DRAFT CONFERENCE EDITION PE OVERTY NVIRONMENTN...

POVERTY

ENVIRONMENT

NEXUS

Sustainable approaches to poverty reductionin Cambodia, Lao PDR and Vietnam

Environmental and Social Development UnitEast Asia and Pacific RegionTHE WORLD BANK1818 H Street, N.W.Washington, D.C. 20433Telephone: 1-202-473-1000www.worldbank.org/eapenvironment

The World Bank in Cambodia113 Norodom BoulevardPhnom Penh, CambodiaTelephone: 855-23-217-304www.worldbank.org/cambodia

The World Bank in Lao PDRPatou Xay - Nehru RoadP.O. Box 345 code 01004Vientiane, Lao PDRTelephone: 856-21-414-209www.worldbank.org/lao

The World Bank in Vietnam63 Ly Thai To StreetHanoi, VietnamTelephone: 84-4-934-6600www.worldbank.org/vietnam

THE WORLD BANK

PO

VE

RT

YE

NV

IRO

NM

EN

TN

EX

US

TH

E W

OR

LD

BA

NK

Sustainable approaches to poverty reductionin C

ambodia, Lao PD

R and V

ietnam

D R A F T C O N F E R E N C E E D I T I O N

POVERTY

ENVIRONMENT

NEXUSSustainable approaches to poverty reductionin Cambodia, Lao PDR and Vietnam

THE WORLD BANK


June 2006

This publication is available online at <<www.worldbank.org/eapenvironment>>.

Front cover photos: Uwe Deichmann, World Bank Photo LibraryCover design: Circle Graphics, Jostein Nygard

Environment and Social Development DepartmentEast Asia and Pacific RegionThe World BankWashington, D.C.June, 2006

This volume is a product of an expert team from Cambodia, Lao PDR and Vietnam and the staff ofthe World Bank. The findings, interpretations, and conclusions expressed in this paper do not nec-essarily reflect the views of the Executive Directors of The World Bank or the governments they rep-resent. The World Bank does not guarantee the accuracy of the data included in this work. Theboundaries, colors, denominations, and other information shown on any map in this work do notimply any judgment on the part of The World Bank concerning the legal status of any territory orthe endorsement or acceptance of such boundaries.

The material in this publication is copyrighted. Copying and/or transmitting portions or all of thiswork without permission may be a violation of applicable law. The World Bank encourages dissem-ination of its work and will normally grant permission to reproduce portions of the work promptly.

For permission to photocopy or reprint any part of this work, please send a request with completeinformation to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923,USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com.

All other queries on rights and licenses, including subsidiary rights, should be addressed to the Officeof the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail [email protected].

P O V E R T Y A N D E N V I R O N M E N T N E X U S iii

ACKNOWLEDGMENTS v

ABBREVIATIONS AND ACRONYMS ix

FOREWORD xiii

PART I: STUDY CONTEXT AND METHODOLOGY 1Introduction 3

1 Study Framework and Methodology 21

PART II: COUNTRY STUDY FINDINGS 412 Poverty-Environment Nexus in Cambodia 433 Poverty-Environment Nexus in Lao PDR 754 Poverty-Environment Nexus in Vietnam 123

PART III: SUB-REGIONAL PERSPECTIVES 1755 Sub-regional PEN Findings and Conclusions 177

Contents

This report is the result of a collaborativeresearch effort by Laotian, Vietnamese, Cambo-dian and international staff.

In Cambodia, Hang Chuon Naron and RosSeilava from the Ministry of Economics andFinance and Kim Saysamalen from the Ministryof Planning made the first recommendations forthe scope of the PEN work. A case study onaccess to natural resources and poverty was leadby Chea Sarin, World Bank consultant. Furtherconsultations have been held with Khieu Muth,Pak Sokharavuth, Long Rithirak and ChuonChanrithy from the Ministry of Environment,Ponn Narith, Peou Samy, Hang Samoeun, KhunSokha and Ross Sovann from the National Com-mittee for Disaster Management, Mao Saray,Chan Darong and Bouy Kim Sreang from theMinistry of Rural Development, Susan Novak,Phin Sopheap and Solieng Mak from the Mekong River Commission, M S Shivaku-mar, advisor, Andrew Cock from the NGOForum of Cambodia, Sik Boreak from WFPPhom Penh, Declan O’Leary from Danida, Sam-reth Sok Heng and Claire Liousse from theUrban Resource Centre Cambodia, Peter Feld-man from Partners for Development, ChanSophal from Cambodia Development ResourceInstitute, Ingrid Cyimana and Han Phouminfrom UNDP Phnom Penh, and Anthony J. Jude,Chamroen Ouch and Paul Van Im from ADB.

In Lao PDR, an inter-ministerial projectteam was assigned to work on the PEN study bythe Committee of Planning and Investment(CPI) under the direction by SingkhamKhongsavanh. The work has been managed byMonemany Nhoybouakong, National ProjectDirector, and supported by KhampadithKhammounheuang from the EnvironmentResearch Institute (ERI) under the Science,Technology and Environment Agency (STEA).Phouthone Sophatilath together with Soun-thone Ketphan and Kamphay Manivong fromthe National Agriculture and Forestry ResearchInstitute (NAFRI) under the Ministry of Agri-culture and Forestry (MAF) implemented thestudy on Non-Timber Forest Products (NTFP)and poverty; Bounmy Souksavath in coopera-tion with Pho Ngeun Soulsavath from theDepartment of Roads under the Ministry ofCommunication, Transport, Post and Con-struction (MCTPC) implemented the study onRoad and Poverty; Soulideth Souvannavongtogether with Lamphoukeo Kettavong fromERI implemented the study on UnexplodedOrdinance (UXO) and poverty; TayphasavanFengthong with Onechanh Keosavanh andVilayvone Mangkhaseuam from the Environ-ment Hygiene unit of the Ministry of Healthimplemented the study on Rural Water and Sani-tation and poverty; and Chanthala Phimmachack

P O V E R T Y A N D E N V I R O N M E N T N E X U S v

Acknowledgements

together with Bounno Fongkhamdeng andVongsak Malivanh from the Urban ResearchInstitute (URI) implemented the study onUrban Water and Sanitation and poverty. Seng-many Keolangsy and Somphath Souvannavongfrom the National Statistical Centre assisted incase study design and national poverty andenvironment analyses.

In Vietnam, TranThi Minh Ha, the Interna-tional Cooperation Department of the Ministryof Natural Resources and Environment(MoNRE), coordinated the overall PEN initia-tive with assistance from Nguyen Thi ThuyDuong on land administration and povertyissues and Nguyen Viet Thang on industrial pol-lution and poverty. Implementation of the casestudy on land administration and poverty wasled by Tran Nhu Trung from the ConsultancyServices and Technology Development Com-pany for Natural Resources and Environment(TECOS). In addition, the TECOS teamincluded Tran Chuong Huyen, Dao TrungChinh, Nguyen Van Truong, Ho Thi Yen Thu,Ngo The An, Mai Van Trinh, Nguyen ThiHong Nhung, Hoang Hong Hue, NguyenThuy Phuong and Nguyen Anh Quan. Imple-mentation of the case study on industrial pollu-tion and poverty was led by Nguyen Van Taifrom the Environment Department underMONRE. A team from the Research Center ofEnergy and Environment (RCEE), whichincluded Nguyen Trung Thang, Nguyen ThiThuy Duong, Hoang Van, and Nguyen DucMinh, implemented the study. The case studyon pesticide use and poverty in the MekongRiver Delta was lead by Nguyen Huu Dungfrom the University of Ho Chi Minh City andKhuc Xuyen from the Center of Occupationaland Environmental Health. A study on environ-mental health and poverty involved Truong VietDzung and Tran Duc Thuan from the Depart-ment of Science and Technology of the Ministryof Health together with Tran Van Dan, consul-tant. Le Mihn Duc and Nguyen Tuan Anh fromthe Science, Tehcnogology and Environment

Division of MPI coordinated a case study onmulti-PEN subjects in the Cau River (sub-)basin, while study implementation was led byVu Tuan Anh from the Socio-Economic Devel-opment Centre (SEDEC) and also includedNguyen Xuan Mai from the Vietnamese Acad-emy of Social Sciences, Nguyen Thi Bich Vanfrom the Centre for Public Health and Devel-opment in Hanoi, and Nguyen Duc Minh fromRCEE. Nguyen Van Pham from the GeneralStatistical Office led a team that provided statis-tical updates for the study.

The World Bank team comprised mostly ofstaff of the Environment and Social Develop-ment Division of the East Asia and PacificRegion, which is managed by Maria TeresaSerra. Specific directions were provided byMagda Lovei, Sector Manager of the Environ-ment Team. The World Bank team that sup-ported the three national teams in their casestudy preparations, drafted the main report textand managed databases. The team includedJostein Nygard (task team leader), Nat Pinnoi,Uwe Deichmann, Andrew Murray, Marija Kuz-manovic (Bank staff), Bjorn Larsen, ClaudeSaint-Pierre, Bruce McKenney, AJ Bostian andPamila Aggarwal (consultants). Bjorn and Claudedrafted the initial PEN guidelines adopted in eachcountry, assisted the local case study teams andperformed report writing on the Lao PDR andVietnam chapters. Bruce and Bjorn finalized theCambodia chapter. Administrative assistance tothe Bank team was provided by Indra Raja,Natalya Gosteva, Giang Thanh Huong Le, HoaThi Phuong Kieu, Vatthana Singharaj andPoonyanuch Chockanapitaksa.

Initially, the report built upon a Phase I PENreport finalized by David Wheeler, Susmita Das-gupta, Uwe Deichmann and Craig Meisnerfrom the Bank’s Development Research Groupin 2002. Susmita and Craig have also coordi-nated the pesticide–poverty case study in Viet-nam referred to in this report.

First drafts of the PEN case studies for eachcountry were presented and discussed at the

A C K N O W L E D G E M E N T S

P O V E R T Y A N D E N V I R O N M E N T N E X U Svi

national consultation workshops in Hanoi inJune 2005, in Vientiane in August 2005 and inPhnom Penh in December 2005. Conferenceand workshop information can be obtained on the following website: www.worldbank.org/eapenvironment/news. Local PEN workshopswere held in several case study provinces in 2004to finalize PEN implementation guidelines anddiscuss case study design.

Peer reviewers included Jan Bojö, BenuBidani, Robert Swinkels (World Bank staff) andBerit Aasen (Norwegian Institute for Urban andRegional Research). Additional inputs, com-ments and reviews were provided by Dan Biller,Phillip Brylski, Rob Crooks, Giovanna Dore,Paavo Eliste, Patchamuthu Illangovan, JulienLabonne, Santanu Lahiri, Samuel Lieberman,Magda Lovei, Lars Lund, William Magrath,Thomas Meadley, Stephen Mink, Glenn Morgan, Nalithone Phonyaphanh, Ian Porter,Kaspar Richter, Klaus Rohland, Jan WillemRosenboom, Steven Schonberger, JitendraShan, Susan Shen, Maria Teresa Serra, Priya

Shyamsundar and Jian Xie, all World Bank staff.Elske van de Fliert from FAO-Hanoi, JohnPatterson Vietnam-Canada Environment Proj-ect; Duong Hong Dat Vietnam Plant Protec-tion Association; Sarah Bales MoH in Vietnam,Frank Radstake, consultant; Craig Leisher WWF,and Kiju Han Korea Institute for IndustrialEconomics and Trade all provided comments.

The report was edited by Robert Livernash,consultant. Circle Graphics designed and man-aged the desktopping. Production was super-vised by Jaime Alvarez. Photos provided byUwe Deichmann and the World Bank PhotoLibrary.

Finally, we would like to express our gratitudeto the Government of Norway, which providedthe main trust funds (TFESSD) to carry out thestudy and to the Government of the Netherlands(BNPP) for providing funds for further publica-tions and dissemination of study results. Thestudy was also supported by the World Bank’senvironmental mainstreaming funds.

E X E C U T I V E S U M M A R Y

P O V E R T Y A N D E N V I R O N M E N T N E X U S vii

ADB Asian Development BankADI Acute Diarrheal IllnessARI Acute Respiratory IllnessBOD Biological Oxygen Demand CAS Country Assistance StrategyCCB-NREM Commune and Community Based Natural Resource and Environment

ManagementCEMMA Committee for Ethnic Minorities and Mountainous Areas (Vietnam)COD Chemical Oxygen DemandCPI Committee for Planning and Investment (Lao PDR)CPR Common Property ResourcesCPRG Comprehensive Poverty Reduction and Growth Strategy (Lao PDR)CSES Cambodia Socio-Economic SurveyCV Craft VillageDANIDA Danish International Development AgencyDECRG Development Research Group (of the World Bank)DFID UK Department for International DevelopmentDHS Demographic and Health SurveyDoSTE Department of Science and Technology Education (Vietnam)EASES Environment and Social Sector Unit of the East Asia and Pacific Region

(of the World BankERI Environmental Research Institute (Lao PDR)FA Forest Administration (Cambodia)FAO Food and Agriculture OrganizationGDP Gross Domestic Product GNI Gross National IncomeGPS Global Positioning SystemGSO General Statistical Office (Vietnam)HH Household

P O V E R T Y A N D E N V I R O N M E N T N E X U S ix

Abbreviations and Acronyms

IBRD International Bank for Reconstruction and DevelopmentICEM International Council for Educational MediaIDA International Development AssociationIDS Institute for Development StudiesIFPRI International Food Policy Research InstituteIFSR Independent Forest Sector Review (Cambodia)INEST Institute of Environmental Science and Technology (Vietnam)IPM Integrated Pest Management ISO International Organization for StandardizationIT Information TechnologyJICA Japan International Cooperation AgencyKEI Korea Environment InstituteKIET Korean Institute for Industrial Economics and TradeLECS Lao Expenditure and Consumption Survey (Lao PDR)LTC Land Tenure CertificateMAF Ministry of Agriculture and Forestry (Lao PDR)MARD Ministry of Agriculture and Rural Development (Vietnam)MCTPC Ministry of Communication, Transport, Post and ConstructionMDG Millennium Development GoalsMoC Ministry of Construction (Vietnam)MOH Ministry of Health (Cambodia, Lao PDR, Vietnam)MoI Ministry of Industry (Vietnam)MoNRE Ministry of Natural Resources and Environment (Vietnam)MoST Ministry of Science and Technology (Vietnam)MPI Ministry of Planning and InvestmentMRC Mekong River CommissionMRD Mekong River DeltaNAFRI National Agriculture and Forestry Research Institute (Lao PDR)NDSP National Strategic Development Plan (Cambodia)NEAP National Environmental Action Plan (Lao PDR)NGO Non-Governmental OrganizationNGPES National Growth and Poverty Eradication Strategy (Lao PDR)NPEP National Poverty Eradication Program (Lao PDR)NPRS National Poverty Reduction Strategy (Cambodia)NSC National Statistic Center (Lao PDR)NSDP National Strategic Development Plan (Cambodia)NSEDP National Socio-Economic Development Plan (Lao PDR, Vietnam) NTFP Non-timber Forest ProductsPEN Poverty and Environment NexusPPA Participatory Poverty AssessmentPRSP Poverty Reduction Strategy PaperRCEE Research Center for Energy and Environment (Vietnam)RRD Red River Delta

A B B R E V I A T I O N S A N D A C R O N Y M S

P O V E R T Y A N D E N V I R O N M E N T N E X U Sx

SEDEC Socio-Economic Development Center (Vietnam)SFE State Forest EnterpriseSIDA Swedish International Development Cooperation AgencySME Small and Medium EnterpriseSPE Seriously Polluting EstablishmentSTEA Science, Technology and Environment Agency (Lao PDR)TECOS Consultancy Service and Technology Development Company for

Natural Resources and Environment (affiliated with MONRE)TVIE Township and Village Industrial EnterpriseUNDP United Nations Development ProgramUNICEF United Nations Children’s FundUNPF United Nations Population FundURI Urban Research Institute (Lao PDR)UXO Unexploded OrdinanceVDG Vietnam Development GoalsVEPA Vietnam Environmental Protection AgencyVND Vietnamese Currency (Dong)VNHS Vietnam National Health SurveyVNLS Vietnam National Living Standard SurveyWASA Water Supply Authority (Lao PDR)WBG World Bank Group’WFP World Food ProgramWHO World Health OrganizationWSS Water Supply and Sanitation

A B B R E V I A T I O N S A N D A C R O N Y M S

P O V E R T Y A N D E N V I R O N M E N T N E X U S xi

This is a draft edition of the Poverty Environ-ment Nexus (PEN) report for Cambodia, LaoPDR and Vietnam, which will be presented atthe sub-regional PEN conference in Vientiane,Lao DPR on June 21–22, 2006.

The purpose of this conference edition is topresent the findings from the studies that havebeen undertaken in each country over the lastthree years as well as to obtain relevant com-ments and feedback from the conference partic-ipants that could be included in the final editionof the report.

We realize that the material presented in thisreport requires time to study and digest. Some ofthe linkages between poverty and environmentare complex and difficult to capture. You, as thereader of this report, are, therefore, encouragedto give yourself sufficient time to begin to dis-cern the main storylines.

The material presented in this report is basedupon comprehensive case studies as well asnational analytical work performed in eachcountry. We have omitted this backgroundmaterial from this conference edition to avoidoverwhelming the readers. This work will beincluded, however, in a CD-ROM attachmentin the final published version.

As the team of authors is still working on theoverall conclusions and recommendations fromthe report, we have decided to postpone theinclusion of a comprehensive summary to thereport until its final publication.

Wish you good reading of this edition andlooking forward to receiving your comments.

PEN Report Authors June 2006

P O V E R T Y A N D E N V I R O N M E N T N E X U S xiii

Foreword to the Conference Edition

STUDY CONTEXT

AND METHODOLOGY

I

P E N S T U D Y I N T H E L O W E R M E K O N G S U B - R E G I O N 3

DEVELOPMENT OF THE PEN STUDY

Past World Bank PEN Work in the Sub-region

In order to understand the extent and significanceof the PEN in the lower Mekong sub-region andto expand the dialogue between the World Bankand the three countries on the issue, the WorldBank launched the first phase of the PEN studyin 2001.

Based upon secondary data sources, the studyexamined seven main environmental indicators1

for mainly province-level correlations betweenenvironment (deforestation, fragile lands, indoorair pollution, lack of access to water, lack of accessto sanitation, outdoor air pollution, and numberof diarrhea cases) and poverty and found that thenexus is differs between countries.

According to the study findings, the nexus inCambodia is confined primarily to ‘brown’ issuessuch as indoor air pollution, water contamina-tion and lack of adequate sanitation, in Lao PDRit spans across all seven environmental issues,whereas in Vietnam, the study results were some-what eclectic, suggesting a possible existence of anexus for fragile soils and indoor air pollution, butrelatively weak correlations for the remainingindicators (see Figure I.1).

Data availability, particularly to the lack ofdistrict-level and more micro-level data, such

as village and commune data, however, con-strained the study. Moreover, the study foundthat the PEN structures are often masked atmore aggregated levels and may be revealedonly through study at more local levels. Thesefindings and challenges were taken into con-sideration in the shaping of the second phaseof the PEN study that is summarized in thisreport.

Motivation and Objectives for the Study

This second phase of the PEN study is designedto address several of the unsolved issues facedby the first phase and within the existing PENliterature on Cambodia, Lao PDR and Viet-nam. Since the earlier study found that thenexus may often be identified at disaggregated,or local, levels, the analysis in the second phaseof the study was conducted both at the regional,provincial and district levels for the national stud-ies (macro-levels) and on the district, village, com-mune, and household levels in the (case-) studies(micro-level). This combination of macro andmicro approaches results in more substantivefindings, which are elaborated in detail in theindividual country chapters. The intention is thatthese elaborations may then lead to a betterscope for defining policy interventions.

Introduction

�

The study seeks to reflect the geographicdiversity of the poverty and environmentalchallenges, so it is sufficiently large in bothgeographic and sectoral scope (see discussion inthe individual country chapters 2–4). It is alsodesigned to be policy- and response-driven, soit not only seeks to identify referred PEN hot-spots, where the poverty-environment dynam-ics is particularly strong, but also to providebasis for mainstreaming the PEN findings intopoverty reduction programs in the three coun-

tries. The gradual conversion of the joint strate-gies for environmental protection and povertyreduction into actual operations is the mainlong-term development objective of this ana-lytical work.

In summary, the study had three main objec-tives: (i) further identify and elaborate uponPENs in the three countries, (ii) review policyoptions vis-à-vis these identified PENs and (iii)establish links to ongoing PEN project initia-tives in the sub-region (see Table I.1).

I N T R O D U C T I O N

P E N S T U D Y I N T H E L O W E R M E K O N G S U B - R E G I O N4

M A P I . 1 Provinces and Socio-Economic Indicators for the PEN Study in the Sub-region

Sources: World Development Report 2005; Poverty Assessment Papers for Cambodia, Lao PDR and Vietnam

Regional and Country Context

Cambodia, Lao PDR and Vietnam face manycommon challenges in their quest for economicdevelopment and poverty reduction (see figureI.2). An important shared feature is that povertyand environment issues in Lao PDR, Cambodiaand Vietnam fall into one of two broad cate-gories: environmental health and natural resourceuse. This feature is not specific only to these coun-tries, as there is a body of research on poverty andenvironment linkages that gives this finding globalsignificance.2 However, it is a particularly rele-vant classification, as the national policy frame-works in the three countries already incorporateboth the issues of clean water and sanitation andof sustainable use of natural resources into devel-opment and poverty reduction strategies.

Based upon both the specific and generalenvironmental and poverty features in the sub-region, several poverty—environment initiativeshave been generated by donors over the last years(see table I.1). While most of them originallyaddressed poverty—natural resource managementissues, newer initiatives have also increasinglyaddressed poverty—environmental health issues.Some of these newer studies have also had multi-sectoral PEN approaches. However, this studymay be among the first that tries to addresspoverty environmental issues throughout thenatural resource management and environmen-tal health subjects at both the national and sub-regional levels. As such, this study strives tomake a contribution towards a comprehensiveunderstanding of the complexity of poverty—environment linkages in the sub-region.



-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

DeforestationRate

Fragile Land(Slope>16%)

Indoor AirPollution(wood/

charcoal use);Vietnam: ARI

No Access to Clean

Water

NoSanitation;Vietnam:notoilets

Number ofDiarrhea

Cases

Outdoor AirPollution

(number ofdeaths from

PM10)

Indicator

Ran

k co

rrel

atio

n w

ith

po

vert

y

Cambodia Lao PDR Vietnam

F I G U R E I . 1 Correlations Between Poverty and Different Environmental Indicators Foundfrom Previous PEN Work in Cambodia, Lao PDR and Vietnam

Specific to VietnamAbsolute number of poor concentrated in low lands. Strong governmentinstitutions

Widely diversified economyreflecting many sectors with PEN features.

Shared by Laos and Vietnam Upland poverty: Indochina Cordillera Poverty reduction policy targeting the uplands Mainly planned economy in transition

Specific to Lao PDR Rural-to-rural migration at core of Government policy

Shared by Vietnam & Cambodia

Urban poverty

Specific to Cambodia Post-war migration Flood regime in the central region Mainly market economy

Shared by Lao PDR and Cambodia

Weak institutional & legal capacity

Shared by the three countriesAccess to natural resources critical to the poorest’s basic livelihood conditions Disproportionately low access to clean water and sanitation by the poor. Rural poverty traps in environmentally-fragile uplands and marginal areas Growing urban poverty with environmental health issues Migration

What Laos, Cambodia and Vietnam also havein common is fast-growing economies that allowfor rapid overall poverty reduction, but also“poverty traps” remaining in the rural sector.These poorer areas are often, although not always,environmentally sensitive areas. High economicgrowth rates occur jointly with the growth ofurban centers3. Urban poverty rates are still low inVietnam and Laos and decreasing in Cambodia(in geographically comparable areas of PhnomPenh, poverty dropped from 11.4 percent to 4.6percent between 1993/94 and 2004), but the issueof low-income groups is expected to become sub-stantial in the future. The urban poor populationis already facing specific environmental healthissues in Vietnam, and so are the populationsemployed in heavy industries, but, in general, thenational policy frameworks appear to devoteinsufficient attention to this growing issue.

In spite of the many common geo-politicaland socio-economic circumstances, however,it is important to acknowledge that Cambodia,Lao PDR and Vietnam also confront uniquechallenges in combating poverty and environ-mental problems. Their distinctive natural endow-ments and institutional and policy responses intheir economic development, including thetransition from socialism to more market-oriented economies, have resulted in relativelydistinct development paths, and consequentlydistinct priorities for environmental manage-ment and poverty reduction. The following sub-sections deal with the specific national contextsof the three countries, which will contribute toa better understanding of the differences inscope, approaches and coverage of poverty andenvironment linkage areas between the threecountries.



F I G U R E I . 2 General Features of Environment and Poverty in the Lower Mekong Sub-region



TA

BL

E I

.1R

elev

ant

PEN

Stu

die

s in

th

e Lo

wer

Mek

on

g S

ub

-reg

ion

, 200

5 (t

o b

e co

mp

lete

d)

Stu

dy

&D

on

or

Init

iati

ve t

itle

met

ho

do

log

ies

Sect

or

focu

sC

ou

ntr

y fo

cus

Key

issu

es a

dd

ress

edIm

ple

men

tati

on

Inte

rnat

ion

al O

rgan

izat

ion

sW

B

AD

B/C

IFO

R

AD

B

UN

DP/

DFI

D

Bila

tera

l Ag

reem

ents

DA

NID

A

SID

A

*AD

B—

Asi

an D

evel

op

men

t B

ank;

CIF

OR

—C

ente

r fo

r In

tern

atio

nal

Fo

rest

ry R

esea

rch

; DA

NID

A—

Dan

ish

Inte

rnat

ion

al D

evel

op

men

t A

ssis

tan

ce; D

FID

—D

epar

tmen

t fo

rIn

tern

atio

nal

Dev

elo

pm

ent

(UK

); U

ND

P—U

nit

ed N

atio

ns

Dev

elo

pm

ent

Pro

gra

m

“Po

vert

y En

viro

nm

ent

Nex

us

in C

amb

od

ia,

Lao

PD

R a

nd

Vie

tnam

”

“Map

pin

g t

he

Lin

kB

etw

een

Po

vert

y an

dFo

rest

Dep

end

ence

inth

e M

eko

ng

Reg

ion

for

Imp

rove

d P

ro-p

oo

rPo

licie

s an

d P

roje

cts”

“Har

mo

niz

ing

Po

vert

yR

edu

ctio

n a

nd

En

vi-

ron

men

tal G

oal

s in

Polic

y an

d P

lan

nin

gfo

r Su

stai

nab

le

Dev

elo

pm

ent”

“Map

pin

g o

f M

on

ito

rin

gEf

fort

s o

n P

ove

rty

and

Nat

ura

l Res

ou

rces

Envi

ron

men

tal

Man

agem

ent”

Cas

e st

ud

ies;

map

-p

ing

PEN

co

rrel

a-ti

on

s; w

ork

sho

ps

Cas

e st

ud

ies;

m

app

ing

fo

rest

/p

ove

rty

corr

ela-

tio

ns;

wo

rksh

op

s

Cas

e st

ud

ies;

eco

log

y-ec

on

om

y m

od

els;

anal

yzin

g, m

od

el-

ing

an

d p

iloti

ng

bes

t p

ract

ices

Fore

stry

, ag

ricu

l-tu

re, i

nd

ust

ry,

po

lluti

on

m

anag

emen

t

Fore

stry

Pove

rty

and

air

po

lluti

on

Fore

stry

, ag

ricu

l-tu

re, i

nd

ust

ry,

ener

gy,

po

lluti

on

man

agem

ent

Mek

on

g

Sub

-reg

ion

(Cam

bo

dia

,La

o P

DR

,V

ietn

am)

Mek

on

g

Sub

-reg

ion

(Cam

bo

dia

,La

o P

DR

,V

ietn

am)

Ho

Ch

i M

inh

Cit

y,V

ietn

am

Vie

tnam

Lin

kag

es b

etw

een

po

vert

y an

d

acce

ss t

o n

atu

ral

reso

urc

es; f

ore

stre

sou

rce

man

age-

men

t; a

cces

s to

wat

er a

nd

san

ita-

tio

n; p

esti

cid

e u

se;

and

ind

ust

rial

p

ollu

tio

nFo

rest

& p

ove

rty

link-

ages

; fo

rest

dep

en-

den

ce; i

nn

ova

tive

app

roac

hes

to

po

vert

y al

levi

atio

nth

rou

gh

fo

rest

ryH

ow

to

ad

dre

ss p

oo

rp

eop

le d

isp

rop

or-

tio

nat

ely

affe

cted

by

Air

Po

lluti

on

Fore

stry

an

d a

qu

a-cu

ltu

re; w

ater

su

p-

ply

an

d s

anit

atio

n;

(ru

ral)

ren

ewab

leen

erg

y, s

ust

ain

-ab

le e

ner

gy

use

2003

–200

6

2004

–200

5

2005

–200

7

2005

–200

9

Cambodia

Economic and social context

Cambodia is a small country in Southeast Asiaencompassing a territory of about 181,000 km2and a population of 13.6 million in 2004. Thecountry’s economic development in the pastthirty years has occurred against a backdrop ofpolitical instability. The rule of the Khmer Rougefrom 1975 to 1979 and the persistent civil strifefollowing the regime’s decline have had lastingconsequences on the country’s economic infra-structure and its human and natural capital. Eventhough the country is now enjoying peace andgreater political stability, it is still struggling toovercome its long legacy of conflict.

The 1990s witnessed a major recovery forthe Cambodian economy. Vigorous economicgrowth of about 7.1 percent per year has resultedin declining poverty rates and increasing real percapita consumption. According to the latestWorld Bank’s Poverty Assessment draft report,poverty incidence in Cambodia declined from47 percent in 1993–94 to 35 percent in 2004 ingeographically comparable areas, and the real percapita household consumption rose about 32 per-cent from 1993 to 2004. This economic growthhas also resulted in improvements in manysocio-economic indicators, such as school enroll-ment, housing quality, etc.4

However, poverty reduction and consump-tion increases were not uniform across the



M A P I . 2 Poverty and Population Density in the Sub-region

Source: East Asia Poverty Reduction and Economic Management, World Bank

country. Rural areas experienced significantlyslower growth rates, even though they startedoff with the lowest average real consumptionin 1993/94. This has resulted in a growingincome disparity between the rural and urbanareas. The Gini coefficient, which measuresincome inequality, is estimated to have risenfrom 0.35 to 0.40.

Urban manufacturing, construction, andtourism have been the primary engines of growthin the past decade, and they have concentratedpredominantly in Siem Reap and Phnom Penh.Since studies have also shown that the poor havebenefited little from remittances,5 it is not sur-prising that this growth has not trickled downmore towards the rural areas. Only about 13 per-cent of all households received such support, andit consisted only of about 8–10 percent of thetotal household consumption.

The lack of strong development in rural areasis also attributable to the slow growth in agricul-tural production. Cambodia’s crop yields, laborproductivity, and land productivity have beenand still are among the lowest in Asia. Agricul-tural production is both dependent upon andsignificantly impacted by the floods that are aunique feature of the Lower Mekong and TonleSap region. Drought occurrence in the country isalso common. Vulnerability to natural disasters is,thus, another critical issue in poverty reductionefforts in Cambodia.

Environmental trends

The loss of forest cover, estimated at a rate of70,000 to 100,000 ha per year, is one of the mostalarming environmental trends in Cambodia.The country’s forests comprise more than 50 per-cent of the country’s territory and are, therefore,a major source of capital for the Cambodianeconomy and an important livelihood resourcefor the poor (see Map I.3). In addition to timber,which is the primary source of fuel for cookingand heat (fuelwood and charcoal comprise about80 percent of the total energy supply), forestsprovide the poor with non-timber forest prod-

ucts (NTFPs), which are used for subsistence andincome generation.

Cambodia’s poor plateau areas have large areasof fragile lands that are very susceptible to infer-tility. Their sandy soils have little organic contentand limited water retention capacity. Therefore,degradation resulting from inappropriate land use,land clearings, and natural disasters further exac-erbates the existing problem of infertile lands



M A P I . 3 Forest Cover in the Sub-region

Source: Global Land Cover 2000 database

and maintains very low agricultural yields. Asthe agricultural sector currently accounts about75–80 percent of the country’s population andaccounts for about 40 percent of the nationalGDP, such pressure on agricultural land resourcesmay have severe impacts on the overall agro-basednational economy.

Cambodia is very rich in water resources, mostimportant of which are the Mekong and TonleSap Rivers and Tonle Sap Lake. Even though theseresources are critical to the country’s continuedeconomic growth, they are increasingly stressedby human activity. Increasing demand for watercombined with pollution from agricultural runoff,industrial effluents, domestic sewage and solidwaste, are severely threatening the quantity andquality of water supply. Groundwater is simi-larly impacted. A declining water table is an issuein many areas.

Access to safe drinking water and sanitation isvery limited. Only 17.7 percent of the total pop-ulation has access to sanitation6 (see Map I.4).Limited capital prevents the establishment ofinfrastructures like sanitation and sewage systems,and, if available, limited management preventsmost poor people from access and connections.Diarrhea and cholera have become endemic dis-eases in the country.

Air quality has been declining in Phnom Penhand other urban centers. Steady increases in sulfurdioxide and nitrogen dioxide have been observedalong with increasing incidence of respiratoryproblems in children. The primary pollutionsources are industrial emissions from fossilfuels, growing transportation fleet and wasteincineration.

Growing solid and hazardous waste is also amajor issue. Urbanization, lack of integrated solid



M A P I . 4 Access to Clean Water in the Sub-region

waste management system, absence of toxic andhazardous waste disposal facilities and impropermanagement are the main culprits for these trou-bling trends.

Lao PDR


Lao PDR is a small, sparsely populated, land-locked country with extensive natural resources.It covers an area of 236,800 km2, and its popu-lation of about 5.7 million people is growingrapidly. Similarly to Cambodia and Vietnam,the country’s economic development in the pastdecades represents to some extent recovery fromthe destruction of past conflicts. Lao PDR sufferedsignificant economic losses from the IndochinaWars, particularly the heavy US bombing ofthe Ho Chi Minh trail. The consequences of thebombings are still visible today. Unexplodedordinances (UXOs), for example, represent amajor issue in the eastern and southern part ofthe country in particular and were covered bythe PEN study.

Lao PDR still has limited infrastructure, suchas roads, water, electricity and telecommunica-tions. However, it is well-endowed with naturalresources, including forests, minerals and waterresources.

Much of the country’s economic and socialcharacter has been defined by its topography(see Map I.5). The upland areas, covering alarge proportion of the northern part of thecountry and border areas in the southern part,are characterized by low potential for sustain-able production of annual crops, lack of accessto roads (about 30 percent of the villages arenot accessible during the wet season, and only20 percent can be reached by truck at any timeof the year),7 high poverty incidence and dis-proportionately high ethnic minority popula-tion. The lowlands, on the other hand (especiallyaround the capital Vientiane), have benefitedgreatly from their developed transportationinfrastructure and their proximity to markets.

With the market-oriented reforms of the 1990s,they saw increased economic growth and signif-icant poverty reduction.

On average, the country has achieved impor-tant strides in its fight against poverty. Growthhas averaged 6 percent per year between 1991and 2003 and the incidence of poverty has fallenfrom 46 percent in 1992 to 33 percent in 2003.However, as mentioned above, these benefitshave not reached the poorest, most marginalizedpopulations. Many of these communities rely onagriculture as their main source of income, eventhough their land may not be fit for cultivation.Shifting cultivation remains an important landuse system, and its application appears to haveboth its pros and cons as well as different viewsabout its possible impact on the environmentand poverty. As a matter of fact, most shiftingcultivation today is rotational agriculture andappears sustainable in districts where rural pop-ulation densities remain (e.g. below 20 inhabi-tants/km2). However many districts are alreadybeyond those levels, and land use systems poseissues of sustainability in terms of environment,household incomes, or both.


The loss of forest resources is one of the mainenvironmental problems that affect Lao PDR. Inthe 1940s forest cover was estimated at about70 percent of the total land area, while todaythat figure has decreased to about 40 percent,which is, in fact, a comparatively high rate (seeMap I.3). A related and significant environ-mental challenge is the depletion of NTFPs,which, like in Cambodia, are a widespread andsignificant income source in rural areas, includ-ing for the poor. The habitat loss for many ani-mal and plant species is a particularly worrisometrend, considering that Lao PDR is one of the most biodiversity-rich countries in South-east Asia. Loss of topsoil due to erosion isanother important issue potentially affectingall sloped land.



Even though Lao PDR has a large per capitavolume of renewable water resources, the qual-ity of both surface and groundwater is declin-ing. In addition, access to clean water andsanitation is a major problem. In rural areas,access to clean water is about 60 percent,whereas access to sanitation is estimated at about36 percent (see Map I.4). There is a high incidenceof diarrhea, dysentery and other waterborne dis-eases. Municipal solid waste is increasing rapidly,and disposal and pollution issues will featureprominently on the government’s environmen-tal agenda.

Vietnam


Vietnam is the largest and most populous coun-try in the Mekong sub-region. Its 82.2 millionpeople live on an area of about 331,700 km2,which includes lowlands in the Mekong Riverdelta in the South and the Red River delta in theNorth, and a large proportion of uplands. As inLao PDR, the different topographical features ofthe country influence its social and economiccontext (see Map I.5).

Even though Vietnam has also suffered sig-nificantly from its legacy of conflict, includingits involvement in all Indochina Wars, it hasbeen moving much faster on the path of eco-nomic recovery and consequently has the mostdeveloped and diversified economy in the sub-region.

With the second-fastest growing economy inthe world after China and estimated growth ratesof about 7 percent per year, Vietnam has man-aged to achieve impressive poverty reduction. Adecade ago, nearly 60 percent of the populationlived below the poverty line. Today that figurehas declined to less than 30 percent. Almost athird of the population—the equivalent of morethan 20 million people—was lifted out ofpoverty in less than a decade.

Progress has also been made in many socialindicators. The Vietnam Development Goals

(VDGs), which are the localized version ofthe Millenium Development Goals (MDGs),demonstrate consistent improvements in schoolenrollment, infant mortality rates, etc.

While poverty continues to decline, however,the pace of reduction has slowed, and there areincreasing disparities between urban and ruralareas. In rural areas, unemployment and under-employment are high, compelling many youngpeople to join the work force. Moreover, povertyreduction has been significantly lower amongethnic minorities compared to the rest of thepopulation.

The sustained economic expansion has pri-marily been driven by the agricultural sector (theMekong and Red River deltas are among themost productive and intensively cultivated areasin Asia), including the expansion of cash crops



M A P I . 5 Land Elevation in the Sub-region

particularly coffee and sugar cane. However,the country has also built up a relatively strongindustrial base, which has created opportunitiesfor many low-skilled workers in the cities.


Unlike in Cambodia and Lao PDR, forest coverin Vietnam has been increasing in the pastdecade, a trend that is taking place together witha partial shift of the forestry sector towards envi-ronmental objectives instead of productionobjectives, with intensification of farming sys-tems and with land tenure on sloped land (seemap I.3). At the same time, however, forestquality has continued to decline. Closed-canopyforest has dropped from almost 100 percent offorested areas a century ago to about 13 percenttoday.

Unsustainable use of fragile lands in Viet-nam is a major environmental threat, and steepslopes in upland areas make soils more suscep-tible to erosion when pressure from agriculturaluses increases. Salinization in the Mekong deltaaffects the quality of drinking water in the area.Groundwater contamination resulting fromthe salt-water intrusion impacts drinking andindustrial water as well as ecosystems and agri-cultural production

As in Cambodia and Lao PDR, access toclean water is a major problem, even though thecountry is well-endowed with water resources(see map I.4). In 2000, only 52.2 percent of thepopulation had access to safe water. With waterdemand increasing rapidly due to growingurbanization and industrial demand, access toclean water can become an even more criticalissue in the future.

Industrial pollution affects water and air qual-ity, and thus has serious health implications forthe population that may disproportionately affectthe poor. Pollution primarily affects surfacewaters. Many of Vietnam’s rivers that flowthrough urban and industrial areas are heavilypolluted. Biochemical-oxygen demand (BOD)and Ammonia-Nitrogen levels, which indicate

the extent of organic pollution in the rivers, exceednational water quality standards by several times.Lakes, small streams, and canals are also severelypolluted.

Urban air pollution is increasing due to agrowing motor vehicle fleet, and growing indus-trial emissions. Fossil fuel consumption by powerplants, which generate about 60 percent of thecountry’s energy needs, emits high levels ofpolluting gases such as sulphur dioxide, carbonmonoxide, nitrogen oxides and fine particulates.Other industrial sectors, such as metallurgy andthe chemical and cement industries, also gener-ate significant levels of pollution.

PEN STUDY DEVELOPMENT AND SUBJECTS

Study Selection and Implementation

The study selection process included not onlythe selection of poverty and environment link-age areas to be examined, but also the appropri-ate geographic locations for each study. Thelocation of the studies was chosen in order to sat-isfy two major objectives.

First, the PEN study sought to cover suffi-cient geographic area so as to adequately repre-sent the geographic diversity in each country(map I.6). For example, studies were located inupland, lowland and piedmont areas (Figure I.3),as well as in the northern, and southern areas inLao PDR and northern, central and southernareas in Vietnam, which have many geographic,social (including ethnic), economic and politi-cal diversified features.

Second, the study locations also sought toexplore the contrast between high- and low-income provinces and between other high- andlow-income administrative units within theprovinces. In Lao PDR, the geographical focus ofthe PEN studies was determined to a significantextent by the Laotian government’s selectioninputs, which reflected its own poverty prioritylocations (ex. 47 poverty priority districts).





MA

P I

.6A

cces

s to

Cle

an W

ater

in t

he

Sub

-reg

ion

In addition, in both Lao PDR and Vietnam,considerations were also made about addressingparts of the country That had not necessarily beenapproached by international agencies (donors)before. In fact, it was recognized early on that thepoorest parts of the countries had not necessarybeen the main recipients of donor assistance.8

The choice of the poverty-environment link-age areas covered by the studies was subject to awider range of criteria. The environmental prior-ity issues identified in the countries’ developmentand poverty reduction strategic plans providedcritical guidance in the selection of the PEN stud-ies. In addition, a set of selection criteria weredeveloped in the dialogue with the governments,which identified sectors affecting the highestnumber of poor people (table I.2). Other selec-tion criteria included: i) the severity of environ-mental impact (on income, health, vulnerability),whether the impact was alleviated by nationaleconomic growth, the local institutional andfinancial capacity to deal with the problem; and

ii) the cost effectiveness (including administrativecost) of interventions.

Therefore, the final decision about studyareas drew a balance between analytical researchobjectives as well as policy objectives by the hostcountries.

After careful consideration of the selectioncriteria in a dialogue with multi-ministerial rep-resentatives in each country, the following envi-ronment and poverty linkage areas were chosenfor the studies:

In Cambodia, the study examined the linkagesbetween poverty and access to natural resources,including agricultural land, forest products, graz-ing land, fisheries and drinking water.

In Lao PDR, the PEN studies examined link-ages between forest resource management (NTFP,roads and UXOs) and poverty, and between watersupply and sanitation (rural WSS and urban WSS)and poverty.

The PEN studies in Vietnam included: landadministration, environment and poverty; Cau



Cumulative share of land area by elevation zone

010

2030

4050

6070

8090

100

< 10

0

300

500

700

900

1100

1300

1500

1700

1900

2100

2300

2500

elevation

per

cen

t

Cambodia

Lao PDR

Vietnam

F I G U R E I . 3 Elevation Distribution in Cambodia, Lao PDR and Vietnam

Source: East Asia Update—Strong Fundamentls to the Fore—Regional Overview, April 2004, East Asia and Pacific Region, The World Bank.

LaoCambodia PDR Vietnam

< 100 61.5 1.4 35.5101–500 33.1 36.5 31.3501–1000 4.9 40.0 23.21001–1500 0.5 20.0 7.91501–2000 0.0 1.9 1.7> 2000 0.0 0.1 0.4



TA

BL

E I

.2Es

tim

ated

Po

pu

lati

on

Aff

ecte

d b

y En

viro

nm

enta

l/H

ealt

hIn

dic

ato

rs

Cam

bo

dia

Lao

PD

RV

ietn

am

Esti

mat

ed N

um

ber

of

Peo

ple

Aff

ecte

dEs

tim

ated

Nu

mb

er o

f Pe

op

le A

ffec

ted

Esti

mat

ed N

um

ber

of

Peo

ple

Aff

ecte

d

Envi

ron

men

t &

% o

f n

on

-po

or

% o

f p

oo

r%

of

no

n-p

oo

r%

of

po

or

% o

f n

on

-po

or

% o

f p

oo

rH

ealt

h in

dic

ato

r:p

op

ula

tio

np

op

ula

tio

np

op

ula

tio

np

op

ula

tio

np

op

ula

tio

np

op

ula

tio

n

Dec

line

in F

ore

st

reso

urc

es*

Use

of

frag

ile la

nd

Fuel

wo

od

/ind

oo

r ai

r p

ollu

tio

nLa

ck o

f w

ater

su

pp

ly**

Lack

of

san

itat

ion

Ho

use

ho

lds

alw

ays

bo

ilin

g

dri

nki

ng

wat

erD

rou

gh

t***

Flo

od

ing

***

UX

O/M

inefi

eld

co

nta

min

atio

n**

**

* C

amb

od

ia: F

rom

200

5 su

rvey

on

co

mm

un

es r

epo

rtin

g s

tatu

s an

d t

ren

ds

in n

atu

ral r

eso

urc

es, c

om

bin

ed w

ith

CSE

S 20

04 c

on

sum

pti

on

su

rvey

dat

a. *

* Th

is is

use

of

op

en,

un

pro

tect

ed w

ater

so

urc

es (

rive

r, p

on

ds

and

str

eam

s) in

Cam

bo

dia

an

d L

ao P

DR

), a

nd

op

en, u

np

rote

cted

wat

er s

ou

rces

an

d d

ug

wel

ls w

ith

nea

rby

po

lluti

on

so

urc

es in

Vie

tnam

. ***

The

dat

a ar

e n

ot

com

par

able

acr

oss

co

un

trie

s d

ue

to s

ub

stan

tial

dif

fere

nce

s in

defi

nit

ion

s. *

***

Esti

mat

ed b

y co

mb

inin

g d

ata

fro

m h

ou

seh

old

co

nsu

mp

-ti

on

su

rvey

s an

d U

XO

/min

efiel

d s

urv

eys.

