Privatization and turnover of irrigationschemes in Sudan: A case study of theWhite Nile pump schemes.” Draft finalreport. Colombo, Sri Lanka: InternationalIrrigation Management Institute.Duplicated.

Sarin, M. 2001. “Disempowerment in the nameof ‘participatory’ forestry? –Village forestjoint management in Uttarakhand.” Forest,Trees and People Newsletter 44: 26-35.

Schafer, J., and R. Bell. 2002 “The State andCommunity-based Natural ResourceManagement: the Case of the MoribaneForest Reserve, Mozambique.” Journal ofSouthern African Studies 28 (2): 401-420.

Shackleton, S., and Campbell, B. 2001.Devolution in natural resource management:Institutional arrangements and power shifts: Asynthesis of case studies from Southern Africa.Bogor, Indonesia: Center for InternationalForestry Research.



Shah, T., B. Koppen, D. Merrey, M. Lange, andM. Samad. 2002. “Institutional Alternativesin African Small Holder Irrigation: Lessonsfrom International Experience with IMT.:Research Report 60. Colombo, Sri Lanka:International Water Management Institute.

Shah, T., V. Ballabh, K. Dobrial, and J. Talati.1994. “Turnover of state tube wells tofarmer cooperatives: Assessment ofGujarat’s experience, India.” Paperpresented at the International Conferenceon Irrigation Management Transfer,September 20–24, Wuhan, China.

Shivakoti, G., and E. Ostrom, eds. 2002.Improving Irrigation Governance andManagement in Nepal. Oakland, CA.:Institute for Contemporary Studies.

Shrestha, K. 2001. “The backlash- Recent policychanges undermine user control ofcommunity forests in Nepal.” Forests, Treesand People Newsletter 44:62-65.

Singh, N. 2002. “Federations of communityforest management groups in Orissa:Crafting new institutions to assert localrights.” Forests, Trees and People Newsletter46: 35-45.

Sjoholm, H., and S. Luono. 2002. “The greenforest pastures of Suledo-Maasaicommunities organize to save their forestsand secure their livelihoods” Forests, Treesand People Newsletter 46: 13-20.

Songorwa, A. 1999. Community-based WildlifeManagement (CWM) in Tanzania: Are theCommunities Interested? WorldDevelopment 27 (12): 2061-2079.

Sonko, K. and Camara, K. 1999. “Communityforestry implementation in the Gambia: itsprinciples and prospects.” In Proceedings ofthe International Workshop on CommunityForestry in Africa: Participatory forestmanagement; a strategy for sustainableforest management in Africa. April 26–30,Banjul, the Gambia.

Srivastava, L. P., and Jeffrey D. Brewer. 1994.“Irrigation management transfer at PaliganjCanal, Bihar, India.” Short Report Series onLocally Managed Irrigation, no. 7. Colombo,Sri Lanka: International IrrigationManagement Institute.

Sundar, N. 2000. “Unpacking the ‘Joint’ in JointForest Management.” Development andChange 31: 255-279.

Svendsen, M. 1992. “Assessing effects of policychange on Philippine irrigationperformance.” Working Papers on IrrigationPerformance, no. 2. Washington, D.C.:International Food Policy Research Institute.

Svendsen, M., and D. Vermillion. 1994.“Irrigation management transfer in theColumbia Basin: Lessons and internationalimplications.” Research Paper 12. Colombo,Sri Lanka: International IrrigationManagement Institute.

Tang, S. Y. 1992. “Institutions and CollectiveAction: Self Governance in Irrigation.”

San Francisco, CA: Institute forContemporary Studies

Ternstrom, I. 2003. “The Management ofCommon Pool Resources: Theoretical Essaysand Empirical Evidence.” The Economic

107

References


Research Institute of the Stockholm Schoolof Economics. Stockholm.

Turral, H. 1995. “Devolution of management onpublic irrigation systems: Cost shedding,empowerment and performance, a review.”Working Paper 80. London: OverseasDevelopment Institute.

Uphoff, N. 1992. Learning from Gal Oya:Possibilities for participatory development andpost-Newtonian social science. Ithaca, NewYork: Cornell University Press.

Varughese, G., and E. Ostrom. 2001. “TheContested Role of Heterogeneity inCollective Action: Some Evidence fromCommunity Forestry in Nepal.” WorldDevelopment 29 (5): 747-765.

Vermillion, D. L. 1997. “Impacts of IrrigationManagement Transfer: A Review of theEvidence.: Research Report 11. Colombo,Sri Lanka: IIMI

Vermillion, D. L. 1992. « Irrigation managementturnover: Structural adjustment or strategicevolution?” IIMI Review 6 (2): 3–12.

Vermillion, D. L. 1989. “Small scale irrigationturnover program.” Vol. 3 of Efficientirrigation management and system turnover:Final report. Colombo, Sri Lanka:International Irrigation ManagementInstitute. Duplicated.