Cam

bo

dia

: Min

efiel

d a

nd

Clu

ster

bo

mb

co

nta

min

atio

n f

rom

200

2 su

rvey

. Lao

PD

R: U

XO

co

nta

min

atio

n f

rom

199

7 su

rvey

.

66 p

erce

nt

Mo

re d

ata

nee

ded

90 p

erce

nt

16 p

erce

nt

60 p

erce

nt

60 p

erce

nt

12 p

erce

nt

24 p

erce

nt

19 p

erce

nt

70 p

erce

nt

Mo

re d

ata

nee

ded

99 p

erce

nt

22 p

erce

nt

80 p

erce

nt

35 p

erce

nt

16 p

erce

nt

23 p

erce

nt

28 p

erce

nt

Mo

re d

ata

nee

ded

< 20

per

cen

t95

per

cen

t

20 p

erce

nt

40 p

erce

nt

0.2

per

cen

t5

per

cen

t20

per

cen

t

Mo

re d

ata

nee

ded

< 45

per

cen

t98

per

cen

t

30 p

erce

nt

70 p

erce

nt

1.7

per

cen

t1

per

cen

t28

per

cen

t

Mo

re d

ata

nee

ded

Mo

re d

ata

nee

ded

80 p

erce

nt

15–2

0 p

erce

nt

10–1

2 p

erce

nt

80 p

erce

nt

Mo

re d

ata

nee

ded

1.3

per

cen

t

Mo

re d

ata

nee

ded

Mo

re d

ata

nee

ded

> 9

0 p

erce

nt

25–3

0 p

erce

nt

30–3

5 p

erce

nt

60 p

erce

nt

Mo

re d

ata

nee

ded

1.1

per

cen

t

River basin, poverty and environment; pesticideuse and poverty, industrial pollution and poverty;and water supply and sanitation and poverty. Themore diversified PEN study structure for Vietnamis a result of this country’s more diversified econ-omy and the consequent greater variety of envi-ronmental issues with poverty implications.

Policy Orientation

The expert team utilized the national five-yearplans, national poverty eradication strategies andpoverty reduction strategy papers, which arejointly prepared by the national governmentsand the World Bank, as an important frameworkfor the development and implementation of thePEN study. They served not only as indicators ofgovernment priorities for interventions in povertyand environment issues, but also as ways to assesswhere the biggest gaps and shortcomings are intargeting and addressing poverty-environmentissues, in order to ensure that the outcomes of thePEN study could be utilized to fill those gaps (seeTable I.3).

These documents demonstrate that there issome recognition within the governments ofCambodia, Lao PDR and Vietnam of the poten-tial for jointly addressing poverty and environ-ment issues. However, there are many issues thatdeserve greater attention and more immediateinterventions, and the PEN study seeks, amongother things, to highlight those issues and bringthem forward on the governments’ agendas.

Moreover, the PEN work has already pro-vided some valuable inputs into these docu-ments. The most substantive contributions weremade towards the Lao PDR National PovertyEradication Program and the Lao PDR PovertyAssessment. In both of these documents, a sep-arate section (NPEP) and a chapter (PA) weredevoted to findings from the PEN study. InVietnam, PEN approaches were also incorpo-rated in the Ministry of Planning and Invest-ment work related to the Five Year Plan. Lastly,the Poverty Assessment in Vietnam used theinitially-developed PEN approach of grouping

the population in quintiles (strata) to distin-guish between the different levels of access towater, sanitation and health coverage.

Inclusive Approach and Country Ownership

High country ownership was one of the mostimportant considerations in the design processfor this study. The World Bank staff workedclosely with local experts and government offi-cials from the initiation of the study to its com-pletion in order to ensure sufficient input fromthe local counterparts as well as shaping of thestudy and its outcomes to fit the government’sobjectives for poverty reduction.

The local counterparts were involved signifi-cantly into the selection of the study locationsand were responsible for implementation of localsurveys.10 They remained involved throughoutthe completion of the study through continueddialogue and data sharing with the World Bankstaff as well as through consultation workshopsand conferences that provided opportunities forknowledge-sharing and enhancing the under-standing of how the PEN findings fit into thegovernments’ poverty reduction agendas.

Partnership and Cooperation

The focus on country ownership provided basisfor continuous dialogue and cooperation betweenthe World Bank staff and the counterparts inCambodia, Lao PDR and Vietnam throughoutthe completion of the study. Cooperation wasestablished with numerous government institu-tions and local non-governmental organizations(NGOs).

In Cambodia, the authors worked closely withthe National Institute of Statistics in obtainingdata from the Population Census. In Lao P.D.R,important poverty data was supplied by theNational Statistics Center. In Vietnam, the dis-trict-level data were assembled by MoSTE, GSOand the Ministry of Health.





TA

BL

E I

.3Po

vert

y-En

viro

nm

ent

Issu

es in

Rel

evan

t A

nal

ytic

al a

nd

Pla

nn

ing

Do

cum

ents

9

Rel

evan

t co

un

try

Key

rep

ort

sta

tem

ents

/tar

get

s

do

cum

ents

Cam

bo

dia

Lao

PD

RV

ietn

am

Five

-Yea

r Pl

ans

Pove

rty

Red

uct

ion

Stra

teg

y Pa

per

s

Nat

ion

al S

trat

egic

Dev

elo

pm

ent

Plan

200

6–20

10M

ain

po

vert

y-en

viro

nm

ent

targ

ets:

incr

ease

acc

ess

to s

afe

wat

er t

o45

% in

ru

ral a

nd

85%

in u

rban

po

pu

lati

on

; in

crea

se a

cces

s to

san

itat

ion

to

22

per

cen

t o

f ru

ral

and

67

per

cen

t o

f u

rban

po

pu

la-

tio

n; i

ncr

ease

fo

rest

co

ver

to

60 p

erce

nt

of

the

tota

l are

a; a

nd

red

uce

fu

el w

oo

d d

epen

den

cy in

ho

use

ho

lds

fro

m a

bo

ut

85.5

per

-ce

nt

to a

bo

ut

61 p

erce

nt

of

tota

lp

op

ula

tio

n.

Nat

ion

al P

ove

rty

Red

uct

ion

Str

ateg

y“I

ncr

easi

ng

op

po

rtu

nit

ies

for

the

po

or

is t

he

key

elem

ent

in p

ro-

tect

ing

en

viro

nm

ent

sin

ce n

atu

ral

reso

urc

es a

re e

rod

ed w

hen

th

ep

oo

r d

o n

ot

hav

e al

tern

ativ

em

ean

s. N

atu

ral r

eso

urc

e d

egra

da-

tio

n is

in p

art

du

e to

exp

loit

atio

nb

y th

e ru

ral p

oo

r, w

ho

are

see

kin

gto

sat

isfy

th

eir

bas

ic n

eed

s. R

edu

c-in

g r

ura

l po

vert

y is

th

eref

ore

esse

nti

al t

o a

chie

vin

g s

ust

ain

able

dev

elo

pm

ent

in C

amb

od

ia.”

Five

-Yea

r So

cio

-Eco

no

mic

Dev

elo

p-

men

t Pl

an 2

006–

2010

Key

en

viro

nm

ent-

po

vert

y ta

rget

s:in

crea

se in

fo

rest

co

ver

to 5

3–55

%;

incr

ease

in t

he

pro

visi

on

of

clea

nw

ater

to

ru

ral a

reas

to

ab

ou

t 65

% o

fru

ral p

op

ula

tio

n; r

edu

ctio

n in

mo

r-ta

lity

rate

s o

f ch

ildre

n u

nd

er 5

to

less

th

an 3

0%; e

limin

atio

n o

f sh

ift-

ing

cu

ltiv

atio

n b

y 20

09–2

010;

ass

ign

-m

ent

of

lan

d a

nd

fo

rest

ed a

reas

to

3782

vill

ages

an

d 2

68,2

45 h

ou

se-

ho

lds

bet

wee

n 2

006

and

201

0

Nat

ion

al P

ove

rty

Erad

icat

ion

Pr

og

ram

me

“Po

vert

yer

adic

atio

n a

ctiv

itie

s sh

ou

ldn

ot

deg

rad

e [t

he]

qu

alit

y an

d q

uan

-ti

ty o

f th

e en

viro

nm

enta

l res

ou

rces

that

th

e p

oo

r d

epen

d o

n, a

nd

(ii)

Imp

rovi

ng

en

viro

nm

enta

l co

nd

itio

ns

can

lead

to

su

stai

nab

le p

ove

rty

erad

-ic

atio

n. T

arg

eted

po

vert

y/en

viro

n-

men

t lin

kag

e ar

eas:

def

ore

stat

ion

and

ag

ricu

ltu

ral l

and

; use

of

frag

ilela

nd

; wat

er r

eso

urc

es a

nd

en

viro

n-

men

tal h

ealt

h; b

iod

iver

sity

; nat

ura

ld

isas

ter

vuln

erab

ility

.”

Five

-Yea

r So

cio

-Eco

no

mic

Dev

elo

p-

men

t Pl

an 2

006–

2010

Mai

n p

ove

rty-

envi

ron

men

t ta

rget

s:in

crea

sin

g f

ore

st c

ove

r b

y 43

%;

incr

easi

ng

pro

tect

ed a

reas

(es

p.

wet

lan

ds)

by

11.2

%; i

ncr

easi

ng

acce

ss t

o c

lean

wat

er t

o 9

5% o

fu

rban

an

d 7

5% o

f ru

ral p

op

ula

-ti

on

; im

pro

vin

g r

esid

enti

al a

nd

ind

ust

rial

so

lid w

aste

co

llect

ion

to

80–

90%

; im

pro

vin

g in

du

stri

alw

aste

wat

er t

reat

men

t an

d t

reat

-m

ent

of

haz

ard

ou

s w

aste

.

The

Co

mp

reh

ensi

ve P

ove

rty

Red

uct

ion

an

d G

row

th S

trat

egy

“Har

mo

nio

usl

y co

mb

ine

soci

o-

eco

no

mic

dev

elo

pm

ent

wit

h

envi

ron

men

tal p

rote

ctio

n a

nd

imp

rove

men

t an

d e

nsu

re t

hat

all

peo

ple

are

ab

le t

o li

ve in

a h

ealt

hy

envi

ron

men

t. A

ctiv

ely

inte

gra

teen

viro

nm

ent

issu

es in

pla

ns,

pro

-g

ram

s, a

nd

pro

ject

s an

d m

ake

it a

crit

erio

n f

or

eval

uat

ing

dev

elo

p-

men

t m

easu

res.

Inte

gra

te f

ully

th

een

viro

nm

ent

and

nat

ura

l res

ou

rce

issu

es in

to t

he

mas

ter

pla

n f

or

soci

o-e

con

om

ic d

evel

op

men

t in

pro

vin

ces,

dis

tric

ts, a

nd

en

sure

that

th

e d

evel

op

men

t is

su

stai

n-

able

an

d d

oes

no

t ca

use

deg

rad

a-ti

on

in n

atu

ral r

eso

urc

es.”



Co

un

try

Ass

ista

nce

St

rate

gie

s

Pove

rty

Ass

esm

ents

The

seco

nd

ob

ject

ive

of

the

Cam

bo

-d

ia C

AS’

s fi

rst

pill

ar is

to

imp

rove

nat

ura

l res

ou

rce

man

agem

ent.

Thro

ug

h le

nd

ing

an

d A

AA

, th

eB

ank

exp

ects

to

ach

ieve

th

e fo

l-lo

win

g o

utc

om

es:

a) im

pro

ved

pro

vin

cial

lan

d u

sem

anag

emen

t;b

) ac

cele

rate

d la

nd

dis

trib

uti

on

&in

crea

sed

nu

mb

er o

f se

cure

lan

dti

tles

;c)

tra

nsp

aren

cy in

gen

erat

ion

of

pu

blic

& c

om

mer

cial

ben

efits

fro

m f

ore

sts;

d)

exp

ansi

on

of

com

mu

nit

y fo

rest

ry;

e) im

pro

ved

po

licie

s &

inst

itu

tio

ns

for

pro

tect

ed a

reas

man

agem

ent.

“Th

e Po

vert

y A

sses

smen

t h

igh

ligh

tstr

end

s in

po

vert

y in

th

e p

ast

dec

ade

wit

h s

om

e sp

ecifi

c fo

cus

on

ed

uca

tio

n, h

ealt

h, a

nd

ag

ricu

l-tu

re a

nd

ru

ral d

evel

op

men

t. H

ow

-ev

er, a

sep

arat

e ch

apte

r d

evo

ted

to p

ove

rty-

envi

ron

men

t lin

kag

es—

sim

ilar

to t

he

on

e p

rep

ared

in t

he

Lao

PD

R P

ove

rty

Ass

essm

ent—

has

no

t b

een

incl

ud

ed. T

he

inte

nti

on

of

Ch

apte

r 2

of

this

rep

ort

is t

op

rovi

de

such

info

rmat

ion

”

Un

der

th

e fi

rst

ob

ject

ive

of

the

CA

S,th

e B

ank’

s lo

ng

-ter

m a

nd

hig

her

ord

er o

utc

om

es f

or

Lao

PD

R in

clu

de:

a) im

pro

vin

g g

ove

rnan

ce a

nd

en

viro

n-

men

tal m

anag

emen

t o

f n

atu

ral

asse

ts;

b)

stre

ng

then

ing

th

e co

llect

ion

an

du

se o

f re

ven

ues

fro

m n

atu

ral

reso

urc

es t

o h

elp

red

uce

po

vert

y;c)

mai

nta

inin

g p

rod

uct

ive

fore

st c

ove

ras

par

t o

f ru

ral l

ivel

iho

od

s;d

) re

du

cin

g d

efo

rest

atio

n r

ates

;e)

imp

rovi

ng

ru

ral l

ivel

iho

od

s.

The

late

st p

ove

rty

asse

smen

tfo

r La

oPD

R h

igh

ligh

ts t

he

stro

ng

po

vert

yre

du

ctio

n in

th

e co

un

try

in t

he

pas

td

ecad

e, b

ut

also

rec

og

niz

es n

um

er-

ou

s p

ove

rty-

rela

ted

are

as, w

her

em

ore

wo

rk is

nee

ded

. It

hig

hlig

hts

the

stro

ng

geo

gra

ph

ic f

eatu

res

of

po

vert

y in

th

e co

un

try,

th

e d

isp

ro-

po

rtio

nat

ely

hig

her

po

vert

y am

on

get

hn

ic m

ino

riti

es a

s w

ell a

s th

e sl

ow

pro

gre

ss in

imp

rovi

ng

man

y so

cial

ind

icat

ors

. Th

e re

po

rt s

tate

s th

atp

ove

rty

is li

nke

d t

o e

nvi

ron

men

tal

con

dit

ion

s an

d d

edic

ates

a c

hap

ter

toth

e en

viro

nm

ent,

wh

ich

is m

ost

lyb

ased

on

fin

din

gs

fro

m t

his

PEN

stu

dy.

Th

e N

RM

sec

tio

n o

f th

e ch

ap-

ter

cove

rs f

ore

st r

eso

urc

es, l

and

reso

urc

es, U

XO

s an

d n

atu

ral d

isas

-te

rs. T

he

envi

ron

men

tal h

ealt

h s

ec-

tio

n f

ocu

ses

on

acc

ess

to c

lean

wat

eran

d s

anit

atio

n a

s w

ell a

s in

do

or

air

po

lluti

on

.

The

dra

ft v

ersi

on

of

the

late

st C

AS

for

Vie

tnam

ad

dre

sses

bo

th n

at-

ura

l res

ou

rce

man

agem

ent

and

envi

ron

men

tal h

ealt

h is

sues

. Pri

-m

ary

focu

s ar

eas

in N

RM

incl

ud

efo

rest

ry la

nd

, co

mm

un

ity-

bas

edla

nd

man

agem

ent,

an

d t

he

con

-ve

rsio

n o

f ag

ricu

ltu

ral l

and

; allo

ca-

tio

n o

f la

nd

tit

les,

par

ticu

larl

y in

up

lan

ds;

bio

div

ersi

ty c

on

serv

atio

n.

Wo

rld

Ban

k p

roje

cts

in e

nvi

ron

-m

enta

l hea

lth

in u

rban

are

as w

illfo

cus

on

was

te w

ater

an

d w

aste

man

agem

ent

(PC

B a

nd

haz

ard

ou

sw

aste

man

agem

ent,

oth

er in

du

s-tr

ial p

ollu

tio

n) a

nd

car

bo

n fi

nan

ce.

The

Vie

tnam

Dev

elo

pm

ent

Rep

ort

2004

, wh

ich

fo

cuse

d o

n p

ove

rty,

do

es n

ot

hig

hlig

ht

envi

ron

men

tal

issu

es a

s co

ntr

ibu

tors

to

po

vert

y,b

ut

do

es d

iscu

ss e

nvi

ron

men

tal

hea

lth

issu

es, s

uch

as

acce

ss t

ocl

ean

wat

er a

nd

san

itat

ion

, as

imp

ort

ant

asp

ects

of

po

vert

y an

din

equ

alit

y in

th

e co

un

try.

Th

en

ewes

t u

pd

ate

on

th

e Po

vert

yu

pd

ate

in V

ietn

am (f

rom

200

6)sh

ow

s fu

rth

er r

edu

ctio

ns

inp

ove

rty,

eve

n if

th

ey a

re s

till

rela

tive

ly u

neq

ual

ly d

istr

ibu

ted

bet

wee

n t

he

dif

fere

nt

reg

ion

s.Th

e st

atis

tics

sh

ow

a r

edu

ctio

n in

the

urb

an-r

ura

l eth

nic

div

ide,

bu

tth

e et

hn

ic m

ino

rity

po

pu

lati

on

lag

s fa

rth

er a

nd

far

ther

beh

ind

inp

ove

rty

red

uct

ion

. Eth

nic

min

ori

-ti

es c

om

pri

se 3

9% o

f th

e to

tal

nu

mb

er o

f p

oo

r.

Cambodia’s Ministry of Environment and theMinistry of Planning in Lao are expected to playa key role in the mainstreaming the PEN findingsin their respective National Poverty ReductionStrategy (NPRS) processes. In Lao PDR, theScience Technology and Environment Agency(STEA) and the Committee for Planning andCooperation (CPC) will be the key partnersin integrating the PEN results in the NationalPoverty Eradication Program (NPEP). Therecently set up Ministry of Natural Resourcesand Environment (MoNRE) and the Ministryof Planning and Investment in Vietnam willprovide support in incorporating the PEN find-ings in the CPRGs.

To sustain the dialogue and cooperationthroughout the process of completion of thestudy and to share findings and knowledge fromthe three countries’ PEN work, the World Bankand its partners organized a series of sub-regionalworkshops (see table I.3).

In completing the study, the World Bank staffalso worked closely with international organiza-tions such as the World Food Program (WFP),the Mekong River Commission (MRC), theWorld Health Organization (WHO), the UnitedNations Population Fund (UNPF), the UnitedNations Children’s Fund (UNICEF), and theUnited Nations Development Program UNDP).

IV. Structure of the Report

The paper is organized as follows. Part I consistsof one chapter that provides a detailed descrip-tion of the methodology and data sourcesapplied to the PEN studies in each country. Part

II consists of the three country chapters, whichpresent the findings from the studies in eachcountry. The sub-regional findings and conclu-sions are presented in Part III of the paper.

The separate studies reflecting findings fromeach of the study areas (map I.6), national PENanalyses and additional supporting material isprovided in the CD-ROM attachment.

Endnote/Reference:To follow:1. The study itself included 5 environmental indicators as

access to water, sanitation and health impacts wereclustered into one indicator.

2. Examples are the jointly DFID, EC, UNDP and WorldBank-prepared “Linking Poverty Reduction and Environ-mental Management—Policy challenges and Opportuni-ties—A contribution to the World Summit on SustainableDevelopment Process,” the World Development Report2003 “Sustainable Development in a Dynamic WorldTransforming Institutions, Growth, and Quality of Life,”and Poverty—Environment material issued in theWorld Bank Environmental Economics Series.

3. There are 4 large urban centers in the sub-region: Hanoi,Ho Chi Minh City, Phnom Penh and Vientiane.

4. Cambodia Poverty Assessment Draft Report.5. Cambodia: Poverty Assessment, presentation of pre-

liminary findings, October 13, 2005.6. The World Bank’s Cambodia Environmental Monitor

for 2003.7. NSC-Lao PDR 1999 (Social Atlas of the Lower

Mekong Basin).8. This fact appears also to be the case for Cambodia—

refer to draft Poverty Assessment.9. The PEN study contributed to the following docu-

ments: Lao PDR PRSP and Poverty Assessment andVietnam’s Five-Year Plan and Poverty Assessment. TheVietnam PRSP was prepared prior to the PEN study’scompletion.

10. Refer to comprehensive reports for local surveys in theCD-ROM attached to the report.




INTRODUCTION

There is broad agreement in the developmentliterature that there is a close link betweenpoverty—and human welfare more generally—and environmental quality. The existence ofsuch a link is intuitively credible, given thatdegradation of natural systems adversely affectsthe livelihoods of people who depend on theseresources for agriculture, forestry, and fisheries.Similarly, environmental degradation in the formof air, land, and water pollution is often higher inmarginal settlements in undesirable areas. Thesetend to be areas with a high concentration ofpoor residents. Contaminated air and water leadsto health problems, which in turn reduce theability to pursue income-earning or livelihood-enhancing activities.

The strength and direction of the relationshipsbetween poverty and environmental conditionscontinue to be ill-defined. It has been difficult toconfirm the existence and nature of these links inempirical studies. In a review of forestry andpoverty, Angelsen and Wunder (2003) call suchrelationships a “controversial link.” At a macrolevel, some researchers believe that in early stagesof development some environmental degradationis unavoidable—a necessary trade-off that will bereversed with increasing wealth. The implicationis that countries should accept increased pollutionand environmental degradation as the price to pay

for faster growth and poverty reduction. Greaterwealth will then enable a society to restore envi-ronmental quality later. Empirical evidence tosupport the existence of these dynamics and thusthe benefits of such strategies has been mixed(Ansuategi and others 1998; Dasgupta and others2002; Perman and Stern 2003).

In fact, much of the work on the poverty-environment nexus shows that the dual develop-ment objectives of maintaining environmentalquality and increasing living standards do nothave to be contradictory, even at earlier stages ofdevelopment. The full costs of environmentaldegradation include loss of productivity andincreased demands on social systems due to theenvironmental health burden. A potential long-term stream of future benefits from sustainablymanaged natural resources remains unrealizedwhen such resources are irreparably damaged.

Typically, poor people, especially in rural areas,depend directly on the natural resource base andtherefore carry the largest burden from pollutionand degradation of natural resources (Ekbomand Bojo 1999). Deterioration or destruction ofnatural resources tends to be triggered by demandor profit-seeking from wealthier people throughlarge-scale logging, depletion of water resources,soil degradation, land degradation due to over-grazing, or pollution from agro-chemicals. Thepoor often fail to obtain a fair share of the bene-

Study Framework and Methodology1

�

fits of rapid industrialization and natural resourceexploitation. Where the poor contribute to landdegradation, it is often due to population pres-sure that forces overexploitation of a shrinkingresource base, frequently in already marginalareas. In many areas, rising populations are due tovoluntary or prescribed in-migration rather thannatural growth of population. Such situationsoften coincide with a breakdown of traditionalresource management systems that helped main-tain an environmental balance. While develop-ment in the form of higher living standards inthe long run can help reduce direct pressure onnatural systems, the short-run objective is to findnew regimes that regulate access to resources forsustainable use or effective protection. Such sys-tems need to be developed by or evolve in closecollaboration with local communities. It is thequality of this response that determines whetherpopulation growth and economic dynamics leadto further degradation or whether the outcomecombines improved living standards with betternatural resource management.

In the context of the World Bank povertyreduction and environmental strategies, a mainobjective is therefore to ensure that poverty allevi-ation activities should not damage the environ-ment on which poor people depend, as this wouldessentially trade off gains in one area for losses inanother (World Bank 2005a). More pro-actively,poverty reduction strategies should also aim atimproving environmental conditions that can helpreduce poverty (Bojö and others 2001; 2004). Inthis context, the Southeast Asia PEN study aims tocontribute to the debate by providing empiricalevidence on relevant linkages in a variety of envi-ronmental sub-sectors and geographic settings.The overall goal is to help identify strategies andinterventions that improve people’s lives andmaintain environmental quality.

This chapter is composed of three main sec-tions. First, it provides a brief review of povertyconcepts and main environmental issues in South-east Asia. The second section describes conceptualand methodological issues pertinent to poverty-

environment analysis—especially as these relateto geographic aspects. Some of the commonelements in the PEN methodology used in thenational and regional analyses are highlighted inthe final section. The country chapters containadditional information on data collection andimplementation issues that are specific to eachof the case studies.

Poverty and Environment Concepts

Poverty

Poverty is a multi-dimensional concept and thereis continuing debate about how poverty shouldbe defined and measured. A comprehensiveoverview of poverty concepts is provided in theWorld Bank’s World Development Report onAttacking Poverty (2001a). This report stressesthat poverty is not just a matter of material depri-vation, but also relates to broader notions of risk,vulnerability, social inclusion and opportunities.Poverty encompasses all forms of deprivationthat prevent a person from achieving their aspi-rations. This wider range of poverty dimensionsalso has policy implications as it acknowledgesthe mutually beneficial interactions from inter-ventions in a broad set of welfare dimensions. Inpractice, it is the nature of the problem that isstudied that will determine which welfare mea-sure is most appropriate. This section will brieflyreview a number of poverty concepts that are rel-evant for applied policy research on poverty-envi-ronment linkages.

The concept of poverty that most closelymatches most people’s notions is absolutepoverty—the ability of people to satisfy theirbasic needs in terms of nutritional intake, shelter,basic amenities, and the ability to take advantageof opportunities such as education. Living stan-dards are determined as the consumption expen-diture equivalent of market and non-marketgoods that are required to satisfy these basic needs.This level is called the poverty line. Poor house-holds are those whose consumption falls underthe level that is considered sufficient to maintain

S T U D Y F R A M E W O R K A N D M E T H O D O L O G Y


basic needs. The degree to which the household isfalling short is called the poverty gap, the sum ofwhich for the entire population is an estimate ofthe monetary resources required to lift everyoneout of poverty. A further commonly used mea-sure, the squared poverty gap or poverty severity,gives extra weight to people who are far below thepoverty line. This measure therefore better incor-porates inequality among the poor and withinthe population as a whole. Consumption ex-pressed in monetary terms provides perhaps themost objective way of comparing living standards,and methods for data collection and analysis arewell developed (Ravallion 1994; Deaton 1997).Most comprehensive poverty assessmentsemploy this approach, often in combination withother methods. Alternative indicators can providemore direct information on welfare outcomes.For instance, nutritional status as indicated bystunting or wasting is used in Demographicand Health Surveys and in food security assess-ments. Similarly, one could define educationpoverty on the basis of literacy rates or testscore achievements.

In contrast to absolute measures of poverty,relative poverty focuses on inequality across a pop-ulation in the distribution of consumption, assets,education, or any other welfare measure. Thisis relevant for targeting interventions to the pop-ulation groups at the bottom of the welfare distri-bution. But it is also relevant in its own right,because severe inequality is considered detrimen-tal to the overall development prospects of a coun-try (World Bank 2005).

Where detailed poverty data are unavailable,policy researchers sometimes use a simple asset-based indicator of human welfare. This may bean index of ownership or availability of certaingoods or services such as a vehicle, good qualityhousing, agricultural land holdings, or connec-tion to utilities (Filmer and Pritchett 2001). Inthe absence of detailed consumption data, theseasset-based indicators can reflect the nature ofthe problems faced by the communities, espe-cially in rural areas. An asset index can also be

more easily designed to match a particular issueunder study. The problem with service access,asset, or wealth indexes is that there is no universal methodology for creating them. Forinstance, individual components of the indexcould be weighted by perceived importance orsimply added up. This can create a high degreeof subjectivity and limits interpretation andcomparability.

Broader concepts of poverty have been pro-posed to more closely address the issue ofopportunity. Rather than looking at income orconsumption itself, the sustainable livelihoodsapproach, for instance, focuses on the types ofresources that determine a household’s standardof living. These are summarized by five types ofcapital: natural, social, human, physical, andfinancial.1 In this framework, the environment isseen as the sum of the resource stock accessibleto households; that is, its natural capital. Otherforms of capital come into play by substitutingor modifying the relationship with the environ-ment. These livelihood concepts have the advan-tage that they explicitly consider access to assetssuch as resource stocks in determining povertylevels. But they are often more difficult to oper-ationalize than standard poverty measures, sincethe components do not always correspond toeasily measurable and comparable indicators. Assuch, they are most useful as a conceptual modelthat can guide more informal and qualitativepoverty analyses, like those in several of the PENcase studies.

A further aspect of well-being is vulnerability,which captures the risk of being poor or fallinginto poverty at some point in the future. Vul-nerability strongly affects choices and options,for instance with respect to food production orinvestment in productive activities. For example,farmers close to the subsistence minimum arerarely able to risk adoption of unproven butpotentially superior new varieties or techniques.Coping capacity in the presence of fluctuatingincomes also affects the subjective perception ofa person’s well-being. The ability to deal with



sudden shocks and the existence of safety netsprovide a feeling of security, which influencespeople’s actions. These concepts are highly rele-vant to the analysis of human welfare, but inpractice they are difficult to measure systemati-cally. Likewise, vulnerability is also relevant whenapplied to issues such as health. Poorer personswith less access to healthcare will be more vul-nerable to environmental and other health prob-lems than people who have access to safety netsor other coping mechanisms.

The various PEN studies described in the fol-lowing chapters employ poverty concepts thatare most appropriate in the specific context. Theregional analysis in Phase I largely followed thepoverty definitions used in national poverty assess-ments. The core indicator is typically the propor-tion of individuals or households that fall short ofconsumption expenditures that support a foodintake of more the 2,000 calories per person, plusminimum non-food expenditures. This indicatoris called the poverty headcount index and thethreshold level, the poverty line, is defined some-what differently in each country.2 The nationalanalyses in the second phase of the PEN studyused the same concepts where data were available,but augmented the set of indicators with specificoutcome measures such as health status. A keycomponent of the spatial analysis of PEN issues atthe regional and national level were the availablepoverty maps that show welfare indicators at a finelevel of geographic disaggregation.

A typical living standard survey collects com-prehensive information about a relatively small,but nationally representative sample of residents—usually several thousand households. This allowsestimation of statistically reliable poverty rates andother indicators for a small number of regions or,at best, at the province level. World Bankresearchers and others have recently developed amethodology to combine the rich detail of wel-fare surveys with information from populationand household censuses (Hentschel and others2000). A census collects a limited set of indica-

tors, but covers all households in the country. Bycombining these two data sources using statisticaltechniques, reliable estimates of poverty andinequality can be derived for much smalleradministrative units than the province—such asdistricts or even communes. Such “povertymaps” have become an invaluable tool for visu-alization and analysis of poverty patterns and fortargeting of interventions.

The first phase of the PEN study made use ofpoverty maps that were created for Vietnam(Inter-Ministerial Poverty Mapping Task Force2003), Cambodia (MoP and WFP 2002) andLao PDR (WFP 2001). The more recent analy-ses took advantage of revised poverty maps andnew data. These are based on large sample nationalhousehold surveys and the latest available censusinformation.3 Both surveys and censuses are com-plex and expensive undertakings. National statis-tical and census offices therefore carry out suchenumerations only every few years (10 years inthe case of censuses). Some of the data availablefor the three countries are therefore seven toeight years old. While the broad patterns areunlikely to change, much of the region is expe-riencing rapid economic growth that triggersdynamic migration patterns and poverty reduc-tion in some regions. Wherever possible, thePEN study therefore augmented available sta-tistical and census data sets with other infor-mation from smaller surveys, district records,or field data collection.

Table 1 summarizes different sources of socio-economic data that may be suitable for poverty-environment studies. Each has advantages anddisadvantages, but all are useful for policy analy-sis, although researchers and decision makersneed to be aware of the limitations in using anyof these data sources.

Environment

The three countries included in this study face thechallenge of pursuing economic growth to gener-ate adequate living standards for a rapidly grow-



ing population. This process requires a largeexpansion of public services—such as water sup-ply and health care—and puts increasing strainson natural resources. The PEN study distinguishesbroadly between environmental conditions thatdirectly affect human health and quality of lifeversus those related to the maintenance of naturalresources that support fundamental biological sys-tems and rural livelihoods.

Environmental Health and Quality of Life

Environmental health, according to the WorldHealth Organization,

“comprises those aspects of human health,including quality of life, that are determined byphysical, chemical, biological, social, and psy-chosocial factors in the environment. It also



T A B L E 1 . 1 Socioeconomic Data Sources

Data source Pros Cons

Population andhousehold census

Comprehensive,multi-topichousehold survey(LECS, CSES, VNLSS, DHS)

Short, sector-specificsurveys (e.g., villagesurvey)

Facility and service provider surveys

Participatory data collection and rapid appraisal (focus group interviews)

Transactions data (e.g., birth/death register, cadastre, etc.)

Can provide reliable data for verysmall areas; sometimes includesinformation on household assetsthat reflects welfare status.

Often the main source of reliabledata on poverty and other aspectsof household welfare; usually col-lects a very comprehensive set ofindividual, household, and com-munity-level indicators.

Fairly low cost and quick turn-around; can be used for monitor-ing through repeated surveys(e.g., annual).

Provides information on the supply-side of service delivery that canotherwise be difficult to monitor(e.g., health, education); can becombined with user surveys atfacility.

Involves citizens; low cost; providesrich contextual information.

Information base is always up-to-date; serves many administrativefunctions.

Typically only available every 10 years;small number of basic indicators;often, only aggregate data arepublished (province or districtlevel); often a long lag betweencensus-taking and dissemination.

Sample sizes are too small to yieldgeographically detailed indicators.

Often not designed to meet statisticalrequirements (sampling, etc.); smallsample sizes; focused on limitednumber of issues, making morecomprehensive analysis difficult;do not typically cover large areasor entire country.

Frequently include quantitativeindicators only (e.g., number ofdoctors), which do not necessarilyreflect quality of service; user surveys at facilities leave out peo-ple who do not have access forgeographic, social, or economicreasons.

Difficult to maintain data qualityand objectivity; not possible toextrapolate to other parts of thecountry or to country as a whole.

Limited number of indicators; com-plex and expensive to maintainelectronically.

refers to the theory and practice of assessing, cor-recting, controlling, and preventing those factorsin the environment that can potentially affectadversely the health of present and future gener-ations.” (World Health Organization; draftdefinition developed at a WHO consultationin Sofia, Bulgaria, 1993).

In the context of the PEN study, the mostimportant aspects of environmental health arethe effects of inadequate water supply and sani-tation in rural and urban areas; air and water pol-lution from industrial activities in cities, towns,and villages; indoor air pollution, especially inthe upland areas of Lao PDR and Cambodia;and pesticide use in agriculture.

Water-related problems enter the poverty-environment nexus through basic public healthissues. Lack of access to clean water among thepoor is one of the primary health problems,largely through coliform bacteria in drinkingwater that cause diarrhea. Lack of proper sanita-tion affects hygiene conditions, increases the risk

of disease, and contaminates drinking water; ifuntreated sewage is released into streams, it canalso affect natural systems. The impact on infantsand small children is particularly severe, but ill-ness among adults also contributes to a largesocial and economic burden due to infectiousdiseases from inadequate water supply, sanita-tion, and hygiene. Although many countries havemade progress in extending safe drinking watersupplies to an increasing share of their popula-tion, coverage in many rural areas is still low. Inpart, this is because the population is widely scat-tered and the cost of supplying infrastructure ishigh, although other issues such as a lack of voiceamong rural populations are equally importantfactors to consider. In towns and cities, problemsare often localized in low-income neighborhoods,where residents may not be able to afford con-nection fees.

Water supply is also threatened by direct indus-trial and municipal pollution of freshwaterresources. Facing limited regulations and enforce-ment, many factories release toxic waste directlyinto streams and rivers, a problem particularlyacute in the densely populated booming industrialcenters in Vietnam. Such pollution originatesfrom large, formal industrial plants, but also fromsmaller cottage industries, many of which arelocated in smaller towns or villages. Effluent pol-lution problems are not limited to manufacturing.In the agricultural sector, farmers have greatlyincreased production, in part with the help ofagro-chemicals such as fertilizers and pesticides.The use of agricultural inputs is not strongly con-trolled; inappropriate application of chemicals,some of which are banned elsewhere but importedillegally, pollute land and water, and pose a signif-icant health risk to farm workers who apply them.

Air pollution is largely a problem in urbanareas where industrial firms and power plantsoften operate with higher pollution intensities(pollution per unit of output) than modernplants that employ state-of-the-art technology.Accelerating motorization and the popularity ofmotorcycles with two-stroke engines also leads



0

10

20

30

40

50

60

70

80

90

100

110

120

1950

1960

1970

1980

1990

2000

2010

2020

2030

2040

2050

Po

pu

lati

on

(m

illio

n)

Vietnam

Cambodia

Lao PDR

F I G U R E 1 . 1 Historical and Projected PopulationGrowth in Lao PDR, Cambodia and Vietnam

Source: UN World Population Prospects 2004, medium variant, UN Popu-lation Division, New York.

to high concentrations of particulates that arethe main cause of pollution-related respiratoryproblems. Whether air pollution affects the poormore severely than the non-poor depends on thetype and location of pollution sources. Whilevehicle emissions affect everyone in a city equally,industrial plants tend to be located in undesir-able parts of a city that also have low-incomeneighborhoods. Low-skilled industrial workersalso tend to be located in more urban areas andmay face a high occupational health risk due topollution. Other types of pollution such asuntreated effluents are also a bigger problem inpoor neighborhoods.

In contrast to outdoor air pollution, indoorair pollution is a significant health problem inrural areas, where traditional cooking methodswith wood or charcoal in poorly ventilated housesgenerate high particulate concentrations. Espe-cially in the cooler upland regions of Vietnam andLao PDR, indoor air pollution is thought to haveconsiderable effects on the health of women andchildren and may be most severe in poor house-holds. Much of the evidence in this area is fromindirect health studies of the incidence of respi-ratory disease, since systematic pollution moni-toring programs and large-scale epidemiologicalstudies have been scarce.

Natural Resource Management

Natural resource management issues have greatrelevance in the subregion. All three countriescontain forest areas with high economic and eco-logical value. About 40 percent of the land area ofLao PDR, for instance, is covered by forests. Esti-mates of the forest sector’s contribution to GDPvary between 4.5 percent—if only official com-mercial benefits are included—to 7 to 10 percentif the contribution of forest resources to sub-sistence households and illegal logging opera-tions are added (World Bank 2001). Commercialtimber extraction is capital- and skill-intensive,requires large economies of scale, and relies onspecialized marketing channels.4 The poor rarely

benefit from logging beyond occasional employ-ment opportunities, local economic stimulus fromtimber operations, and possible pro-poor govern-ment programs funded through logging proceeds.The habitat destruction from large-scale logging,however, can damage the natural resource baseon which poor farmers, hunters, and gatherersdepend.

Logging operations are often poorly regulated,and illegal logging of high-value timber is a severeproblem in the region. Much of the demandcomes from neighboring countries, especiallyafter the logging ban in China in 1998 led tradersthere to look elsewhere in the region for timbersupplies. Overall, deforestation processes in theregion reflect the problems globally. The driversinclude a range of mutually reinforcing factorssuch as economic and institutional conditions,national policies and enforcements, and proximateand remote influences that encourage agriculturalexpansion, wood extraction, and infrastructureextension (Geist and Lambin 2002).

Remaining forest areas in Southeast Asia con-tain an extraordinary degree of biodiversity, par-ticularly in the remaining montane forest regions.One way for the poor to benefit from this speciesrichness is through non-timber forest products(NTFPs). NTFPs can be important for the pooras a livelihood strategy and safety net. A metaanalysis of empirical studies worldwide showedthat forest environmental income provides anaverage of 22 percent of household income, abouthalf of which is cash income (Vedeld and others2004). But few such products have the potentialfor sustainable scaling up of collection and com-mercialization, and therefore have limited poten-tial for supporting socioeconomic advancement.Consequently, NTFPs are often a symptom ofpoverty rather than a solution. Yet, especially inLao PDR, NTFP dependence by the rural poormay be higher than in many other countries.

Some forest products—such as sticlac, car-damom, bamboo, benzoin, rattan, and pineresin—are legally harvested NTFPs with com-mercialization potential (WFP 2004). A more



sinister form of NTFP utilization is the illegaltrade in threatened plant and animal species.Trade in wildlife that originates in or passesthrough Vietnam, for instance, is well docu-mented (Song 2003; World Bank 2005c). Mostof the demand is from urban areas in the regionand from China. The sources are mostly the pro-tected areas of Vietnam, Lao PDR, Cambodia,and Myanmar, where the last remaining popula-tions of many threatened species live. With risingliving standards in the region, demand for wildanimal and plant products will increase further.This is likely to lead to the extinction of manyunique species. This kind of trade is difficult tostop given the scarce resources of wildlife agenciesand the vast profit margins available to traders.

In the upland areas of Lao PDR and Vietnamand the lower lying regions of Cambodia, forestsprovide vital ecological services to rural commu-nities as well as to downstream towns and cities.Ecological services such as regulation of hydro-logical functions, carbon storage, and preserva-tion of genetic diversity are forest functions forwhich compensation or benefit-sharing schemescould be designed. Direct payment schemes forsuch services have been tested in various parts ofthe world, but it is no trivial matter to implementecological service payment systems that benefitthe poor (Pagiola and others 2005; Landell-Millsand Porras 2002). Still, successful experienceswith payments for environmental service schemesespecially in Latin America may hold promise forsimilar projects in Southeast Asia. Indirect bene-fits from ecological endowments such as tourismare increasingly important in the economies of thethree countries. Future expansion of those rev-enues is threatened by large-scale deteriorationof the most valuable regional ecosystems.

Perhaps the most important role that forestsplay in providing livelihoods for the poor is as aland reserve that can be converted to agriculturaluses on a temporary (rotating or shifting cultiva-tion) or permanent basis (Angelsen and Wunder2003). For the poor, this may often be the mosteconomically beneficial use of previously forested

land, but can also contribute to large-scale removalof forest cover and negative impacts on otherbeneficiaries of forest resources and on environ-mental functions. Due to its high labor require-ments and low productivity, shifting cultivationis often done by poor farmers. It is a sustainableuse of forest areas when sufficient fallow periodsfollow land conversion (Fox and others 2000).But various factors may lead to reduction of fal-low periods to unsustainable levels. Populationgrowth—often through in-migration rather thannatural growth—increases the number of farm-ers dependent on a given land area, and large-scalelogging or setting aside of forest areas for otheruses further contribute to a shrinking resourcebase available to the poor for forest-based agricul-ture. With shorter rotation cycles, fragile soils—especially in steep upland areas—lose their abilityto support traditional agricultural practices.Where fallow periods are too short for systems torecover, sustainable shifting cultivation turnsinto unsustainable slash-and-burn agriculture thatmay lead to complete depletion of soil fertility(Casetella and Quang 2002; Palm and others2005). This process can be reversed, but requiresconsiderable investment in restoration of plantnutrients, improved land management, or plant-ing of trees. Such long-term investments willmake sense to farmers only if they have sufficientlysecure tenure to their land.

Intensified permanent agriculture—usingmodern farming techniques adapted to the ter-rain and climate—can be successful in providinga livelihood base to farmers. But in areas of rapidpopulation growth, rural strategies based onintroduction of modern technologies alone areunlikely to be successful in absorbing surplus laborwithout simultaneous expansion of cropping areasin previously forested regions. However, scalingup of technology transfer and extension services,combined with provision of education and othersocial services, infrastructure and effective regu-lation, can have important spillovers for job cre-ation in the rural non-farm sector, where mostof the future rural employment growth needs to



occur (Müller and Zeller 2002; Cuong andothers 2003).

GEOGRAPHIC SCALE AND SCOPE OF THE PEN ANALYSES

Natural resources are distributed unequallywithin a country. Similarly, industrial activity is

usually concentrated in urban centers and withinparts of larger cities, so the burden of air, land,and water pollution will also vary by geographiclocation. Countries also often show large varia-tions in living standards, even when adjusting forcost-of-living differences. Apart from this geo-graphic heterogeneity, the three countries also dis-play great complexity in the nature and severity of



Poverty-environment linkages are varied and complex. Consequently, the range of environmentalindicators that need to be evaluated to study various aspects of the nexus is also very broad. Somerelevant information can be extracted from socioeconomic data sources (reviewed earlier in thischapter). Examples are health outcomes that may have environmental causes or household behav-ior with respect to the use of natural resources. For many other aspects, however, informationneeds to be collected using specialized tools.

Remote sensing—capturing images of features on the ground from space-based systems orairplanes—is one of the most useful tools for collecting natural resources data for large areas atregular intervals. Optical remote sensing systems such as Landsat or Spot have been invaluable innational or regional assessments of land cover, including forests, wetlands, and agricultural areas(see Figure XX). Some of these systems have been in operation for several decades, so archivedimages can provide historic views for change analysis. Remotely sensed data is usually the methodof choice for forest cover inventories and monitoring because they provide data collected using aconsistent methodology for large areas. This circumvents the problem of subjective evaluation incompiling ground-reported information from local observers. Other remote sensing platformsgenerate more specialized information, such as topography (elevation and slope) from radar- orlaser-based systems, climatic and atmospheric data, and even particulate pollution concentration.Lately, commercially available high resolution (60cm–1m) satellite data have become available.These products can be interpreted visually, reducing the need for complex image analysis proce-dures that are required for the coarser resolution data (15m–1km). However, the cost of coveringlarge areas and the time required for visual interpretation mean that high-resolution images haveso far not been used extensively in natural resources applications.

Not all environmental information can be captured remotely. Obtaining high-quality informa-tion on water and air pollution, flood levels, rainfall, or temperature requires a network of ground-based monitoring stations. Those recording stations can be permanent or mobile. Best known areweather stations that report continuously and yield long-term climatic averages. Unfortunately,for many environmental indicators permanent monitoring networks are very sparse and recordingperiods are often interrupted. An example is the scarcity of river gauge data that severely hampersthe analysis of flooding that might be linked to land use changes upstream. With prices falling formany kinds of monitoring devices, targeted data collection campaigns are increasingly feasible. Forinstance, Dasgupta and others (2004) deployed air pollution monitoring devices in a large numberof households for 24-hour periods to study indoor air pollution exposure in Bangladesh. Such dataalso add context and substantive evidence to the analysis of household survey data on health out-comes, especially if it includes medical information such as lung capacity measurement. The benefitsfrom combining different, multidisciplinary data sets in analyzing human-environment interactionsare clearly very large.

Collection of environmental data is most useful if it is part of a long-term information manage-ment strategy and if the outputs are widely disseminated to policy makers, researchers, and the pub-lic. Since ecological processes do not follow political boundaries, comprehensive regional approachesshould play an essential role in improving natural resource management. Examples are the Brazilianinitiatives on rainforest monitoring, which are based on remote sensing and include public disclosureof real-time information on forest cover changes. An institutional template for such efforts in South-east Asia is the Mekong River Commission, but any effort focused on terrestrial ecosystems shouldalso include the parts of mainland Southeast Asia that are not part of the Mekong watershed.

B O X 1 . 1 Sources of Environmental Information

environmental and social problems. While thereare many similarities, such as the natural resourcemanagement issues affecting the poor in theuplands of Vietnam and Lao PDR, each countryalso has its specific set of problems. The boomingcities of Vietnam face pollution problems thathave not reached the much smaller urban centersof Lao PDR. Indoor air pollution is a problem inthe uplands of these two countries, but less so inCambodia where lower elevations and differentcultural preferences mean that most cooking isdone outside. The PEN study accounts for thesecomplexities by considering a very broad (thoughby no means exhaustive) range of poverty-envi-ronment issues. It also accounts for geographicheterogeneity explicitly by analyzing problems atmultiple scales and by employing spatially refer-enced data whenever possible.

Multi-Scale Analysis—from Regional to Local

The study applies a multi-scale approach com-bining analysis of aggregate data at nationaland regional levels with very detailed case stud-ies that reach down to the village and individ-ual household level. The first phase of the projecttook a regional perspective by analyzing similarvariables for the three countries, typically at thelevel of provinces (Dasgupta and others 2005).What emerged was that for many environmen-tal and welfare-related questions, the provincelevel provides insufficient detail to yield use-ful answers. Within-province variation is oftengreater than that between provinces. In the sec-ond phase, national-level analysis, data werecollected and analyzed for all parts of the coun-



Source: NASA Landsat Geocover 2000. Vegetation in green, bare soil/agriculture in red/purple.

F I G U R E 1 . 2 Land Cover in Northern Lao PDR

try, and therefore relied as much as possible ondistrict-level information. Finally, core contri-butions of the PEN study are the topic- andplace-specific case studies that support thelocal analysis. Here, information is typicallycollected at the subdistrict, village, and house-hold level.

Any specific poverty-environment issue shouldbe analyzed at the scale that is most relevant tothe characteristics of the topic. For instance, airpollution from cement mills may only be dan-gerous in the urban area where the facility islocated, while acid rain may affect natural sys-tems far away from the pollution source. In theCau River Basin case study in Vietnam, pollu-tion from riverside industrial facilities has animpact downstream, so a focus only on the imme-diate pollution source would miss the widercontext of the nexus. Other applications—forinstance, using province-level poverty data—provide many insights into the spatial distribu-tion of welfare levels when looking at thecountry as a whole. Investigation of specificregional linkages between human welfare andenvironmental processes, however, requires“zooming in” to the district level or below inorder to capture variations in patterns. Forinstance, most behavioral aspects of poverty-environment interactions need to be assessed atthe household level whenever possible. Anexample is hygiene behavior, which often variesbetween economic or ethnic groups and hasstrong relations to health outcomes in areas ofpoor access to water and sanitation. However,when looking at water supply issues, the mostsuitable level of analysis may be the commu-nity or village as the basis for provision ofnetworked services, or even the district whereinfrastructure investment targeting decisionsare made.

In practice, a major barrier to multi-scaleanalysis is the availability of data of sufficientquality. Apart from census information, mostsocioeconomic indicators are derived from sam-ple surveys, which are usually not reliable at the

district or subdistrict level. Similarly, forestryinformation is sometimes collected using satel-lite remote sensing with a resolution of 500mor 1km. This is too coarse to detect deforestationdue to shifting cultivation or selective logging.Even where reliable data by administrative unitsare available, the analyst often faces challengesin interpreting the information (see box 1).

Use of Spatial Data to Account forGeographic Heterogeneity

In addition to looking at PEN issues across multi-ple scales, the PEN study relied as much as possi-ble on geographically referenced information. Itconsists of statistical information and derived indi-cators that can be linked, for instance, to a digitalmap of administrative units, as well as inherentlyspatial data such as digital forest cover, slope, orelevation maps. These maps are used to display thegeographic distribution of the data, combine oroverlay different maps to generate “spatial cross-tabulations,” and to extract new variables that areused in descriptive and statistical analysis.

For instance, among the most dominant geo-graphic patterns in the subregion are large dif-ferences between the lowland and upland areas,especially in Lao PDR and Vietnam. More than60 percent of the Lao PDR land area is above500 meters, and more than 20 percent above1,000 meters. For Vietnam, the corresponding fig-ures are 30 percent and 10 percent. Lowland areasalong the major river systems in the region—theMekong and the Red River—were traditionallymore suitable for agriculture and have long sus-tained higher population densities than the oftensteeply sloped upland areas. Higher elevationareas, in contrast, contain much of the remain-ing forest cover and a high degree of biodiversity.The nature of the poverty-environment nexus istherefore very different in the two areas. With theexception of the plantation agriculture in the Cen-tral Highlands in Vietnam, most intensive com-mercial agriculture is in the lowlands, where issuessuch as overuse of pesticides are thus most severe.



Similarly, severe industrial pollution is mostlyconcentrated in the densely populated urbanareas of Hanoi and Ho Chi Minh City. In theuplands, on the other hand, forest-related issuesare very significant. This includes maintenanceof ecological functions, the use of non-timberforest products, and intensification of tradition-ally sustainable forms of mountain agriculturesuch as shifting cultivation.

The contrast between the natural resourcemanagement issues in the uplands and pollution-related environmental problems in the lowlandsis mirrored by a distinct difference in the natureof poverty between the two regions. With very few

exceptions, poverty is most severe in the uplandareas, where the districts with the highest propor-tion of poor are located. These areas, however,have low population densities, so the absolutenumber of poor is relatively small. Lowland areashave seen the fastest decline in poverty rates. Buttheir much higher population size means that theycontinue to contain by far the largest share of thepoor, especially in Vietnam. Fig. 4

The upland-lowland dichotomy is also relevantin the context of public infrastructure provision.Provision of networked infrastructure such aselectricity and water, but also roads, is much morecostly in sparsely settled rural areas. In the case of



Many socioeconomic indicators and some environmental data are available for administrativeunits such as districts or provinces. Analyzing such data geographically or statistically often posesmajor challenges. The most common of these is when data are only available at fairly coarse levelsof aggregation. Data at the province or district level may hide considerable heterogeneity that isvisible at the subdistrict level (see Fig XX below). For instance, relatively rich provinces might contain pockets of poverty that require urgent policy intervention.

Two related problems that are frequently encountered are that the units do not match thephenomenon being studied and that boundary changes over time make it difficult to performchange analysis. The first of these problems can occur in poverty-environment studies whenadministrative units were drawn with no consideration of underlying geographic features suchas river basins, areas with rural versus urban characteristics, or the distribution of ethnic groups.Depending on how these reporting units are drawn, the summary values of socioeconomic or bio-physical indicators may change significantly. In principle, just by re-aggregating existing data, onemay find completely different results in correlation studies. The larger the units, the bigger thisproblem will be. A related issue is that different levels of administration may have very differentinfluence on policies relevant to the topic studied. An example is Cambodia, where elections havebeen held only at national and commune level. The intermediate province and district levels there-fore have much less relevance for policy implementation. Consequently, the PEN study found thatresearch at the district level in Lao PDR provides useful insights into poverty-environment inter-actions, while such patterns are nonexistent at the province level. For Vietnam, both levels are useful, while for Cambodia it is necessary to go to the subdistrict level to analyze these topics.

The second problem occurs when boundaries change significantly. This tends to happen beforeeach major data collection effort, such as a census. It has also been observed as a consequence ofdecentralization, where units might be created largely as a way to increase the number of politicalappointees that might benefit from newfound decision-making power (Fitrani and others 2005).Analysis of possible causal relationships requires observations of changes over time, where changesin one variable during a certain period of time have an impact on changes in another variable dur-ing a subsequent period. But often it is impossible to reconcile data sets collected in different peri-ods. To address both problems, an important task for data collection agencies is to maintain anddisseminate information at the lowest possible level of aggregation that does not compromise dataconfidentiality concerns—for example, at the enumeration area level, which will include a few hun-dred households. This allows for aggregation of indicators to any set of units that are appropriatefor the problem being studied. Increasingly, census and statistical agencies in many countries aremoving in this direction.

B O X 1 . 2 Working with Data Recorded by Administrative Units

the upland regions, the terrain also contributes tohigh costs in relation to the size of the popula-tion served, and local governments may have lesscapacity for effective management of investmentprograms. In more densely settled urban or high-intensity agricultural areas, in contrast, there arelarge economies of scale in service provision, whichis reflected in much higher coverage rates.

Analytical Methods Used in the PEN Study

The PEN analyses in the three countries followedslightly different approaches, since the nature ofspecific issues and data availability vary to someextent. However, for all three countries there is anoverall unifying framework anchored in a tightlinking of a national analysis using secondary datawith locally relevant in-depth studies. In Lao

PDR, the sector-specific studies were very closelycoordinated and implemented by local expertteams that interacted frequently among them-selves and with a group of advisers. Consequently,all five case studies followed a very similar path. InCambodia, the comprehensive natural resourcemanagement case study was implemented by alocal policy research group, and in Vietnam sec-toral policy research was conducted by differentteams of national and international experts.

Tight Linking of National/Local Analysis

For all countries, the sector-specific, local studiesare embedded in a comprehensive analysis at thenational level. This analysis relied on a variety ofinformation sources, including published reportsand statistical data sets. For each country, thenational population and household census, and to



Data source: MoP and WFP (2002)

F I G U R E 1 . 3 Poverty Rates for Provinces, Districts, and Communes

a lesser extent the agricultural and economic cen-sus, provided a core set of indicators at all levelsof the administrative hierarchy. The census yieldsdemographic information, as well as some indica-tion of access to water and sanitation and availableassets. As described previously, each country alsohas implemented comprehensive household sur-veys with the objective of collecting detailed infor-mation on general living conditions or on healthstatus for a large sample of the population. Whilethese samples are not large enough to yield reliabledata at the district level, analysis of the individual-level micro data provides valuable insights, espe-cially on environmental health issues. Whereverpossible, indicators at various levels of aggregationwere presented in map form for exploratory visu-alization and presentation of key results.

Statistical analysis of secondary data revealsinitial correlations between environmental and

welfare indicators. While correlation does notnecessarily imply causality, the resulting infer-ences provide some indication of possible rela-tionships, which in turn suggest appropriateinterventions. More formal testing of causallinks is not possible in most cases, because mostof the required information is available for onlyone time period. Fundamental questions—suchas whether poor people are mostly victims ofenvironmental degradation, or whether povertyitself causes a decline in the natural resourcebase—cannot be answered with statistical cer-tainty without fairly extensive time series data.Furthermore, in attempting to identify causallinks, there is also a problem of intervening vari-ables that affect the poverty-environment inter-actions. For instance, children in poorhouseholds are more likely to suffer from water-borne diseases, but there is a significant influence



F I G U R E 1 . 4 Relation Between Poverty and Elevation in Vietnam

from factors such as the education of themother, which can be both a cause and conse-quence of poverty. For these reasons, the studiespresented in this report do not claim to show sta-tistically sound proof of specific relationships.Rather they are designed to provide solid empir-ical evidence of relevant interactions and corre-lations that point to an urgent need for policyintervention.

In the first phase of the PEN study, thenational-level analysis assigned districts or prov-inces into four categories in each of a numberof dimensions of the poverty and environmentnexus: poverty incidence, deforestation rate, steep-ness of slope as an indicator of soil vulnerability,wood/charcoal use, unsafe water source, andprevalence of childhood diarrhea. The compiledinformation was presented as a color-coded tablethat clearly highlights coincidence of high or lowrankings (Fig. XX).

Selection of Case Studies

The national analysis in the first phase was exten-sively updated and expanded in the second phase.This process helped refine the development of amatrix of PEN issues and focused the identifica-tion of key knowledge gaps to be considered in the in-depth analyses. In discussions withnational counterparts, the potential focal studieswere then narrowed down to a subset for eachcountry, as discussed below. The geographicalselection of these studies was guided by the fol-lowing main criteria:

– Poverty incidence, geographic distribution ofpoverty, and poverty persistence (so-called“poverty traps”).

– Government priority regions, provinces anddistricts for poverty reduction, economic devel-opment, and environmental management.

– Representation of the distinct socioeconomic,environmental, and renewable resource usepatterns in the countries (e.g., upland/lowland,industrial concentration and “hot spots”, pes-ticide intensity, etc.).

– Estimates of the number of poor and non-poor affected by environmental issues.

For Lao PDR, three studies relate to naturalresource management issues: use of non-timberforest products by upland communities; investi-gation of the link between road access, poverty,and environmental quality; and the impact ofunexploded ordnance on poverty and naturalresource management. Two closely related stud-ies, one in urban centers and the other in ruralvillages, focused on access among the poor towater supply and sanitation and the sustainabilityof water and sanitation services in poor villages.In each of these, a subset of districts was chosenbased on criteria that ensured variations in condi-tions across the sample. For each selected district,the study teams assembled a detailed statisticalprofile by combining nationally available indi-cators with information collected locally fromdistrict officials. Further sampling proceeded byselecting villages within each district and house-holds within each village. The selection aimedat achieving a pre-defined stratification—forinstance, based on distance from roads or degreeof in-migration. This approach is not intended toproduce a scientifically accurate sample, as is thecase in a large comprehensive household survey.Rather it supports more qualitative informationcollection through community questionnairesand interviews of focus groups and key stake-holders. The outcome of this process in eachin-depth study is a contextually rich package of information that helps identify the complexinteractions between human and natural systemsin the focus regions.

The Cambodia study focuses on the status,trends, and impacts of access to natural resourcesamong poor and non-poor households. Thestudy selected two provinces, each of which hada protected area, different types of land accesssituations, and high migration and poverty rates.Households in six villages were selected for astructured interview. Additional informationwas collected from key informants, such as gov-ernment officials and focus group interviews. As





Province Poverty rate

Deforest-ation

Slopedland

Wood/ Charcoal

Unsafe Water

Child Diarrhea

Luangphrabang Huaphanh Oudomxay

Savannakhet

Champasack Xiengkhuang Phongsaly Saravane KhammuaneXayabouri Vientiane Municipality Vientiane Luangnamtha Bokeo Attapeu BorikhamxaySekong

Xaysomboon

1 – worst 2 3 4 – best Source: PEN I final report

Source: PEN I final report

F I G U R E 1 . 5 Poverty and Environmental Problems in Lao PDR Provinces

which life depends—is reflected in the in-depthstudy of the Cau River Basin. The Cau Riverexperiences significant pollution from industrialand municipal sources. The study aimed at iden-tifying the distribution of the impacts and find-ing possible interventions for remediation. Incontrast to most of the other studies, the geo-graphic unit of analysis is the watershed, whichdoes not completely match the administrativehierarchy for which most socioeconomic data

in the Lao PDR case studies, poor and non-poorhouseholds were identified using a simple asset-based indicator approach.

The in-depth studies carried out in Vietnampresent a more varied mix of PEN situations,reflecting the more complex set of problems facedin a country with a large population and rapidlygrowing economy. This balancing act of creatinga dynamic economy that contributes to povertyreduction—while preserving natural systems on

area available. The area of interest falls into twoprovinces, in which structured interviews with700 households were conducted. This informa-tion was augmented by direct observation andinterviews with decision makers and stakehold-ers. The poverty measure employed is based onincome data from the questionnaire, which facil-itates ranking of households by economic status.

Industry is also the focus of another in-depthstudy that analyzed the spatial relation betweenpoverty, industrial concentration, air and waterpollution, and environmental health risk. Thisanalysis was accompanied by rapid local-levelsurveys of a number of urban hot spots and craftindustry villages, highlighting that the pollutionproblem is not restricted to large industrial estatesbut also affects life in smaller towns and evenvillages. The comprehensive Vietnam NationalHealth Survey of 36,000 households providedadditional insights on possible outcomes of expo-sure to pollution such as respiratory illness and

contamination of shallow dug wells, on whichthe poor rely disproportionately. The same sur-vey also yields insights at the national level onpublic health concerns related to substandardwater supply and sanitation, as well as possibleeffects of pesticides in agriculture.

Environmental health implications of pesti-cide use are the topic of a case study in theMekong Delta, which is the largest region ofintensive rice cultivation in the country located ina fragile ecosystem. A team of medical researchersimplemented a detailed survey of about 600 farm-ers in 10 districts that collected socioeconomicinformation as well as medical tests for the expo-sure to toxic chemicals. The land policy casestudy, finally, combined an analysis of a land-related database of 610 districts in Vietnam witha survey in three provinces in the upland regions.Within these, districts with contrasting condi-tions were selected. Asset-based poverty indi-cators aided the household stratification for



T A B L E 1 . 2 Data Sources Used at Different Administrative Levels

Nation Province District Commune Village Household

CambodiaCensus of Population 1998 � � �Cambodia Socioeconomic � �

Survey (CSES) 1997Case studies � �

Lao PDRCensus of Population 1995 � � �Lao Expenditure and � �

Consumption Surveys (LECS)II-1997/98, III-2002/03

Case studies � �

VietnamPopulation and housing � � � �

census 1999Vietnam Living Standards � �

Surveys (VNLSS 1998, 2002);Vietnam National Health � �

Survey (VNHS) 20XXNationwide land inventories �INEST/MOST (CV)Case studies � �

analysis of structured interviews and focusgroup discussions.

In summary, the process of analysis andconsultation resulted in the following thematicstratification of in-depth studies of poverty-environment linkages in the three countries:

– PENs with a potentially high number of poorbeing affected by renewable natural resourceissues and environmental health concerns:rural water supply and sanitation in Lao PDRand Vietnam, NTFPs and road developmentin Lao PDR, land use in Vietnam, and naturalresource management in Cambodia.

– Development trends and emerging PENs inurban areas: small town water supply and san-itation in Lao PDR, industrial pollution inrapidly growing cities in Vietnam.

– Previously unexplored PEN issues where sig-nificant knowledge gaps exist: UXOs in LaoPDR, pesticides in Vietnam, indoor air pollu-tion in Vietnam.

– Geographic priorities and cross-sectoral regionalissues: Son Cau River basin and pollution issuesin craft villages in Vietnam.

Endnotes1. http://www.undp.org/pei/pdfs/PEIPhase1Summary-

Paper2.pdf2. This information also forms the basis for estimation of

cross-nationally comparable welfare measures, such asthe proportion of the population living on less than $1or $2 a day.

3. Vietnam National Living Standards Surveys (VNLSS1997/98); Cambodia Socioeconomic Survey (CSES1997); Lao PDR Expenditure and Consumption Survey(LECS II 1997/1998; LECS III data from 2002/03 werealso utilized in Chapter 3). Census enumerations in the2000 census round were carried out in Vietnam in 1999,Cambodia in 1998, and Lao PDR in 1995. Data from the2005 census in Lao PDR were not yet available.

4. See Angelsen and Wunder (2003) for a discussion ofpoverty-forest linkages on which this section draws.

ReferencesAngelsen, A., and S. Wunder. 2003. “Exploring the forest-

poverty link: Key concepts, issues and research impli-cations.” CIFOR Occasional Paper No.40. Bogor,Indonesia: Center for International Forestry Research.

Ansuategi, A., E. B. Barbier, and C. A. Perrings. 1998. “Theenvironmental Kuznets curve.” In J. C. J. M. van denBergh and M. W. Hofkes, eds. Theory and Implementa-tion of Economic Models for Sustainable Development.Dordrecht, NL: Kluwer.

Bojö, J, K. Green, S. Kishore, S. Pilapityia, and R. Reddy.(2004). “Environment in PRSPs and PRSCs.” WorldBank Environment Department Paper No. 102. Wash-ington, DC: World Bank.

Bojö, Jan, Julia Bucknall, Kirk Hamilton, Nalin Kishor,Christiane Kraus, and Poonam Pillai. 2001. PovertyReduction Strategy Papers Source Book. (Chapter 11,Environment). Washington, DC: World Bank.

Casetella, Jean-Christophe, and Dang Dinh Quang, eds.2002. “Doi Moi in the Mountains. Land use changesand farmers’ livelihood strategies in Bac Kan Province,Vietnam.” Hanoi: The Agricultural Publishing House.

Chantirath, K. 1998. “Forestry resources and the underlyingcauses of deforestation and forest degradation in LaoP.D.R.” In IGES Forest Conservation Project. “A Steptoward Forest Conservation Strategy.” Interim Report.www.iges.or.jp/en/fc/phase1/interim-contents.htm.

Cord, L., H. Lopez, and J. Page. 2003. “When I Use aWord . . . Pro-poor Growth and Poverty Reduction.”Processed. Washington, DC: World Bank.

Cuong, Pham Manh, Manfred Zeller, Regina Birner, andDaniel Mueller. 2003. “Improvement of access to avail-able agricultural technologies: did it cause land-usechanges and poverty reduction in the North-westernupland of Vietnam?” Paper presented at the Conferenceon International Agricultural Research for Development,Göttingen, Germany, October 8–10.

Dasgupta, S., B. Laplante, H. Wang, and D. Wheeler. 2002.“Confronting the environmental Kuznets curve.” Jour-nal of Economic Perspectives 16: 147–168.

Dasgupta, Susmita, Mainul Huq, M. Khaliquzzaman, KiranPandey, and David Wheeler. 2004. “Indoor Air Qualityfor Poor Families: New Evidence from Bangladesh.” Pol-icy Research Working Paper 3393, DevelopmentResearch Group. Washington, DC: World Bank.

Dasgupta, Susmita, Uwe Deichmann, Craig Meisner, andDavid Wheeler. 2005. “ThePoverty-EnvironmentNexus in Cambodia, Lao PDR, and Vietnam.” WorldDevelopment 33 (4): 617–638.

Deaton, A. 1997. The analysis of household surveys. Balti-more, Maryland: Johns Hopkins University Press.

DFID, EC, UNDP, and World Bank. 2002. Linking PovertyReduction and Environmental Management: Policy Chal-lenges and Opportunities. Washington, DC: World Bank.

Duraiappah, Anantha K. 1998. “Poverty and Environmen-tal Degradation: A Review and Analysis of the Nexus.”World Development 26 (12): 2169–2179.

Ekbom, Anders, and Jan Bojo. 1999. “Poverty and Environ-ment: Evidence of Links and Integration into the Coun-try Assistance Strategy Process.” Discussion Paper No. 4,



Environment Group, Africa Region. Washington, DC:World Bank.

Filmer, D., and L. Pritchett. 2001. “Estimating wealtheffects without expenditure data–or tears: An applica-tion of educational enrollment in states of India.”Demography 38(1):115–132.

Fitrani, F., B. Hofman and K. Kaiser (2005), Unity in diver-sity? The creation of new local governments in a decen-tralizing Indonesia, Bulletin of Indonesian EconomicStudies, 41, 1:57–79.

Fox, Jefferson, Dao Minh Truong, A. Terry Rambo, NghiemPhuong Tuy En, Le Trong Cuc, and Stephen Leisz.2000. “Shifting Cultivation: A New Old Paradigm forManaging Tropical Forests.” BioScience 50 (6): 521–528.

Geist, Helmut J., and Eric F. Lambin. 2002. “ProximateCauses and Underlying Driving Forces of TropicalDeforestation.” BioScience 52 (2): 143–150.

Hentschel, J., P. Lanjouw, J. Lanjouw, and J. Poggi. 2000.“Combining Household Data with Census Data toConstruct a Disaggregated Poverty Map: A Case Studyof Ecuador.” World Bank Economic Review, 14 (1):147–166.

Inter-Ministerial Poverty Mapping Task Force. 2003.Poverty and inequality in Vietnam. Spatial patterns andgeographic determinants. Hanoi: Ministry of Agricultureand Rural Development, International Food PolicyResearch Institute, Institute of Development Studies.

Landell-Mills, N., and I. Porras. 2002. Silver Bullet or Fool’sGold? A global review of markets for forest environmentalservices and their impacts on the poor. London: Interna-tional Institute for Environment and Development.

MoP and WFP. 2002. Estimation of poverty rates at the com-mune level in Cambodia. Phnom Penh: Ministry of Plan-ning and UN World Food Program.

Müller, D., and M. Zeller. 2002. “Land use dynamics in thecentral highlands of Vietnam: a spatial model combin-ing village survey data with satellite imagery interpreta-tion.” Agricultural Economics 27: 333–354.

Pagiola, S., A. Arcenas, and G. Platais. 2005. “Can Paymentfor Environmental Services Help Reduce Poverty? AnExploration of the Issues and the Evidence to Date fromLatin America.” World Development 33 (2): 237–253.

Palm, Cheryl A., Stephen A. Vosti, Pedro A. Sanchez, andPolly J. Ericksen. 2005. Slash-and-burn agriculture. Thesearch for alternatives. New York: Columbia UniversityPress.

Perman, R., and D. I. Stern. 2003. “SP3Evidence from panelunit root and cointegration tests that the environmentalKuznets curve does not exist.: Australian Journal of Agri-cultural and Resource Economics 47 (3):325–347.

Ravallion, M. 1994. Poverty comparisons. Chur, CH: Har-wood Academic Publishers.

Song, Nguyen Van. 2003. “Wildlife Trading in Vietnam:Why it Flourishes.” EEPSEA Research Report 2003 -RR6. Singapore: Economy and Environment Programfor Southeast Asia. (http://www.idrc.ca/uploads/user-S/10705050241SongRRlayout.doc, last accessed2/13/06).

UNDP & EC. 2000. “Attacking poverty while improvingthe environment: Towards win-win policy options.”Poverty and Environment Initiative. New York: UNDPand European Commission.

Vedeld, P., A. Angelsen, E. Sjaastad and G. Kobugabe-Berg.2004. “Counting on the Environment: Forest Incomesand the Rural Poor.” Environment Department PaperNo. 98. Washington, DC: World Bank.

World Food Programme (WFP). 2001. “Lao PDR PovertyAnalysis.” Asian Development Bank, Lao National Sta-tistics Center, United Nations World Food Programme.Analysis and report prepared for the Asia and PacificForum on Poverty, Manila.

World Food Programme (WFP). 2004. “Analysis of Vul-nerability to Food Insecurity District Level, Vulnerabil-ity Analysis and Mapping.” Vientianne: United NationsWorld Food Programme.

World Bank. 2001. Lao PDR Production Forestry Strategy.2 vols. Washington, DC: World Bank; SIDA; Ministryof Foreign Affairs, Government of Finland.

World Bank. 2003. World Development Report 2003: Sus-tainable Development. Washington, DC: World Bank.

World Bank. 2005a. Environment strategy for the WorldBank in the East Asia and Pacific Region. Washington,DC: World Bank.

World Bank. 2005b. World Developing Report 2005: Equityand Development. Washington, DC: World Bank.

World Bank. 2005c. “Going, going, gone . . . The illegal tradein wildlife in East and Southeast Asia.” Environment andSocial Development East Asia and Pacific Region Discus-sion Paper. Washington, DC: World Bank.



COUNTRYSTUDY

FINDINGS

II


This chapter examines poverty-environmentrelationships in Cambodia through an analysisof national data and case study findings. Webegin with an overview of the national contextand trends for poverty and natural resources,providing a synthesis of existing studies and infor-mation across Cambodia’s regions and provinces(Map 2.1). Next, we conduct a range of newanalysis to identify factors associated with poverty,and to investigate specific poverty-environmentrelationships pertaining to natural resource depen-dence, drinking water sources, sanitation, nat-ural disasters, and mine/UXO contamination.To complement national analysis of poverty andnatural resource relationships, we then presenta case study on poverty and access to naturalresources conducted in two provinces. Finally,we provide a summary of key findings and dis-cuss policy implications.

POVERTY: CONTEXT AND TRENDS

Poverty incidence in Cambodia fell from about47 percent in 1993–94 to 35 percent in 2004,according to findings from the 2004 CambodiaSocio-Economic Survey (CSES).1 However,rural poverty incidence (37.8 percent) remainsconsiderably higher than urban areas (17.6 per-cent) (Table 2.1). Since about 85 percent of

Cambodia’s population is located in ruralareas— 11 million out of a total population of13 million—poverty remains largely a ruralproblem. Indeed, of the total population livingbelow the poverty line in Cambodia, morethan 90 percent are located in rural areas, equalto more than 4 million people.

In addition to urban-rural differences in pov-erty, poverty rates are quite uneven across Cam-bodia’s regions and provinces (Map 2.2). Thepercentage of the population living below thepoverty line in coastal areas is 27 percent, butthis rises to 52 percent in the plateau/mountainsregion. Although poverty severity is highest in therural plateau/mountains region, this region is lesspopulated than the plains and Tonle Sap regions,which together account for about 75 percent ofthe population in poverty.

Trend estimates across geographically compa-rable areas from CSES 1993–94 to 2004 indicatepoverty has declined in all regions except for therural plateau/mountains. Overall, most provincesmade progress in reducing poverty over thepast decade. Poverty rates only increased in twoprovinces—Kampong Thom and KampongSpeu, though these increases are within the statis-tical margin of error. Poverty reduction was con-siderably less on a percentage basis in the ruralareas during this period than in the urban and

Poverty and Environment in Cambodia

2

�

semi-urban areas. The main reason is becauseinequality in the distribution of per capita con-sumption increased significantly.2 In decompos-ing inequality, approximately 86 percent of thetotal increase can be attributed to a rise in inequal-ity within rural regions, while only 14 percentcan be attributed to an increase in rural-urbaninequality (World Bank 2005a).

Nationwide, poverty rates of householdsheaded by ethnic minorities (defined as non-Khmer) are only slightly higher (35.8 percent)than Khmer-headed households (34.6 per-cent). However, because Chinese and Vietnamesehouseholds are typically wealthier than Khmerhouseholds, these minority groups may offsethigher poverty rates among smaller ethnic groupsconcentrated in the upland areas of the north-eastern provinces—about 17 ethnic groups com-prising over 100,000 people (Map 2.3).

P O V E R T Y A N D E N V I R O N M E N T I N C A M B O D I A


M A P 2 . 1 Regions in Cambodia (as applied in CSES Survey 2004)

Source: World Bank 2005b.

T A B L E 2 . 1 Poverty Estimates by Region, Urban andRural, 2004

Percentage of Populationpopulation below distribution of poverty

poverty line (% of all poor)

Region Urban Rural Total Urban Rural Total

Phnom Penh 1.1 8.9 4.6 0.2 1.0 1.1Plains 13.7 32.9 32.1 0.7 39.0 39.7Tonle Sap 28.2 45.4 42.8 3.7 33.4 37.0Coastal 20.4 30.1 26.8 1.6 4.6 6.1Plateau/ 32.6 56.3 52.0 1.8 14.2 16.1

mountainsCambodia 17.6 37.8 34.7 7.9 92.1 100.0

Source: CSES 2004.

NATURAL RESOURCES: CONTEXT AND TRENDS

Most of Cambodia’s rural population dependson one rice crop per year, fish and other aquaticresources, forest products, and wage labor. CSES(2004) data indicate the income of the averagerural household is comprised of crop cultivation(30 percent); livestock (10 percent); commonproperty resources, especially fish and forest prod-ucts (25 percent); and wage labor, nonagriculturalactivities, other (35 percent). Although rice farm-ing is the dominant livelihood activity, oftenhouseholds are active in all of these income-generating pursuits, diversifying in order to opti-mize labor resources during different seasons, andsafeguard against the risks of agricultural failure(McKenney and Prom 2002; Chan and Acharya2002; Prom and Ballard 2005).

Cambodia enjoys one of the highest naturalresource endowments per capita in the region.This includes considerable unused land, relativelyabundant forest resources, and highly productivefreshwater fisheries. A lack of (credible) data,however, makes it difficult to characterize the cur-rent status of natural resources with precision.

Fisheries

The national catch has been estimated in the rangeof 300,000 to 400,000 metric tons (360,000 met-ric tons in 2003), though the challenges ofcollecting catch data suggest a wide range ofuncertainty about this estimate. About 80 per-cent of the catch is from inland fisheries, mostfrom the Tonle Sap Lake, as well as the Mekongand Tonle Bassac River areas. Marine fisheriesalong Cambodia’s 435-km coastline provide theother 20 percent of the catch, important to coastalareas of Kampong Som, Kampot, and Koh Kongprovinces.

Although fish catch and composition dataare lacking, there is a strong public perceptionthat fisheries resources are becoming scarcerdue to overfishing and various destructive fishingapproaches. In a recent opinion poll of communeleaders in Cambodia’s rural areas (1,500 com-munes), 86 percent reported that the volume offish catch has decreased compared to five yearsago (Seila and Danida 2005). A major concern isthat ongoing habitat destruction and fishing pres-sure will reduce fish stocks and diversity to thepoint where fish production consists of a limited



M A P 2 . 2 Poverty Incidence by Region & Rural Poverty Incidence by Province

Source: CSES 2004.

number of small, low-value species, while larger,higher-value fish species become a diminishingproportion of the catch. This occurs becauselarger fish species typically require more than aone-year cycle for reproduction, which makes

them more vulnerable to fishing pressure. Theresulting change in the species compositionreduces catch values per unit of fishing effort(Degen and others 2000).

Forests

The Forest Administration estimates that forestcover increased from 58 percent of the countryin 1997 to 61 percent in 2002 (Map 2.4). Thisestimate stands in stark contrast to numerousother forest sector studies, which indicate thatlogging and clearing of forests have been rampantduring this period (Independent Forest SectorReview 2004; Fraser Thomas and others 2000;www.globalwitness.org). For example, the Inde-pendent Forest Sector Review (IFSR 2004) esti-mates 55 percent of forest area has been disturbedor heavily degraded (Map 2.5). The IFSR alsohighlights the problem of uncontrolled clearing offorests, which has been occurring at an accelerat-ing pace over the past decade. Whereas 0.3 mil-lion ha of forest were cleared from 1993 to 1997,1.3 million ha were cleared from 1997 to 2002(Map 2.6). This finding is consistent with other



M A P 2 . 4 Forest Cover of Cambodia, 2002

Source: Forest Administration 2003.

M A P 2 . 3 Minority Ethnic Population & Uplands/Lowland in Cambodia

Source: Population Census 1998.

estimates indicating that the rate of deforestationincreased from an average annual rate of about0.5 percent prior to the mid–1980s, to about1 percent from the mid–1980s to mid–1990s,and then to 1.0–1.7 percent from the mid–1990sto 2000 (IFSR 2004; McKenney and others2004). The most recent indication of forestdecline comes from a national opinion poll ofcommune leaders, which finds that 72 percent ofcommune leaders in rural areas report forest cov-erage has declined compared to five years ago(Seila and Danida 2005).

Land

Approximately 40 percent of the rural popula-tion lives off of less than 0.5 hectares of agricul-tural land, an area too small to yield sufficient riceto meet the needs of the average rural household.Landlessness appears to be on the rise, a resultof population pressure, land sales under duress(for example, health emergencies), and specula-tive land purchases (usually by Phnom Penh res-idents). About 13 percent of the rural populationwas estimated to be landless in 1997 (CSES1997), rising to 16 percent in 1999 (CSES 1999)and to 20 percent in 2004 (CSES 2004).

The chief means of acquiring new land forcultivation is forest clearance. This, along withuncontrolled logging, has resulted in rapid ratesof forest loss in recent years. But with relativelylarge areas of forest remaining, it is not surpris-ing that only 28 percent of commune leadersreported less access to land for cultivation com-pared to five years ago. Along with uncontrolledforest clearing, a key management concern is thequality of cleared land for agriculture over thelong term. If cleared land does not support pro-ductive agriculture due to poor soil fertility andwater availability (and much of Cambodia’s landhas low/medium soil productivity, Map 2.7),forest clearance could result in an ongoing cycleof poverty and environmental loss. Indeed, cur-rent estimates are that more than half the landcleared from 1997 to 2002 had low-to-medium



M A P 2 . 5 Forest Disturbance by Roads and Villages in Cambodia, 2002

M A P 2 . 6 Forest Cover Clearance in Cambodia,1993–1997 and 1997–2002

Source: IFSR 2004.

Source: IFSR 2004.

soil fertility (Dümmer 2004). Agricultural yieldsin these areas are likely to be poor. In turn, thismay lead to greater pressure on natural resourcesand more land clearing in these areas—creatingan ongoing cycle of forest clearing, poor (anddeclining) agricultural yields, greater pressure onnatural resources, and more forest clearing.

Establishing land use management regimes tobring logging and forest conversion for agricul-ture under control will be a significant challenge.Approximately one-third of the Cambodian pop-ulation (about 4 million people) live within 5 kmof forest, and about 10–15 percent of the popula-tion (about 1.5 million people) live within 5 kmof high-value forest (evergreen/semi-evergreen),which provide richer timber resources and otherforest products (Map 2.8). Moreover, the com-bination of in-migration and population growthis resulting in a rapid increase in new villagesbeing established in rural forested/cleared areas.Whereas the establishment of new villages largelyoccurred along the agriculture-forest frontierprior to 1998, villages established from 1998 to

2001 have penetrated much more deeply intoforest areas (Map 2.9).

Water

With an estimated annual runoff of 475 billioncubic meters from the Mekong system drainingover 85 percent of the country, Cambodia’s waterresources are plentiful (World Bank 2005b). Butrainfall is concentrated during a six-month wetseason, with little infrastructure for capturing,storing, and regulating this runoff. Flooding isan annual occurrence, damaging infrastructureand crops, but also providing benefits to thefloodplain by bringing soil nutrients, water, andcreating areas for fish spawning/habitat. Duringthe dry season, river levels drop substantially, andsurface water resources are largely limited torivers (Mekong, Sap, and Bassac) and the TonleSap Lake. Groundwater resources are largelyuntapped; they appear to be uneven geographi-cally (most in close proximity to rivers), but morestudy is needed. Irrigation coverage in Cambodia(7 percent) is far less than the neighboring coun-tries of Vietnam (34 percent), Thailand (26 per-cent), and Laos (18 percent) (World Bank 2005a).As a result, most agriculture is highly dependenton the timing and levels of rainfall.

The Management Vacuum

Over the past decade, the Cambodian govern-ment’s chief approach to natural resource man-agement has been to contract out large areas toCambodian and foreign investors as forest, fish-eries, and agricultural/economic concessions. Thishas proven to be a highly ineffective managementapproach. It has failed to galvanize economicgrowth in rural areas; generated only limited levelsof government revenue through taxes and fees;and often resulted in conflict between local com-munities and concession operators when commu-nities have lost access to resources (McKenney2002; World Bank 2005b).



M A P 2 . 7 Soil Productivity of Cambodia

Source: Forestry Administration 2003.

Recent policy interventions suggest the gov-ernment recognizes some of these problemsand is seeking alternative approaches to naturalresource management. Support for “social con-cessions”—a mechanism that could promoteland distribution—is one acknowledgment bythe government that a focus on large land-holdings as “economic concessions” has, withfew exceptions, attracted little investment (WorldBank 2004). In the forest sector, several conces-sions have been canceled, and remaining conces-sionaires are being required to submit strategicforest management plans while a ban on log-ging and log transport is in effect. However, itremains unclear whether this process will resultin tangible improvements in forest manage-ment, or to what extent alternative managementapproaches will be promoted. A number of for-est sector studies call for a broader shift in focusfrom industrial approaches to the explorationof various commercially oriented communityforestry approaches (IFSR 2004; McKenney andothers 2004). In the fisheries sector, the govern-ment acted to improve access to fisheries in2001 by canceling commercial fishing lots cov-ering more than half the fishing concession area(McKenney and Prom 2002). However, manyof the remaining fishing lots control the mostproductive fisheries.

As a result of the changing management ap-proaches, much of Cambodia’s natural resources(perhaps 30 to 40 percent of total area, Figure 2.1)are now under no clear management regime(World Bank 2005b). Canceled forest conces-sions have reverted to the “forest reserve,” butmanagement arrangements have not been defined(Table 2.2). Likewise, the roughly 500,000 ha offisheries concessions released in 2001 have notbeen brought under clear management regimes.Finally, an estimated 1.75 million ha of “unused”land has not yet been declared to be under anyspecific ownership, use, or control. Until thesemanagement vacuums are addressed, there willbe the potential for misappropriation and over-exploitation of resources (World Bank 2005b).



M A P 2 . 8 Villages within 5 km of Evergreen/Semi-Evergreen Forests in Cambodia, 2002

M A P 2 . 9 Cambodian Village Locations 1998, and New Villages 1998–2001

Source: Adaptation of data from IFSR 2004.

Source: IFSR 2004.

Natural Resources and Moving Out of Poverty

With much of Cambodia’s natural resourcesunder no clear management, and existing man-agement regimes often weakly enforced, naturalresources are frequently accessed and used byhouseholds as common property resources (CPR).The importance of these resources as a safety netfor rural households has been established in anumber of studies (Prom and Ballard 2005; CSES2004; Chan and Acharya 2002; McKenney andProm 2002). Indeed, the majority of rural poorhouseholds depend on CPR for some amount ofconsumption and income generation. But a recentstudy entitled Moving Out of Poverty Study (Cam-bodia Development Resource Institute 2006)attempts to take the analysis a step further byinvestigating the extent to which CPR may helphouseholds move out of poverty over time.3 Thestudy’s findings suggest the role of CPR has beenlimited. Key conclusions include the following:

• CPR household income declined for all mobil-ity categories due to declining availability andaccess. For some villages, forest access has been

restricted or prohibited; others located in theMekong and Tonle Sap areas have seen a sharpdecline in fish catch.

• CPR played little or no role for householdsthat moved out of poverty.

• Total income for the “chronic poor” group hasbecome more dependent on labor (up from 28to 38 percent), and less dependent on agricul-ture (33 to 23 percent) and CPR (28 to 18 per-cent). This reflects agricultural problems withfloods and droughts, as well as declining accessand availability of CPR.

• CPR access still played a crucial role as a safetynet in sustaining rural poor households cop-ing with shocks (floods, droughts, and healthemergencies).

FINDINGS: ANALYSIS OF POVERTY-ENVIRONMENT RELATIONSHIPS

In this section we present findings from our analy-sis of poverty-environment relationships, drawingon a number of recently developed datasets.4 Weidentify major factors associated with poverty and



Protection,concession,cultivation

55%

No clearmanagement

38%

Towns,infrastructure,

land mines7%

F I G U R E 2 . 1 Managed/Unmanaged Areas


T A B L E 2 . 2 Cambodia Land Tenure/Use Estimates, 2004

AreaCategory (million ha) % Share

Forests not under 4.7 26.1concessions

Protected areas/forests 3.3 18.3Cultivated areas 2.7 15.0Forest concessions 3.1 17.2Scrub land, nonwooded 1.7 9.4

landTowns, infrastructure 1.1 6.1Released fishing 0.5 2.8

concession lotsAgricultural concessions 0.8 4.4Land mines contami- 0.1 0.6

nated areasTotal 18.0 100.0


investigate specific poverty-environment relation-ships pertaining to natural resource dependence,drinking water sources, sanitation, natural disas-ters, and mine/UXO contamination.

To set the stage for this analysis, Table 2.3 pre-sents major factors with a statistically significantrelationship to poor and non-poor households inCambodia. While most of these relationships arewell-established, some are surprising (floods asso-ciated with non-poor households). Likewise,Table 2.4 summarizes the disparity between poorand non-poor households for a number of keypoverty and environment indicators.

Natural Resource Dependence

Nationwide, a substantial proportion of house-holds are engaged in collecting fuelwood andother wood forest products (72 percent), collect-ing non-wood forest products (21 percent), andcatching fish/seafood (53 percent) (Figure 2.2).Not surprisingly, it is much more common forrural households to be engaged in these activitiesthan urban households. Focusing on rural areas,the relationship between poverty and engagingin these natural resource-dependent activities isclear—more than twice as many households inthe poorest quintile engage in these activitiescompared to the richest quintile (Figure 2.2).This relationship was noted in a previous analy-sis of CSES data by Knowles (2005):

“Two variables referring to sources of house-hold income are positively and significantlyrelated to food poverty, i.e., the “householdcatches fish, shrimp, crabs or oysters” and the“household collects firewood, charcoal, timberor other forest products.” The fact that thesevariables are positively related to food povertydoes not mean that they cause food poverty.Instead, the causal link most likely runs in theopposite direction, i.e., food-poor householdsare more likely to use these activities to obtainadditional income. This result underlines theimportance of access to common resources as acoping mechanism for the poor.”