Vermillion, D. L., and C. Garcés-Restrepo. 1996.“Results of irrigation management in twoirrigation districts in Colombia.” IIMIResearch Report 4. Colombo, Sri Lanka:International Irrigation ManagementInstitute.

Vermillion, D. L., M. Samad, N. Murthy, S. K.Raheja, U. Amarasinghe, and M. Svendsen.1996. “A standard methodology to assessthe impacts of irrigation managementtransfer.” Colombo, Sri Lanka: InternationalIrrigation Management Institute, Colombo.Draft (duplicated).

Vermillion, D. L., and C. Garcés-Restrepo. 1994.“Irrigation management transfer inColombia: A pilot experiment and itsconsequences.” Short Report Series onLocally Managed Irrigation, no. 5. Colombo,Sri Lanka: International IrrigationManagement Institute.

Vermillion, D., W. Xinyuan, Z. Xiying, and M.Xuesen. 1994. “Institutional reform in twoirrigation districts in North China: A casestudy from Hebei Province.” Paperpresented at the International Conferenceon Irrigation Management Transfer,September 20-24, Wuhan, China.

Vermillion, D., and S. H. Johnson III. 1990.“Turnover and irrigation service fees:Indonesia’s new policies to achieveeconomically sustainable irrigation.”Irrigation and Drainage Systems 4 (3): 231–247.

Virtanen, P. 2003. “Local Management ofGlobal Values: Community-Based WildlifeManagement in Zimbabwe and Zambia.”Society and Natural Resources 16: 179-190

Wade, R. 1988. Village republics: economicconditions for collective action in south India.Cambridge, U.K.: Cambridge UniversityPress.



Wainwright, C. and W. Wehrmeyer. 1998.“Success in Integrating Conservation andDevelopment? A Study of Zambia.” WorldDevelopment 26 (6): 933-944.

Wester, Philippus, A. During, and J. Oorthuizen.1995. “Locally managed irrigation in theSenegal River valley in the aftermath ofstate disengagement.” Short Report Serieson Locally Managed Irrigation, no. 9.Colombo, Sri Lanka: InternationalIrrigation Management Institute.

White A. and A. Martin 2002. Who owns theworld’s forests? Forest tenure and publicforests in transition. Forest Trends andCenter for International EnvironmentalLaw. Washington DC: Forest Trends.

Wijayaratna, C. M. 1993. “Irrigation turnoverand self-management: The Philippineexperience.” Colombo, Sri Lanka:International Irrigation ManagementInstitute. Duplicated.

Wijayaratna, C. M., and D. L. Vermillion. 1994.“Irrigation management turnover in thePhilippines: Strategy of the NationalIrrigation Administration.” Short ReportSeries on Locally Managed Irrigation, no. 4.Colombo. Sri Lanka: International IrrigationManagement Institute.

Wily, L. 1997. “Villagers as Forest Managers andGovernments: ‘Learning to Let Go’ –theCase of Duru-Haitemba and Mgori Forestsin Tanzania.” Forest Participation SeriesNo.9. International Institute forEnvironment and Development.

Wily, L. 2002a. “The political economy ofcommunity forestry in Africa- Getting the

power relations right.” Forest, Trees andPeople Newsletter 46: 4-12.

Wily, L. 2002b. Participatory forest managementin Africa: An overview of progress andissues. Draft manuscript. www.cbnrm.net.

Wily, L. 2001. “Forest Management andDemocracy in East and Southern Africa:Lessons from Tanzania.” Gatekeeper SeriesNo.95. International Institute forEnvironment and Development.

Wily, L. 1999. “Moving Forward in AfricanCommunity Forestry: Trading Power, NotUse Rights.” Society & Natural Resources 12:49-61.

Wily, L., A. Akida, O. Haule, H. Haulle, S.Hozza, C. Kavishe, S. Luono, P. Mamkwe,E. Massawe, S. Mawe, D. Ringo, M. Makiya,M. Minja, and A. Rwiza. 2000. “Communitymanagement of forests in Tanzania: A statusreport at the beginning of the 21st century.”Forests, Trees and People Newsletter 42: 36-45.

Wily, L., and P. Dewees. 2001. From Users toCustodians: Changing Relations between Peopleand the State in Forest Management inTanzania. Policy Research Working Paper2569. Washington, D.C.: World Bank.

Woodhouse, P. 1997. “Governance and LocalEnvironmental Management in Africa.”Review of African Political Economy 74: 537-547.

Wunder, M. 1997. “Of Elephants and Men:Crop Destruction. CAMPFIRE and wildlifemanagement in Zambezi Valley,Zimbabwe.” Doctoral Dissertation. AnnArbor, Michigan: University of Michigan.

109

References


Yap-Salinas, L. H. 1994. “Converging factors inthe successful transfer of irrigationmanagement responsibilities to water users’associations in the Dominican Republic.”

Paper presented at the InternationalConference on Irrigation ManagementTransfer, September 20-24, Wuhan, China.

Public Disclosure Authorized Devolution of Resource Rights...

Documents

Transcript of Public Disclosure Authorized Devolution of Resource Rights...