T A B L E 2 . 4 Poverty and Environment Indicators andPopulation Affected

Estimated number ofpeople affected

% of non-poor % of poorPoverty & environment indicator: population population

Lack of water supply (using open 16 22water sources)*

Using unprotected dug wells* 8 16Lack of sanitation (no latrine)* 60 80Households never boil drinking 24 48

water*Mine field and Cluster bomb 19 28

Contamination (2002)**Household use of Solid fuels/ 90 99

indoor air pollution*Flood (≥3 years in last 5 years)* 24 23Drought (≥3 years in last 5 years)* 12 16Urban environment*** <20 <10No All-weather Road* 19 28

*Estimated from CSES 2004. **Estimated from a combination of CSES 2004and the National Level 1 survey of minefields. ***This is simply the percent-age of non-poor and poor living in urban areas, and therefore potentiallyaffected by urban environmental issues such as outdoor air pollution.

T A B L E 2 . 3 Factors with Statistically Significant Relationship to Poverty

Factors associated with Factors associated withpoor households non-poor households

Household in Pleateau/Mountain Household in urban areaor Tonle Sap region

Provincial capital not in close Household in Phnom Penhproximity

Household size All-weather road in closeproximity

Female-headed household Head of household has some schooling

Drought-affected in past Head of household agefive years (older)

Mines severely affect Flood-affected in pastamount of available five yearsagricultural land

Cluster bombs dropped in/near village

Source: Estimated from CSES 2004 and National Level 1 Survey 2002.

To take this analysis a step further, we assess thepotential relationship between poverty, house-holds engaged in natural resource-dependentactivities, and areas where resources may be indecline. First, to understand trends in the nat-ural resource base, we draw on recent opinionpoll data from a survey of rural commune leaders

(Seila and Danida 2005).5 Specifically, we focuson responses to four poll questions (Table 2.5).

Map 2.10 indicates where these problems maybe most concentrated, showing the percentage ofcommunes within each province that responded“decrease” for at least three of the four responses.Problems of natural resource decline are mostacute in the northern most provinces, with about60 percent of the communes reporting naturalresource decline in Ratanakiri, Stung Treng, andOddar Meanchey. Regionally, the relationshipbetween natural resource decline and povertyappears strongest in the plateau/mountain region(Figure 2.3).

The poor are likely to be disproportionatelyaffected by the negative impacts of naturalresource decline (Figure 2.4). By examining onlythe households engaged in natural resource-dependent activities in communes reportingresource declines, it becomes clear that the poorare considerably more dependent on these dete-riorating resource bases than the non-poor.Indeed, about twice as many households in thepoorest quintile depend on these activities com-pared to the richest quintile.



0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

National Rural Urban

Collects fuelwood/forest products Collects NWFPs Catches fish/seafood

Rural

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Poorest 2 3 4 Richest

Collects fuelwood/forest products Collects NWFPs Catches fish/seafood

F I G U R E 2 . 2 Households Engaged in Natural Resource-Dependent Activities (left), by Household Consumption Quintile in Rural Areas (right), 2004

Source: Estimated from CSES 2004.

T A B L E 2 . 5 Commune Opinion Poll Responses: Natural Resource Decline

Communes respondingPoll question decrease

Volume of fish catch compared 86 percentto 5 years ago?

Forest coverage compared to 72 percent5 years ago?

Number of people with access 28 percentto land for cultivation compared to 5 years ago?

Number of people able to secure 19 percentlivelihood from natural resources in 2010?

Source: Seila and Danida 2005.

Drinking Water Sources

The proportion of Cambodian households withaccess to safe or improved drinking watersources ranges from about 65 percent (dry sea-son) to 75 percent (wet season) (Figure 2.5). Forthis analysis, sources generally considered safe orimproved include water piped in dwelling or onthe premises, tube/piped wells or boreholes,protected dug wells, and rainwater. Water fromtanker trucks and vendors may also be safe, butas this market is unregulated it is difficult toconfirm. Unsafe sources include unprotecteddug wells and ponds, rivers, and streams. Over20 percent of the population uses rainwater fordrinking water during the wet season, switch-ing to this source mainly from rivers, ponds,and streams, tanker trucks and vendors, andtube/piped wells.

Drinking water sources: non-poor and poor households

The proportion of non-poor Cambodian house-holds with access to safe/improved drinking watersources ranges from about 70 percent (dry season)to 80 percent (wet season), compared to 55 per-cent (dry season) to 65 percent (wet season) forpoor households (Figure 2.6). Water piped indwelling is much more common for the non-poor(17 percent) than the poor (2 percent). Andalmost one-quarter of non-poor householdsuse rainwater as drinking water during the wetseason compared to only 19 percent of poorhouseholds.

A strong relationship between poverty andunsafe drinking water sources (unprotected dugwells and ponds, rivers, and streams) is furtherrevealed when examining across household con-sumption quintiles (Figure 2.7). Nationally,households in the poorest quintile are more thanthree times as likely to use unsafe drinking watersources as the richest quintile. This disparity islower in rural areas, where the poorest quintile isonly twice as likely to use unsafe water sources.But in urban areas, households in the poorest



0% 10% 20% 30% 40% 50% 60%

Phnom Penh

Plains

Tonle Sap

Coastal

Plateau/Mountain

Rural Poverty Incidence % Communes Reporting NR Decline (3 of 4 responses)

M A P 2 . 1 0 Percentage of Communes within a ProvinceReporting Decreases in Natural Resources (3 out of 4 responses of decrease—fish, forest cover, land access, and long-term sustainability)

F I G U R E 2 . 3 Natural Resource Decline and RuralPoverty Incidence by Zone

Source: Estimated from Seila and Danida 2005.

Source: Estimated from CSES 2004 and Seila and Danida 2005.

quintile are 8–10 times more likely to use unsafewater sources compared to the richest quintile.

Drinking water sources: rural and urban households

The proportion of rural households with accessto safe/improved drinking water ranges from

60 percent (dry season) to 70 percent (wet season),compared to over 85 percent for urban households(Figure 2.8). Looking at specific water sources,water piped in dwelling is far more prevalent forurban households (over 40 percent) comparedrural households (4 percent). Use of rainwaterduring the wet season is more common for ruralhouseholds (25 percent) than urban households(13 percent).

Analyzing rural and urban households accord-ing to non-poor or poor status sheds further lighton relationships between poverty and drinkingwater sources (Figure 2.9). First, while rural non-poor households have somewhat greater accessto safe/improved drinking water sources thanrural poor households, this is largely due to morepurchases from tanker trucks and vendors duringthe dry season and greater use of rainwater duringthe wet season. The disparity is greater in urbanareas, where about 90 percent of non-poor house-holds have access to safe/improved drinkingwater, compared to only about 65–70 percentof poor households. Moreover, non-poor house-holds in urban areas are far more likely to havethe convenience of water piped in dwelling (50 percent) than poor households in urbanareas (7 percent).

Factors associated with accessing drinkingwater sources: regression results

Several variables show a statistically significantrelationship with access to drinking water sources.First, access to unsafe drinking water sources isassociated with households that are poor, locatedin the coastal or plateau/mountain regions, andlocated in rural areas far from district and provin-cial capitals and all-weather roads. Access tounsafe water is also associated with male headsof households who have no schooling (female-headed households are associated with usingsafe/improved drinking water sources at a statisti-cally significant level).

During the wet season, approximately one-quarter of households nationwide switch to theuse of rainwater. Variables associated with using



0%

10%

20%

30%

40%

50%

60%

70%

80%


Fuel wood/wood products collection and decline in resource

NWFP collection and decline in resource

Fish/seafood catching and decline in resource

National

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Dry season Wet season

Pond, river or stream

Unprotected dug well

Tanker truck, vendor

Rainwater

Tube/piped well/borehole orprotected dug well

Piped in dwelling or onpremises

F I G U R E 2 . 4 Households Engaged in Natural Resource-Dependent Activities in Communes Reporting Resource Decline,By Household Consumption Quintile

F I G U R E 2 . 5 Drinking Water Sources, Dry and Wet Season, 2004

Source: Estimated from CSES 2004 and Seila and Danida 2005.



Dry Season

0%

20%

40%

60%

80%

100%

non-poor poor




Rainwater

Tube/pipedwell/borehole orprotected dug well

Piped in dwelling oron premises

Wet Season

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

non-poor poor




Rainwater



F I G U R E 2 . 6 Drinking Water Sources, Non-Poor and Poor, Dry and Wet Season, 2004

F I G U R E 2 . 7 Unsafe Drinking Water Sources, by Household Consumption Quintiles, 2004


National

0%

10%

20%

30%

40%

50%


Wet Season Dry Season

Rural

0%

10%

20%

30%

40%

50%



Urban

0%

5%

10%

15%

20%

25%

30%

35%

40%




rainwater include non-poor households, female-headed households, and older heads of house-holds with some schooling. Rainwater use is alsoassociated with households located in the coastalor plateau/mountain regions, in rural areas withall-weather roads, and in areas close to the provin-cial capital. Similarly, variables associated with theuse of drinking water from tanker trucks andvendors include non-poor households, female-headed households, and heads of households withsome schooling, as well as households located incloser proximity to district and provincial capitals.

Boiling or otherwise treating drinking water

Nationwide, approximately 75 percent of non-poor households boil or otherwise treat theirdrinking water compared to only about 52 per-cent of poor households (Figure 2.10). This rela-tionship between poverty and boiling drinkingwater is illustrated even more clearly when look-ing across household consumption quintiles, withabout twice as many households in the richest

quintile boiling water compared to the poorestquintile. In rural areas, boiling drinking waterby non-poor households (71 percent) and poorhouseholds (51 percent) tracks rates at the nationallevel. But in urban areas, boiling drinking wateris much more common, especially by non-poorhouseholds (nearly 90 percent). Note that boilingis by far the most common method for makingdrinking water safe; few households treat theirdrinking water by other methods.

Analyzing by drinking water source revealsthat those households accessing unsafe watersources are the least likely to be boiling their water(Figure 2.11). That is, the households that mostneed to boil their water are the least likely to doso. This is especially true for poor households,where boiling is done by less than 25 percent ofhouseholds using water from unprotected wellsand only about 40 percent of households usingwater from ponds, rivers, and streams. But povertyis not a sufficient explanation, as only a modestproportion of non-poor households using theseunsafe drinking water sources boil their water—40 percent using water from unprotected wellsand 57 percent using water from ponds, rivers,and streams.

Regression analysis indicates several variablesassociated at a statistically significant level withhouseholds boiling drinking water. These vari-ables include households being non-poor, locatedin urban areas, and located in the plains or plateau/mountain regions. Boiling drinking water is alsoassociated with older heads of households whohave had some schooling, as well as with house-holds that have separate kitchens and use cleanfuels for cooking.

Sanitation

Nationwide, only about 30 percent of Cambo-dian households have access to basic sanitation—they are connected to sewerage, have a septic tank,pit latrine, or other facility without septic tank,or use a public or shared toilet (Figure 2.12). Therest of the population (70 percent) has no accessto sanitation services.



Rural & Urban

0%

20%

40%

60%

80%

100%

Rural &Dry

Season

Rural &Wet

Season

Urban &Dry

Season

Urban &Wet

Season




Rainwater



F I G U R E 2 . 8 Drinking Water Sources, Rural and Urban,Dry and Wet Season, 2004


Basic Sanitation: non-poor and poor pouseholds

Nearly 40 percent of non-poor Cambodian house-holds have access to basic sanitation (primarilyconnection to sewerage and septic tanks) com-pared to only about 8 percent of poor households(Figure 2.13). This relationship between povertyand lack of basic sanitation is illustrated evenmore clearly when looking across household con-

sumption quintiles, with about three times asmany households in the poorest quintile lackingbasic sanitation compared to the richest quintile.

Basic sanitation: rural and urban households

Access to basic sanitation is far more common inurban areas (65 percent of households) than ruralareas (20 percent of households) (Figure 2.14).



Dry Season & Rural

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor




Rainwater



Dry Season & Urban

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor




Rainwater



Wet Season & Urban

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor




Rainwater



Wet Season & Rural

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor




Rainwater



F I G U R E 2 . 9 Drinking Water Sources, Rural and Urban, Non-Poor and Poor, Dry and Wet Season, 2004



National

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor

No, never

Sometimes

Yes, always

Rural

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor

No, never

Sometimes

Yes, always

Urban

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor

No, never

Sometimes

Yes, always

Boiling by Household Consumption Quintile

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%


Always Sometimes

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Piped

water

Tube w

ell/b

ore h

ole

Prote

cted w

ell

Rainwat

er

Tanke

r tru

ck/ve

ndor

Unpro

tecte

d well

Pond,

rive

r, str

eam

Non-Poor Poor

National

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1

None

Open land

Public or shared toilet

Pit latrine or other w/outseptic tank

Septic tank

Connected to sewerage

F I G U R E 2 . 1 0 Boiling/Treating Drinking Water, Non-Poor and Poor, by National (top-left), HouseholdConsumption Quintile (top-right), Rural (bottom-left) and Urban (bottom-right), 2004


F I G U R E 2 . 1 1 Boiling or Otherwise Treating Drinking Water by Source, Non-Poor and Poor, 2004

F I G U R E 2 . 1 2 Access to Basic Sanitation, 2004



Connection to sewerage—30 percent of urbanhouseholds compared to less than 2 percent ofrural households—accounts for much of thedifference. A closer examination of the lack ofsanitation for rural households shows how theproblem is nationwide, with more than 80 per-

cent of rural households lacking sanitation in allbut five provinces.

Analyzing rural and urban households accord-ing to non-poor or poor status sheds further lighton relationships between poverty and sanitation.Only 7 percent of rural poor households have



F I G U R E 2 . 1 3 Sanitation by Non-Poor and Poor (left) and Household Consumption Quintile (right), 2004

F I G U R E 2 . 1 4 Sanitation by Rural and Urban (left); Lack of Sanitation for Rural Households by Province (right), 2004

Non-Poor & Poor

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Non-Poor Poor

None

Open land



Septic tank


Lack of Basic Sanitation by Household Consumption Quintile

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%




Rural & Urban

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Rural Urban

None

Open land



Septic tank


access to basic sanitation, compared to 26 per-cent of rural non-poor households (Figure 2.15).Greater use of septic tanks by the rural non-pooraccounts for much of this difference. In urbanareas, about 20 percent of poor households haveaccess to sanitation, compared to 73 percent ofnon-poor households. Here, the urban non-poorhouseholds are 10 times as likely as urban poorhouseholds to be connected to sewerage.

Factors associated with access tosanitation: regression results

Several variables show a statistically significantrelationship with access to basic sanitation. Thesevariables include households that are non-poor,located in urban areas, with access to all-weatherroads and in close proximity to district and provin-cial capitals. Access to sanitation is also associatedwith older heads of households who have hadsome schooling. Female-headed households aremore likely to lack access to sanitation.

Natural Disasters

Analysis of the relationships between poverty andhouseholds affected by natural disasters is based

on CSES 2004, which provides data on house-holds affected by floods and droughts between1999 to 2003. Over this five-year period, 53 per-cent of households report being affected by floodand 63 percent report being affected by droughtin at least one year. To isolate the householdsexperiencing more severe problems with thesenatural disasters, we identified households thathad been affected in three of more of the fiveyears, either by flood (23 percent of households)or drought (14 percent of households), for fur-ther analysis.

For these flood-affected households, the poor-est quintile was less often affected by floods thanricher households (Figure 2.16). While it mayseem counterintuitive that richer householdsappear located in more flood-prone areas thanpoor households, there may be two explanationsfor this finding. First, for rice farming purposesthere are both “good” floods (providing waterand replenishing soil nutrients) and “ bad” floods(large unmanageable amounts of water). It is notclear whether the survey (and respondents) madethis distinction. Second, the fact that richerhouseholds are more affected by floods may sim-ply reflect their ownership of higher quality riceland nearer to water sources.

For households affected by drought in threeor more of the five years, poorer householdsare more often affected than richer households(Figure 2.16). From 1999 to 2003, about 17 per-cent of households in the poorest quintile wereaffected by droughts in three or more of thefive years. This percentage declines as house-hold consumption increases, with only about13 percent of the richest quintile affected bydrought.

Controlling for other factors, both types ofnatural disasters show a statistically significantrelationship to poverty. For households affectedin three or more of the five years, drought is asso-ciated with poor households, whereas flood isassociated with non-poor households (Table 2.3above).



0%

20%

40%

60%

80%

100%

Urban &Non-Poor

Urban &Poor

Rural & Non-Poor

Rural & Poor

None

Open land



Septic tank


F I G U R E 2 . 1 5 Sanitation, Rural andUrban, Non-Poor andPoor, 2004


Mine and Unexploded Ordinance(UXO) Contamination

Analysis of the relationship between poverty andmine/UXO contamination is based on CSESpoverty data (2004) and the National Level 1 Sur-vey (2002) conducted in nearly 14,000 villages toidentify mine/UXO contamination. The NationalLevel 1 Survey distinguishes between villagesaffected by mines, cluster bombs (a class of UXOthat may affect villages in a similar way as mines),and spot UXO. Villages are more likely to expe-rience socio-economic impacts, such as reducedaccess to agricultural land, from mine and clus-ter bomb contamination.

Analysis of affected households shows astrong relationship between mine and cluster-bomb contamination and poverty (Figure 2.17).The poorest quintile is almost four times aslikely to be affected by mine contamination,and twice as likely to be affected by cluster bombcontamination, as the richest quintile. Like-wise, mine contamination is much more likelyto reduce access to agricultural land wherehouseholds are poor. Indeed, controlling forother factors, the presence of mine contami-

nation limiting access to agricultural land shows a statistically significant relationship with beingpoor.

CASE STUDY FINDINGS: ACCESS TO NATURAL RESOURCESAND POVERTY6

The PEN II case study was undertaken to inves-tigate relationships between access to naturalresources and poverty (Box 2.1: Case StudyMethodology). Household survey findings indi-cate that across the six villages studied the per-centage of poor households range from 30 to60 percent. More than half of poor householdsexperience significant annual rice deficits—theymust buy rice for more than half of the year tosupport consumption. The average amount ofagricultural land for poor households is only1.1 ha compared to 2.4 ha for non-poor house-holds. Overall, the poor are more likely to live ina thatch (rather than wooden) house, own fewerand lower-value livestock, and have less involve-ment in petty trade and small business activity.While the proportion of poor and non-poorhouseholds hiring out their labor is similar, a



Affected by Flood in 3 or more of past 5 years

0%

5%

10%

15%

20%

25%

30%


Affected by Drought in 3 or more of past 5 years

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%


F I G U R E 2 . 1 6 Housholds Affected by Flood (left) or Drought (right) in Three or More of the Five Yearsfrom 1999 to 2003, by Household Consumption Quintile


higher percentage of the poor hire out labor toclear degraded forests, while non-poor tend to bemore involved in timber extraction.

Nearly all rural households across the studyareas, whether poor or better-off, depend on nat-ural resources for their livelihoods. In most cases,households are cultivating rice (83 percent), rais-ing livestock (68 percent), collecting forest prod-ucts (98 percent), and fishing (75 percent). Theproportion of households engaged in these liveli-hood activities is quite similar to rural householdsacross Cambodia (CSES 2004), though some-

what more households in the study areas col-lect forest products and somewhat fewer raiselivestock.

At the commune and village level, officialsreport a number of changes in natural resourcesover the past decade. First, significant forestclearing for rice and other agricultural crops hasoccurred, consistent with national trends. At thesame time, progress in improving agriculturalproductivity has been mixed, with some areasreporting increases, some decreases, and someno change. Across the study areas, households



Mine Contamination - National

0%

2%

4%

6%

8%

10%

12%

14%


Mine and/or Cluster Bomb Contamination - National

0%

5%

10%

15%

20%

25%

30%

35%


Mine Contamination Severely Affecting Access to Agricultural Land

0%

1%

2%

3%

4%

5%

6%

7%


Cluster Bomb Contamination - National

0%

5%

10%

15%

20%

25%


Spot UXO Contamination - National

0%

5%

10%15%

20%25%

30%35%

40%45%

50%


F I G U R E 2 . 1 7 Housholds Affected by Mines (top left), Cluster Bombs (top right), Mines/Cluster Bombs(middle left), Spot UXO (middle right), and Mines Limiting Access to Agricultural Land(bottom left), by Household Consumption Quintile

Source: Estimated from CSES 2004 and National Level 1 Survey 2002.



B O X 2 . 1 Case Study Methodology

In the design of PEN phase II for Cambodia, interviews were conducted with a range of key stake-holders to identify poverty-environment issues of national priority. This included consultations witha number of government partners, including the Ministry of Environment, Ministry of Economy andFinance, and Ministry of Planning. These interviews revealed several recent and ongoing studiesaddressing critical issues associated with poverty and natural disasters, land, water, and health andsanitation. In addition, a number of studies were focused on poverty issues in floodplain areasaround the Tonle Sap Lake and in the Mekong Delta region (Prom and Ballard 2005; So and Ballard2005). However, one issue of increasing concern for Cambodia—the relationship between access tonatural resources and poverty—was not the explicit focus of any of these studies.

To address this concern, PEN II examines access and poverty issues for a variety of naturalresources, including agricultural land, forest products, grazing land, fisheries, and drinking water. As the study sought to avoid floodplain areas (in order to avoid repeating the work of other studies),fisheries concerns are not covered in depth, whereas forest and land issues receive more comprehen-sive coverage. For the purposes of this study, access to natural resources is defined as householdshaving the ability to collect/obtain the specified resource or product. In practice, it was found that allhouseholds with access to resources had also made use of this access in the past year.1 Lack of accessmay be attributed to a lack of means (transportation, resource extraction equipment), physical barri-ers such as poor road conditions, and/or legal and institutional barriers such as restrictions on accessto land or specified areas/resources.The case study is guided by the following research questions:

• To what extent do the poor differ from the non-poor in terms of current access to naturalresources and environment in Cambodia?

• What trends in access to natural resources and environment can be observed, and do thesetrends differ between the poor and the non-poor?

• What are the environmental impacts from different types of natural resources and environmentutilization by the poor and non-poor? What are the key determinants of the environmentalimpacts? To what extent can these environmental impacts be economically evaluated?

• What are the specific and practical policy options leading to poverty reduction and sustainableutilization of natural resources and environment?

PEN II case studies are intended to deepen current knowledge about poverty-environment relation-ships and to support broader investigation of potential poverty-environment linkages in nationaldata. In Cambodia, two provinces were selected for study of natural resource access issues—Kampong Speu and Kampong Thom. These were chosen because each fulfills the criteria of havinga “protected area,” diverse types of concessions, ongoing in-migration, and a substantial share ofpoor districts/communes. A total of six villages (three in Aoral district of Kampong Speu provinceand three in Santuk district of Kampong Thom province) were then selected for more focused studybased on their proximity to protected areas, proximity to roads, and population densities.

Data were collected through (a) semi-structured interviews with officials at the provincial, district,commune, and village levels; (b) a household survey of 120 households (20 households randomlyselected in each of the six villages); and (c) focus group discussions to capture additional informationon the potential effects of wealth differences and gender. Data analysis required classifying house-holds into “poor” and “non-poor” categories. This study relied on methods developed and employedby GTZ in its rural development programs (including in the villages in Kampong Thom). Key criteriain establishing poor and non-poor status included amount of residential land, amount of productiveland, sources of family income, livestock, housing, means of transport, electric appliances, andperiod of time in which the household lacks food within a year.

It should be noted that problems were encountered in applying these selection criteria. First,proximity to a protected area was revealed to be unimportant, as parks tend to lack enforcement.Proximity to (quality) forest resources is a more meaningful variable to consider. Second, proximityto roads was not as meaningful as proximity to markets. For example, in Kampong Thom, whilethe three selected villages have differing levels of road access, all are less than one hour from amarketplace. Likewise, the three selected villages in Kampong Speu have differing road access, butall are very far from a marketplace. Finally, although provinces were selected according to their

(continued )

and officials report that availability of naturalresources (especially forest and fisheries resources)has decreased substantially over the past decade,with the sharpest declines occurring in the pastfive years. Officials blame the declines on in-creased demand, overexploitation, and uncon-trolled activity.

Access to Agricultural Land

Gaining access and claim to land in the studyareas involves little more than clearing it. As aresult, in some villages there has been significantin-migration of landseekers (including landless

and better-off “newcomers”). Newcomers oftenhire local people to clear land, usually for agri-culture but also for speculative purposes. Insome areas, there is plenty of access to land avail-able, but the land is unproductive. For instance,migrants were drawn to one village by the promiseof three ha of land. While this was provided,they have found yields to be very low due to poorsoils and a lack of rain. Most are now trying tomake a living in charcoal production.

Approximately 25 percent of the householdssurveyed report being landless. The landlesshouseholds are almost all poor, and most arehighly dependent on access to natural resources



B O X 2 . 1 Case Study Methodology (Continued )

low and high population density, at the village level it was found that all potential rural villageshad a population density well below the national average.

A number of limitations of the PEN II study require noting here. First, the sample size (20 house-holds per village for six villages) for the case study is small; the intention was for the survey toprovide on-the-ground insights regarding the poverty and environment nexus, but not necessar-ily to provide data for analysis of statistical significance. Second, as noted above, the classification of“poor” households is not based on a robust poverty line. This may lead to a small margin of error inthe classification of “poor” and “non-poor” households, and in the comparison of case study findingsto national data. Third, at the time of writing, PEN II had only limited access to data and findingsfrom the CSES, which suggests potential scope for identifying additional relationships between casestudy findings and the national context and trends beyond those discussed in this chapter.

Finally, analysis of spatial associations between poverty and environment was hampered becausemost credible data (such as CSES) are only presently available at the provincial level, with some dataalso available at the commune level (for example, Seila/Ministry of Planning commune database2003). For analysis in Cambodia, Vietnam, and Laos, PEN II has targeted poverty-environment spatialanalysis at the district level. While this may be an appropriate administrative level for analysis inVietnam and Laos, it has a number of shortcomings in Cambodia. First, Cambodia has only had elec-tions at the national and commune level.1 As a result, most policy (and donor) support targetinglocal issues focuses on communes and their elected Commune Council. Policy development is lessdriven at the district and provincial levels because they are administered by unelected officials.

Second, from a geographical standpoint, aggregating data at the district level has not beenfound to be highly relevant because most rural districts include semi-urban areas (district markets).This mix of rural and urban areas makes district-level data less meaningful for study of poverty thanmore disaggregated units. Moreover, districts may include floodplain and upland areas,2 and thusvillages dependent on different types of natural resources (e.g., fisheries, forestry) for their liveli-hoods, all of which complicate assessments of poverty-environment relationships at the districtlevel. To achieve more meaningful results, with greater opportunities for policy mainstreaming atthe local level, future analysis of poverty-environment issues in Cambodia would be better targetedat the commune level.

1When the term “access” is translated into Khmer, it implies the household has the internal ability to obtain aresource. Households with the ability to obtain resources also have experience doing so, given that these resourcesare relied on and often constitute a major part of livelihoods.2Cambodia’s governing levels include national, provincial, district, commune, and village.3For instance, many districts surrounding the Tonle Sap Lake also include forested areas far from the lake.

(especially forest products) for their livelihoods.Many of these households express concern thatincreasing overexploitation of resources maymake it difficult for them to sustain their liveli-hoods in the future. In addition to landlessness,there are indications of increasing concentrationof land ownership. For example, the averagenon-poor landholding household in Santuk owns2.77 ha, compared to 0.73 ha for poor land-holding households.

Access to Forest Resources

Availability of forest products has decreasedmarkedly over the past several years in all studyareas. Timber, resins (which are produced bylarge trees), and wild animals are among theproducts that have declined most dramatically.Wood for charcoal production is also on a rapiddeclining trend. Decreasing at a somewhat slowerrate are wild vegetables, vinery, and other minorproducts. As low-value products, these resourcesare likely under less extractive pressure than tim-ber and other wood resources.

Access to forest resources in the study areas islimited by legal restrictions, taxes, and ForestAdministration (FA) control of forest concessions(in Santuk district, where three forest concessionsoperated until recent suspension of activities).Households in Santuk district recall that whenconcessionaires tightened forest boundaries andplaced armed guards to restrict access severalyears ago, they became poorer. The hardship con-tinued during the years of concession operation,and has become worse as the forest resources havecome under FA management. The FA not onlyrestricts forest access but also either charges fees tovillagers collecting forest products, or confiscatesthe products altogether. The restrictive regimeappears to be selectively enforced, with more non-poor households gaining access to timber (46 per-cent) than poor households (11 percent).

At present people have the right to access the foreststo collect forest by-products. However, the author-ities (forest wardens, soldiers, and police) wait to

very unjustly and cruelly take a share from theproducts collected by the innocent people. Forinstance, when people transport the products theyworked very hard to collect to the market or home,these people charge very high informal fees frompeople without shame or confiscate their transportmeans, making people lose their last resort of liveli-hood. The community can only educate people tolove nature, but cannot protect ourselves from suchabuses. (A village chief in Santuk district)

Given the restrictions, some households havefound creative ways in which to collect timberwhile avoiding fees and confiscation. For instance,villagers often cut up large logs (worth consider-able sums) into low-value firewood to avoidrestrictions. Likewise, villagers may cut timber forhouse construction and wait for a year or more forthe wood to become “old” looking before trans-porting it. This way they appear to be movingtheir house, rather than construction materi-als, and can avoid paying officials on the road.Such delays and inefficiencies caused by the cur-rent enforcement regime represent a sub-optimaluse of forest resources, but villagers see no otheroptions, given the current restrictions.

In Aoral district, many households are involvedin charcoal production. Although access to for-est resources is open, a fee of R10,000 ($2.50)must be paid to authorities for each unit of char-coal produced (1 meter by 1 meter). Charcoalproducers consider this to be a high fee, as it isequal to their labor costs per unit. Authoritiesenforce against timber extraction selectively, stop-ping villagers from logging, but also allowingbands of loggers to operate as long as they payfees twice per month. Households in the studyareas complain about what they perceive to beunjust restrictions and selective enforcement.They believe the presence of enforcement author-ities, or designation of protected areas, has littleeffect if outside logging interests are involved.

“The whole village is in protected area but log-ging is happening every day.” (A village chief inAoral district)

As charcoal production has been ongoing for sev-eral years in Aoral district, the forest surrounding



the study villages has all been cut. The forest isnow 6–10 km away, with the frontier movingfarther away each year. Early in the morning,these villagers depart on koyun7 to go to the for-est to collect firewood for charcoal and timberfor fences. Most villagers use axes; chainsaws areonly owned by powerful people in the area. In thecurrent forest area, villagers are in the process ofmaking a “second cut” (the best wood has alreadybeen collected). While villagers express concernabout the future, they also say they will make athird cut (to extract roots) if they have no otheroption.

Access to Grazing Land

Access to grazing land is important to rural liveli-hoods because raising cattle and buffalo providesdraft animals for agriculture as well as supple-mental income if sold. On average, 65 percent ofthe households in the study villages raise cows orbuffaloes, with poor households owning feweranimals than non-poor households. Likewise, theaverage value of animals owned by poor house-holds is only half that of non-poor households.

Where grazing areas are available, access isgenerally open to all. Most households report adecrease in grazing areas compared to 10 yearsago due to conversion of these areas into crop-land. However, some households (usually non-poor) no longer require grazing land, as they haveopted for machinery (koyun) to replace draft ani-mals, so they have less need to raise cattle andbuffalo.

Access to Fisheries

Cambodia’s fisheries play a major role in rurallivelihoods, supporting consumption and pro-viding the chief source of protein in the ruraldiet. Even though the six surveyed villages areprimarily forest-dependent, 75 percent of house-holds benefit from catching a small amount offish, frogs, and other aquatic resources. Most vil-

lagers fish during the wet season and early dryseason, when there is enough water in these areasto support more abundant aquatic life. Catchesare primarily for household consumption.

There are no restrictions on access to fishingareas for the study villages. Although roughly thesame proportion of poor and non-poor house-holds fish, non-poor households catch twice asmuch fish as poor households. This occurs becausethe non-poor can afford better fishing equipment.Households report a decline in their overall catch,especially over the past five years. They attributethis decline to overexploitation, including theillegal use of destructive fishing equipment andmethods.

Access to Drinking Water

Access to drinking water is similar for poor andnon-poor households in the study areas. How-ever, all villages complain about the quality ofwater and the difficulty of collecting it. Drinkingwater in Santuk district is mainly drawn fromwells, but many households note that the watersmells and tastes bad. They believe there is limein the well water, and therefore they prefer othersources when available. Households in Aoral dis-trict use a variety of sources, including rain water,streams, and hand pumps. Although householdsunderstand that water from streams is generallynot clean, they often use it. Where possible, house-holds in both study areas use water from streamsfor irrigating vegetables, bathing, washing clothes,and feeding animals.

Less than one-quarter of households boilwater before drinking. Villagers note that it isquicker and easier to drink un-boiled water.Most households report that they do not washtheir hands before having meals and there is alack of other hygienic practices. However, thereis a degree of awareness among most householdsthat hygienic practices and drinking boiled waterhelp reduce health problems and illness.



Summary of Case Study Findings

Restrictions on timber access appear to benefit non-poor households; no restrictions were identified forother natural resources. Although this case studydid not find restrictions on access to most naturalresources, access to timber was restricted for for-est areas currently managed by the Forest Admin-istration in Kampong Thom. This restrictionappears to be selectively enforced, with morenon-poor households gaining access to timber(46 percent) than poor households (11 percent).In Aoral, charcoal producers pay an informal taxto authorities on each unit of charcoal produced;in effect, the tax doubles their cost of production.For other natural resources studied (agriculturalland, grazing areas, fisheries, and drinking water),no notable differences in the proportion of poorand non-poor households with access were iden-tified. This appears to be the case because nearlyall households in the six villages studied, whethervery poor or better off, depend on access to theseresources.

Non-poor households make more effective use ofopen access. The study did find differences in howeffectively (and profitably) non-poor householdsmake use of their access to natural resourcescompared to poor households. This appears tooccur for two reasons. First, non-poor house-holds tend to have better means of transporta-tion and equipment for exploitation of forestand fisheries resources. This allows them to collect more, over greater distances. Second,because the non-poor are able to collect prod-ucts more efficiently, they are better able tocover the fixed costs associated with travel toresource-rich areas, pay fees to officials, andstill make modest earnings. Poorer householdsmay require longer travel times, collect fewerproducts for their labor output, and have dif-ficulty covering fee charges. These differencesin the productive use of access tend to com-pound over time, as non-poor households usesome of their earnings to invest in better trans-

portation and equipment for extraction, sup-porting even more extraction, while the poorcontinue at a level of subsistence.

Poverty reduction associated with opening ofaccess tends to be short-lived, likely to be followed byincreasing poverty over the long term. Althoughboth non-poor and poor households may initiallybenefit from the opening of access to naturalresources (for example, cancellation of forest con-cessions or fishing lots), this poverty reduction islikely to be short-lived. Opening access—with noclear management structure to fill the vacuum—results in a period of overexploitation, ofteninvolving destructive practices, that will likely befollowed by increasing poverty over the long termas the resource base declines (Figure 2.18).

For example, rapid exploitation of charcoalresources by villages in Aoral district appears tohave led to short-term poverty reduction. Manyhouseholds used initial profits from charcoal tomake home improvements and invest in koyuntransportation (to support farming and the trans-port of charcoal). It is only more recently that, asthe forest frontier has moved farther away, char-coal is becoming less profitable and charcoal pro-ducers are worried about the future. Due to poorfarming conditions (drought and low soil fertil-ity), they do not see options other than continu-ing charcoal production. Thus, it can be expectedthat poverty will increase as resources becomescarcer and farther away, unless other opportu-nities arise for the community members to maketheir living.

Although Figure 2.18 reflects this case study’sfinding for villages in forest areas, the same appearstrue for fishing villages in Cambodia. Non-poorfishing households tend to have better boats andgear, providing them a greater ability to benefitfrom open access fisheries. As noted in the interimfindings of the Tonle Sap Participatory PovertyAssessment:

“[Fishing] villagers defined three different cat-egories of well-being: medium, poor, and des-titute. Medium households are differentiated



from poor and destitute households based onownership of different fishing assets, both interms of quantity and quality. The mediumhouseholds tend to have larger boats with largerengines, which enables them to go further outin the lake to fish. These same households alsotend to have much better quality fishing gearthan do the poor/destitute households” (Promand Ballard 2005).

Like the overexploitation of open access to for-est resources, open-access fishing leads to a rapiddecline of fisheries resources. A study of fishinglots released in 2001 for open access fishingnotes: “Initial positive impacts from improvedaccess to fisheries are now perceived to be declin-ing due to unregulated competition for resourcesand widespread use of illegal fishing gear” (Com-munity Fisheries Department Office 2004).

Strict controls on previously accessed resourcescan increase poverty and cause sub-optimal use ofresources. This impact has been noted in numer-

ous rural livelihood studies (Oxfam GB 2000;Chan and Acharya 2002; McKenney and Prom2002; IFSR 2004; World Bank 2005b; www.ngoforum.org.kh). This case study found thatstricter controls on forest product collectionimposed by concessionaires, and then the ForestryAdministration (FA), have caused considerablehardship to local communities in Santuk dis-trict. These measures reduce the income of localhouseholds, to the benefit of concessionaires, FA,and a few non-poor households (who still havethe means to harvest timber and simply bribeofficials and guards when they are caught). More-over, selective enforcement of restrictions (forexample, outsiders are allowed access whilelocals are not) likely intensifies the scramble forresources, as local people act to benefit fromrapidly declining resources in any manner theycan. For instance, in Santuk district, logging byoutsiders continues despite restrictions affectinglocal households. At the same time, locals have



Increase income

Subsist

Other options or migration

Resource depletion

Invest in increased resource extraction

Non-Poor

Poor Poverty

Time

F I G U R E 2 . 1 8 Poor vs. Non-Poor—Open Access and Natural Resource Dependent Livelihoods

Poor: subsistence or some short-term gainsNon-poor: larger short-term gains

Poor: continuing poverty, return to poverty, or alternativesNon-poor: alternatives or possible move into poverty

identified a number of informal ways of benefit-ing from timber resources as well, but in a highlyinefficient manner.

The abundance and quality of natural resourcesin the areas surrounding the six villages is in significant decline, suggesting the potential forincreasing poverty in the future. This is especiallythe case for forest and fisheries resources. Vil-lagers indicate that the decline has worsened inrecent years. While resource decline has nega-tively affected both poor and non-poor house-holds, it was not possible to distinguish if onegroup has been disproportionately impacted.Compared to poor households, non-poor house-holds tend to collect more resources, often ofhigher value. As a result, a decline in resourceswould likely reduce the total income of non-poor households by a greater amount than poorhouseholds. But this may not move these house-holds into poverty, as they may still earn an ade-quate income from natural resources or be ableto redirect their labor to alternative income-gen-erating activities. Poor households earn less fromnatural resources than non-poor households, butthese earnings are often a substantial proportionof their total income. As a result, any reductionin earnings from natural resources can causesevere hardship, especially since poor householdsoften lack alternative income-generating options.

SUMMARY AND POLICY IMPLICATIONS

Natural resource dependence and poverty

1. The poor are disproportionately dependenton natural resources for their livelihoods.And this holds true in areas reportingdeclines in the resource base. Problemsof poverty and resource decline appear to be most pronounced in the plateau/mountain region. In analyzing householdsthat are both engaged in natural resource-dependent activities and located in com-munes reporting resource declines, we findtwice as many households in the poorestquintile depend on these activities com-pared to the richest quintile. This suggeststhat if these resource bases continue to dete-riorate, the poor will bear a disproportion-ate burden.

Policy implication: Targeting support oflocal resource management, land use plan-ning, and agricultural and off-farm assis-tance should consider communes reportingdeclines in forest cover, fisheries, and landaccess, and concerns about future resource-based livelihoods, as well as poverty levelsand other indicators (such as soil quality,water resources).



B O X 2 . 2 Potential Policy Links: Ongoing Initiatives on Natural Resources

• Policies to promote institutionalizing land and natural resource management at the communelevel (e.g., Seila Program and Commune, Community Based Natural Resource and EnvironmentManagement (CCB-NREM)).

• Initiatives aimed at improved productivity and diversification of agriculture, off-farm opportuni-ties, and land, forest, and fisheries reform under the Rectangular Strategy and the National Strate-gic Development Plan (NSDP). The NSDP consolidates a number of past planning processes anddocuments, including the National Poverty Reduction Strategy (World Bank), Socio-EconomicDevelopment Plan (Asian Development Bank), and the UN Millennium Development Goals.

• Reform efforts aimed at improving land, forestry, and fisheries laws; regulations; and enforcementmechanisms. Of particular importance are policy interventions for forest and fisheries concessions(especially addressing possible reform/termination, monitoring, and enforcement) and landdistribution mechanisms (for example, social concessions).

• Programs using an array of criteria for identifying target areas and communes (Seila, CCB-NREM,and various other rural development and conservation NGOs).

2. Natural resources provide an importantsafety net for the poor, but not often a path-way out of poverty. The poor are dispropor-tionately dependent on natural resources forincome/consumption, and these resources playa critical role in helping households cope withcrises, such as floods and droughts. Despite theimportant role of natural resources in rurallivelihoods, income generation activities (fish-ing and forest product collection) appear toplay little or no role in moving householdsout of poverty. Indeed, findings from our casestudy in Kampong Speu on charcoal produc-tion indicate households often use initial earn-ings from resource extraction to invest in moreefficient extraction (equipment, transporta-tion), which in turn can lead to an accelerationin the decline of the resource base. Both non-poor and poor households may benefit fromthis period of rapid extraction, though non-poor households tend to be better positionedto do so. As resources are depleted, however,many households may return to poverty, espe-cially where other livelihood options are lack-ing, as is the case in Kampong Speu. The factthat natural resource-dependent activities playlittle role in moving households out of povertyis unlikely to change, as income from theseactivities is falling as a proportion of house-hold income, and this is occurring across allhouseholds, rich and poor. The fall in incomereflects an overall decline in the resourcebase, and in some areas increasing controlover resources by private concessions andpowerful interests (as illustrated by our casestudy site in Kampong Thom).

Policy implication: Natural resource man-agement should be focused on developinglocal management regimes suitable for main-taining resources and providing access for thepoor/vulnerable, not aimed at industrial-level extraction, which has a history of dismalresults for the environment, local communi-ties, and government revenue. In establishing

local management regimes, consider target-ing interventions where rapid resource extrac-tion is either beginning, and/or the resourcebase is not yet highly degraded, and encour-age a combination of sustainable resourcemanagement and investment of extractionearnings in agricultural and off-farm oppor-tunities (rather than even more extraction).

3. Natural resource management at the “ex-tremes” is not pro-poor; both highly restric-tive regimes and open access managementappear to result in negative outcomes forthe poor compared to the non-poor. Clearly,restricting the rural population’s access to nat-ural resources (as has been the case for manyforest, fishing, and economic/land concessionareas) takes away productive resources thatsupport rural livelihoods. While this can makethe surrounding rural population worse off,the impact can be especially severe for food-poor households, which tend to be moredependent on these resources.

But the converse to restrictive regimes—“open access” (or a management vacuum)—may not be much of an improvement for thepoor over the long run. First, the poor tend tobe less able to take advantage of open access toexploit resources as profitably as non-pooractors, because they lack the capital means(equipment, transportation). Second, “openaccess” generally is not fully open, but ratherinvolves paying a range of informal fees foraccess—a disproportionate burden for thepoor. Lastly, the poor tend to be more depen-dent on natural resources and have fewer alter-native options for income generation. As aresult, when open access exploitation leads toresource decline, the impact on livelihoods isagain felt disproportionately by the poor.

Policy implication: The managementvacuum created by past cancellations of for-est concessions and fishing lots needs to beaddressed with greater support for naturalresource assessments, setting priorities for



management (for example, rural poor, conser-vation), and promoting locally appropriatemanagement regimes. Remaining concessionsin Cambodia still claim a large proportion ofthe richest resource areas. These (selectively)restrictive management regimes should bereformed to improve appropriate access forlocal communities (with ongoing indepen-dent monitoring), or the concessions shouldface termination. It is critical, however, thatreforms and terminations do not lead to man-agement vacuums (as has been the case in thepast), which invite ad hoc control by powerfulinterests. There must be support for resourceassessments, consultation and priority-setting,and establishment of fair and enforceablemanagement.

Drinking water wources and poverty

4. The poor are disproportionately depen-dent on the use of unsafe water sources.Our analysis indicates households in thepoorest consumption quintile are threetimes as likely to use unsafe water sources(unprotected dug wells and ponds, rivers,and streams) as households in the richestquintile.

Policy implication: In targeting serviceprovision for drinking water, consider factorsidentified in this analysis as associated (at sta-tistically significant levels) with the use ofunsafe drinking water sources. These includehouseholds located in the coastal or plateau/mountain regions, as well as rural areas farfrom district and provincial capitals and all-weather roads. Additional factors includehouseholds that are poor and have heads ofhouseholds with no schooling.

5. Households accessing unsafe water sourcesare the least likely to be boiling their water.This is especially true for poor households,where boiling is done by less than 25 percentof households using water from unprotectedwells and only about 40 percent of house-

holds using water from ponds, rivers, andstreams. However, only a modest proportionof non-poor households using these unsafesources boil their water—40 percent usingwater from unprotected wells and 57 per-cent using water from ponds, rivers, andstreams.

Policy implication: Consider education andawareness programs to encourage the boilingof drinking water, especially for householdsusing unsafe sources. Although poor house-holds using unsafe sources are more at riskdue to their lower levels of boiling, even aconsiderable proportion of non-poor users ofunsafe sources do not boil. This suggestspoverty itself is not a sufficient explanationfor the lack of boiling, and there may be apotential for significantly reducing drinkingwater-related health impacts through greatereducation and awareness programs targetingthis segment of the population.

Sanitation and poverty

6. Access to basic sanitation is lacking acrossall rural areas. In urban areas, 65 percent ofhouseholds in urban areas have access to basicsanitation, with 30 percent connected to sew-erage. But such sanitation infrastructure/services are dearly lacking in rural areas, whereonly 20 percent of households have access tosanitation and less than two percent are con-nected to sewerage. The problem is even worsefor the rural poor—only seven percent haveaccess to basic sanitation and less than 0.5%are connected to sewerage.

Policy implication: With so much of therural (poor) population lacking access tobasic sanitation, targeting based on povertyis unlikely to be meaningful. To makeprogress at such a broad scale, improvingbasic sanitation in rural areas needs to beelevated as a national policy priority, withappropriate technical and resource supportfor action.



Natural disasters and poverty

7. Both droughts and floods show a statisti-cally significant relationship to poverty.But whereas droughts are associated withpoverty, floods are associated with notbeing poor. Droughts and floods are a com-mon problem for Cambodian households.For the five-year period from 1999 to 2003,63 percent of households report beingaffected by a drought and 53 percent reportbeing affected by a flood in at least one year.Whereas poorer households tend to be moreoften affected by drought than richer house-holds, the reverse is true for householdsaffected by floods. This may reflect the chal-lenges of defining “flood” (which dependingon its nature may be viewed as good or badfor rice farming), as well as the likelihoodthat richer households own higher qualityrice land (in flood-prone areas).

Policy implication: In targeting problemsaffecting the poor, developing appropriatedisaster responses to drought problems shouldbe elevated to a higher priority. This is not todiminish the importance of addressing floods,but only to recognize that droughts dispro-portionately affect the poor and thereforedeserve due attention in poverty reductionstrategies.

Mine/UXO Contamination and Poverty

8. Mine and cluster bomb contaminationshow a strong relationship with poverty.The poorest household consumption quin-tile is almost four times as likely to be affectedby mine contamination, and twice as likely tobe affected by cluster bomb contamination, asthe richest quintile.

Policy implication: In targeting areas fordemining, village poverty rates (as identified inthe CSES 2004 data) should be considered asan additional element of the targeting criteriacurrently used. Where villages in contaminated

areas have higher poverty rates, demining theseareas should be given higher priority.

Endnotes1. The CSES is the largest and most extensive multi-objec-

tive survey undertaken in Cambodia, involving a sampleof 15,000 households. Cambodia’s baseline poverty lineconsists of a single national food poverty line and threeregional nonfood allowances. Poverty estimates are basedon per capita consumption.

2. A sharp and statistically significant increase in the Ginicoefficient occurred in rural areas—from 0.27 in 1993–94 to 0.33 in 2004.

3. The study assesses how poverty has changed for 1,000 households in nine villages between 2001 and2004. The villages were selected to represent Cambo-dia’s different eco-regions. To capture mobility, thestudy groups households into seven categories: (1) com-fortably rich, (2) climbing into wealth, (3) escapingpoverty, (4) static middle, (5) chronic poor, (6) deepen-ing poverty, and (7) falling into poverty.

4. Key datasets include CSES 2004, Seila and Danida2005, and National Level 1 Survey 2002.

5. Rural communes account for nearly 1,500 out of justover 1,600 communes in Cambodia.

6. This section summarizes key case study findings; moredetailed case study findings, data, tables, and informa-tion can be found in the Annex.

7. “Koyun” is a Khmer word referring to a machineintended to be used for pulling a plow, but it is also com-monly used for transportation. To a degree, it is a mech-anized substitute for draft animals and an oxcart.

ReferencesAndersen, Henny. 2003. “Poverty and Natural Resources

and the Environment in Cambodia.” Phnom Penh:Royal Danish Embassy (Danida).

Asian Development Bank. 2001. Participatory PovertyAssessment in Cambodia. Manila: ADB.

Cambodia Development Resource Institute. 2006. “MovingOut of Poverty Study: Cambodia National SynthesisReport.” Phnom Penh: CDRI.

Chan, Sophal, and Sarthi Acharya. 2002. Facing the Challengeof Rural Livelihoods: A Perspective from Nine Villages inCambodia. Working Paper 25. Phnom Penh: CambodiaDevelopment Resource Institute (CDRI).

Chan, Sophal, Tep Saravy, and Sarthi Acharya. 2001. LandTenure in Cambodia: A Data Update. Working Paper 19.Phnom Penh: CDRI.

Community Fisheries Department Office. 2004. “Impactsof Fisheries Policy Reforms: Kampong Cham, Pursat



and Takeo Provinces.” Phnom Penh: DFID andIMM, Ltd.

Conway, Tim. 2005. “Identification and Targeting of thePoor—Basic Principles.” Presentation to the NationalForum on Identification of Poor Households, PhnomPenh, February 15.

Council for Social Development. 2005. “Report on Pro-ceedings and Results of the National Forum on Identi-fication of Poor Households.” Draft (March 4, 2005).Phnom Penh: Ministry of Planning.

Dasgupta, Susmita, Uwe Deichmann, Craig Meisner, andDavid Wheeler. 2002. The Poverty-Environment Nexusin Cambodia, Lao PDR, and Vietnam. Washington, DC:World Bank.

Degen, P. F., N. Van Acker, N. van Zalinge, N. Thuok, andD. Loeung. 2000. “Taken for granted: Conflicts overCambodia’s freshwater fish resources.” Paper presentedat 8th IASCP Conference, Bloomington, Indiana, May31 to June 4.

Dümmer, Ignas. 2005. “Information and Knowledge BasedSystem for Land Suitability Analysis in Cambodia.” Pre-sented at the Land Resource Assessment Forum, PhnomPenh, September 14–17.

Fraser, Thomas, GFA-AGRAR, and ANZDEC. 2000.“Cambodia Forest Concessionaire Review Report.”Phnom Penh: Sustainable Forest Management Project,Asian Development Bank.

Global Witness. www.globalwitness.org.Independent Forest Sector Review. 2004. The Forestry

Sector in Cambodia. Phnom Penh: Joint CoordinatingCommittee—Donor Working Group on NaturalResources and the Forestry Administration.

Knowles, James C. 2005. A New Set of Poverty Estimatesfor Cambodia, 1993/4 to 2004. Report to the EASCountry Units of the World Bank. Washington, DC:World Bank.

McKenney, Bruce. 2002. “Questioning Sustainable Con-cession Forestry in Cambodia.” Cambodia DevelopmentReview 6 (1): 1–7.

McKenney, Bruce, and Prom Tola. 2002. Natural Resourcesand Rural Livelihoods in Cambodia: A Baseline Assess-ment. Working Paper 23. Phnom Penh: CDRI.

McKenney, Bruce, Yim Chea, Prom Tola, and Tom Evans.2004. Focusing on Cambodia’s High Value Forests: Liveli-hoods and Management. Phnom Penh: CDRI andWildilfe Conservation Society.

Ministry of Planning. 2005. National Strategic DevelopmentPlan for Poverty Reduction 2006–2010: Process for Prepa-ration. Phnom Penh: Ministry of Planning.

Murshid, K. A. S. 1998. Food Security in an Asian Transi-tional Economy: The Cambodian Experience. WorkingPaper 6. Phnom Penh: CDRI.

NGO forum. www.ngoforum.org.kh.OXFAM Great Britain. 2000. The Cambodia Land Study

Project (Vols. 1–3, 1999; Vol. 4, 2000). Phnom Penh:Oxfam Great Britain.

Prom, Tola. and Brett M. Ballard. 2005. “Interim Report 1for the Tonle Sap Participatory Poverty AssessmentProject.” Phnom Penh: CDRI.

Royal Government of Cambodia. 2004. “The RectangularStrategy for Growth, Employment, Equity and Effi-ciency in Cambodia.” Paper presented by Prime Minis-ter Samdech HUN SEN and the Third Legislature ofthe National Assembly at the Office of the Council ofMinisters, Phnom Penh, July 16.

Seila and Danida. 2005. A National Opinion Poll: Communecouncil’s perception of its natural resource base and liveli-hood options. Phnom Penh: Danida.

World Bank. 2004. Cambodia at the Crossroads: StrengtheningAccountability to Reduce Poverty. Report No. 30636-KH.East Asia and the Pacific Region: World Bank.

World Bank. 2005a. From peace to prosperity: an assessmentof poverty in Cambodia. Phnom Penh: World Bank.

World Bank. 2005b. Cambodia Rural Sector Strategy Note:Towards a Strategy for Rural Growth and Poverty Reduc-tion. Phnom Penh: World Bank.

Zohir, Sajjad. 2005. Report on CMDGR 2005 Update Exer-cise: Undertaken during 25 May–15 June 2005. PhnomPenh: UNDP.




GEOGRAPHIC DIMENSIONS OF POVERTY

Over the last decade, Lao PDR’s economicgrowth averaged nearly 6 percent per year andwas accompanied by significant poverty reduc-tion. Poverty incidence declined from 46 percentin 1993 to 33 percent in 2003.1 At the same time,there was a substantial narrowing in poverty inci-dence among provinces and the highest provin-cial poverty incidence came down from over70 percent in 1993 and 1998 to less than 55 per-cent in 2003 (Map 3.1).

Lao PDR is still, however, one of the poorercountries in Asia, with a GDP per capita of $425in 2004 (World Bank, 2005). Substantial urban-rural divides remain, and there are large geo-graphic variations in poverty and development.More than 85 percent of the poor live in rural areaswith rural poverty incidence at 38 percent in 2003,in contrast to 20 percent in urban areas (LECS III).

Poverty, as discussed above, is defined as con-sumption or expenditure below a minimal level.This provides a quantifiable measure useful formonitoring progress over time and for identifyinggeographic areas with persistent poverty in needof focused attention and interventions. Povertycan also be defined in broader perspectives. TheGovernment of Lao PDR defines poverty as thelack of basic necessities—such as inadequate food

and clothing, permanent housing, access totransportation, and inability to meet health andeducation expenses. In this context the Govern-ment identified 46 districts of first priority forpoverty reduction.2 Many of these districts arealso heavily contaminated by unexploded ordi-nance (UXOs) from the Vietnam War (Map 3.1).In the eyes of the poor themselves, povertyincludes rice insufficiency, lack of large animals,and vulnerability to ill health (ADB 2001).From these broader perspectives, improving ourunderstanding of poverty-environment linkagesis vital for the development of more sustainablelivelihoods and an improved quality of life, bothin terms of natural resource management andenvironmental health risks.

The poorest districts in Lao PDR are char-acterized by very sloped land, relatively lowpopulation density, and—particularly in theSouth—by ethnic minorities. Many communi-ties in these districts are small and remote, withlimited access to roads and markets and improvedwater supply and sanitation, and a high relianceon natural resources for their survival. In thecentral and southern part of Lao PDR, it is thepoorest districts that also have the highest for-est cover (Map 3.2).

The National Growth and Poverty Eradica-tion Strategy (NGPES) of the Government ofLao PDR establishes a link between poor districts

Poverty and Environment in Lao PDR

3

�

and upland districts, but there is no official defi-nition of upland districts in Lao PDR and nodirect correlation between elevation and poverty.Some 55 percent of the 142 districts in Lao PDRhave a maximum elevation between 500 m and

200 m, but only 28 percent of the 46 prioritypoor districts are among them. Slope, however, isa variable that may describe poor upland districtsbetter since about half of the districts with higherpoverty incidence than the national average have

P O V E R T Y A N D E N V I R O N M E N T I N L A O P D R


Provincial poverty District poverty Government priority Percent of villages Incidence 2003 incidence 1998 with high UXOs districts

M A P 3 . 1 Poverty Incidence and Government Priority Districts for Poverty Eradication

Sources:

M A P 3 . 2 Geographic and Demographic Characteristics of Lao PDR

Sources:

more than two-thirds of their land area with slopeshigher than 16 percent (Maps 3.1–3.2).

Grouping the districts in Lao PDR by povertyincidence provides some perspectives on poverty-environment linkages (Table 3.1). The two groupsof districts with the highest poverty incidence areless populous, but have the largest number ofpoor.3 In these districts, forest cover was lower anddeforestation higher during the 1993–97 period.4

Per capita agricultural land and rice productionwas substantially lower, and significantly less flatland is available for cultivation.5 The districts’share of population with sanitation facilities wassubstantially lower, and reliance on open unpro-tected surface water was higher.6 Crude mortalityrates were also higher, and literacy rates consider-ably lower. Access to roads was also lower in thesepoor districts.7

The government of Lao PDR has identified46 priority districts—or nearly one-third of thecountry’s districts—for poverty reduction anddevelopment (NGPES). Many of these dis-tricts are in the poorest provinces—Oudomxay,

Phongsaly, Houaphanh, and Saravane—andalong the border with Vietnam in the South(Map 3.1). A comparison of these districts to the46 districts with highest poverty incidence in1998 is revealing.8 The government’s priority dis-tricts have (a) the lowest rate of water supply cov-erage and highest reliance on open unprotectedsurface water; (b) the highest reliance on fuelwood; (c) the highest share of sloped land (slopegreater than 16 percent); (d) the highest crudemortality rate; (e) the lowest literacy rates; and(f) the lowest rice production per capita. Theonly environmental indicator that is higher in46 poorest districts is the deforestation thatoccurred from 1993 to 1997.

The priority districts show high consistencywith the government’s definition of poverty andoverlap closely with many environment-povertypriorities. NGPES, which initiated the firstpoverty reduction program with a geographicalfocus on 46 priority poverty districts, identifiesthe environment as one of the key inter-sectoralpriorities in the fight against poverty. For forest



T A B L E 3 . 1 District-Level Poverty Profile

District poverty incidence (1998) 0–25% 25–41% 41–64% 64–96% All Lao PDR*

Number of districts 32 32 32 32 128Population (million) 1995 Census 1.4 1.2 0.9 0.9 4.4Population (million) 2000 Estimate 1.7 1.4 1.0 0.9 5.1Population in poverty 1998 (million) 0.27 0.47 0.52 0.70 1.96Percent of total area with slope >16% 41.9 50.5 66.0 70.3 59.5Rice production per capita (kg/person) 280.4 264.5 229.4 188.1 247.3Large animals per household 1.84 2.75 2.52 2.59 2.36Forested land % of total land 1997 39.4 43.8 37.7 34.6 38.6Deforestation rate of forested land 1993–97 2.9 2.4 2.7 3.2 2.8Agricultural land % of total land 1997 12.2 8.3 3.2 1.7 5.5Percent of households relying on open 24.6 53.8 63.4 74.2 49.9

unprotected surface water in 1995Percent of households without toilet 52.0 76.3 76.6 84.3 69.7

facilities in 1995Crude death rate in 1995 6.8 10.6 12.1 14.7 10.5

(per 1000 population)Percent of population with education 72.9 58.3 49.7 40.3 57.7

in 1995

Sources:*ADB (2001) only presents poverty estimates for 128 districts (of a total of 142 districts).

resource degradation and issues related to agri-cultural land, however, it is necessary to con-sider a broader focus than the priority group ofdistricts.

POVERTY-ENVIRONMENT ISSUES

A set of environmental issues in relation topoverty can be estimated from available data, per-taining to environmental health, natural disas-ters, and the urban environment (e.g. outdoor airpollution). The disparity between poor and non-poor households is particularly substantial withrespect to lack of potable water supply, basic san-itation, and UXO contamination. While a lessernumber of people are affected by malaria anddrought, the poor are also disproportionatelyaffected by these problems.9 Flooding seems toaffect a larger share of the non-poor populationthan the poor. Solid fuels are the main sources ofhousehold energy for cooking both among thepoor and non-poor population. As more than85 percent of the poor reside in rural areas, andurban poverty incidence is “only” 20 percent,urban environmental problems, such as outdoorair pollution, are more an issue for the non-poorpopulation. Urbanization is still low in Lao PDR,

and urban air pollution remains moderate andnowhere close to the levels in many large cities inSouth-East Asia. Access to a road, an importantdeterminant of, for instance, environmental ser-vices, such as water supply and sanitation, is sig-nificantly lower among the poor population.

Sufficient data is not available to provide anestimate of the number of poor and non-pooraffected by environmental issues related to nat-ural resources such as forests, non-timber forestproducts, or land. In agriculture, however, ricesufficiency and livestock holdings are defined bythe poor themselves as principal indicators ofpoverty in Lao PDR (ADB 2001). Illness andlack of health services were also raised by thepoor as central issues.

In the context of this study, poverty dimen-sions are those factors that affect rice sufficiencyand livestock holdings, which include, but arenot limited to, quality and quantity of naturalresources. Indicators for groups of districts bypoverty incidence confirm the importance ofrice sufficiency (Table 3.1). However, the analy-sis does not apply to animal numbers: poorerdistricts have higher overall livestock numbers,even though the poor within these districts mayoften be those with fewer or no animals.



T A B L E 3 . 2 Number of Poor and Non-Poor Affected

Estimated number of people affected

Poverty & environment indicator: % of non-poor population % of poor population

Lack of water supply (using open water sources)* 20 30Lack of sanitation (no latrine)* 40 70UXO Contamination (1997) 20 28Household use of Solid fuels/indoor air pollution* 95 98Malaria Incidence (confirmed cases 1999–2001)** 0.2–0.4 1.0–2.0Flood (1995–97 and 2002)*** 5 1Drought (1995–97)*** 0.2 1.7Urban environment na <10–15No Access to Road* 15 30

Source: Estimated using a variety of statistical techniques from the following data: *LECS III 2002/03, **WHO data,***Natural Disaster Management Office data. ****Less than 15 percent of the poor reside in urban areas (LECS III2002/03).

PEN Study Focus Areas

The PEN study set out to deepen the under-standing of poverty and environment linkages inLao PDR by examining their relationship at theprovince, district, and household levels and tomainstream the findings into the national policyagenda. The scope of the study was designed to fitinto the national policy framework for povertyreduction. This includes the National Growthand Poverty Eradication Strategy (NGPES) andthe National Poverty Reduction Strategy Paper(PRSP), published in 2003.

The main national partners of the PEN studyin Lao PDR were (a) the Committee for Planningand Investment (CPI); and (b) the EnvironmentalResearch Institute (ERI) of STEA (Science, Tech-nology & Environment Agency), which is thenational agency in charge of environment directlyunder the prime minister. In addition, a numberof other government agencies participated in theimplementation of the study and provided criticaldata for the study (Table 3.3 and Box 3.1).

Dialogue with these partners—and findingsfrom the initial PEN analysis (e.g. an extensivePEN study background report and Table 3.2)—led to the selection of five focus areas, three ofwhich are related to natural resources and two inthe area of water supply and sanitation (Table 3.3).Four of the focus areas relate to rural develop-ment and one to urban development, reflectinga rural population share of more than 75 per-cent in Lao PDR.

The PEN focus areas cover three of the fourNGPES policy sectors: transport, health, andagriculture and forestry. None of the focus areasdealt directly with education. However, it becameevident during the course of the PEN study thateducation—and more broadly information- andawareness-building—is an essential ingredientin all five PEN focus areas. One of the focusareas relates to the Lao UXO program, which isone of three poverty-related national programsincluded in NGPES.

A two-pronged approached was followed inthe study of the PEN focus areas. A nationwideanalysis was carried out using secondary data atprovince, district, village and household level,and field surveys were undertaken in four dis-tricts. The nationwide analysis and field surveysinformed each other, strengthened the confidencein the PEN findings, and provided wider geo-graphic relevance. The field surveys were imple-mented by a study association led by ERI/STEA,comprised of research groups from the nationalministries and their agencies.

On the recommendation of the Governmentof Lao PDR, the field surveys took place in twoprovinces in the far north of the country and intwo provinces in the far south. The governmentselected one district in each province for the ruralfield surveys and the two smallest provincial cap-itals for the urban field surveys (Table 3.4).

All four districts where the rural focus stud-ies took place are located in the upland districts.



T A B L E 3 . 3 PEN Focus Studies and Implementing Agencies

Poverty and environment focus Implementing agency

Natural Resources NTFP management National Agro-Forestry Researchand Livelihood Institute MAF)

Road access and development Environment and Social Divisionof MTCPC

UXO contamination Environmental Research Institute(STEA)

Environmental Health Rural water supply| and sanitation Department of Hygiene (MoH)Urban water supply and sanitation Urban Research Institute (MTCPC)



B O X 3 . 1 Methodologies and Data

The methodology employed in each of the five focus studies entailed the combination of nationalmapping and correlation analysis of poverty and environment indicators with village and house-hold surveys. The research teams, especially those working on the rural field surveys, were con-fronted with substantial challenges in carrying out local research in remote locations, so themethodology was carefully adapted to overcome these challenges (see chapter 1 for more detail).In addition, the study team took great care to ensure consistency between the indicators collectedin the national databases and those collected through the surveys.

The study utilized precise sampling criteria for villages and households. Poverty levels in thesurveys were assessed through a scoring system based on both productive and non-productivehousehold assets (asset indicators, in turn, were collected at the household and village levels).Three of the PEN focus studies conducted comparative analysis of poorer and less-poor house-holds, while the two others completed a simple comparison at the village level only.

For some of the variables, information could not be updated from the first phase of the PENstudy conducted in 2003–04. For example, the 1993–97 forest data were used, even though pat-terns of deforestation may have changed significantly in the last eight years. Selected variables atdistrict level (ADB 2001) from the LECS II (1997/98) surveys were analyzed in correlation with vari-ables from the PEN database. While the LECS III data provided further insights into some of thesame indicators at household and province level, reliable district level data have not been pro-duced from LECS III. Moreover, the new Population Census 2005 can be an important source ofanalysis of PEN linkages.

Partner Data assistance

National Statistical Center (NSC) LECS III, and economic and demographic dataMCTPC Road network data at district levelLao UXO program office UXO contamination data from the national

UXO survey in 1997NAFRI NTFP database for 39 districts (the PEN focus study)URI Urban water supply for several provincesEnvironmental Health Division (MoH) Water supply breakdowns in rural villages

T A B L E 3 . 4 PEN Field Survey Districts and Towns

District and towns Location Poverty incidence (1998)

Notou district Phongsaly province (North) 79%Phongsaly town Phongsaly province (North) naNamor Oudomxay province (North) 93%Kaleum district Sekong province (South) 63%Lamam town Sekong province (South) naPhouvong Attapeu province (South) 71%National average 39%

Source: District poverty incidence is from ADB (2001) based on LECS II. Poverty incidence is not available (na) for thetwo towns.

They are among the poorest within the group of46 priority poor districts identified in NGPES.The proportion of ethnic minority people ishigher than 98 percent in three of the selecteddistricts, and is 72 percent in the fourth. TheMon-Khmer ethno-linguistic group accountsfor more than half of the district population inthree of the districts.

All four districts border either China or Viet-nam, and two of the borders are located at highelevations. The marginalization of these districtsis decreasing, however, as international roads andborders are opening up.

As the PEN field surveys focused on the uplanddistricts listed as priority poor districts in NGPES,the surveys have a strong de facto linkage to the on-going national policies of rural development basedon “focal areas” and of stabilization of shiftingcultivation. These policies are discussed in thecontext of individual PEN focus areas and at theend of the chapter.

POVERTY AND NATURAL RESOURCE MANAGEMENT

Close to 40 percent of land area in Lao PDR wasclassified as forest in 1997, according to satelliteimagery data. Agricultural land, including fallowland, occupied 13 percent of the total territory.Shrub and grassland constituted 43 percent oftotal land area (differentiating between fallowand shrub and grassland, however, is subject todebate). There are substantial differences in landuse and vegetation across the three regions inLao PDR. Forest land was 65 percent of totalarea in the South, 48 percent in the Center, andonly 21 percent in the North. Evergreen forestconstituted 36 percent of total forest (Map 3.3).

Deforestation in 1993–97 was highest amongthe districts with the highest poverty incidence,and, geographically, in the northern provinces,which have some of the highest poverty incidenceand lowest forest cover. A total of 1,118 km2 offorest was lost during these four years in the



M A P 3 . 3 Forest Cover 1997 and Deforestation 1993–97

Sources:

North, representing more than 5 percent ofremaining regional forest cover in 1997. Thereare at least two important dimensions to this sit-uation. First, these findings alone are not suffi-cient to conclude that poor people’s livelihoodsare significantly affected by this deforestation,and even less so that the poor are responsible forsome of this deforestation. Second, the findingsrefer to 1993–97, and patterns of deforestationmay have changed significantly in the last eightyears. For instance, losses and scarcity of ever-green forest are a particularly growing concernarising from the PEN focus study of NTFPs.

Demographic, socioeconomic, and geographicfactors seem to have influenced the long-termtrends in forest cover in Lao PDR.10 Populationdensity and agricultural land area are associatedwith lower forest cover. Road access and road den-sity is not found to have a negative effect. Steep-ness of land is found to be negatively associatedwith forest cover. This is particularly the case forland with steepness above 30 percent. No asso-ciation between overall forest cover and flat landis found beyond the influence of populationdensity and agricultural land. Conversely, flat-ness of land is more influential in explaininglower evergreen forest cover than populationdensity and agriculture.

Compared to the Central region, the North-ern region is found to have significantly less andthe Southern region more forest cover thanexpected from steepness of land, populationdensity, and agriculture. No “border effect” wasfound for districts bordering Thailand or Viet-

nam in terms of total forest cover, that is, dif-ferences in overall forest cover between borderdistricts and the rest of the country are allexplained by the aforementioned factors. Dis-tricts bordering Vietnam do however have moreevergreen forest than can be explained from thefactors discussed above, and districts borderingThailand have less evergreen forest.

The higher household poverty incidence indistricts bordering Vietnam could be a particularconcern for PEN linkages. These districts stillhave higher rates of forest cover than the averagedistrict in Lao PDR. A poverty reduction strategythat opens the Vietnam border districts with roadlinkages to Vietnam may want to pay particularattention to forest conservation objectives.

These findings confirm the existence of apoverty and environment nexus between poverty,forestry, and land in Lao PDR: there are stronglinkages between indicators, poverty groups, andregions, but causal relationships that would helpexplain these factors are not identified throughthis analysis.

Many factors potentially affecting povertyare directly or indirectly related to naturalresources (Figure 3.1). The largest difference inpoverty incidence (i.e. substantially lower inci-dence) is associated with literacy, road access,non-agriculture as main income source, andurban household location. Poverty is also lowerin communities with development projects andland and forest allocation programs, and thosecommunities that have less UXO contaminationand more irrigated land. Proximity to district cen-



T A B L E 3 . 5 Poverty & Environment Nexus in Natural Resource Management

PEN areas: Indicators: Poverty and geographic analysis:

Poverty Poverty incidence 2002/03 Somewhat higher in the NORTHHighest in districts bordering Vietnam

Forest Deforestation (% and km2) 1993–97 Highest in the NORTH and POOREST DISTRICTSEvergreen forest cover 1997 High districts bordering Vietnam

Land Agricultural land per capita 1997 Lowest in the NORTH and POOREST GROUPSloped land (% of total land) Highest in the NORTH and POOREST GROUP

ter and residing in the provincial capital district isalso associated with lower poverty incidence. Incontrast, poverty incidence in the Government’s46 priority districts and in districts bordering toVietnam are higher than can explained by thefactors discussed above.11

As many of the factors are associated withurban-rural household location, an analysis wasundertaken for rural households only. The analy-sis confirmed that all the factors are associatedwith poverty incidence even within rural areas.

The lower poverty incidence among house-holds that have irrigated land and that are notdependent on agriculture as their main source ofincome deserves further analysis. The first indi-cates a relationship between more intensive agri-culture and poverty reduction, while the secondindicates a relationship between diversification of

incomes out of agriculture and poverty reduction.Poverty in relation to UXO contamination androad access is further discussed later in the report.

New information about villages with landand forest allocation programs and developmentprojects, and other factors was also revealed fromthe analysis:

• More than 40 percent of the villages in theLECS III survey reported that land and forestallocation programs had been implementedthere. Poverty incidence in these villages issignificantly less than in villages without allo-cation programs, even when comparing villageswithin rural areas. Conversely, the ParticipatoryPoverty Assessment (ADB 2001) indicates thatthere are significant trade-offs between protec-tion efforts and poverty reduction in the short



0% 10% 20% 30% 40% 50% 60%

Urban

Landallocationprogram in village

Irrigated land

Main income source(non-agriculture)

Road access

UXO problem (<50% ofvillages in district)

Read (>50%ofHHmembers)

Development projectinvillage

YES NO

Source: Cross-tabulations from household and village data in LECS III 2002/03.

F I G U R E 3 . 1 Poverty Incidence in Relation to Select Indicators

to medium term in many districts. Many com-munities depend heavily on the forest for theirlivelihood, and many villages reported thatthey no longer have sufficient access to forestresources after the implementation of allocationprograms.

• Nearly 50 percent of surveyed villages reportedhaving an ongoing development project at thetime of the survey. Poverty incidence in thesevillages was lower than in villages without aproject. However, a development project hada somewhat stronger association with higherhousehold consumption than lower povertyincidence. This could suggest that develop-ment projects may be benefiting non-poorhouseholds more than poor households, butfurther assessment of this issue is needed.

• Last but not least, ethnicity, village migration,and market access were not on their own asso-ciated with lower or higher poverty incidence.

These findings confirm the complex linkagesbetween poverty and natural resource manage-ment. Poverty and renewable natural resourcelinkages must be seen in the context of broadereconomic development. These issues are highlycomplex and interdependent. They involve thelong-practiced traditional agricultural productionsystems, forest protection and agricultural landallocation, availability and access to NTFPs, thevillage consolidation program, and developmentand participation in a market economy. Naturalresources are used not only by rural households,but also by other stakeholders. No informationcan be derived on the latter from the datasets ana-lyzed. These issues all deserve further quantitativeanalysis in combination with local research.

Non-Timber Forest ProductResources and Poverty

Non-timber forest products (NTFPs) are impor-tant to the vast majority of rural households in theuplands of Lao PDR. The Lao government views

NTFPs as having substantial potential to improverural livelihoods in the poorest districts. There is atwo-way beneficial relationship between forestconservation and sustainable use of NTFPs. Mar-keted NTFPs, in addition to self-consumed ones,have historically been important export productsand remain so today. In contrast, the regulatoryframework is vague, and only anecdotal quantifiedinformation on NTFPs is available. There is anacknowledged risk of resource decline. The PPA(ADB 2001) reported that the quantity of NTFPswas declining in many village areas. Pilot rehabili-tation or cultivation projects are being initiated,but a clear policy statement is missing.

The government’s interest in NTFPs is closelyrelated to the stabilization policy for shifting cul-tivation. NTFPs are often cited as an alternativeto shifting cultivation—for example, in villagesconsolidated into “focal areas.” The relationshipbetween sustainable NTFP management andswidden agriculture12 or village consolidationappears, however, to be a complex one, as thereis contradictory evidence from field research.Some researchers argue that most NTFP speciesrequire some type of human intervention formaintenance. Swidden cultivation in the formof rotational agriculture would, therefore, have a positive impact on several NTFP species.Others, on the other hand, argue that swidden-ing in a context of population growth inevitablyimpacts NTFP resources through a reduction inold-growth forest. NTFPs are, therefore, an entrypoint to research poverty, environment, andshifting cultivation.

The PEN study of NTFPs was designed toanswer two questions. First, the study lookedat two potential NTFP development strategies:(1) sustainable harvest of natural NTFPs, and(2) NTFP cultivation. The objective was to seekevidence about the viability of these strategies,especially for poor households. Second, the PENstudy explored the relationship between villagemigration and consolidation policies and the pres-sure on NTFP resources in poor communities.



NTFPs in Lao PDR are not incorporatedinto a statistical collection system. A collectionof data and information was therefore under-taken from the districts. A total of 39 districtsresponded with detailed data and informationon predominant types of NTFPs, trends in theiravailability in district forests over time, trends inmarket demand, and other valuable informa-tion. Data is available for most districts in threeprovinces in the North, one in the Center, and onein the South. Only 15 of the respondent districtsare among the 46 priority districts (Map 3.4).13

The large number of NTFP species used inLao PDR is an important reason for the lack ofclarity in the sector. The survey led to the iden-tification of 13 major species, on which theanalysis subsequently focused. Finally, NTFPincomes and the importance of swidden prac-tices are two topics that are difficult to address inrural surveys, since households may be wary oftaxation and the swidden control policy. Proxyindicators were assembled in addition to these.The analysis was, however, able to use surveyfindings on NTFP incomes and upland riceareas after data consistency was confirmed.

NTFPs of National and Regional Importance

The districts reported altogether 37 differentmarketed NTFP species, with three species onaverage considered as very important (Map3.5). Among these, eight are important mar-keted species in at least two among the threeregions (North, Center, and South) and five inone region (Table 3.6). Cardamom was citedas important in almost 70 percent of the dis-tricts, and rattan in almost 60 percent. Onlybenzoin, an important local species, was leftout of this analysis, because the producing dis-tricts in northeastern Lao PDR did not answerthe survey.

While NTFP species are harvested from threeclearly differentiated domains—old-growthnatural forests, secondary forests, and youngfallows—classifying individual species into onecategory is not easy. Only three of the 13 mainspecies were reported as fully harvested from the“natural forest” domain. On average, four mainNTFP species were reported as forest species andtwo as fallow species.



M A P 3 . 4 NTFP Survey Districts in Relation to Ethnic Minorities and 46 Priority Districts

All main species identified, except for broomgrass and paper mulberry, are mostly harvestedfrom natural forest. A marked contrast is observedbetween forest species and fallow species. Thedistricts reported a decline in resources of mostNTFP species, particularly forest species, andonly reported an increase for the two fallowspecies (Map 3.6). Declining quality was alsoreported, especially for species. District respon-dents primarily attributed resource degradation tooverharvesting and shifting cultivation.

Northern districts reported an active markettrend, especially close to the China border. Priceincreases were noted for forest species, withdeclining populations and decreasing marketedquantities. Conversely, the two fallow speciesshowed an increase in population, marketedquantities, and prices.

Some domestication, predominantly in theNorthern provinces, was reported for seven of the13 main species and for five secondary species.However, only one species—eaglewood—was



NTFPs on Fallow Land NTFPs in Forest

M A P 3 . 5 NTFPs of National and Regional Importance Reported by Each District

Source: PEN study (NAFRI).

T A B L E 3 . 6 Commercially Important NTFPs in 39 Districts

8 Species of national relevance 5 species of regional relevance

Cardamom, rattan, orchid, bong bark, dammar, Malva nut, galangal, sugar palm fruit,eaglewood, bamboo shoots, broom grass meuak bark, paper mulberry

reported as partly cultivated by a larger majorityof the districts (about 90 percent). Paper mul-berry, which grows spontaneously in fallows,ranked second in cultivated species. Cardamomand galangal, two species that are easy to culti-vate, are harvested from the wild in around70 percent of districts.

NTFPs in the Household Economy

About 96 percent of households in the northerndistricts and 72 percent in the South reported cashincome from NTFPs every year. Householdsreported average annual NTFP sales of US $73 inthe Namo District in the North and US $22 in thesouthern district. This is respectively 16 percentand 5 percent of average cash expenditures in theprovinces in the LECS III survey. However, thefact that households ranked NTFPs as their sec-ond highest income source—after food crops andbefore animal sales—would indicate much higherincomes. Reported incomes, however, provide auseful basis for comparison between households.

Poverty was not a primary factor in NTFPcollection. The lowest household income fromNTFPs was less than US $5, while the highestwas 1,000 times more. NTFP income was, how-ever, 40 to 50 percent higher in households in thelowest wealth category than in the medium- andhigh-wealth categories in the northern district.14

There was a reverse, but not statistically signifi-cant, relationship in the southern district.

Other than poverty, variables that were signif-icantly correlated to NTFP incomes were the ageof the village, cultivation of upland rice, and mar-keting channels. In the northern district, house-holds with more than one hectare of upland ricearea had 2.5 times more income from NTFPsthan households with less than 0.5 hectares. In thesouthern district, the households in the surveyhad resettled to a new paddy basin, upland riceareas were small, and the difference in incomebetween households with less than and those withmore than 0.5 hectares of upland rice was mini-mal (Figure 3.2).

Distance to NTFP collection sites increasedmarkedly during the last 10 years, from lessthan 5 km to almost 10 km on average in thenorthern district, and from less than 7 km to16 km in the southern district. This is an indi-cation of resource degradation in the North,and of village resettlement in the South. How-ever, walking distance is not an absolute obsta-cle to NTFP collection. Only 9 percent ofhouseholds, all of them in the North, cultivatesome NTFPs.

More than 90 percent of households in theNorth and slightly fewer in the South were awareof customary NTFP management. NTFPs forself-consumption are an open resource, whileNTFPs for sale are reserved for the local com-munity. No improved management system wasobserved in any of the villages surveyed, the newestof which was two years old and the oldest thirty.The older villages have significantly more incomefrom NTFPs. The only new village with a sub-stantial NTFP income has the largest upland ricearea per household, and as much as 35 percentof the surveyed households are members of amarketing group. Members of the NTFP market-ing group have NTFP incomes more thanthree times higher than non-members. There



Increase Decrease

M A P 3 . 6 Number of Increasing/Decreasing Main NTFP Species

was an absence of extension of this successfulorganization to other villages.

Discussion

NTFPs are important for the rural population ofupland areas in Lao PDR regardless of ethnicity,gender, wealth classes, or distance to forests.Households are more likely to use NTFPs as anincome source when they are engaging in shift-ing cultivation, because this gives them knowl-edge about resources, closer access, and time forcollection. These households may be poorer, anduse NTFP income to compensate food deficits.This is the case in some communities, but notin others.

The existence of a declining NTFP resourcetrend is obvious for species growing in matureforests and calls for urgent action. The disappear-ance of some of the NTFP resources wouldsubstantially affect broad numbers of uplandcommunities, where poverty incidence remainshigh. Domestication has hardly started, so NTFPsremain by and large a common property resource.NTFP cultivation is, however, an option only forsome NTFP species. Community-based resourcemanagement will remain the only option for otherspecies. Communities’ customary regulations areweak in the face of high market demand and in theabsence of a formal regulatory framework.

The existence of regulations and the interest ofboth communities and local governments in moresustainable resource management indicates, how-ever, that community-based NTFP managementcan be a viable option. Market-related commu-nity groups appear to work well and to integratepoorer households successfully. Such groups havemore potential to limit harvest levels than quotas.Local governments do set up quotas based onrough estimates of resource demand and supply,but these are difficult to manage.

The small number of marketable species ofnational relevance offers an opportunity for poli-cies and programs to address individual NTFPspecies in a specific manner. Efforts to help culti-vation take off could concentrate on those forestspecies, for which domestication is technically fea-sible and not on open-space species. Community-based resource management programs couldfocus on other forest species that cannot be culti-vated. Public access to open-space species doesnot threaten resources and is an appropriateoption for the poorest.

Land use systems are evolving as a combinedoutcome of migration trends to roadsides andareas with potential for paddy field developmentand government policies. Stabilization of shift-ing cultivation through land allocation gener-ally shortens rotation cycles. One can expect anincrease in the availability of fallow NTFPs, but



0

200

400

600

800

1000

1200

Kip

(00

0) p

er y

ear

<0.5 0.5-1.0 >=1.0

Namo District

Upland Rice Farming (ha per household)

0

200

400

600

800

1000

1200

Kip

(00

0) p

er y

ear

<0.5 >=0.5

Phouvong District

Upland Rice Farming (ha per household)

F I G U R E 3 . 2 NTFP Income in Relation to Upland Rice Farming

these have provided limited incomes so far.Encouragement of migration to new settlementsalong roadsides with no option for paddy fieldswill enhance this trend, and so will the mergingof several small villages into a larger one underthe village consolidation policy.

Migration to roadsides, where paddy fields areavailable, may reduce households’ NTFP incomesin the short term and impact the livelihoods ofpoorer households, but longer-term trends areunclear. In the absence of such policies, however,the more valuable NTFP resources will continueto decline together with old-growth forest.

NTFPs are valuable observation lenses toremind policy makers of the complexities andtime dimensions of policies related to land usechange in upland areas, where swidden agricul-ture remains an important livelihood source. Theprevailing sentiment that there is a simple causalrelationship between shifting cultivation and thedecline of NTFP resources may overlook othercritical factors behind the ongoing decline ofNTFP resources.

What is needed is neither a rapid shift ofland use systems out of swidden cultivation northe status quo. Instead, communities need gov-ernment support in managing their forest re-sources, including NTFPs, particularly in thecontext of an active market. Communities liv-ing in remote locations and away from roadsare equally in need of such support. The stabil-ity of villages and slow land use change may beimportant factors for success.

Roads, Environment, and Poverty

The Government of Lao PDR faces the challengeof building a sustainable road network in a coun-try where both poverty and natural resources areconcentrated in upland districts. The NGPESacknowledges a high correlation between the lackof road access and poverty. The policy frame-work of the transport sector in Lao PDR followsa step-by-step approach. The government andthe donor-funded projects started to direct theirefforts toward rural roads in the second half of

the 1990s, after the national and provincial net-works had been largely completed (MRCDP1996). The mid-2000s appear to be a turningpoint. All districts are now connected at least byearth roads, projects have demonstrated the via-bility of village feeder road construction, andvarious options to fund and ensure village roadmaintenance are in place or being piloted.

In the next stage, the government may eithercontinue to follow a step-by-step approach byensuring all-weather links to district towns, oropt for a feeder road network serving a broadnumber of villages. The first option is stated asa priority in NGPES for poor districts. Policychoices need to take into account the environ-ment: the opening of roads is widely regarded asan important factor behind deforestation (IUCN1997). Policy choices are also related to the gov-ernment’s overall policy of supporting migrationto roadsides in upland areas. “Focal areas” arestated in NGPES as a priority for road develop-ment next to district towns.

A central question—whether to build a ruralroad network with fewer roads of higher quality,or to expand the network to a broad number ofvillages—is an important area of policy decisionmaking. The latter option might be viable, pro-vided that lower, cheaper standards are selected.The PEN study seeks to compare opportunitiesand constraints of these two options in terms ofpoverty reduction and preservation of environ-mental resources.

The PEN study encountered significant diffi-culties both in national analysis of district-levelindicators and in the village survey. The percent-age of villages within 6 or 11 km from a main orsecondary road is a useful road access indicator,15

but no update was available. Actual village accessmay be more valid in a country where villagelocation evolves rapidly. NGPS compiled datafor the selection of priority poor districts andNSC prepared the population census, but thetwo datasets are largely inconsistent. This maybe related to the difficulties of tracking numbersof villages. Defining road access also is difficult.16



In addition, local governments may sometimeshave limited incentives to report accurate infor-mation on village roads. Road density, used inconjunction with density of population and vil-lages, would conversely be a convenient indica-tor. Unfortunately, rural roads are not recordedseparately from paths.

These difficulties have limited the scope ofthe national data analysis but have shed light onthe obstacles that the government is facing inmonitoring and planning its local road network.The field surveys provided a valuable compre-hensive picture about the actual status of localroad and transportation systems in two marginaldistricts. Interview findings were mostly assem-bled at the community level.

The survey team found it difficult to avoidselecting villages recommended by local author-ities. The villages surveyed mostly turned out tobe recent or upcoming beneficiaries of road con-struction. This has limited the scope for usingthe PEN study’s methodology of comparativeanalysis between poorer and less-poor villagesand households, but it has provided valuableinsights into the processes used by local govern-ments when selecting beneficiary villages in theirroad programs.

Nationwide Perspective

LECS II (1997–98) demonstrated that poor dis-tricts had substantially lower access to roads, asmeasured by percentage of villages with main orsecondary roads within 6 kilometers. LECS IIIconfirmed this correlation by showing thatpoverty incidence is almost twice as high amonghouseholds in villages with no access to a roadcompared to households with access. However,correlation between road access and povertyincidence is not a simple causal relationship. Forexample, the road network is being built at amarkedly slower pace in upland districts, wherevarious poverty factors are combined.

Road access and road density were not foundto have a negative effect on overall district forest

cover, but is associated with lower district ever-green forest cover.17 This indicates that openingroads may lead to some deforestation, and thatthis potential impact needs to be managed. Thisis particularly important in districts bordering toVietnam where most of the remaining evergreenforest in Lao PDR is located.

Overall, 72 percent of villages in Lao PDRalready have at least seasonal road access. Villagesmay be served by these roads, by districts roads, or by rural roads linking villages only. Around 36 percent of the 30,000 kilometers of roads arerural roads; an unknown proportion of these arepaths that are not suitable for motorized access.Village access to roads is uneven among districts.Average road density in the country is slightlyabove 12 km per 100 km2, but it ranges from 1 to35 in the rural districts. In 22 provinces, morethan 50 percent of the villages have no road access.

In short, the country has already achievedsome coverage of village road access despite lowroad densities. A simple theoretical assessment ofvillage densities provides a rationale for this. Ruralpopulation density was below 17 people/km2 inhalf of the districts in 1995, and below 10 in halfof the NGPES priority districts.18 Villages in ruraldistricts19 have an average population of not lessthan 380 people nationwide and 325 people inpriority poor districts. Only 23 districts have anaverage village size of less than 250 people. Thismeans that a district with 10 people per km2 hason average less than 3 villages per 100 km2. Roadinvestment cost per village served would increaserapidly in correlation with lower village densi-ties, if villages were equally distributed over thenational territory.

This, however, is not the case. In the four dis-tricts surveyed, only the southern half of theNamo District in the North has a traditional set-tlement of villages scattered over hilly terrain.Elsewhere, and in many districts, judging fromthe available detailed topography maps, mostvillages are distributed in strings along paths orrivers, which are the traditional transport modes.In many such districts, 10 kilometers of rural



roads per 100 km2 may be sufficient to offer fullroad access. This is equivalent to 10 meters ofroad per inhabitant, or $150 per capita based ona rough estimate of $15,000 per km of earthroads, a figure that can be smaller if local com-munities provide part of the labor.

There is also a direct link between low com-munity density and limited environmentalimpact. Should road opening affect a corridor of50 meters on each side of the road, the result-ing impact would still be relatively limited as apercentage of total land. Environmental riskanalysis, therefore, relates more to areas of spe-cific environmental value than to overall forestcover rate.

A District and Village Perspective

Kaleum District in the South and Nhot Ou Dis-trict in the North have both very low populationdensity and very limited village road access (Table3.7). Both districts cover large territories of morethan 300 km2, which is 70 percent above thenational average. Nhot Ou is crossed by an all-season road from the South to an internationalborder with China. Only one village other thanvillages along this main road has a motorableroad, and plans for future road construction arelimited to one other village. In Kaleum, the roadto the district town itself is hardly passable outsidethe dry season. This road serves only two rural vil-lages. All other villages are cut off from the districttown during the rainy season floods. However,

an increasing number of villages are accessible bytrucks during the dry season. The governmenthas not mentioned specific plans for buildingvillage access roads.

The road network maps available at thenational level are misleading. They mention sev-eral rural roads with their registration number.In Nhot Ou, these are the traditional paths thatcontinue to serve for travel by foot or with horseanimals. A substantial proportion of villages aredistributed along or close to these paths. InKaleum, many of the paths are not used much,since most villages in Kaleum are located alongrivers, and boats are the traditional travel mode.

Enterprises are intensively involved in roadconstruction in both districts. A Chinese enter-prise has built the new village road in Nhot Ou.The main road was also reportedly built througha Chinese enterprise, although no confirmationcould be obtained. In Kaleum, the enterprisecontrolling the metal scrap market from UXOsin the district has built a dry-season road to accessvillages in the eastern part, and a State ForestEnterprise (SFE) from Vietnam reportedly builtanother road.20 Unfortunately, enterprises do notbuild bridges, so the rainy season access problemseither remain unsolved or increase with roaddegradation from the enterprise trucks. Informa-tion regarding road construction by enterprises,even location, was difficult to access in the dis-tricts and is not available at the national level.

Village discussions provided detailed accountsof how a direct road link makes a critical difference



T A B L E 3 . 7 Rural Road Network Indicators in the Two Districts Surveyed

Road Villages with % of Villagesdensity no road villages with no dry Rural population

Km/ access/total with no season cart Village density density100 Km villages road access per 100 Km2 people/Km2

Kaleum 3 53 / 60 85% 23 / 60 1.9 4Nhot Ou 4 62 / 86 72% 57 / 86 2.8 8

Source: national database and in-district checks

to villages today, now that active market outletsare within reach.21 Frequent market trips withbuses become feasible. They increase both incomeopportunities and social links. Farm product salesand job opportunities increase substantially in therainy season, which remains a food deficit seasonfor many. Increased incomes allow purchase oftransportation means and access to media, whichin turn reinforce all other advantages. Informa-tion exchange and support from the local govern-ment increase as well.

Consensus on the difference that road accessmakes for a village has become so strong that allparties, from government to villagers, tend toequate poverty with absence of road access. Dif-ferences between poorer and less-poor villagesamong the ten villages studied indicate, albeit ona non-representative village sample, that roadsare one of the key factors that influence the inter-action. Ethnicity, not roads, stands out as thefirst causal factor for poverty in the survey.

Villages that are gaining road access are makingsubstantial efforts to receive government approval

and expedite the process of completion. In NhotOu, in the first village community to obtain aroad, villagers started the work by themselves andbrought small tractors in spare parts before theroad was built. The second village awaiting a roadappears to be using its opium producer status andposition right on the border to be recognized as apoor village and a focal area, therefore qualifyingfor road construction.

It is, therefore, not surprising that the vil-lages interviewed made strong statements thatthey were confident in their ability to managetheir forest resources after a road was openedand to restrict access from outsiders. Data doesshow some increase in NTFP collection for salein villages along roads, but this increase was notsignificant. Spontaneous answers during inter-views also tend to validate the NTFP studyfindings regarding the potential of communityparticipation in natural resource managementsystems.

Preferred transportation tools have regionalspecificities. In Kaluem, mostly transportationtrucks are available, which require larger roadsand high maintenance costs. Motorbikes andsmall tractors, available through the Chinese mar-ket, would allow the construction of narrowerand cheaper roads. Tractors have a dual advan-tage; they can be used for both plowing andtransportation.

Discussion

The case of Kaleum illustrates the validity of thepriority given in NGPES to completion of the all-weather road network to district centers. How-ever, both Kaleum and Nhot Ou demonstrate aneed for further road outreach to the rural areas.Roads are one element among multiple povertyfactors. Direct road access to existing villagesmakes a critical difference to communities inupland areas because it gives them access to anactive market. Communities in remote locationsare willing to invest tremendous energy in secur-ing road access. Market access is not only a means



T A B L E 3 . 8 Selected Village Indicators inRelation to Poverty

Category of village Poorer Less Poor

% Lao Loum households 0 55% Households with means 5 34

of Transportation% Villages located by 17 50

All-Season Road% Households with 22 55

Media Equipment% Villages with Seasonal 50 75

RoadHours Walk to All-Season 2,8 3,5

Main road% households with NTFP 74 75

Income% Villages with Land Use 83 75

PlanningHours Walk to Old Growth 2,3 0,8

Forest

Source: From PEN survey in 10 villages (MCTPC).

to improve cash incomes, but also an importantsocial link in isolated areas.

The district surveys shed some light on the bar-gaining process between local governments andvillage communities in applications for a new vil-lage road. This is a very competitive process, giventhe modest level of on-going road constructionprojects. In this process, the successful communi-ties are generally those that have developed spe-cific strategies. These villages may be better off,stronger communities, or communities with spe-cial features that make them qualify as focal areas.

This trend raises the question of how to orga-nize and rank applications. Donor-funded roadsector projects have piloted prioritization througha scoring system based on community participa-tion and cost in relation to population served. Thesurvey highlights this need for prioritization in dis-tricts that are located on borders that have largeinvestment needs and/or have a multi-ethnic pop-ulation with contrasting socioeconomic status.Nhot Ou has all these features.

Full village coverage by the road network is adistant prospect in marginal districts. This maypartly explain the absence of a strategy in futurerural road network development at the national orlocal level. No comprehensive financial frame-work is in place. Such a strategy is needed, how-ever, and could be designed to address povertyand environment at the same time. Buildingcheaper tracks instead of roads does not appearto meet current local expectations. This mightchange rapidly, however, as smaller transportationmeans have become more available through bor-der trade. The widespread view that building aroad network in a sparsely populated territory is achallenge further limits active planning of a ruralroad network. Figures show instead that a well-distributed network may serve strings of villagesas soon as the country’s GDP makes investmentsof less than $200/capita economically viable.

The information flow between national andlocal government levels regarding the status ofthe road network is incomplete and not up-to-date. This is especially the case for roads built by

enterprises. The role of private enterprises otherthan construction companies in developing theroad network is already important in poor dis-tricts located on the borders. The role of govern-ment remains critical in planning and monitoringas well as in ensuring additional investments,for example in bridges. In contrast, there is sur-prisingly little information available on theseenterprises.

Linkages between road construction andincreased natural resource extraction are somewhatambiguous. Communities appear highly moti-vated to manage their resources well, providedthey are given the responsibility to do so. Con-versely, the involvement of large enterprises inbuilding roads against logging contracts is likely togenerate impact. What is needed may not so muchbe to question current practice that allows them tolog a limited area on both sides of the road as it isto ensure sound implementation and monitoring.

UXOs, Environment and Poverty

Unexploded ordinance (UXOs) dating fromthe mid-1960s and mid-1970s affect 15 of LaoPDR’s 18 provinces. Linkages between UXOcontamination and poverty are obvious, but theissue has not achieved high visibility. Deconta-mination started in 1996/97 under the nationalLao UXO program, but it is a very expensiveprocess and is proceeding at a slow pace.

Little is known about linkages between UXOcontamination and the environment. Some arguethat farmers are coping with UXOs by farming onslopes, while others claim that UXOs could limitencroachment on forests and exploitation of tim-ber and biodiversity degradation. In the lattercase, UXO contamination would have an unfor-tunate effect on communities’ health and liveli-hood but positive linkage to the environment,while in the former case the negative linkage tothe environment would be an additional reason tosupport UXO-affected communities.

The Government of Lao PDR has created thenational Lao UXO program—which provides a



channel for funding decontamination and publicawareness initiatives—into the NGPES. The pro-gram is described as relating strongly to povertyreduction and to the shifting cultivation stabiliza-tion policy. Decontamination is to be planned inareas identified as priority under these policies.

The PEN study sought to analyze poverty andenvironment linkages of the UXO issue by ascer-taining the impacts of UXO presence on house-holds’ agriculture and overall land use systems.And it sought to analyze poverty and environ-ment linkages of the current prioritization processin activities related to the UXO decontamination,and to identify potential environmental linkagesspecific to forested areas that might arise fromdecontamination.

The study benefited from the national LaoUXO program database. The database includesfindings of the comprehensive 1996/97 surveyof UXO impact (Handicap International 1997)as well as the updated status of village operations.These data were analyzed by linking district levelUXO contamination and LECS II–III povertydata. The methodology for field research tookinto account substantial obstacles, which wereexpected. The research only took place in theKaleum district.

The national Lao UXO program and thelocal government provided logistical supportand translation from the local Mon-Khmerlanguages. Information was mostly collected atcommunity level through focus group discus-sions. The national research team mostly usedhousehold interviews to confirm informationderived from group discussions.

The limited possibilities of directly observingdetailed land uses could not be fully overcome.However, quantitative indicators were assembledfor the six villages surveyed, and discussions withelder people, who could remember the pre-warsituation, indicated qualitative trends. A realisticset of findings was derived after supplementingthis information with the national database analy-sis and with district interviews.

UXO Contamination

The national survey of UXO contaminationrevealed in 1997 that at least 22 percent of all vil-lages in Lao PDR had a problem with UXOs,affecting close to 25 percent of the total popula-tion.22 In the Central region, nearly 50 percentof all villages are affected (Map 3.7).

UXO contamination is affecting the poor dis-proportionately. An estimated 28 percent of thepoor and 20 percent of the non-poor lived in vil-lages with a UXO problem in 1997/98.23 Morethan 70 percent of villages are affected in thepoorest districts of the southern region. The 1997survey ascertained that more than 450 UXO-related accidents had occurred annually between1974 and 1996. Casualties are only decreasingslowly: the Lao UXO program has recorded 110accidents annually from 1999 to 2004, but thesefigures only cover 59 districts.

Both LECS II and LECS III show that UXOcontamination in the South and Central regionsis associated with higher poverty, less productiveassets, and lower food security. In addition, LECSIII shows that living in districts of the South andCentral regions, where more than 50 percent ofvillages are affected by UXOs, is associated with



UXO problem High UXO problem

M A P 3 . 7 Percent of Villages in District Having UXO Problems

T A B L E 3 . 9 UXO Contamination and Key Indicators2002/03 in the Southern Region

UXO > 50% UXO < 50%

Poverty Incidence 2002/03 41% 22%Irrigated land (percent 4% 4%

of households)No Rice (months per year) 3.6 2.6Fuel wood collection 0.9 0.6

(hours per day)Water collection 1.2 0.8

(hours per day)No toilet facility 72% 64%No water supply (using 20% 7%

open water source)Emigration rate 1.6% 1.1%

(last 12 months)Literacy (> 50% of house- 63% 83%

hold members can read)Death rate (last 12 months) 0.6% 0.5%Development project 48% 62%

in village

UXO > 50 (or < 50): households living in districts where more (or less) than50 percent of villages are affected by UXOs. Source: The indicators are gen-erated from LECS III 2002/03. UXO data are from the national UXO survey1997.

lower improved water supply and sanitation cov-erage rates, more household time spent on fuelwood and water collection, and lower educationlevels and health status. This difference betweendistricts highly affected by UXOs and other dis-tricts is especially high for access to safe water inthe South (Table 3.9).

Villages in highly contaminated areas are alsoless likely to have ongoing development projects,and some villages are experiencing a higher rateof out-migration. As seen from the NTFP study,there is also an indication that districts with highUXO contamination have less income fromNTFPs, a source of income that could poten-tially compensate for the lower availability ofpaddy rice and irrigation.

Two positive findings were observed: (1) roadaccess is equally available in districts with lowand high UXO contamination, and (2) forestresources are relatively abundant in the Southand Central regions.

The Government’s 46 priority districtsaccount for 49 percent of highly affected vil-lages and 46 percent of recorded deaths since1999. They only account for 30 percent of the



0%

10%

20%

30%

40%

50%

60%

Percentage ofVillages Affected

12% 32% 56%

North Center South0%

10%

20%

30%

40%

50%

60%

70%

80%

Southern Region

%of Villages Affected 52% 50% 52% 73%

Least Poor Districts

Less Poor Districts

Poor Districts Poorest Districts

Sources: Left: calculated from the national UXO survey 1997; right: Data are from LECS II 1997/98 and UXO national survey 1997.

F I G U R E 3 . 3 Left: UXO Contamination in Lao PDR (1997). Right: Percent of Villages Affected by UXOs inRelation to Poverty in the Southern Region (1997/98)

agricultural land decontaminated between 1999and 2004.

The geographical distribution of UXO con-tamination in the six southern provinces is notunrelated to the distribution of Mon Khmer eth-nic groups. Mon Khmer people form the vastmajority of villages in the Indochinese Cordillera.UXO-impacted villages are widely distributed inthe six provinces except in the Mekong corridor,so there is no correlation between the proportionof Mon Khmer people and proportion or num-ber of villages affected by UXOs. It is, however,likely that several of the small Mon Khmergroups mostly or fully live in districts with a highproportion of UXO-affected villages. The map ofAgent Orange missions shows close concentra-tion of impacts on the Cordillera (Lao UXO pro-gram 2004 & Chaignon 2000).

No association was found between forest coverand percent of villages with UXO contaminationin a district. This finding is perhaps not surpris-ing, since affected areas range from the denselyforested Cordillera to the Xieng Khouang plateau,where forest cover is low. However, in districtswith high UXO problems, forest cover is lowerthan can be explained by region, population den-sity, agriculture, or topography. This findingdeserves further analysis.

Kaleum District

The distribution of villages affected by UXOs inKaleum derives from two factors: (1) the initialbombing impacts, and (2) active migration trendssince the 1990s. All quantitative data in Kaleumshould be regarded as tentative.24 Most villageswith UXO impact are located today in the west-ern third of the district closer to the districttown. Several of the seven severely affected villagesare right next to the district town. Villages fromthe eastern part of the district are strongly encour-aged to migrate to the west. The bombings fol-lowed the Ho Chi Minh trails that extendedthroughout much of the district. One villagesurveyed was created from the migration of two

of the 22 contaminated eastern villages and hasbeen quite successful with support from an inter-national project. The number of villages that hadmoved out is unclear.

Village consolidation in Kaleum is also active,with or without migration.25 The 1997 UXOsurvey mapped 54 villages affected by UXOs outof a reported total of 115 villages. The researchteam recorded 60 villages. One village surveyedwas formed by merging three nearby villages.The new location was severely contaminated,and reportedly no decontamination support wasmade available. There is also out-migration toother districts in the province, both spontaneousand government-sponsored. District populationreportedly has declined by 7 percent in the 1990s(Chaignon 2000).

Bombings within villages precisely targeted theHo Chi Minh trails. This has clearly resulted inhigher UXO contamination along rivers and val-ley floors and in village centers. People are exposedto UXOs in daily activities such as cooking or fish-ing. Paddy cultivation is especially dangerous,both when reclaiming new paddy fields and whenplowing the land. UXO decontamination is a pre-requisite in construction works for roads and com-munity buildings like schools. Since 2000, areas ofpaddy reclamation range from zero to only threehectares in all villages except for one, which hasbenefited from a donor-supported program. Thisold village is easy accessible from the districttown and had been selected for decontamina-tion for paddy field opening in the coming year.Altogether the Lao UXO program has cleared19 hectares of paddy fields and 30 hectares forconstruction works between 1999 and 2004.

Swidden agriculture is comparatively one ofthe least dangerous activities, because UXOsexplode spontaneously when the fallow is burned,and crops are sown without plowing. Swiddenagriculture nowadays takes place close to the vil-lages with very short rotations as a result. It is notpossible to determine to what extent this is a cop-ing strategy due to UXOs, or is the result of theshifting cultivation stabilization strategy. The fact



that hunting nowadays also takes place close tothe villages—and that only one out of six villagesreports substantial NTFP collection—points tothe former explanation. The fact that all villageshave a land use plan points to the influence ofgovernment policy, and so does the fact that localcommunities have been coping with high UXOpresence on their own from 1973 until 1999. Forexample, they have become accustomed to col-lecting UXOs either to store them away or morerecently to sell metal scrap. The latter has becomea major income source in the district. Villagersstated that they saw several thousand—up to13,000—UXOs in their own village within oneyear and that many more are still unidentified.

Households appear to be adjusting their liveli-hoods in several ways. Interestingly, only one outof six villages is developing “classical” improvedagriculture based on paddy and livestock. Allother five villages are reporting alternatives andplans that range from migration to innovativeincome sources (Table 3.10). Villagers seem to bemoving around without limitation from UXOs.Inhabitants in all villages collect NTFPs and goto the market regularly by foot. Decreasing ani-

mal assets may be an indicator of difficulties inadjusting. Half of the villages had lost large ani-mals since the war, and another 50 percent (notnecessarily the same ones) have decreasing or sta-ble animal numbers today.

Discussion

The PEN study confirms that UXOs do notprotect forests from encroachment and thatdecontamination has no potential for negativeconsequences on forest cover. First, decontam-ination is proceeding at a very low pace. About5,500 hectares have been decontaminated be-tween 1999 and 2004, an average 16 hectaresper district per year. Second, UXO contamina-tion is a powerful factor limiting reclamation ofpaddy fields and decrease of swidden agricul-ture. The industrial forestry sector is importantin Kaleum and elsewhere in the IndochineseCordillera26 but, except for NTFPs, it is mostlyunrelated to households and was not covered bythe study. UXOs are certainly an issue for this sec-tor, too, and its linkage to environment deservesanalysis.



Village 1 2 3 4 5 6New or old village Migrated Old Migrated Migrated Migrated Old village

from east, village from from recentlyconsolidated nearby nearby from

village village village nearby,consolidated

village

UXO impact

Animal assets Since warSince 2000

Paddy developmentsince 2000

Livelihoods Today

Tomorrow

Source: PEN village survey (ERI). HH = household.

T A B L E 3 . 1 0 Past & Current Trends in Six Villages Surveyed

Low/Moderate

DecreaseIncreaseLarge area95% HHPaddy & live-

stock dvt

Moderate

IncreaseDecreaseSmall area10% HHSeasonal

migrationSawmill

Moderate

DecreaseStableNoneMetal scrapTourism

High

DecreaseIncreaseSmall area95% HHNTFPs Out-

migration

Severe

IncreaseDecreaseSmall area20% HHMetal scrap

Fruit trees

Moderate/High

IncreaseStableSmall area10% HHGold

panningPaddy dvt

UXOs may be contributing to rotational agri-culture with much shortened cycles together withthe government’s policy of shifting agriculturestabilization. The combination of both issues maycreate unique difficulties for the local population.Making the new farming systems sustainable ischallenging. Few locations in UXO-affected areasare appropriate for opening sufficient areas ofpaddy fields. There is no indication of UXOs lim-iting animal husbandry, but the difficulties ofadjusting livelihoods appear to result in severalcommunities’ selling out their animal assets.

A clear national policy statement is needed tolink UXO decontamination to shifting cultiva-tion stabilization, but actual results appear to belimited in the district surveyed. Very few villagesare prioritized for the opening of paddy fields.Access is obviously one of the conditions of villageselection. Within a village, some communities optfor sharing the new paddy fields, while in otherspaddy fields are operated by a few families only.

UXOs are not so much a natural resourcemanagement issue as a strongly limiting factorto overall rural development. The PEN studydemonstrates that even villages with no UXOsare indirectly affected by them in districts withheavy UXO contamination. This is especiallythe case in districts where UXO impact is con-centrated closer to the district town.

The livelihood strategies observed in the vil-lage survey also point to the limitations of agricul-ture in most villages and to the need for support ofa more diversified rural economy. However, pri-oritization of decontamination efforts may bemore critical for construction works than foragriculture. A high proportion of the districtsuse decontamination mostly for purposes otherthan agriculture.27 This is perhaps why in 1998,roads—and not UXOs—were cited as the mainproblem by village leaders (Chaignon 2000).The private sector is also playing its share indecontamination.

Prioritization of UXO-related support is,therefore, especially relevant at the district leveland may relate to all sectors. In 2004, the Lao

UXO program was planning to work only in dis-tricts listed as poor in NGPES, including two-thirds of very poor districts (Lao UXO program2004). With widely spread UXO contamina-tion, it is understandable that the more denselypopulated lowland and urban areas may haveattracted support in the past. Yet, within districts,UXO-related activities—with the importantexception of awareness raising, which is alreadycovering broad numbers of villages—do need tobe targeted for expansion.

Targeting villages with easy road or riveraccess may be the only solution, but doing so inthe framework of the village resettlement andconsolidation policy is particularly questionable.Kaleum is not the only district where UXOs areconcentrated in areas targeted for resettlement.28

Local governments in marginal districts do nothave the capacity to identify safe locations fornew settlements and to plan timely decontami-nation. Their energies would be better employedin planning infrastructure for existing villagesand supporting communities’ initiatives for liveli-hood diversification.

POVERTY AND ENVIRONMENTAL HEALTH

Water Supply & Sanitation

Achieving equitable access to sustainable watersupply and sanitation services is essential for pro-tecting the health of the poor and improving theirquality of life. Poor households tend to have moreyoung children and elderly than non-poor house-holds. These age groups are most vulnerable to dis-ease. Poor households have less access to qualityhealth services, and can least afford the cost ofmedical treatment and income losses from disease.

Increasing household access to improved watersupply and basic sanitation is viewed by theGovernment of Lao PDR as an important partof poverty alleviation and socioeconomic devel-opment. The Sixth National Socio-EconomicDevelopment Plan 2006–10 aims at increasing



national water supply coverage rates by severalpercentage points in the next five years.

Substantial progress has been achieved in pro-viding the population with safe and accessiblewater supply in the last decade. Improved watersupply coverage increased from less than 20 per-cent in 1995 to 55 percent of the populationin 2003.29 The population share with protectedgroundwater tripled during the period and theshare with unprotected groundwater declined bynearly 10 percentage points. The most strikingachievement was the reduction in the popula-tion share relying on open, unprotected watersources, such as rivers, ponds and streams, whichdeclined from nearly 50 percent to 22 percent.Other types of water supply, including gravity-fed systems, increased from practically nonexis-tent to 18 percent of the population.

The increase in sanitation provision has beensubstantially slower than the increase in improvedwater supply, but is still significant. The popula-tion share with no sanitation facilities (“no toilet/latrine”) declined from 70 to 50 percent from1995 to 2003, and pour-flush toilets now serve37 percent of the population.

The poor have not benefited enough fromthese improvements. Of particular concern forenvironmental health is the population continu-ing to rely on open, unprotected water sources,such as surface water. In 2003, about 30 percentof the poor relied on surface water, compared to20 percent of the non-poor (LECS III). While thisrepresents a substantial decline from 1995, therewas no relative gain for the poor.30

The situation is more pronounced for basicsanitation. Nearly 70 percent of the poor did nothave toilet facilities in 2003, compared to slightlymore than 40 percent of the non-poor The non-poor population with no toilet facilities declinedby 21 percentage points, while for the poor, thedecline was only 13 percentage points from 1995to 2003. The gap between the poor and the non-poor therefore seems to have increased.31

The inequity in access to improved watersupply and sanitation is not only an urban-rural

disparity. More than a third of the poorest ruralhouseholds depend on surface water comparedto 23 percent of the richest rural households.32

Almost 80 percent of the poorest rural house-holds do not have basic sanitation, compared to43 percent of the richest (Figure 3.4).

The reliance on open, unprotected watersources (surface water) and the lack of sanitationvaries greatly across provinces and for non-poorand poor households. More than 50 percent ofthe poor in Luangphrabang and Houaphanhprovinces use surface water, and more than80 percent of the poor have no toilet/latrinein Phongsaly, Luangphrabang, Khammouane,Savannakhet, and Saravane. Access also varies dur-ing the rainy and dry seasons. In the South, thereare significant differences across the two seasons.

In 1995, district population water supply andsanitation coverage rates were strongly associatedwith urban population share, villages’ access toroads and their education level, and to someextent UXO contamination.33 This continues tobe the case today. About 28 percent of the ruralpopulation relies on surface water, compared toonly 4 percent in urban areas. Population withtoilet facilities exceeds 85 percent in urban areas,



No water supply* No sanitation facilities

M A P 3 . 8 Population with No Water Supply and Sanitation

Source: LECS III 2002/03. *Population with drinking water directly fromrivers and other open, unprotected water sources.

but is only 40 percent in rural areas. Of the pop-ulation with access to roads, less than 20 percentuse surface water and nearly 60 percent have toi-let facilities. Of the population without access toroad, more than 40 percent use surface waterand only 20 percent have toilet facilities. Havingno water and sanitation services is also stronglyassociated with household literacy level. Only 18 percent of households rely on surface water inwhich more than half of the members can read,while nearly 35 percent of households use sur-face water in which less than half of memberscan read.34 Rural areas lag far behind the urbanwith respect to literacy levels.

An analysis was undertaken to assess the roleor association of each of these factors withhousehold use of surface water and lack of a toi-let facility.35 The analysis enabled a comparisonof households that are generally similar exceptwith respect to one factor of the assessment. Theanalysis resulted in the following conclusions:

• Rural households are 7.7 times more likelythan urban households to rely on surfacewater.

• Households in villages with no access to roadsare 1.9 times more likely to rely on surfacewater than households with road access.



0%

5%

10 %

15 %

20 %

25 %

30 %

35 %

40 %

S u r f a c e W a t e r 3 5 % 3 3 % 2 8 % 2 7 % 2 3 %

1 2 3 4 5

0%

10%

20%

30%

40%

50%

60%

70%

80%

No Sanitation 7 9 . 0 % 7 2 . 4 % 6 0 . 8 % 5 5 . 2 % 4 2 . 5 %

1 2 3 4 5

Source: Calculated from LECS III, grouping households from poorest (=1) to richest (=5)

F I G U R E S 3 . 4 Reliance on Surface Water and Lack of Sanitation in Rural Areas in 2003

• Households in which more than half of themembers cannot read are 20 percent more likelyto rely on surface water than households inwhich more than half of the people can read.

• The presence of UXO problems in a village isassociated with a 34 percent higher likelihoodof using surface water.

• Controlling for the above factors, being poorincreases the likelihood of relying on surfacewater by 15 percent.36

There is an even stronger association betweenthe same factors and lack of sanitation; that is,no toilet/latrine:

• Rural households are 8 times more likelythan urban households to have no sanitationfacilities.

• Households in villages with no access to roadsare 3.4 times more likely to lack sanitation facil-ities than households with road access.

• Households in which more than half of themembers cannot read are 4.3 times more likelyto lack sanitation than households in whichmore than half of the members can read.

• Having UXO problems in the village is asso-ciated with a 7.8 times higher likelihood ofnot having sanitation.

• In addition, controlling for the above factors,being poor increases the likelihood of not hav-ing sanitation by 60 percent.

These factors have supply and demand aspects.Household or community location and roadaccess are likely to have influenced supply deci-sions in the water and sanitation sector, with pri-ority given to urban areas, easily accessible ruralareas, and areas with low risk of UXO accidentsand/or low cost of UXO clearance. These supplydecisions, among other issues, may be driven bycost considerations as well as concerns about thesustainability of water and sanitation services inremote rural areas.

On the demand side, road access, educationand literacy, and poverty may influence the capac-

ity of communities and their perceptions of pri-orities to obtain and sustain water and sanitationservices. Remote communities often have lesscontact with and less access to water and sani-tation decision makers. Moreover, less edu-cated communities may be less skilled incommunicating needs. They may also have lessappreciation and/or knowledge of health risksassociated with lack of water supply and sani-tation. Poverty also influences demand for waterand sanitation, as poor communities havefewer resources to contribute to constructionand operation/maintenance in rural areas andto water connection fees and water tariffs inurban areas.

Urban Water Supply and Sanitation

The Water Supply Authority Lao PDR (WASA),responsible for urban water supply, emphasizes theimportance of social fairness, quality of services,consumer satisfaction, and financial sustainabilityin the water sector (Annual Water Sector Perfor-mance Report 2003).

There are more than 140 urban centers inLao PDR, corresponding to district capitals.Half of these centers have a population less than4,000–5,000. In addition, there are many ruralvillages with urban characteristics. Nearly 60 per-cent of the urban population resides in the18 provincial centers, of which more than three-fourths live in the largest five.37

Nearly 75 percent of urban households haveimproved water supply, while 25 percent con-tinue to rely on unprotected well water and openwater sources. However, according to a 2002report entitled Small Towns Water Supply andSanitation Initiative in Lao PDR, more than 100of the district centers do not have a piped watersupply with house connections.

By household living standard, 60 percent ofthe poorest urban households and more than80 percent of the richest urban households haveimproved water supply (Figure 3.5). The dispar-ity is, however, larger for piped water supply



with a house connection. Nearly 30 percent ofthe poorest households have house connection,while more than 45 percent of the richest haveconnection.

Over 90 percent of urban households havebasic sanitation facilities. Nearly 80 percent ofthe poorest households and more than 95 per-cent of the richest households have basic sanita-tion (Figure 3.6). Quality and hygienic safety ofthese facilities vary substantially within andbetween urban areas and across levels of livingstandard.

A Look at Two Small Towns

A PEN study was carried out in Phongsaly andLamam towns to gain a better understanding ofthe needs and priorities in the water and sanita-tion sector for the poor in small towns in LaoPDR (Box 3.2). The towns are provincial capi-tals of two of the poorest provinces in the coun-try, and have populations less than 10,000. Bothtowns have a piped water supply network serv-ing a majority of the households, but no sewagenetwork.



0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

1 2 3 4 5

P i p e d W a t e r P r o t e c t e d g r o u n d w a t e r "Other" water

U n p r o t e c t e d g r o u n d w a t e r O p e n w a t e r s o u r c e s

Source: Estimated from LECS III (2002/03). Note: Improved water supply is here defined as piped water supply, pro-tected groundwater, “other” water (mainly gravity-fed systems), and rainwater (1 to 2 percent of urban households;not included in the figures above). Note: Living standard quintiles are consumption-based estimates from LECS III,with quintile 1 being the poorest households.

F I G U R E 3 . 5 Urban Household Water Supply by Living Standard Quintile

0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

T o i l e t F a c i l i t y 7 8 .2 % 8 5 . 4 % 8 8 .8 % 9 0 . 7 %

1 2 3 4

Source: Estimated from LECS III (2002/03). Note: Living standard quintiles are consumption-based from LECS III, withquintile 1 being the poorest households.

F I G U R E 3 . 6 Percent of Urban Households with Toilet Facility by Living Standard Quintile



B O X 3 . 2

The Urban Research Institute (URI) in Vientiane implemented a study on poverty and urban watersupply and sanitation in Phongsaly and Lamam towns.* The study assessed types of household watersupply and sanitation facilities, piped water connection coverage, user satisfaction, and factors influ-encing demand for piped water supply and improved sanitation in poor and non-poor households.A specific aim of the study was to identify policies and interventions to promote equitable access towater and sanitation services that provide health protection and improved living conditions for poorand vulnerable households.

The study included a survey of about 550 households, questionnaire interviews with the districtwater utilities (Nam Papa), and interviews with district authorities and community leaders. Six outof nine urban villages were surveyed in Phongsaly town and six out of six in Lamam town. Assetscores and self-reported living standard levels were used to identify poor households.

*URI. 2006. Water Supply and Sanitation Situation in Relation to Poverty in Two Urban Districts, Lao PDR. PENreport prepared by URI. Vientiane.

Household Living Standard (1=low; 5=high) Household Owns House and Land

0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

1 2 3 4 50 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

N o Y e s

Education Level of Household Respondent Location of House

0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

No schooling Primaryschool

Secondaryschool

Uppersecondary

school

0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

Walking path Small road Secondaryroad

Main road

F I G U R E 3 . 7 Percent of Households with Piped Water Supply in Phongsaly

Piped water connection. In Phongsaly, the pipedwater connection rate is less than 40 percentfor the poorest households but more than80 percent for the richest. Affordability alonedoes not explain this disparity. Householdeducation level, land and house ownership, and

location of the house relative to the main road alsoaffects the water connection rate (Figure 3.7).38

The lowest connection rate is therefore amongpoor households with low education level whodo not own their property and who reside alongthe walking paths.

In Lamam, piped water connections are alsoassociated with living standards and householdownership of house and land. The influence ofeducation levels is weaker than in Phongsaly,and the connection rate is not affected by house-hold location relative to roads.

Piped water user satisfaction. Nearly 80 percentof households in Phongsaly reported satisfactionwith the piped water supply. The highest ratesof satisfaction were in locations closest to thewater distribution point. Only 60 percent ofhouseholds were satisfied in the villages most faraway from the distribution point. Satisfaction isalso lower among households reporting thattime savings were of high importance in theirdecision to connect to the piped water supply,and among households reporting that thepiped water supply is insufficient to providetheir most important water needs. Thesefinding are understandable in light of the lowquantity of water that the piped water networkis capable of providing in Phongsaly. Mosthouseholds therefore need to supplement theirwater demand from other sources, which inmost cases is a time-consuming endeavor.Education level also seems to have a subtantialinfluence on satisfaction with the piped watersupply in Phongsaly (Figure 3.8).39

In Lamam, more than 85 percent of house-holds are satisfied with the piped water supply,although satisfaction reaches only 60 percent inone of the villages. Households that reported timesavings as an important motivation for connect-ing to the water supply were more satisfied thanhouseholds for which time savings was of minorimportance. This is contrary to the findings inPhongsaly, but is consistent with the substantiallyhigher volume of water supplied by the system inLamam. The rate of satisfaction is about the samefor poor and better-off households, and satisfac-tion is marginally higher among households withhigher education level.

It is not clear why satisfaction with a pipedwater supply is associated with household edu-cation level. One reason could be that more edu-cated households have a higher appreciation forthe potential hygiene and health benefits of hav-ing convenient access to increased quantity andquality of water.

Water tariffs. Water tariffs in Phongsaly areamong the highest in the country. Nevertheless,85 percent of the poorest households and70 percent of other households reported thattariffs are reasonable. In Lamam, nearly 50 per-cent of households—poor and rich—reportedthat tariffs are reasonable.



0 %

1 0 %

2 0 %

3 0 %

4 0 %

5 0 %

6 0 %

7 0 %

8 0 %

9 0 %

1 0 0 %

No school ing Primary school Secondary school Upper secondaryschool

F I G U R E 3 . 8 Household Satisfaction and Education Level in Phongsaly

Even though water tariffs per cubic meter inLamam are less than half the tariffs in Phongsaly,household water expenditure is higher in Lamam.This may explain why fewer households inLamam report that water tariffs are reasonable.Household water consumption is several timeshigher in Lamam than in Phongsaly, mainlybecause water is less expensive and supply is moreabundant in Lamam.

Household expenditure on piped water rangesfrom 1 to 3 percent of total household expenditureon all goods and services, and from 4 to 12 percentof non-food expenditure. A large user spends morethan a small user as a percent of the household’s

total expenditure, but less than a small user as apercent of non-food expenditure (Table 3.11).

Water connection fee. The household waterconnection fee in Phongsaly is Kip 500 thousand(nearly $50), and Kip 900 thousand (nearly$90) in Lamam. In Phongsaly, this is nearly10 percent of annual household expenditure onall goods and services for the poorest households,and 2 percent of total expenditure of the richesthouseholds. As a share of household non-foodexpenditure, the connection fee is 47 percent forthe poorest households and 3 percent of therichest households (Figure 3.9). In Lamam, the



T A B L E 3 . 1 1 Household Water Bill in Phongsaly and Lamam

Small user Large user

Phongsaly Lamam Phongsaly Lamam

Per capita water consumption per day (liters) 20 60 100 250Household water bill (000 Kip per year) 60 130 575 770Water bill % of household total expenditure 1.1% 2.4% 2.0% 2.7%Water bill % of household non-food expenditure 5.7% 12.4% 4.0% 5.3%

Note: Calculated based on block tariffs and annual household water consumption data from DCTPC, and urban house-hold expenditure quintiles from LECS III. The lowest quintile (poorest households) is applied in the calculation of waterbill as a percent of consumption for “small user” and the highest quintile is applied for “large user.” Urban VientianeMunicipality is excluded to better reflect urban household expenditure and living standards in smaller towns.

Phongsaly Lamam

0%

20%

40%

60%

80%

100%

1 2 3 4 5

% of total % of non-food

0 %

2 0 %

4 0 %

6 0 %

8 0 %

1 0 0 %

1 2 3 4 5

% of total % of non-food

Note: 1 is lowest living standard quintile and 5 is highest.

F I G U R E 3 . 9 Connection Fee as Percent of Annual Household Consumption

water connection fee is as much as 87 percent ofannual non-food expenditure for the pooresthouseholds and 6 percent for the richest house-holds. This may explain why Lamam has a lowerconnection rate than Phongsaly.

In Phongsaly, 75 percent of the householdswithout piped water, but with network availablein their section of the town, stated that the highconnection fee was the most important reasonfor not connecting to the piped water supply. InLamam, all households without piped watersupply reported that the high connection fee wasthe main reason for not obtaining a connection.

Basic sanitation. More than 95 percent of thesurveyed households had toilet facilities inPhongsaly, ranging from 90 percent among thepoorest households to 100 percent among therichest households. In Lamam, nearly 80 percentof surveyed households had toilet facilities. Lessthan 45 percent of the poorest households hadtoilet facilities, compared to more than 95 percentof the richest households (Figure 3.10).

Living standards, education levels, and thepresence of elderly family members in the house-hold explain much of the difference in the typeof toilet facilities in households in Phongsaly.Pour-flush toilets were much more frequent inricher households, and dry pit toilets more fre-

quent in poorer households. Education levels alsohave some influence on the type of toilet. House-holds with higher education had flush toiletsmore frequently than households with low edu-cation.40 Pour-flush toilets were much more fre-quently used in households with elderly than inhouseholds without elderly household members.In contrast to factors influencing piped waterconnection, the location of a house in relation tothe type of road and ownership of house and landwere not found to influence households’ choiceof toilet facility in Phongsaly.

In Lamam, four factors were identified thatinfluenced whether households had a toilet facil-ity and the choice of facility. Education levelwas found to have a larger influence in Lamamthan in Phongsaly. House location also mat-ters. Households living away from main roadsand along walking paths are less likely to havea pour-flush toilet with septic tank or to have atoilet at all. Households not owning a houseand land had a pour-flush toilet with septic tankless frequently.

Vulnerable croups. Incidence of diarrheal illness,and mortality from diarrheal illness, is generallyhighest among young children and elderlyindividuals. Piped water supply and sanitationfacilities, which makes it easier to practice good



Phongsaly Lamam

0%

20 %

40 %

60 %

80 %

100%

1 2 3 4 5

0%

20%

40%

60%

80%

100%

1 2 3 4 5

Note: Lowest living standard=1 and highest living standard=5

F I G U R E 3 . 1 0 Percent of Households with Toilet Facility

hygiene, are therefore particularly important forhouseholds with young children and elderly.

In both Phongsaly and Lamam, the pipedwater connection rate is about the same forhouseholds with children or elderly as in house-holds without children or elderly. With respectto basic sanitation in Phongsaly, householdswith children have the same type of facilities ashouseholds without children. Households withelderly individuals, however, more frequentlyhave pour-flush toilets than households with-out elderly individuals. In Lamam, householdswith children and elderly are less likely to havea pour-flush toilet and more likely not to havea toilet facility.

Sustainability of Rural Water Supply and Sanitation

Over the last decade, there has been significantprogress in providing improved water supply andsanitation systems to the rural population in LaoPDR. Water and sanitation village committeeswere promoted to improve the sustainability ofthese systems and community and householdhygiene. The water and sanitation systems have,however, not been free of problems. The Ministryof Health (MoH), which is responsible for ruralwater and sanitation, was therefore interested ina study on factors influencing the non-sustain-ability of rural water supply and sanitation. Sucha study was included as a PEN study, with a par-ticular focus on poverty.

A sustainable water supply and sanitation sys-tem is defined as the system providing an accept-able level of services throughout its design life.This includes acceptable water quality, quantity,and continuity, and the physical reliability of thesystem to deliver services. To this definitionshould be added healthy hygienic conditions ofthe system, and a community and householdhygiene that helps provide the maximum healthbenefits of the system.

The PEN study was carried out in 32 villagesin four provinces in the North and South of LaoPDR under the leadership of the Department ofHygiene at the Ministry of Health (Box 3.3).The water supply consisted of gravity-fed sys-tems (GFS) in 23 villages, and of bore holes innine villages (Table 3.12). More than 60 percentof the water supply systems had experiencedbreakdowns in the prior 12 months, and 30 per-cent of the systems had more than five break-downs. In nearly 40 percent of the cases, it tookmore than five days to repair the systems. Theperception of households and village leadersand committees was that poor management wasan important factor in the breakdowns. About75 percent of village leaders and committeesstated that poor management was one of thecauses, and 35 percent of households stated thatpoor management was the main cause of break-downs (Figure 3.11).

The sanitary inspections of the gravity-fedwater supply systems found that proper mainte-nance and precautionary measures to prevent the



B O X 3 . 3

The Department of Hygiene and Prevention at the Ministry of Health implemented a study onsustainability of rural water supply and sanitation in four poor districts in Lao PDR.* The study set outto identify factors influencing non-sustainability, looking at the quality of community management,community participation, inspections and maintenance, and community and household hygiene.Survey questionnaires were developed and administered in 32 villages to over 600 households,32 village committees or village leaders, and several district offices. Sanitary inspections wereconducted for all the water systems, 100 latrines, and more than 250 households.

*Ministry of Health. 2005. Factors influencing the non-sustainability of improved water supply and sanitation programs in rural areas in poor provinces in Lao PDR. PEN report prepared by Environmental Health Division,Department of Hygiene and Prevention, Ministry of Health, Vientiane, Lao PDR.

risk of water pollution or to keep good hygieneconditions were inadequate in more than 30 per-cent of cases. This included accumulation ofstagnant water around the stand post, risk of pol-lution through faulty pipes, unscreened andunfenced water intake installations, and damagesto stand post platforms. Risk of pollution fromupstream human settlements, farm animals, andcrop production was also identified, but muchless frequently (Figure 3.12).

Sanitary inspections of 21 bore holes in thenine villages with bore holes also found conditionsin need of improvement. Inadequate fencingaround hand pumps to keep out animals, stagnantwater accumulations, and other nearby sources ofpollution were observed in more than 40 percentof cases. Latrines within 10 meters of the handpump were also found in a few cases (Figure 3.13).

Soap was available in only 6 out of 100 latrines,and water tanks were absent in 60 percent of the



T A B L E 3 . 1 2 Study of Rural Water Supply and Sanitation Sustainability

Province Phongsaly Oudomxay Sekong Attapeu

District Notou Namor Kaleum Phouvong

Number of Villages 8 8 8 8Number of Households per District 156 160 147 160Type of Water Supply System in villages GFS 7 GFS GFS bore holes

1 bore holeNumber of Latrine Inspections 45 40 0 15Household Hygiene (number of 64 64 64 64

households per district)

Notes: GFS=gravity fed system

0% 10% 20% 30% 40% 50% 60% 70% 80%

Poor management is main cause ofbreak-down

Poor management is one of thecauses of break-down

Break-downs > 5 times per year

Number of days to repair is > 5 days

WS break-down in the last 1 year

F I G U R E 3 . 1 1 Water Supply (WS) Break-Downs

cases. Cleanliness was unsatisfactory in 40 of thelatrines, and 20 to 30 of the latrines sufferedfrom poor physical conditions (Figure 3.14).

Several factors were associated with watersupply breakdowns and frequency of break-downs in the 32 villages. Some of the mostimportant factors included low demand respon-siveness; no financial fund for proper operationand maintenance and repairs; no regular inspec-tion of systems; and low education levels in the

village. Villages with ethnic minorities seemedalso to face more difficulties in avoiding break-downs, and so did poorer villages (Table 3.13).

Demand for the water supply systems origi-nated from the villagers in nearly 70 percent of thecases, or the village in nearly 95 percent of thecases. More than 90 percent of households pro-vided labor and materials. Post-project monitor-ing took place, however, in only 55 to 60 percentof the cases, and regular sanitary inspections of the



0% 10% 20% 30% 40% 50%

Upstream human settlements polluting the water source

Upstream crop production polluting the water source

Damages to water tank

Upstream farm animals polluting the water source

Damages to stand post platform

Unfenced water intake installation

Unscreened intake installation

Risk of pollution through water pipe

Stagnant water around stand post

F I G U R E 3 . 1 2 Sanitary Inspections of the Gravity Fed Water Supply Systems

0% 10% 20% 30% 40% 50%

Latrine within 10 m of handpump

No concrete floor around thehand pump

Other sources of pollutionwithin 10 m of hand pump

Stagnant water within 2 m ofhand pump (faulty drainage)

Inadequate fencing aroundhand pump

F I G U R E 3 . 1 3 Sanitary Inspections of the Bore Holes



0% 20% 40% 60% 80% 100%

Bad condition of latrine floor

Bad condition of latrine house

Bad smells

No water tank in the latrine house

No soap in the latrine house

F I G U R E 3 . 1 4 Household Latrine Inspections

T A B L E 3 . 1 3 Factors Associated with WS System Breakdowns

Low demand responsiveness (e.g. village participation)No financial fund for proper O&M and repairs

No post-project monitoring by government and international agenciesNo regular inspection of WS system

Low education levelEthnic minorities

Poverty

Note: Factors were statistically significant at 95 percent level (Chi-square).

systems were performed in less than 40 percent ofthe villages. The poorest performance was withrespect to having a well-functioning water andsanitation village committee, training of villagecommittees, and household collections for anoperations and maintenance fund (Figure 3.15).

Discussion

Over the last decade, there was significantprogress in providing improved water supply andbasic sanitation. The poor, however, lag farbehind, and no relative gains were achieved. Fourfactors explain much of the low coverage rates in

many poor communities: rural location, no vil-lage access to a road, illiteracy or low education,or UXO contamination. The poor have lowercoverage rates of improved water supply andbasic sanitation in urban areas too. Low educa-tion and unaffordability of services explain someof this disparity. The study in Phongsaly andLamam confirms that these factors are importantin small towns too, pointing to the high waterconnection fees and low education levels as themain reasons for low connection rates amongpoor households. Other factors include house-holds not owning their property or a locationaway from main roads.

Provision of rural water supply and sanitationwill need to be further invigorated if equity inaccess to services is to be achieved. Particularattention is needed for communities with noaccess to a road and in areas with high UXO con-tamination, as these communities are among thepoorest in the country. Sustainability of servicesfor these communities is an important considera-tion. They tend to have higher illiteracy or lowereducation levels, pointing to a strong need fordeveloping capacity for sustainable managementof water supply and sanitation services.

Poor rural communities seem to be moreinflicted with water supply system breakdownsthan better-off communities. Inadequate manage-ment is most frequently reported as the main causeor important cause of these breakdowns. Poorcommunities are also less likely to establish villagefunds for maintenance and repairs. Hygieneimprovements are also much needed to fully ben-efit from improved water supply and sanitation.This includes regular inspection and properhygiene maintenance of water supply systems,appropriate community behavior in relation towater use, and improved household and personal

hygiene. These factors all point to the need forincreased emphasis on community capacity build-ing with proper management training, commu-nity participation, information and awarenesscampaigns, and instituting procedures to establishvillage funds for maintenance and timely repairs.

In urban areas, the household connection feeis of critical importance. Its high level is one of themain reasons for poor households not obtaining awater connection. The fee is a substantial share ofpoor households’ annual expenditure on all goodsand services and non-food expenditure. As finan-cial sustainability of water agencies is an impor-tant objective of the government, any policy toadjust the connection fee would need to strive tobe revenue-neutral.

Household education level is an importantdeterminant of demand for improved watersupply and basic sanitation. Information andpublic awareness programs about the benefits ofimproved services—such as opportunities forimproved household hygiene and potentiallybetter water quality—can therefore be an impor-tant instrument for encouraging household con-nections. This is particularly important among



0% 20% 40% 60% 80% 100%

Money collected for WSS fund

Village Committee participated in training

Village has WSS Village Committee

Village Committee sanitary inspections > 8 times/year

Financial contribution from village/villagers

International agency monitoring of WS after completion

Government monitoring of WS after completion

Demand for WS originated from villagers

Contribution of material from village/villagers

Demand for WS originated from village

Contribution of labor from village/villagers

F I G U R E 3 . 1 5 Community Participation and Water Supply System Management

poor households, because they generally haveless education. Public awareness programs shouldalso target households with young children andelderly household members. These groups aremore at risk of disease.

Satisfaction with the piped water supply sys-tem was found to be higher among householdswith higher education levels. More educatedhouseholds may appreciate the value of havingpiped water, just as their demand for piped wateris higher than households with lower educationlevels. Public awareness programs communicat-ing the values of having improved water supplymay therefore also help increase satisfaction inaddition to increasing connection rates.

Poor households in urban areas are also affectedby network availability. Piped water is sometimesnot available in the area of the town where thepoor live. This is the case in some areas of bothPhongsaly and Lamam towns, and especially forhouseholds living along walking paths and awayfrom main roads. As these households tend to bepoorer than other households, it is important thatplanning of network expansions should considerequity.

For new water and sanitation projects insmall towns in Lao PDR—and for networkexpansions and service improvement in townswith existing piped water supply—an increasedemphasis on a participatory approach, andinformation and awareness campaigns in projectpreparation can help enhance user satisfaction,ensure appropriate service levels, allow a tariffstructure and connection fees that promoteequity, and help ensure that household expecta-tions are realistic and enable households to makeinformed decisions with their scarce financialresources, particularly in poor communities.

OTHER ENVIRONMENTAL ISSUES

Indoor Air Pollution

More than 95 percent of the population in LaoPDR uses fuel wood for cooking (LECS III). Lit-

tle is known about the level of exposure to fuelwood smoke and associated health effects in LaoPDR. Initial research by Laotian institutions41

seems to confirm concerns raised by WHO thateye and lung irritation from fuel wood smokeappears to be a substantial issue, especially amongwomen and young children.

Poor people do not have sufficient funds to buysimple energy saving wood stoves. About 95 per-cent of the non-poor population and as much as98 percent of the poor population are using woodfor cooking. Similarly, 98.7 percent of all house-holds are using wood for cooking in the group ofpoorest districts compared to 80.9 percent in thegroup of least poor districts. The poor are, there-fore, likely to be disproportionately affected byindoor air pollution. However, the number ofpeople affected depends on factors such as indoor-outdoor cooking practices, type of housing andventilation, and type of fuel wood consumed.

Malaria

According to WHO data, the total number ofconfirmed malaria cases declined from 80,000 in1996 to 27,000 in 2001. Based on regressionanalysis of province-level LECS III data, an esti-mated 1 to 2 percent of the poor and 0.2–0.4 per-cent of the non-poor were affected by malaria.This, however, only relates to confirmed cases ofmalaria and is likely to underestimate the numberof actual cases.

A substantially higher number of people maybe also affected by the threat of malaria in the sensethat they may restrict their activities during thepeak malaria season to avoid contracting the dis-ease. The PPA (ADB 2001) states that 64 percentof participating villages reported malaria to be amain health issue. While six years of data are insuf-ficient to draw firm conclusions about trends, itappears that the non-poor have benefited morefrom the reduction in annual malaria cases dur-ing the 1996–2001 period: 72 percent of malariacases are estimated to have occurred among thepoor in 1996, a figure that increased to 80 per-cent in 2001.



Natural Disasters

The impact of floods and droughts on the poorversus the non-poor was estimated through aregression analysis using 1995, 1996, 1997, and2002 data from the Natural Disaster Manage-ment Office in Lao PDR.42 The results obtainedpoint to a strong indication that floods largelyaffect the non-poor, while drought and otherdisasters largely affect the poor. It also seemslikely that the number of poor affected by droughtis higher than the number of poor affected byfloods, even though the area affected by floodingis substantially higher than the area affected bydrought.

POVERTY-ENVIRONMENT IN THENATIONAL CONTEXT

The findings from the PEN focus studies in theNorth and the South, along the nationwideanalysis, are broadly relevant to upland districts,remote border districts along the Vietnameseborder and the mountainous sections of theChinese and Cambodian borders, districts witha high share of ethnic minority people, and dis-tricts with high UXO contamination. A largepercent of districts in Lao PDR and all 46 prior-ity districts share at least one of these criteria(Table 3.15).

The natural resource management PEN stud-ies, although located in marginal districts, shedlight as a whole on the set of policies through

which the Lao PDR government seeks to sup-port natural resource conservation in its uplandareas. The government has set impressive targets:stabilization of shifting cultivation and eradica-tion of opium production. The land use planningand land allocation policy is a main instrumentin the stabilization of shifting cultivation. The“focal area” policy, whereby government effortsfocus on a small number of villages closer to theroad network, was first designed as the ruraldevelopment policy of Lao PDR. It is presentedas the policy for community-driven rural devel-opment in poor districts in NGPES.

Village consolidation is proceeding at a steadypace. In 2003, the number of villages was downto less than 11,600 villages or 74 percent of the number of villages in 1997.43 Although thedatasets clearly indicate that the policy has beenimplemented more rapidly in lowland areas, it is



T A B L E 3 . 1 4 Estimated Number of Poor and Non-PoorAffected by Natural Disasters

Non-Poor Poor

Affected by flood (number of people) 149 000 16 000Affected by flood (percent of people) 5.1% 1.0%Affected by drought (number of people) 28 000 5 000Affected by drought (percent of people) 1.7% 0.2%

Note: Figures are based on 1995 population and do not take into accountpopulation growth between 1995 and 2002.

T A B L E 3 . 1 5 Number of Districts Sharing PEN Focus Study District Criteria

Villages withRemote Ethnic Minority Mon Khmer High UXOBorder People Above People Contamination

Districts Upland 1/ Districts 50 % Above 43% Above 15%

Lao PDR 74 70 78 39 37NGPES priority 33 24 42 25 18

poor districts

1/ Defined as districts with maximum elevation above 500 m.

also taking place in upland areas, often in combi-nation with the other rural development policies.The PEN studies provide evidence from three dif-ferent perspectives and lead to a single conclusion:while these policies may well reflect a sound visionfor the longer-term future of the uplands, theirrapid pace of implementation runs counter totheir stated objective of poverty reduction, andhas achieved limited gains in terms of resourceconservation.

The Mon-Khmer ethnic groups have a specialposition in the analysis of poverty and environ-ment linkages. They have long been recognized asindigenous groups with a strong tradition of sus-tainable rotational agriculture. It is increasinglyunderstood that they face specific poverty issues.The PEN study findings appear to have high rel-evance for the Khamu people in the northernregion and the various small groups along theVietnamese border in the South.44

The poverty and environment challengesfacing the Mon Khmer people in the South arereflected in the PEN focus studies. They oftenlive in villages with high UXO contamination.They live in sparsely populated, highly forestedareas, in which the national policy frameworkis conducive to proactive resettlement. Yet thePEN focus studies tell vivid stories of commu-nities that are adjusting, diversifying their liveli-hoods, entering the market economy, and lookingahead at a more community-based managementof local natural resources. Innovative solutionsare needed to address poverty and environmentjointly in these circumstances, and the PENstudies only provide some initial insights aboutsuch solutions.

Policy Implications andRecommendations

The following section provides some policyimplications and recommendations that emergefrom the PEN study. While many of these rec-ommendations are not necessarily new, the

PEN study provides further evidence of theirimportance.

Rural water supply & sanitation. Five keyrecommendations arise from the PEN study to achieve equity in access, sustainability ofservices, and health protection of the poor inrural communities:

• Rural communities without access to roads andin areas with high UXO contamination are dis-proportionately lacking improved water supplyand sanitation. These are among the poorestcommunities in Lao PDR. A strategy to reachthese communities needs to be developed. Thestrategy will need to take into account villageconsolidation plans and carefully evaluate costand maintenance implications of water supplyoptions for these communities.

• Sustainability of water supply and sanitationservices is essential for achieving equitableaccess and health protection of the poor. Estab-lishment of well-functioning village commit-tees therefore needs to be further emphasized,and periodic assistance might be needed toensure their continuity and efficiency. This isparticularly important in poorer communities,which seem to have more problems with watersupply system breakdowns.

• Communities with high illiteracy or low edu-cation are more likely to not have improvedwater supply and basic sanitation. Particularattention is required in developing manage-ment capacity in these communities. Thisincludes proper training in water supply sys-tem management, developing mechanisms forcommunity participation, establishing regularinspections and maintenance procedures, andstrengthening of monitoring and evaluation ofwater and sanitation projects.

• Village funds are essential for proper systemmaintenance and timely repairs. Building com-mitment and procedures to institute a fund,even if gradually, therefore needs to be further



emphasized, even in the poorest communities.Further consideration should also be given toevaluating system options in light of mainte-nance and management requirements.

• Good community, household and water sup-ply system and latrine hygiene is essential inorder to fully benefit from improved watersupply and sanitation. Hygiene promotionprograms therefore need to be developed andimplemented. Improved hand-washing andpoint-of-use drinking water disinfection areparticularly effective in protecting health(Fewtrell and Colford 2004).

Urban water supply & sanitation. Avail-ability of improved water supply and sanitationis much more widespread in urban than ruralcommunities. However, much is still to beachieved in small towns, especially for the poor.Three priority recommendations have beenidentified:

• Equitable access to piped water supply in urbanareas requires a review of water connection fees.The fees are unaffordable for many poor house-holds. A revenue neutral policy change could beconsidered whereby water tariffs are raised tocompensate for a reduction in connection fees.This would prevent compromising the finan-cial viability of water agencies and quality ofservices.

• Network expansions should be equitable sothat the poor are not the last households toreceive water supply. This implies that socialconsiderations receive an equal footing withfinancial and technical considerations in net-work expansion. Particular attention is requiredfor so-called “temporary” housing and settle-ments further away from main roads.

• Affordability is not the only reason the poordo not have water connections or basic sani-tation. Education level is equally important.Information and public awareness of thebenefits of improved water supply and sani-

tation is therefore an important ingredientin achieving equity. In planning new pipedwater supply systems in other towns, under-standing perceptions of these benefits is there-fore a prerequisite.

Transport. The PEN road study identified fourrecommendations for poverty reduction andsound environmental management in Lao PDR:

• The NGPES’s focus on ensuring all-weatheraccess roads to all district towns is relevant,but policy and technical decisions to helpimprove the local road network beyond thatstage need to be taken now. A village roadnetwork will have substantial poverty reduc-tion impact, provided an appropriate prior-itization of villages to be served first isensured. This will require long-term plan-ning and management. Experience fromdonor-supported pilot projects deserves to besummarized and mainstreamed. Risks ofnegative environmental impact may be lim-ited overall, but should remain a concern inand around areas of environmental value.Confirming local responsibilities in naturalresource management appears to be a viableand efficient option.

• A comprehensive and up-to-date monitoringsystem of the local road network at the nationallevel is a key tool to manage the poverty andenvironmental aspects of rural road develop-ment. There is room for substantial improve-ment. A combination of improved databasetools and of more transparent informationflows between national and local levels of gov-ernment is needed.

• There is an urgent need to recognize the risksand opportunities of the involvement of privateenterprises from various sectors in road con-struction in Lao PDR’s many border districts.The capacity of enterprises to invest in roads isan opportunity in terms of poverty reduction.The government has an important role in mon-



itoring environmental impacts at the nationallevel. Enterprises need to be encouraged toselect sustainable road construction technology,so that the roads create a long-term povertyreduction effect. This objective may require thecreation of financial incentives.

• Narrow tracks are more cost-efficient andenvironmentally friendly than roads. Build-ing tracks instead of roads may attract moreinterest, as small tractors for farmers andmotorbikes for government staff visits to vil-lages become more widely available. It alsoremains an option worth investigating in thenear future.

Agriculture and forestry. The PEN study providesthe basis for two recommendations related toNTFPs:

• Systematic monitoring of NTFPs in terms ofresources, products, cultivation, and markets isa necessity in order to better manage risks ofresource decline and market collapse. Bothwould disproportionately affect the poor. Mak-ing cultivation of NTFPs a fully viable and rec-ognized economic activity is also needed in thatregard. This objective will require not only con-tinued applied research but also improvementsin the legal and policy framework.

• The land use planning and land allocation pol-icy, as stated in NGPES, needs to be mademore appropriate to the poverty reductionneeds of upland villages. The high motivationof local communities to better manage localresources is an opportunity. Mainstreamingsuccessful local pilots in community participa-tion can help take advantage of this opportu-nity. NTFP harvesting and marketing are anideal entry point for overall community forestmanagement in areas where NTFP collectionfor sale remains an important income source.

Cross-sectoral interventions. The comprehensivepolicy of the Government of Lao PDR for itsupland regions, which includes stabilization of

shifting cultivation, development of largercommunities (preferably along roads), andfocus of government efforts on thesecommunities, may be an appropriate vision inthe very long term. However, all the PENstudies provide evidence that the current highpace of implementation of this policy iscounterproductive, both in terms of sustainablepoverty reduction and environmental pro-tection. A slow, gradual transition in the landuse systems of upland communities appearsto be a much more appropriate option. TheGovernment of Lao PDR should use thisincreasing evidence provided by its researchersand donor-funded projects to turn to morepeople-centered, service-oriented ways ofsupporting communities that wish to migrateand/or intensify their land use practices. Providingsupport to communities off the road is a necessity,if this is to be achieved. This is not an easy task,and innovative programs will be needed.

The UXO study demonstrates that activedonor contribution to UXO decontaminationshould be encouraged. UXO decontaminationis a win-win solution in terms of poverty reduc-tion, improved environmental management,and providing safe water supply and sanitation.Decontamination efforts will, however, requirea very long period of time, and in the mean-time, all coping options for local communi-ties should be supported. In southern Laos, the incompatibility of meaningful support forUXO-affected communities with a proactiveresettlement policy must be recognized andaddressed. Diversified community initiatives,with or without migration and inside or out-side of agriculture, deserve careful attentionand support.

Any joint solution in terms of poverty reduc-tion and environmental protection merits carefulgeographical targeting and improved coordinationbetween stakeholders, especially more intensivecommunity participation. Table 3.16 provides asummary of PEN study findings related to eachrecommendation.





TA

BL

E 3

.16

Lao

PD

R S

um

mar

y o

f Pr

iori

ties

fo

r Pr

og

ram

s w

ith

Jo

int

Posi

tive

Po

vert

y R

edu

ctio

n-E

nvi

ron

men

t Im

pac

t

Co

ord

inat

ion

nee

ds

bet

wee

nD

evel

op

men

t ar

eas

Pro

gra

m p

rio

rity

Geo

gra

ph

ical

tar

get

ing

sect

ors

an

d a

gen

cies

All

sect

ors

Ag

ricu

ltu

re

& f

ore

stry

Inte

gra

ted

live

liho

od

inve

stm

ent

pro

gra

m in

pri

ori

ty p

oo

r u

pla

nd

dis

tric

ts

Sup

po

rt t

o d

iver

sifi

ed d

evel

op

men

t st

rate

gie

sin

UX

O-a

ffec

ted

dis

tric

ts•

Div

ersi

fied

inco

me

gen

erat

ion

incl

ud

ing

lives

tock

an

d o

ff-f

arm

•Sl

ow

tra

nsi

tio

n o

ut

of

swid

den

ag

ricu

ltu

reC

om

mu

nit

y p

arti

cip

atio

n in

nat

ura

l res

ou

rce

man

agem

ent:

•M

ore

par

tici

pat

ory

lan

d u

se p

lan

nin

gal

low

ing

fo

r sl

ow

tra

nsi

tio

n o

ut

of

swid

den

agri

cult

ure

•N

TFP

har

vest

ing

an

d m

arke

tin

g u

sed

as

an e

ntr

y p

oin

t in

co

mm

un

ity

reso

urc

em

anag

emen

tSu

stai

nab

le N

TFP

pro

du

ctio

n:

•C

om

pre

hen

sive

mo

nit

ori

ng

sys

tem

fo

rN

TFP

pro

du

ctio

n a

nd

mar

kets

•Po

licy

stat

emen

t, le

gal

fra

mew

ork

an

dap

plie

d r

esea

rch

pro

mo

tin

g c

ult

ivat

ion

All

villa

ges

wit

hin

dis

tric

ts,

bo

th a

lon

g r

oad

s an

daw

ay f

rom

ro

ads

Dis

tric

ts a

lon

g V

ietn

amb

ord

ers,

wit

h a

fo

cus

on

Mo

n-K

hm

er v

illag

es in

the

Sou

th

Up

lan

ds

dis

tric

tsN

ort

her

n r

egio

n f

or

NTF

Ps

All

sect

ors

, dis

tric

t g

ove

rnm

ent,

par

tici

pat

ion

of

villa

ge

com

mu

nit

ies,

NG

Os

Lao

-UX

O p

rog

ram

, dis

tric

t g

ove

rnm

ent,

par

-ti

cip

atio

n o

f vi

llag

e co

mm

un

itie

s, N

GO

s

Ag

ricu

ltu

re &

fo

rest

ry w

ith

tra

nsp

ort

Dis

tric

t g

ove

rnm

ent,

par

tici

pat

ion

of

villa

ge

com

mu

nit

ies

Nat

ion

al g

ove

rnm

ent

Dis

tric

t g

ove

rnm

ents

Ag

ricu

ltu

re &

fo

rest

ry d

epar

tmen

tR

esea

rch

org

aniz

atio

ns

(co

nti

nu

ed)



Hea

lth

Tran

spo

rt

Ru

ral c

lean

wat

er &

san

itat

ion

ser

vice

s:•

Imp

rovi

ng

th

e su

stai

nab

ility

of

serv

ices

inp

oo

r co

mm

un

itie

s th

at a

lrea

dy

hav

e se

rvic

es•

Co

mm

un

ity

and

ho

use

ho

ld h

ygie

ne

pro

mo

-ti

on

pro

gra

ms

•Ta

rget

ing

of

serv

ice

pro

visi

on

to

po

or,

mar

gin

aliz

ed a

nd

dis

adva

nta

ged

co

mm

u-

nit

ies

(th

e o

nes

th

at a

re le

ast

likel

y to

attr

act

serv

ices

on

th

eir

ow

n)

Urb

an c

lean

wat

er &

san

itat

ion

ser

vice

s:•

Rev

iew

of

wat

er n

etw

ork

co

nn

ecti

on

fee

insm

all t

ow

ns

•Pu

blic

aw

aren

ess

cam

pai

gn

s to

incr

ease

dem

and

fo

r im

pro

ved

wat

er s

up

ply

an

dsa

nit

atio

n in

po

or

ho

use

ho

lds

•D

ue

con

sid

erat

ion

to

eq

uit

y in

net

wo

rkex

pan

sio

n p

lan

nin

gM

ain

stre

am b

asic

ro

ad a

cces

s p

rio

riti

zati

on

and

mo

nit

ori

ng

ap

pro

ach

into

nat

ion

alro

ad c

on

stru

ctio

n p

rog

ram

Reg

ula

tory

fra

mew

ork

an

d m

on

ito

rin

g s

yste

mfo

r en

terp

rise

-bas

ed r

oad

co

nst

ruct

ion

Poo

r d

istr

icts

in t

he

No

rth

and

So

uth

an

d a

lon

g t

he

bo

rder

reg

ion

to

Vie

tnam

Smal

l pro

vin

cial

& d

istr

ict

tow

ns

Mar

gin

al d

istr

icts

Esp

ecia

lly b

ord

er d

istr

icts

Bo

rder

up

lan

d d

istr

icts

Nam

saat

, Min

istr

y o

f H

ealt

h a

nd

su

pp

ort

fro

m n

atio

nal

an

d p

rovi

nci

al g

ove

rnm

ent

WA

SA, N

am P

apa,

urb

an v

illag

e le

ader

s an

dco

mm

un

ity

org

aniz

atio

ns

Nat

ion

al, p

rovi

nci

al, a

nd

dis

tric

t le

vels

wit

hvi

llag

e co

mm

un

ity

par

tici

pat

ion

Ag

ricu

ltu

re a

nd

fo

rest

ry f

or

acco

mp

anyi

ng

com

mu

nit

y-b

ased

nat

ura

l res

ou

rce

man

agem

ent

Nat

ion

al, p

rovi

nci

al, a

nd

dis

tric

t le

vels

wit

hen

terp

rise

s

TA

BL

E 3

.16

Lao

PD

R S

um

mar

y o

f Pr

iori

ties

fo

r Pr

og

ram

s w

ith

Jo

int

Posi

tive

Po

vert

y R

edu

ctio

n-E

nvi

ron

men

t Im

pac

t (C

on

tin

ued

)

Co

ord

inat

ion

nee

ds

bet

wee

nD

evel

op

men

t ar

eas

Pro

gra

m p

rio

rity

Geo

gra

ph

ical

tar

get

ing

sect

ors

an

d a

gen

cies

Costs of Implementation

Implementation of the recommendations pro-posed above involves various levels of costs. Somecosts can be expected to be relatively modest,such as monitoring, awareness and informationprograms, and hygiene improvement programs,but do nevertheless require sustained commit-ment and human and logistic capacity in order tobe effective and sustainable. Other interventionssuch as water supply and sanitation and ruralroads involve capital investments. The level ofthese investments depends, among other things,on service level. While it is difficult to provideexact estimates of investment requirements, somefigures are available to provide a plausible costrange for provision of rural and urban water sup-ply and sanitation.

Available capital cost estimates in Lao PDRfor rural water supply are on the order of US$10 per capita, which include construction mate-rials and construction of facilities. In addition,communities generally contribute about 40 per-cent of local cost in-kind, labor, local materialsand cash. External support cost of constructionmaterials for basic sanitation facilities is estimatedas low as US $2 per capita, with communitiescontributing about 85 percent of local costs.These cost estimates suggests an overall invest-ment requirement of US $22 million to reach an85 percent rural coverage rate of improved watersupply and basic sanitation in Lao PDR by 2020as targeted by the NGPES. This cost is in addi-tion to local community contributions of laborand local materials.

There is more uncertainty as to the cost of pro-viding urban water supply and sanitation. A pri-ority is provision of improved water supply tosmall towns which lag behind main urban centers.The historic per capita cost of urban water supplyin Lao PDR is about US $135. This however islikely to be substantially higher than what isrequired for small towns. The cost in small townsin Vietnam and Cambodia are estimated at US$23–40 per capita, while the capital cost of small-

scale independent providers in a few towns in LaoPDR is in the range of US $5–15. A range of US$10–40 per capita suggests a total investmentrequirement of US $6–23 million to provide allremaining small towns, i.e., about 85 towns, withimproved water supply by 2020, reflecting asmall-town population growth rate of 3.5 percentper year.45

While significant resources are required toimplement many of the proposed recommenda-tions from this PEN study, the effectivenessand sustainability of implementation clearly alsodepend on provincial and local government com-mitment, capacity and contributions; programsto raise awareness and capacity, and for makinginformed choices in local communities; anddemand by these communities and local authori-ties for programs that addresses poverty reductionand natural resource management and providesenvironmental services such as water supply andsanitation. Factoring in these conditions for effec-tiveness and sustainability is needed in regionalpoverty reduction targeting, such as for the Gov-ernment’s 46 priority districts and other districtsin need of poverty-environment interventionsand programs. In this context, as part of the PENstudy, a research paper on the economics ofregional poverty-environment programs withapplication to Lao PDR was prepared to shedsome light on central government or donor allo-cation of resources to provinces and local com-munities. The paper takes into account trade-offsbetween multiple objectives, and differences ininvestment and administrative costs across regionswithin a country.46

Endnotes1. Lao Expenditure and Consumption Survey 2002/03

(LECS III) and 1992/93 (LECS I).2. One of the districts, in Vientiane Municipality, has

“graduated” and there remain now 46 priority districts.3. The district poverty data are from ADB (2001), based

on poverty incidence in 1998 from LECS II.4. Forest data are from satellite imagery for the period

1993–97. These data were the most recent consistent



data available at the time of the PEN analysis presentedin the this report.

5. This does not include land in fallow/re-growth, butonly cultivated land.

6. Linkages between reliance on open unprotected surfacewater and health effects are discussed in the water andsanitation section.

7. Access to roads is measured as the percentage of villageswith main or secondary roads within 6 kilometers.

8. Poverty incidence is from LECS II as reported in ADB(2001).

9. It should be noted that the data on malaria incidencereflect confirmed cases. From the surveys in ADB(2001) households do report however that malaria is amajor threat to their health.

10. The results of a regression analysis of the variations indistrict forest cover in 1997 are provided in the PENbackground report for Lao PDR.

11. All the factors were assessed in a regression analysisusing LECS III household data (combined with UXOcontamination data). They were all found to be statis-tically significant. This means that each factor is statis-tically associated with poverty, after controlling for theother factors. The regression results are available in thePEN background report.

12. Swiddening is used here as a neutral term when refer-ring to the agricultural practice itself, and the termshifting cultivation is used when referring to the policy.

13. This is a reminder of the limitations of analyzingpoverty at a single level, the district. It might also indi-cate a lower importance of marketed NTFPs in the pri-ority districts, or a lower interest in this non-officialsurvey.

14. A wealth indicator was created for each household sur-veyed based on household durable goods and produc-tive assets.

15. Households living in villages within a 6 km radius froma road can generally make a round trip to a road withinthe same day.

16. NGPES has defined low access as presence of at leasttrails accessible by cart during the dry season. Donor-funded projects on basic road access had proposed ahigher but easier to observe standard of motorableaccess 10 months per year (WSP civils 2004).

17. The results are from a regression analysis of district leveldata, controlling for agricultural land, population den-sity, border location, and regional and topographic dis-trict characteristics.

18. Population growth is high and assessed at 2.5 to 2.8 percent, but strong migration trends from moremarginal to less marginal areas result in a stable or evendecreasing population in a fair proportion of districts.The 2005 census will provide more accurate popula-tion trends.

19. 15 districts have an urban population of more than 30 percent and may be defined as urban; there is no

urban population data for the nine districts createdafter the 1995 census.

20. Kaleum borders A Luoi district in Thua Tien Hueprovince in Vietnam. The timber sector there is activein conjunction with the Da nang port. The newly ren-ovated Ho Chi Minh road passes through A Luoi andthis might increase attraction toward villages in Kaleum.

21. The border in Nhot Ou became an official internationalborder in recent years. It crosses to the fast-growingSimao prefecture, Yunnan province. People in Kaleumgo to Saravanh, the next provincial capital.

22. The survey took place in 72 percent of villages in LaoPDR; 31 percent of surveyed villages were found to beimpacted by UXOs.

23. This was estimated with a district level regression equa-tion by combining district poverty incidence data fromLECS II 1997/98 (ADB 2001) and the UXO nationalsurvey data.

24. Even the figure for total land area could not be con-firmed.

25. Village communities have only 32 households on aver-age (20 in the six villages surveyed) and half this num-ber of houses due to extensive occurrence of extendedfamilies.

26. Phouvong, the other PEN study district in the South,is a good example.

27. Non-agricultural land accounts for more than one-third of total land area decontaminated by the LaoUXO program in 29 districts out of 57 where activitieshave taken place.

28. Phouvong, the other district studied in the South, isa national focal area (JICA 2002) where the newlyreclaimed lowlands are heavily affected by UXOs.

29. Population Census 1995 and Lao Expenditure andConsumption Survey 2002/03 (LECS III). Improvedwater supply is here defined as piped water, protectedgroundwater, rainwater, and “other water supply.”“Other water supply” mainly refers to gravity-fed sys-tems, but it is unclear from the LECS III data if thiscategory also includes any types of unimproved watersupply.

30. In 1995, about 60 percent of the poor and 40 percent ofthe non-poor used surface water. The change from 1995to 2003 was therefore 50 percent for both the poor andnon-poor, with no relative gain for the poor. The figuresfor 1995 are estimated from the Census 1995 and LECSII. The figures for 2003 are from LECS III.

31. Estimated from the Census 1995 and LECS II for 1995and LECS III for 2003.

32. Twenty percent poorest and richest as defined by con-sumption quintiles from LECS III.

33. From a district-level regression analysis of data com-bined from the 1995 census, poverty incidence fromLECS II, and UXO data from the national survey in1997.

34. The preceding analysis is from LECS III 2002/03.35. Logistic regression analysis of LECS III household data.



36. Recall that poor households are 50 percent more likelyto rely on surface water than non-poor households.However, the results are after controlling for the otherfactors. Those factors are associated/correlated withpoverty. The result (i.e., 15 percent) is therefore lowerthan the “uncontrolled” result (i.e., 50 percent).

37. WASA, DHUP, and URI. 2002. “Small Towns WaterSupply and Sanitation Initiative in Lao PDR.” Pre-pared with support from WSP-EAP.

38. These results are from a regression analysis of thehousehold data from the URI survey.

39. This result is from a regression analysis of the house-hold data, controlling for household location andimportance of time savings.

40. After controlling for living standard.41. Lao Women’s Union Gender Resource Information

and Development Center, 1999.42. The following equation was estimated, one for floods,

one for drought:

N = a + bp POPp + bnp POPnp + e

where N is number of people affected, and POPp andPOPnp are number of poor and non-poor respectivelyin each province affected by flood or drought. Theconstant (a) was restricted to zero in the regression toreflect that N will be zero if there are no poor or non-poor people. As the data contained no information onthe number of people affected, a conversion fromhectares damaged to affected people was undertakenbased on agricultural land per household and averagehousehold size. This conversion provides an order ofmagnitude. Two regression equations were also esti-mated with slope dummy variables for provinces thatwere not affected by flood or drought, little affectedand medium affected. The estimated coefficients fromthese equations allow for an estimate of the numberof poor and non-poor affected by flood or droughtacross the country.

43. This was the newly reported number of villages for the next population census. An even lower number ofvillages, 10,700, was reported in 2003. The reasonbehind this difference between 2003111111 and 2005is unclear.

44. Similar or closely related groups live on the Vietnameseside of the border. Several of them are officially recog-nized as ethnic groups with special difficulties in Vietnam(chapter 5).

45. The cost estimates of rural and urban water supply andsanitation in Lao were provided by Thomas Meadley,World Bank, Vientiane Lao PDR.

46. See Buys et al (2004). The Economics of RegionalPoverty-Environment Programs: An Application ForLao PDR. Prepared for EASEN, World Bank.

References (To be completed)ADB 2001. Lao PDR Participatory Poverty Assessment.Bouahom, Bouthong and others, eds. 2005. “Poverty reduc-

tion and Shifting Stabilization in the uplands of LaoPDR: Technologies, Approaches and Methods forImproving Rural Livelihoods.” Proceedings of a Work-shop Held in Luang Prabang, Lao PDR, January27–30, 2004. Vientiane: National Agriculture andForestry Research Institute.

Buys et al (2004). The Economics of Regional Poverty-Environment Programs: An Application For Lao PDR.Prepared for EASEN, World Bank.

Chaignon, Jacquelyn. 2000. “UNDP Sekong IndigenousPeople’s Development Program: Inception Report &Extended Program Strategy.” City: publisher.

Goudineau, Yves, ed. 1997. “Resettlement and Social Char-acteristics of New Villages: Basic Needs for Resettledcommunities in the Lao PDR.” An ORSTOM Survey.Vols. 1 and 2. Vientiane: UNDP.

Government of Laos (Lao PDR?). 2003. The NationalPoverty Eradication Program (NPEP): a ComprehensiveApproach to Growth and Development. Eighth RoundTable Meeting, City, September 2003.

Handicap International. 1997. Living with UXO: NationalSurvey on the Socio-Economic Impact of UXO in LaoPDR. Final report. Prepared for the Ministry of Labour& Social Welfare, Lao National UXO Programme.City: publisher.

IUCN. 1997. Manual of EIA Procedures for Road Projects inthe Lao PDR. Prepared for MTRPC with Assistancefrom SIDA. City: publisher.

JICA. 2002. Government’s Efforts to Rural Development inConsideration with Ethnic Minorities in Lao PDR. City:Lao Consulting Group.

Kakwani et al 2002. title. City: publisher.Lao National UXO Program. 2004. Turning Point: UXO

Lao Work Plan 2004. city: publisher.Lao UXO Program. 2002. “UXO Lao Community Aware-

ness In-depth Study, Khammouane.” City: publisher.Lao Women’s Union Gender Resource Information and

Development Center, 1999. title. City: publisher.MTCPC. 1996. “Notes from the Proceedings of the Lao

Road Sector Donor Coordination meeting.” Paper pre-sented at the Lao Road Sector Donor CoordinationMeeting, February 6 and 7, 1996, city.

SIDA/MTCPC. 2004. “Community Road ManagementImplementation Guide.” City: publisher.

World Bank. 2004. “Upland Livelihoods Improvement Proj-ect Concept Note.” Washington, DC: World Bank.

World Bank. 2005. “World Developement Indicators.”Washington, DC: World Bank.

WSP Civils 2004. Technical Assistance report to the BasicAccess Component of the Lao Swedish Road SectorProject 3.




INTRODUCTION

In 2004, some 15.6 million people in Vietnam—19 percent of the population—were still livingbelow the poverty line despite the impressivepoverty reduction gains made in recent years. Thisis equivalent to 80 percent of the total populationof Cambodia and Lao PDR. The Governmentof Vietnam faces the challenge of continuing itspoverty reduction efforts while ensuring preser-vation or restoration of environmental resources.Are there win-win policy options and interventionsthat can jointly address poverty and environment?How can a donor community that supports thegovernment’s growth and poverty reduction strat-egy best contribute to these win-win solutions?These are the questions that the PEN researchseeks to answer.

The PEN research is a collaborative effortbetween a World Bank team and three ministries,two of which are cross-sectoral agencies (Box 4.1).This enables the research to encompass policiesand programs with a multi-sectoral scope. Landadministration and sustainable developmentare two of these horizontal areas. As part of thestrengthening legal framework in Vietnam, theland law is in a process of finalization in the formof a “code.” The Cau River basin is the first areaselected for a pilot sustainable development pro-gram under the national Agenda 21 program.

The methodology of the PEN research, just asin the two other countries, starts with observingcorrelations between poverty and environmentindicators through mapping and correlation analy-sis. Local quantitative and qualitative research onspecific issues is carried out to shed light on someof the causes behind these correlations. Evidencecollected at various levels was then combined toanswer the specific research questions defined ineach problem area. The national research groupshave provided significant help in assemblingnational datasets. They have carried out all localresearch activities.

Vietnam is a highly diverse country, and thePEN research covers as much of this diversity aspossible. Poverty and environment linkages inVietnam can be structured into two broad cate-gories, environmental health and natural resourceuse. Diverse poverty issues and environmentalissues are at stake in each of these two categories.The scope of the PEN research encompasses therural, urban, and industrial sectors. In the ruralsector, the research targets “poverty pockets,”geographical areas trapped in very high povertyrates and deep poverty. It also covers the variouspolicy tools of the national policy framework(Box 4.2).

A set of five detailed research areas remainsinsufficient to cover the diverse dimensions ofpoverty and environment in Vietnam. The spe-

Poverty and Environment in Vietnam4

�

cific issues of protected areas are largely left aside,and so is the detailed analysis of natural resourceuse and poverty in the lowlands, which is a lowerpriority compared to the uplands. Field observa-tions mostly also leave aside border areas and theCentral Highlands. The national datasets, how-ever, allow drawing conclusions on a broaderscale after careful analysis of the relevance of thecase studies in the national context.

The PEN research uses a large recent dataset, the National Health Survey, and has itselfproduced a new database by pooling togetherstatistical data from different administrativelines. Some important datasets have, however,remained outside of the scope of the study. Dis-trict-level data was assembled whenever possi-ble. Since provinces normally assemble districtdata but only provide province-level data to the

P O V E R T Y A N D E N V I R O N M E N T I N V I E T N A M


B O X 4 . 1 PEN Research Partners in Vietnam

– The Ministry of Natural Resources and Environment is in charge of both the environment andthe management of land and other natural resources. MoNRE was the PEN research partner fordetailed research on industrial pollution and on land administration.

– The Ministry of Health was the PEN research partner for detailed research on water supply &sanitation.

– The Ministry of Planning and Investment is in charge of comprehensive advice on socioeconomicdevelopment strategies, programs, and plans. MPI’s National Office for Sustainable Developmentwas the PEN research partner for detailed poverty and environment analysis in the Cau River basin.

The following groups have undertaken research activities: the expert group at Hanoi Medical Col-lege for water supply & sanitation; the Research Center for Energy & Environment for industrialpollution; the Center of Occupational and Environmental Health of the Vietnam Association ofOccupational Health and the University of Economics, Ho Chi Minh City, for pesticides; TECOS underMoNRE for land administration; and the Socio-Economic Development Center for the Cau River.

This chapter includes an analysis of the databases assembled by the research partners, while thediscussion on searching for causes behind poverty and environment correlations is largely basedon the national reports by the research groups.

B O X 4 . 2 Detailed PEN Research Areas and Policy Framework

Environmental health and povertyWater supply, sanitation, and poverty: a national analysis of the recent Vietnam National HealthSurvey. Policy framework: sector policies and area-based programs in health, agriculture, and ruraldevelopment.

Industrial pollution and poverty: a set of case studies combined with analysis of the NationalHealth Survey. Policy framework: 1/ national remediation program for large polluting industries;2/ regional remediation program for household-based polluting industries in the Red River Delta;and 3/ strengthening of the legal framework in the urban sector.

Pesticides, health, and poverty: a case study on pesticide use in the Mekong River Delta. Policyframework: national program for the reduction of agricultural inputs.

Natural resource use and povertyLand administration, environment and poverty: a case study in three provinces on the impact of thenew land law in rural upland regions combined with the analysis of a national land use database.Policy framework: strengthening of the legal framework in the rural sector.

Integrated poverty and analysis researchPoverty and environment in the Cau River basin: a case study in two provinces on natural resourceuse in the upstream section of the river basin and environmental health in the downstream section.Policy framework: National Agenda 21 program.

national level, this has occasionally requiredmanual transcription of district information.At the local level, a complicating factor in ruralsurveys was the dual administrative levels ofcommune and village. Statistical data is avail-able only at the commune level, while povertyissues and environment issues tend to differamong villages within a commune. We usedthe village level when defining samples.

4.1. NATURAL RESOURCE USE AND POVERTY



B O X 4 . 3 National Data Sources for PEN in Vietnam

The Vietnam Living Standard Surveys (VNLS) are large-sample household surveys designed to mea-sure trends in poverty and associated factors. The most recent was released in 2004. Previous sur-veys were carried out in 1997–98 and 1992–93. The surveys are representative at the province level.Small-area estimation techniques use correlations between poverty and statistical data available atmore local levels in order to assess poverty down to the commune level. Unless otherwise indicated,all poverty data in the report are from these sources.

The Vietnam National Health Survey (VNHS) carried out in 2001–02 provides a unique opportu-nity to assess the impact of poverty and various other factors on environmental health in Vietnam.The survey covers 36,000 households and 150,000 individuals in all provinces. It includes informationon drinking water supply, sanitation, household behavior, water pollution sources, prevalence rates,and indicators of severity of diseases in all age groups, as well as living standards, education levels,and housing characteristics. Unless otherwise noted, this survey is the source for all national analysisin water & sanitation and health.

The PEN II land use database was assembled by TECOS from datasets provided by CEMMA, GSO,MoNRE’s 2000 and 2003 land census, land use and detailed land allocation statistics, the Ministryof Transportation, and the Center for Hydro-Meteorology and Environmental Network. Unlessotherwise noted, all variables in the natural resource analysis use this database.

Additional GSO datasets. Unless otherwise noted, all industrial output data is directly compiledfrom GSO data. GSO also contributed critical datasets to complement other sources.

Reliability of statistical data. The research teams have carefully crosschecked data for quality.Some variables, often in relation to more remote provinces or districts, remain unreliable. Whileabsolute quantitative values for specific locations may be questionable, the datasets provide solidinformation regarding overall trends in time and in geographical distribution.

Key findings– Issues of sustainable natural resource use are

linked to poverty in the uplands, not in thelowlands.

– There is a mechanical correlation be-tween poverty, fragile soils and location orloss of forest resources that is largely due to

association of uplands with high povertyincidence.

– Trends in forest resources are a combinedoutcome of how local governments, for-est enterprises, migrants and households,especially ethnic minority groups, inter-act locally.

– Legal framework improvements, espe-cially in land administration, are one ofthe leverage points that have potential toimprove sustainability of natural resourceuse and reduce poverty, but the poorestcommunities are largely by-passed bythese improvements.

– Three questions are equally relevant toidentify win-win poverty and environmentpolicy options: where and how are the poordisproportionately affected by resourcedecrease or degradation? where and how arethey at risk of natural disasters? and whereand how are is their lack of access to thelocal resource base an issue?

4.1.1. Geographical Dimensions

The distribution of poverty in the country, whendefined as the percentage of population livingbelow the poverty line, is strikingly similar to amap of elevations (Figure 4 in chapter 1). Povertyreduction in Vietnam has been very rapid overthe last decade. Poverty incidence had alreadydeclined to 36.5 percent in 1998–99 and furtherdeclined to 19 percent in 2003–04. Accordingto the latest poverty update of the World Bank,poverty incidence in Vietnam in 2004 is only athird of the poverty incidence recorded in 1993.Variations in poverty incidence remain acrossprovinces, ranging from 0 percent poverty in HoChi Minh City to a maximum of 74 percent inLai Chau Province, the northwest province thatborders Laos. Provincial average per capita expen-diture also varies substantially.

Correlation of the spatial distribution of povertyis visible at all available levels of analysis: region,province, district, and commune (Maps 4.1).Historical processes, through which the admin-istrative units in Vietnam have been delineated,have generally followed an upland/lowland limit.Recent changes in administrative units are oftenreinforcing this trend. Thai Nguyen Province inthe Cau River basin in northeastern Vietnam,for example, was divided into two provinces: anupland province, Bac Kan, and a midland pro-vince, Thai Nguyen. The same is often observedwhen larger districts are divided into two.

The large majority of the poor in Vietnamlive below 500 meters of elevation (Figure 4, chap-ter 1). This is an effect of the extreme imbalanceof population densities between the lowlands anduplands (Map 2, introduction). Almost half ofVietnam’s population was concentrated in 1999



M A P S 4 . 1 Poverty Incidence at Various Levels

Left: overall poverty, regional level (2004); center: overall poverty, province level (2004); right: rural poverty, district level (1999).

in the 24 small, flat lowland provinces aroundHanoi and HCM City, constituting only 16 per-cent of the total land area of Vietnam, with a28 percent average poverty incidence. Correlationof poverty incidence with elevation is however notvisible when looking at mean elevations withinadministrative divisions. Elevation alone is notsignificantly correlated to poverty at district level(IFPRI & IDS 2003) or province level. There arelower elevation areas or more fertile areas withinmost upland or hilly provinces and districts. High-elevation plateau regions in the Northeast and inthe Central Highlands regions also have a mix ofhigh and medium overall lower poverty incidence.

Association of high poverty incidence withelevation implies mechanical linkages betweenpoverty and forest resources and between povertyand fragile lands. Correlation does not imply thatcausal relationships between these factors are atplay. There is nevertheless a striking associationbetween poverty and these indicators. At theprovincial level, provinces around Hanoi andHCM City have about 60 percent of their landcultivated. Only about 6 percent of land is forestland in these two regions, compared to the nationalaverage of 34 percent. Hilly and mountainousprovinces have 90 percent of Vietnam’s forestland. A comparison of district-level poverty andenvironment indicators between categories ofdistricts defining officially lowland and uplanddistricts confirms this contrast (Table 4.1 andMap 4.2). It also shows the importance of hillyand mixed districts alongside higher elevationareas in poverty and in forest resources.

Many of Vietnam’s high elevation areas canbe described both as environmentally fragile areasand as poverty traps. Forest resources are spreadacross mountainous and hilly areas. The Cordillera,with its ridge forming most of the border betweenVietnam and Laos, is a rich area of biodiversity(Map 3, introduction). Almost all protectedareas are located in the high elevation regions.Mountainous districts have 55 percent of “com-mercial forests” and 51 percent “environmen-tal forests” (Box 4.4). They are also home to

47 percent of Vietnam’s 10.5 million ethnicminority people.1

The severity of poverty (or poverty depth)is substantial in districts with highest elevation(IFPRI & IDS 2003). Poverty rates have decreasedsignificantly between 1999 and 2004, but moreslowly than elsewhere. The six provinces withmore than 50 percent of the poor population in2004 are the three provinces of the Northwestregion, the three mountainous border provincesin the Northeast region, and Bac Kan Provincein the Cau River basin. The three provinces with40 to 50 percent of the poor population in 2004are the Central Highland provinces other thanLam Dong (the more developed province recentlyincorporated into the region).

Communes in upland areas often have scat-tered villages with contrasting issues poverty andnatural resource use. The commune is the lowestadministrative level, but each commune is com-prised of several villages. Remote villages in theseareas are typically home to more disadvantagedethnic groups. Forest resources also tend to be con-



T A B L E 4 . 1 Poverty and Environment Indicators in Lowland, Upland, and Other Districts

Hilly &District category 1/ Lowland Mixed Mountainous

Number of districts 282 195 134Mean elevation (m) 45 441 925Number of poor 12.4 10.5 5.0

(M people, 1999)% poor 28 43 62Rural density (inhab/km2) 418 185 45% ethnic minority people 4 15 61% agricultural land 65 24 12Paddy per rural capita (m2) 958 471 549% of area with slope < 8° 98 59 32% of area with slope > 30° 0 3 8Forest land area (M ha) 1.6 4.4 5.6% forest land in land area 22 38 37

1/ CEMMA defines six categories of districts. In addition to lowland andmountainous districts, other categories include partly hilly districts (80 dis-tricts), hilly districts (73), hilly and mountainous districts (39), and partlymountainous (3).

centrated in some of the villages—often a legacyof the state forest-farm zoning system. Moun-tainous provinces have more villages in each com-mune. The Northwest, Northeast, North-Centralcoast, and Central Highland provinces have 10 to11 villages per commune, whereas other regionshave 4 to 7 villages per commune. This figurereaches 14 in Sonla and Tuyen Quang provinces.The number of villages has increased between1995 and 2000 by 17 percent in the Northwestregion and by 41 percent in the Central High-lands, which indicates active in-migration.

The lowlands have their own environmentalthreats, but most natural resource managementissues do not appear to impact the poor dispro-portionately. Aquaculture and fishing in thelowlands have developed at a rapid pace, andincreased protection of the marine resourceshas not followed. In 2003, there was only one

protected wetland area in Vietnam. Coastal andmarine degradation was assessed as a mediumlevel poverty-environment priority. The linkagebetween fisheries and poverty, however, is notvery strong. A regression analysis of provincialfishery employment in relation to poverty inci-dence carried out at the beginning of the PENresearch has found that around 2.6 percent of thenon-poor and only 0.5 percent of the poor areemployed in this sector. Typhoons and floodsare an exception, since they affect a number ofpoor coastal provinces.

4.1.2. Forest Resources and Poverty

Deforestation mostly takes place in poorer pro-vinces (Table 4.2). This is partly a simple conse-quence of the geographical superposition of theforestry sector and forestry.



B O X 4 . 4 Forestry Indicators and Databases

Forest cover rate is a widely used indicator in Vietnam, and one of the Millennium DevelopmentGoal indicators. The Government has a strategic goal to recover a 43 forest cover rate (Povertytask force 2003). There are three ways of defining forest cover rate.

– Under the land administration in MONRE, the land census defines forest land and its sub-classifications. This classification is for zoning purposes only.

– MARD regularly carries out forest inventories of forested areas.– FAO data provides actual forest cover from satellite images.

Coupling the MoNRE and MARD databases could provide data on the percentage of land zoned as forest land that is actually under forest cover. In practice the datasets are difficult to reconcile.Trends indicated by the different data sources are equally difficult to reconcile. FAO indicates thatdeforestation in Vietnam may have been as high as 2.3 percent per year from 1980 to 1990, whileforest cover increased marginally from the mid–1990s. The 1994–99 forest land data indicate thatdeforestation in Vietnam continued at about 1.3 to 1.4 percent per year.The datasets are best used to assess geographical patterns, not trends over time.

B O X 4 . 4 (continued) Land Census Definitions

Special use forest Protection forest Production forestPlantation forest Plantation special use Plantation protection Plantation productionNatural forest Natural special use Natural protection Natural production

Equivalent forest management regimePlantation forest Environmental forests: Environmental forests: Commercial timber

plantationNatural forest Protected areas Watershed protection Natural forest with

on-going commercial exploitation

Several indicators point to an important con-trast in linkages between poverty and forestryresources between the northern regions and thecentral regions. During 1994–99, half of theprovinces experienced a net increase in forest landarea, with a total increase of 1.14 million hectares,or 11 percent of the 1994 forest land area. Thelargest losses of forest land were in regions withhigh poverty incidence. Loss of forest land in thecentral third of the country—starting from QuangBinh Province—continued throughout the 1990sand totaled about 1.25 million hectares from 1994to 1999. Poverty rates in all these provinces exceptone are well above the Vietnam average. Con-versely, poverty incidence has been associated withan increase in forest land area in the northern partof the country during the same period (Map 4.4).This fact points to an absence of proof that thepoor are one of the causes of deforestation orthat they are poor because they are significantlyaffected by deforestation. It is however clear thatin the poor provinces, the poor do not benefitenough from forestry resources, whether they arelogged or protected, to raise their incomes abovethe poverty line.

Forest types also differ between the northernand central regions, largely in association to thestructure of the State forestry sector. “Environ-mental forests” are largely dominant in northernupland areas, while natural production forest—an indicator approaching land zoned for com-mercial forestry—is nearly absent in the Northbut remains important in the central regions(Maps 4.2). There is concentration of naturalproduction forest in the central third of thecountry. In 2000, 682 SFEs were managing 5.5 million ha, or 17 percent of the total landarea in the whole country, and undertook a mixof watershed management and commercial log-ging activities (MoNRE 2000). Starting in theearly 1990s, the SFEs that were in charge ofmanagement, exploitation, processing, and dis-tribution of the country’s forest resources haveundertaken a reform process. Enterprises dis-continuing commercial operations to put the

forest they manage under the protection of for-est management boards are more prevalent inthe northern regions, while those that havebecome fully commercial SFEs are more pres-ent in the central regions. This trend is contin-uing in subsequent changes in forest land between2000 and 2003.

While less than 20 percent of forest land isallocated to enterprises in most districts in theNorth, this proportion is higher than 40 percentin around 106 districts, most of them locatedin the Central regions (Map 4.2). More than40 percent of total land area is allocated toforestry enterprises in 21 districts in these regions.The land census records allocation of forestland to various categories of stakeholders. In2000, 3.5 million ha of forest land, or 31 percentof the total land zoned for forestry, were allo-cated to enterprises, just twice the percentage offorest land allocated to households. By 2003,the proportion of forest land allocated to house-holds had increased to 22 percent, while forestland allocated to enterprises was unchanged.Allocation of forest land to enterprises hasdecreased in 20 provinces, but has continued toincrease in 18 provinces.

Finally, dominant types of ethnic minoritygroups differ in the northern and in the centralregions. CEMMA defines 54 minority national-



T A B L E 4 . 2 Disparities in Provinces with DecreasingForest Land in 1994–99

Provinces with Provinces withPoverty incidence 0–30% poverty 50–80% povertyin provinces (1999) incidence incidence

Number of provinces 11 14Loss of forest land (hectares 13 145

per 1,000 people)Loss of forest land 4% 9%

(% of total land area)Loss of forest land (% of 25% 20%

natural forest area)Gross forestry output 15 106

(m3 per 1,000 people)

ities. There is no widely accepted grouping of the54 groups into broader categories. We make atentative grouping can based on ethno-linguisticgroups (Khong Dien 2002). If a “small southerngroups” category is defined by including 24 Mon-Khmer and Malayo-Polynesian groups, a “mainnorthern groups” category (including the fourmain Thai family groups and the Muong) and a“small northern groups” category (including smallTibeto-burman group), a picture with largelydisjointed regions of residence appears (Map 4.3).There is an apparent paradox between the fact thatthe small southern groups, which have tradition-ally been described as having more sustainablesystems of rotational agriculture, live in defor-estation areas, while the northern groups live inreforestation areas.

4.1.3. Soil Resources and Poverty

In almost all upland areas, there is a trend amongsubsistence farming systems toward intensifica-tion with the opening of new paddy fields and anincrease in multiple cropping in the paddy fields.From 2000 to 2003, paddy area in Vietnamdecreased overall by close to 246,000 hectares, or6 percent of the total paddy land area, but itincreased in 21 of the country’s 62 provinces, alllocated in the upland northern and central uplandregions. All provinces in the Northwest and Cen-tral Highland regions increased their paddy area.The increase reached 28 percent in Lai ChauProvince, which has the highest rate of rotationalagriculture in the country. Of the 36 provincesthat had an increase in the cropping index in



M A P S 4 . 2 Regional Contrast in the Forestry Sector

Left: change in forest land area (1995–2000). Center: Percent of protection and special use forest in forest land (2000). Right: Percent ofnatural production forest in forest land (2000).

paddy fields, 16 are in the three upland regions—the Northwest, Northeast, and Central Highlands.This indicates that household land use systemsin the North and in several Central provinces aremostly on a virtuous path of intensification; thatis, away from rotational agriculture toward moresustainable systems centered on paddy fields. Howmuch the poorest communities and householdsbenefit from these improvements is howeverlimited by the fact that poorer communities andhouseholds generally have fewer or even nopaddy fields.

Migration makes linkages between agricultureindicators and poverty and environment muchless visible. In the Central Highlands, the regionwith highest in-migration, the rural population

of 3 million farm close to 550,000 ha of rubberand coffee, an 80 percent increase from 1995. Theimpact of subsistence farming on poverty andenvironment in the region is limited comparedto this large-scale development of perennial crops.Subsistence farming is however changing fast.The increase of paddy area in Kontum Provincereached 37 percent between 2000 and 2003.

Linkages between degradation of fragile soilsfrom farming activities and poverty have beendiscussed over many years in Northwest Vietnam,the region with highest area of sloped farmland.They remain difficult to quantify. All provincesin the three upland regions, as well as in theNorth-central coast region, have fewer paddyfields than sloped farmland (Map 4.9c). This



M A P S 4 . 3 Forest Land Allocation and Ethnic Minorities

Left: forest land allocated to enterprises (percentage of total land area, to update to percentage of forest land) (2000). Center: forestland allocated to households. Right: districts with high share of “main Northern” and “small Southern” ethnic groups.

proportion reaches five in Lai Chau and eight inSonla. Flood events are geographically associatedwith cultivation of sloped land. The Black Riverbasin was the main area in Vietnam where threeor more flood events were recorded in 2002 and2003. The Black River has been at the center offorest rehabilitation policies in the 1990s. It isalso a region where maize is being developed ona broad scale for the feed industry. The threeNorthwest provinces alone, with only 3 percentof rural population in Vietnam, grow 15 percentof the national area of maize. The area of maizein these three regions grew again by 22 percentbetween 2000 and 2003. While commercialmonoculture of maize on steep slopes has beendocumented to generate erosion, there is alsofield evidence that maize introduced in rotationalagriculture beside upland rice can improve over-all sustainability of the rotation.

Whether related to human activity or not, thenatural disasters in Vietnam are often located inareas with a high proportion of poor. In additionto floods in the Black River basin and other basinsin the North, the poor in the Central coastalregion are impacted by typhoons and subsequent

floods. There were 23 coastal districts with povertyincidence of 50 percent in 1999.

4.1.4. Land Administration, NaturalResource Use and Poverty

The geographical linkage between land allocationto poor households and other stakeholders indi-cates how important land administration is inunderstanding the poverty and environment nexusin Vietnam. The national legal framework for landadministration has been modernized extensivelyin the past decade. A land law was created in 1987and was partly or fully updated three times. Thelatest version of the land law—implemented inearly November 2004—is a key version, sincethe following one planned for 2008 would be along-term land code.

The 2004 land law, although primarilydesigned to resolve urgent land administrationissues in urban and lowland rural areas, addresses,two broad fields of special relevance in poorupland areas, land titling and land use planning(Table 4.3). National stakeholders expect the lawto be an instrument to speed up the allocation of



T A B L E 4 . 3 Transition Stages in Land Administration in Vietnam

1. Early transition 2. Mid transition 3. Late transition

1. Land tenure

2. Land use planning

In italics: stages for different areas and types of land at the start of the 2004 land law.Source: summary of stakeholder interviews

Allocation of land userights

Forestry land: not completed

2.1. Decreasing weightof plan targets,increasing mappingand participation

Upland areas: ongoing

1.2. Issuance of LandTenure Certificates

Agricultural land: justcompleted

Forestry land: starting2.2. Shift from control

of productive usecategories to controlof use transfer categories

New Law: controlstransfers out of paddyand environmentalforest

1.3. Fully transferableland titles

Urban land: new Law

2.3.1. Local plans man-aged by communityauthorities to regu-late land market

Urban areas: starting2.3.2. “Unused land”

category cancelledNot started yet

formal titles on forest land, the land category wherethis process is lagging behind. The new law alsointroduces for the first time a controlled transferof paddy areas and of “environmental” forest landto other uses through land use planning.

Limited reliability of quantitative data on landuse and land administration is an important fac-tor to take into account in the analysis of landadministration reform. Discrepancies between theland census and actual land use are well-knownin Vietnam. They reach a dramatic level in thesix communes surveyed in percentage terms forpaddy fields, and in absolute terms for forest land(Table 4.4). Interestingly, there is more forestobserved in the photographs than forest reportedby local officials.

Limitations in data availability also reflect lim-itations in access to information on land admin-istration and handling capacity among local cadresin marginal areas. Each commune has at least onestaff member dedicated to land administration,but these staff members had not received infor-mation about the new land law. Land census sta-tistics are compiled at the district level, but werenot available for the commune level. Data onforest land allocation was available at the village

level, but did not match commune-level figures.Two out of the six communes could not reportnumbers of households with paddy land allocated,while the others simply reported 100 percent,a figure that is not consistent with the surveysample. None of the local officials could recall the2000 situation for land use rights. Capacity oflocal staff was assessed on a scale of 1 to 5 depend-ing on their education level and other indicators,with 1 indicating highest capacity. Capacity scoreswere 2.7 and 3 in less-poor communes and vil-lages respectively, and 3.3 and 4 in poorer com-munes and villages.

MoNRE is facing difficulties in receivinginformation from the provinces. Land titling datacould only be assembled at province level during



B O X 4 . 5 A Land Administration Survey in Six Upland Districts

Are land administration instruments an appropriate and viable answer to natural resource uselinkages to poverty and environment? How does conversion of land uses, especially paddy fieldsand forests, affect poor households versus non-poor households and what is its impact on theenvironment? Can the environmental and social impacts of the new law be monitored, and how?

A survey carried out in three upland provinces where implementation of the new law startedearlier—Tuyen Quang in the Northeast region, Nghe An in the North-central coast, and Binh Dinhin the South-central coast—was undertaken to answer these questions. One poor commune wassurveyed in each of six contrasted districts in terms of land use trends and structure of the forestrysector. Implementation of the land law was reviewed at district and commune level through statis-tics and local interviews. A 174-household survey allowed a poor-non poor comparative analysisusing poverty scores based on household assets. Aerial photographs allowed the research team tocompare actual land use with reported data.

A survey fully or partly based in the lowlands, the main area targeted under the new law, wasinitially considered but rejected in favor of comparative analysis of various upland areas for tworeasons. First, global research indicates that land use transfers from agriculture to urban or indus-trial uses have no net impact on poverty, since farmers move from agricultural to manufacturing jobs.Second, ongoing studies are focusing on lowland issues with poverty and environment implications:land consolidation, and loss of paddy areas to aquaculture.

The full report of the national research team is available on a separate CD Rom (TECOS 2005).

T A B L E 4 . 4 Actual Land Use/Statistical Land Use

Average 6communes Minimum Maximum

Total land area +0.4% −0.4% +1.2%Forest land area +38% −0.4% +104%Paddy land area +140% +23% +296%

the PEN research, not district level, and only 21out of 61 provinces provided full information.The information gap is highest for forest land titlesissued to enterprises. Even data on land titlingprogress for paddy fields are inconsistent. Thedatasets remain useful to monitor time andregional trends. Paddy land titling is well advancedin the Northeast and Northwest regions andlagging behind in all Central Highland provinces.Forest land titling is less advanced for householdsthan titling for enterprises. Twenty-eight provincesdo not report any household Land Tenure Cer-tificates (LTCs) on forest land, whereas only fourprovinces do not report any enterprise LTC onforest land.

Conversion from paddy to other uses, thefirst control point for land use transfers in thenew land law, does occur in upland areas, but sofar linkages to poverty are indirect only. Paddyland area decreased by 300–400 ha in of each ofthe three surveyed provinces between 2000 and2003, but in only two of the six surveyed districtswith higher population density. Paddy conver-sion takes place in less-poor communes. We alsocollected oral testimony regarding the value ofpaddy fields to both the poor and less-poor uplandfarmers. Paddy area per capita, from 240 m2 in onecommune in Nghe An to 750 m2 in one communein Tuyen Quang, is not smaller than the provin-cial average, but paddy remains an important foodsecurity element. Paddy conversion—totalingonly 2 ha—was reported in two communes andmostly relates to fish pond development by a fewbetter-off households in one village. Things mightchange, however. Of the households surveyed,3 percent indicated their interest in turning paddyfields into perennial crop plantations.

Changes in forest land zoning are activelytaking place, albeit at a lower pace than in the1995–2000 period. Transfers from environmen-tal forests to other uses, which are due to be con-trolled through land use plans, are limited. Mostchange relates to gradual rezoning of land zonedas unused, generally rotational land and relatedfallow. This land is mostly rezoned production

forest land, a small proportion being transferredinto agricultural land. Unused land decreased by60 percent in one of the northern districts between1995 and 2000. The decrease is slowing down,but was still 10 percent in 2000–2003 in oneof the central districts. The forest land domain,already covering at least 40 percent in the com-munes, and up to 90 percent in one north-centralcommune, continues to increase.

Forest rezoning creates contrasting local situ-ations. In the northern province, although norezoning took place in the two communes studied,some other communes have had large land trans-fers from special use forest to protection forest,bringing income opportunities from forest pro-tection contracts but raising questions of bio-diversity protection. Other communes have hada substantial decrease in production forest andan increase in protection forest, reducing oppor-tunities for sloped land development. In NgheAn, land rezoning as production forest is substan-tial in some communes and negligible in others.In Binh Dinh, forest land is being rezoned intoagricultural land in some communes in the Kinhdistrict, while in the ethnic minority district com-munes rezone land into production and protec-tion forest.

Actual use of sloped land zoned as productionforest appears to be unrelated to land allocationor land titling (Table 4.5). Land allocation mayhave contributed to more decision making in onlyone surveyed village in Tuyen Quang. In thisvillage, 58 percent of households have forest landallocated, and they are developing orange pro-duction, unlike the other village with a state forestenterprise and only 4 percent of households withforest land allocated. It is local rapport—amongstakeholders, state forest enterprises, national parkstaff, and community/households—and the pres-ence or absence of market opportunities thatappears to explain why households use or do notuse the forest land they have been allocated. Theethnic dimension is a visible factor in this stake-holder rapport. Provincial and district policiesalso play an important role, for example, with



regard to rotational agriculture. One district inNghe An promotes regulated rotational agricul-ture on allocated forestland, while the neighbor-ing one restricts it.

Among the three land instruments analyzed—land use planning, land allocation, and land tenurecertificates—new procedures for land use planswere found to be largely not taken into accountyet. Land use plans have largely remained eco-nomic planning instruments used by technicalbureaus at the provincial level with limited par-ticipation from more local levels. Land use plansare in operation in only four of the six districtssurveyed. Only Nghe An, an upland provincethat is generally fairly advanced in land admin-istration, had land use plans in both districts sur-veyed. None of the six communes had a land useplan in place.

Progress in forest land allocation is highlydependent on the different provincial policies.Allocation of land titles on forest land has startedin the three provinces, although not in the com-munes surveyed. All households in the Nghe AnProvince communes have land use rights onforest land or are about to receive land titles,though this land only accounts for 11 to 19 per-cent of total forest land (Figure 4.1). In TuyenQuang, proportions of households with forest

land allocated between 4 and 60 percent in thevillages studied. In Binh Dinh, no forest landallocation is reported in the villages surveyed.Economic organizations, mostly state forestenterprises, are important stakeholders in thethree provinces, albeit at a variable degree. Thearea of forest LTCs issued to economic organi-zations is three times that issued to householdsand six times in Binh Dinh province.



T A B L E 4 . 5 Specific Uses of Sloped Land in the Six Communes Surveyed

Commune Uses of sloped land Important stakeholders % Poor Ethnic groups

TQ 1 Orange in village 1 SFE in village 2, 66% 84% MNGwith conflict

TQ 2 RA forbidden by local Protected area 87% 53% MNGgovernment but practiced

NA 1 Tourism: 1 village resettled Protected area 69% 94% MNGRA restricted by localgovernment

NA 2 RA organized by district District 67% 95% MNG & SSGwith land allocation

BD 1 Cassava, cashew SFE, without conflict 53% NoBD 2 RA, cashew SFE, with conflict 80% 97% SSG

RA = rotational agriculture; MNG = “Main Northern ethnic groups”; SSG = “Small Southern ethnic groups” (seedefinition in national analysis).

0

40 000

80 000

120 000

160 000

TQ1

TQ2

NA1

NA2

BD1

BD2

Other forest land

Forest landallocated to HH

Other forestallocated to EO

Natural productionforest allocated toEO

F I G U R E 4 . 1 Allocation of Forest Land in the Six Districts Surveyed

Source: PEN Case Study Database.TQ = Tuyen Quang, NA = Nghe An, BD = Binh Dinh

No correlation between more secure landtenure and poverty is visible in the survey, eitherat the household or village level.2 Poverty scoresof households with and without forest land allo-cated are equal. While correlation of large areasof forest land to wealth status is a frequent occur-rence in Vietnam (see Cau River section), landallocation itself is not associated with less poverty.For the most part, the households surveyed areaware of the possibility of applying for LTCs inthe near future. Only 4 percent mentionedadministrative difficulties among the reasonswhy they did not use their sloped land. Landallocation even with formal titles is not perceivedas secure land tenure. Only forty-five percent ofhouseholds indicated they were planning somechange of land use on their forest land. Only19 percent stated they were planning to developperennial crops. 55 percent said that the issuewas insufficient knowledge to do so, and 82 per-cent that they were not planning to use theirland titles to access credit.

The prevalence of large land managers onsloped land, and of agriculture and forestryadministration in decision making, suggests thatforest land allocation has only created sometenure security when provincial and district gov-ernments encourage decision making by thehouseholds. When this support is absent, defor-ested sloped land remains largely under-used, or

can be spontaneously developed with environ-mental impacts.

Policy implications

In the absence of innovative approaches to landadministration in marginal rural areas, a double-tracked system in the land administration legalframework is being created. Urban areas areaccessing a more comprehensive legal frameworkwith the new land law, while poorer, marginalareas may continue to operate largely outside theland administration framework and to lag behindin land administration improvements.

Capacity-building among local cadres is crit-ical to avoiding this risk of a modernized landadministration bypassing marginal rural areas,and the new law offers an opportunity for a sub-stantial information and training program. Thiscapacity issue is compounded by the fact that vil-lage leaders are legally in charge of handling landuse matters. Community participation is recog-nized in Vietnam as a means to overcome theabsence of reliable statistics and to improve localdecision making, but local capacity to use theseapproaches needs to be enhanced as part of aninformation and training program.

A transition from top-down land use plans toparticipatory, problem-solving oriented land useplans has the potential to enhance local decisionmaking in poor communities vis-a-vis other stake-holders. Land use plans are also needed to avoidenvironmental impacts from sloped land develop-ment. A balance needs to be found betweenidentifying critical land-use control points, pos-sibly other than those in the land law, and avoid-ing the detailed directions of land use that stillprevail and are in themselves a powerful deter-rent to local initiatives. Improved land useplanning will require disentangling the dis-crepancy issue between statistics and actualland use, and modern tools such as aerial pho-tographs and GPS may be an opportunity todo so.

Formal land use titles are a necessary condi-tion for land tenure security, but not a sufficient



0 20 40 60 80 100

21-25

26-27

28

29-30

31-35

Po

vert

y sc

ore

%

With (82 HH)

Without (92 HH)

F I G U R E 4 . 2 Households with and without Forest Land Allocated by Poverty Level

one. The perception of security is as important aslegal land titles in ensuring that title holders canuse their sloped land to secure or improve theirlivelihoods. Land use plans can support thisprocess, and local access of the poor to credit, agri-cultural inputs, and markets are a pre-condition—still often missing in marginal rural areas.

Monitoring poverty and environment impactsof the Land law can be done through better use ofexisting instruments and local experience ratherthan by creating a new monitoring system. Theland census, with its detailed information requestson land use types and stakeholders receiving landuse rights, is a perfect tool. What is needed is a bet-ter dialogue between central level and provinces toensure that reliable data are assembled and used.The wealth of experience being assembled byprovinces and districts through pilot landadministration schemes deserves an evaluationprocess and the exchange and dissemination of lessons learned. Monitoring and evaluationof the land law can only be successful if itincorporates the impact of forestry policy andreform through improved linkages betweenthe two agencies.

4.2. ENVIRONMENTAL HEALTH AND POVERTY

4.2.1. Lack of Safe Water andSanitation and Poverty



Key findingsIndustrial pollutionThere has been a trade-off between povertyreduction and environmental quality in thelast decade in particular from the non-statesector

The urban poor are disproportionatelyemployed in industry with high occupa-tional health risk and exposure to pollution,particularly in small-scale industry.

Young children and elderly—the agegroups most susceptible to health effectsfrom pollution—are disproportionatelyconcentrated in poor households in theurban industrialized areas.

Rural and peri-urban craft industrieshave rapidly raised household income incraft villages, but with serious pollutionand health effects. Non-household laborhired in craft industries is disproportion-ately from low-income households, andare exposed to serious air pollution andtoxic substances. Dug wells, mostly usedby low-income households, are at risk ofcontamination from untreated wastewatergenerated by craft industries.

Drinking water and sanitationThere is a significant level of inequity inaccess to safe water and sanitation through-out Vietnam.

Significant numbers of lowland poorusing dug wells or surface water are affectedby lack of access or pollution of drinkingwater.

The role of ethnicity in higher incidenceof diseases related to water and sanitation isvisible but not fully explained.

Pesticide usePesticide use has become prevalentthroughout Vietnam including in poorprovinces.

The poor are more exposed to toxicpesticides than the non-poor.

The poor as aware as the non-poor ofthe risk related to pesticide use but haveless access to information that could helpthem avoid these risks.

The Vietnam National Health SurveyHow do the poor differ from the non-poorin access safe water and sanitation? Whatis the level of health risk, particularly fordiarrheal illness, of various forms of drink-

Vietnam has achieved impressive gains in watersupply and basic sanitation coverage rates. About80 percent of the population has access to cleandrinking water. The two main unsafe watersources—surface water and wells, both of whichhave high risk of pollution—declined respectivelyfrom 20 percent to 10 percent and from 50 per-cent to less than 35 percent between 1992–93and 2001–02. At the same time, the share of thepopulation with drilled-well water increased from5 percent to over 20 percent. While access to watersupply and sanitation facilities has improved sig-nificantly, it still remains a major issue for the

poorest segments of the population, which havemuch lower coverage rates than the rest of thepopulation. More than 17 percent of the poorest20 percent of the population use surface water asdrinking water, while only 3 percent of the richest20 percent use surface water.

Beyond this direct link between poverty andaccess to clean water and sanitation, no obviousgeographical linkage between poverty and indi-cators of environmental health in relation to watersupply. The overall difference between rural andurban areas in coverage of clean water supply islimited. The disparity is within rural and urbanareas. About 8 percent of people in urban areasand 22 percent in rural areas are without safedrinking water; that is, they use surface water ordug wells with a nearby pollution source.

The disparity between rural and urban areasin access of the poor and the non-poor to toiletfacilities is substantial. In urban areas, almost allof the richest households have toilet facilities, whilethis proportion is 3 percent in the rural areas. Inthe urban areas, 24 percent of the poorest do nothave toilet facilities, while in rural areas this pro-portion raises to 38 percent.

Unsafe water supply combined with lowsanitation and poor hygiene behavior is highly



ing water supply and sanitation in Viet-nam? How many people are at risk of ill-ness, especially in relation to the poorersegments of the population? What are theresulting priorities in water and sanitationimprovements, and should they have ageographic or socioeconomic focus? Thesequestions can be answered by the NationalHealth Survey database. All data, unlessotherwise mentioned, were obtained froma multivariable logistic regression analysisusing detailed information for the morethan 150,000 individuals in the survey.

0

10

20

30

40

50

60

Tap water Drilled well Rainwater Dug well Surfacewater

Othersource

Drinking water source

%

1992 - 93

1997 - 98

2001 - 02

F I G U R E 4 . 3 Drinking Water Supply in Vietnam

Ministry of Health defines clean drinking water as piped city water, drilled well water, rain water, protected moun-tain spring water, and dug well water without a pollution source within a 7-meter radius. A pollution source isdefined as, for instance, a toilet, sewage, or a cattle shelter.

correlated to diarrheal illness in many countries(Curtis and Cairncross 2003; Fewtrell and Colford2004; Esrey and others 1991). Hand washingalone is often found to reduce diarrheal illness byas much as 45 percent (Curtis and Cairncross2003). The poor are generally disproportionatelyaffected by disease, because they tend to have lessaccess to medical services (Figure 4.4), less finan-cial resources to pay for quality services, moreincome losses from illness, and are more oftenself-employed or work in the informal sector.They often have fewer resources to undertakeaverting behavior (actions taken by individualsand households to reduce the risk of exposureto health risks). Averting behavior observed in Vietnam includes purchase of bottled drink-ing water or point-of-use drinking water treat-ment; use of pesticide protection equipment;residence in less polluted neighborhoods; andhygienic behavior.

Unsafe drinking water sources and lack oftoilet facilities are causing substantially higherrates of diarrheal illness. While this finding appliesto many countries, a specific feature in Vietnamis that two very different types of unsafe drinkingwater, surface water and polluted dug wells, con-tinue to be an important source of drinking water(Figure 4.5).



31%

24%

20%

17%

9%

25%

17%

11%

6%

3%0%

5%

10%

15%

20%

25%

30%

35%

Poor Near Poor Average Better-off Rich

Rural Population Urban Population

38%

17%

12%

7%3%

24%

13%

7%

2%0.3%0%

5%

10%

15%

20%

25%

30%

35%

40%


Rural Population Urban Population

F I G U R E 4 . 7 Population without SafeDrinking Water (left) or Toilets (right)

SP10

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%


F I G U R E 4 . 4 Percent of Households Living in Commune with Doctor

wells are the main water source in various low-land areas, except in the Mekong River Delta.The use of dug wells exceeds 75 percent in someprovinces, but is below 5 percent in the MekongRiver Delta. There are six provinces with morethan 25 percent of the population with a pollutionsource near their dug well. This figure reachesmore than 40 percent in some of the provinceswith the highest reliance on dug wells.

The prevalence rate of diarrheal illness variesconsiderably across provinces. An index of theburden of diarrheal illness, created from the sur-vey data and reflecting both prevalence rate andduration of illness, shows that disease burden ishighest in the Northwest and Central Vietnam, aswell as in a few of the provinces in the Northeastand the Mekong River Delta. The correlationbetween the disease burden and poverty incidenceis pronounced (Figure 4.6).

While pollution of water sources may arisefrom human or other sources, such as sewage and



Rain water

Tap waterDrilled well water

Dug well (no pollution source)

Bought water

Surface water

Dug well (with pollution source)

Piped mountain spring water

Other water

F I G U R E 4 . 5 Population Drinking Water Supply 2001/02

ADI = acute diarrheal illness

Percent of population using Percent of Population Percent of dug wells with surface

using dug wells nearby pollution source water

M A P 4 . 4 Population Using Unsafe Water Sources

Use of surface water is common, not only inmarginal areas but also in more developed areas.Surface water is the main water source for morethan 50 percent of the pollution in six Northwestprovinces and in the Mekong River Delta. Polluted

animal waste, the lack of toilet facilities is asso-ciated with both low living standards and higherincidence and longer duration of diarrheal illness.In six provinces in the northern and central partsof the country, more than 50 percent of the pop-ulation lack toilet facilities.

Poverty is associated with significantly higherdiarrheal illness in children, but not in adults.Young children in poor households are at higherrisk not only of illness, but also of diarrheal mor-tality. In the overall population, the use of open,unprotected water sources for drinking water isfound to increase the risk of diarrheal illness by51 percent in children under five years of ageand by 33 percent in the population over five(compared to the average for all other sources ofdrinking water). Children in households havingno toilet facility have a 34 percent higher risk ofillness than children in households with a toiletfacility. There is a 23 percent increase in risk forthe population over five. Children under five aresubstantially more affected by diarrheal illness.At the national level, the diarrheal prevalence ratein children under the age of five is nearly twiceas high in “poor” compared to “rich” households.Comparing across age groups, the prevalence rateis almost four times higher in children under

five years than in older children in the pooresthouseholds, but “only” twice as high in the richesthouseholds. For children under five, the averageduration is 3.3 days in the poorest householdsand 2.8 days in the richest households.

Polluted dug wells are a critical element of thepoverty and environmental health linkage. Nearly35 percent of the population relies on dug wells.Around 20 percent of all dug wells, which corre-spond to nearly 8 percent of the total population,have a pollution source within a 7-meter radius.Of the population that use dug wells, 27 percentof the poorest people have a pollution source nearthe well, while only 13 percent of the richest havea pollution source near the well. The risk of diar-rheal illness in all age groups is 15 percent higherfor individuals using drinking water from dugwells with a nearby pollution source. The risk ofdiarrheal illness is 36 percent higher for thosewithout a toilet facility compared to those witha toilet facility.

Once households have access to simple facil-ities for drinking water and toilets, there is noadvantage—in terms of environmental health—of moving toward improved facilities. Risk ofdiarrheal illness between tap water, clean dugwells, drilled wells, rain water, or piped spring



ADI = acute diarrheal illness

ADI prevalence by Age Group and Quintile

8.2%7.1%

6.6%5.4%

4.2%

2.2% 2.3% 2.1% 2.2% 1.9%

0%

2%

4%

6%

8%

10%

Poor Near poor Average Better-off Rich

Under 5 5 and older

F I G U R E 4 . 6 Prevalence of Diarrheal Illness by Age and Living Standard



Index of disease burden Population with Pollution Population without toiletsSource near drinking water

M A P 4 . 5 Index of Disease Burden and Pollution Sources

27.1% 26.1%23.2%

20.3%

13.1%

0%

5%

10%

15%

20%

25%

30%


F I G U R E 4 . 7 Population Using Dug Well with Nearby Pollution Source

water is similar. The same applies to types oftoilet facilities, such as simple pit latrines, pour-flush, and modern flush toilets except, again,when drinking water with a nearby pollutionsource is used. This finding is worth noting, sincehouseholds have high demand to shift to moremodern toilets. The use of flush toilets increasedfrom less than 10 percent to over 20 percent,while the category “other toilet,” which includestoilets without safe drainage, declined to less than10 percent over the last 10 years. This demandcontinues as the use of simple toilets remainswidespread (at over 30 percent of the popula-tion) and sharing a toilet is common (a almost15 percent).

Three options are available to householdswith a polluted local water source to increasesecurity of their drinking water. These options, bydecreasing frequency, are boiling drinking water,the use of rainwater, and chemical or mechanicaltreatment of water. The poor are less likely toaccess any of these three options. An overwhelm-ing majority of the population boils drinkingwater. Nationally, 78 percent of households reportthat they always boil their drinking water, rangingfrom well over 90 percent of households in mostof the provinces in the Red River Delta and someof the Northeast provinces, to less than 50 percentof the households in many of the provinces in

the Mekong River Delta. 63 percent of the poor-est households always boil their drinking water.This is lower compared to 87 percent of therichest households but still a remarkably highproportion.

Rainwater is increasingly used as an alterna-tive to polluted water sources. More than 50 per-cent of households use rainwater in several ofthe provinces in the Red River Delta, and in theMekong River Delta rainwater use ranges from 10to 30 percent. The poor are less likely to have thisoption. Only 6 percent of the poorest householdsare using rainwater compared to 14 to 16 percentof the “middle-income” groups. No elevated riskof illness was found in this study from rainwatercompared to other improved or protected sourcesof drinking water. More than 20 percent of thehouseholds report that they apply chemical ormechanical treatment to their drinking water.Of the poorest households, 13 percent report theyuse such drinking water treatment. About 27 per-cent of the richest households apply chemical ormechanical treatment to their drinking water.

Behavior, and particularly behavior among eth-nic minority people, appears to be a critical factorexplaining diarrheal illness. Two factors likely tobe associated with behavior are the mother’seducation and ethnicity. Children whose mothershave not completed primary education have a



Population always boiling drinking water (quintile)

62.9%75.1% 79.1% 83.3% 87.3%

0%

20%

40%

60%

80%

100%


F I G U R E 4 . 8 Boiling of Drinking Water in Relation to Poverty

27 percent higher risk of illness than childrenwith mothers who have completed either a pri-mary or a higher level of education. Similarly, inthe population above five years, lack of primaryeducation is associated with a 24 percent higherrisk of illness. “Poor” households are more likelythan “rich” households to rely on surface waterfor drinking water, to have no toilet facility, andto have a lower level of education. No such influ-ence is found for the over-5 population.

While mothers in ethnic minority groups arelikely to have received less education (if any in somegroups), ethnicity in itself is found to explain noless than 44 percent of higher risk of illness amongthe over-5 population. This finding controls forall factors commonly associated with ethnicity,including education. One possible explanationfor the fact that poverty quintile, not ethnicity,is revealed in the analysis for children under 5 isthat poverty reflects the influence of ethnicity ofparents, as ethnic minority people tend be poorerthan the majority of the population in Vietnam.There is, of course, no simple linkage betweenethnicity and behavior. The proportion of house-

holds that take the risk of not boiling water,although small, is clearly located in the provinceswith a high share of ethnic minority populationand in the southernmost provinces of Vietnam.

The relative share of behavior and lack ofaccess to improved water cannot be fully assessed.This derives from the fact that, although such a high share of the population boils its drink-ing water, the use of open, unprotected watersources (such as surface water) is still associatedwith an elevated risk of diarrheal illness. Insuf-ficient water quantity for hygiene purposes,such as hand washing and domestic cleanliness,is likely to result in increased risk of illness,even if drinking water quality after boiling is ofsufficient quality.

Policy implications

Safe water and sanitation programs face a chal-lenging issue of targeting. Sound policies couldsimply give priority to providing safe watersupply and basic sanitation to all those withoutthese services. In reality, competition to accessimproved services arises between poor and non-



Children Under 5 Years Population Over 5 Years

51%

34%

27%

-14%

-20%

-10%

0%

10%

20%

30%

40%

50%

60%

Usingsurfacewater

No toilet Mother notfinishedprimary

Quintile

33%

23% 24%

44%

0%

10%

20%

30%

40%

50%

Usingsurface water

No toilet Unfinisedprimary

Ethnicminority

F I G U R E 4 . 9 Relative Risk of Diarrheal Illness

poor households or communities in a contextof limited budgets. The 135 program allowstargeting at the commune level. Appropriate tar-geting of poorer villages within these communesis critical.

Pro-poor solutions to the dug-well pollutionissue will require affordable technical innovations.Pollution from local domestic and human sourcescan be addressed through improved waste man-agement. It is not so, however, for diffuse pollu-tion of agricultural and, increasingly, industrialorigin. Making sure that the poor have access toaffordable alternative solutions to dug wells willbe critical.

Improving safe water and sanitation in areaswith substantially lower coverage requires closecoordination between the agencies in charge ofsetting up the infrastructure, health services, theorganizations in charge of information and edu-cation, and in some cases the organizations incharge of ethnic minority affairs.

4.2.2 Industrial Pollution, Poverty and Health

Industrial pollution and poverty

Industrial development in Vietnam has been adriving force in the impressive reduction in povertyand the growth in living standards in the lastdecade. The higher the industrial output percapita, the lower the poverty incidence and thehigher the per capita expenditure (Figure 4.10).On the one hand, industry, and especially indus-trial growth, contributes to poverty reduction andan increased standard of living. On the other hand,local communities, which may or may not directlybenefit from the industrial growth in their area,are affected by pollution. Industrial pollutionpresents a serious health threat to workers andthe public, particularly for low-income groups,which can least afford ill-health.

In line with their high industrial concentra-tion, the highest pollution loads are found in two



Percent of Households AlwaysBoiling Drinking Water

Percent of Household Rarely or NeverBoiling Drinking Water

M A P S 4 . 6 Percent of Households Boiling Drinking Water


0%

10%

20%

30%

40%

50%

60%

70%

80%

0 5 10 15 20 25 30

Industrial Output per Capita

Pov

erty

Inci

denc

e 20

04

0

2000

4000

6000

8000

10000

12000

0 5 10 15 20 25 30

Industrial Output Per Capita 2004

Per

Cap

ita E

xpen

ditu

re 2

002/

03

F I G U R E 4 . 1 0 Living Standards and Industrial Output by Province

Poverty incidence is from VLSS 2004, per capita expenditure from VNHS 2002/03.

B O X 4 . 6 Methodology of the Industrial Pollution and Poverty Case Study

The PEN study involved a “macro” and “micro” approach and incorporated existing national, provin-cial, district, commune, village and household data on industry, poverty, water supply, and health.

At “macro” level, the Vietnam National Health Survey (VNHS 2002) was used to analyze healthconditions, employment and demographic characteristics, and household water supply. The nationalresearch team collected district level data from offices in 14 provinces, assembled a national data setof seriously polluting establishments (SPEs) at the commune level, and categorized “craft” industryvillages at the commune level in the Red River Delta from a study by MoST/INEST.

At “micro” level, the national team, with initial support from the Korean Institute for IndustrialEconomics and Trade (KIET), conducted field observations and interviews in several cities and“craft” villages.

Linkages taken into account reflect the complexity of determinants of impacts of poverty onthe poor. The following relations between industry, pollution, human exposure, natural resourcedamage, and impacts on the poor were all taken into account in the analysis:

Industrial Structure

Technology

PoorHouseholds

Non-PoorHouseholds

Exposure

Pollution Determinants:Industrial growthHeavy industryCraft Industry

Process technologyPollution abatement technologyOperation and maintenance

Impact Determinants:Industrial locationPopulation concentrationHousehold microenvironmentsOccupationWater supply and agricultureAverting behavior

Vulnerable groupsHealth statusSocio-economic statusMedical careIncome protection

regions, the Red River Delta and the Southeastregion around Ho Chi Minh City. Inter-regionaldifferences in industrial output are not pronouncedwhen presented as output per capita. The South-east and Red River Delta have a slightly higherper capita output. However, when looking at totalindustrial output, a more likely indicator of poten-tial pollution and health issues, about 70 percentof national industrial output is concentrated inthe Red River Delta and the Southeast. Nearly50 percent of this output comes from the South-east. Little of the industrial output is producedin the Northwest and the Central Highlands(Figure 4.11).

The poor are more likely to be industrial work-ers with higher occupational health risk. In theRed River Delta and the Southeast, 30 to 35 per-cent of urban low-income workers are employedin industry, compared to 15 to 25 percent fromhigh-income households. The difference is evenmore pronounced in the six provinces with thehighest industrial output in Vietnam (Figure 4.12).In Hai Phong, 55 percent of low-income work-ers are employed in industry, compared to only13 percent of the high-income group.

Within the two most industrialized regions ofVietnam, the Red River Delta stands out forhaving rapid industrial growth, which is likely to

generate both more rapid poverty reduction andmore environmental problems. Vietnam’s indus-trial production has increased about 15 percentper year between 1995 and 2004 (Maps 4.7).Regional industrial output growth during thatperiod was 17 percent in the Red River Delta.Eight of the ten provinces with the fastestindustrial growth are in the RRD, Northeast,and Northwest regions. The lowest growth wasin the Central Highlands and the MekongRiver Delta.

Linkages between poverty and geographicaldistribution of industrial development and pol-lution strongly relate to the type of industrialsector—State versus non-State. Non-state indus-trial output is strongly associated with higherliving standards and lower poverty. Provincialpoverty incidence is 0.2 percent lower for every1 percent higher industrial output per capita fromthe state sector, and 0.4 percent lower for every1 percent from non-state industry. Similarly, theprovincial per capita expenditure level is 0.1 per-cent higher for every 1 percent higher industrialoutput from the State sector, and 0.15 percenthigher for every 1 percent from the non-Statesector. Compared to the state sector, the largerinfluence of the non-State industrial sector onprovincial living standards might be due to the



0

5

10

15

20

25

30

Red RiverDelta

North East North West NorthCentralCoast

SouthCentralCoast

CentralHighlands

South East MekongRiver Delta

Output per Capita Output per Urban Capita

F I G U R E 4 . 1 1 Regional Share of National Industrial Output in 2004

Figures are Million Dong (1994 constant prices).

non-State sector’s generally smaller scale andhigher labor intensity.3

The non-State sector is at the same time thesector with the highest correlation to poverty red-uction, and the sector in which more polluting—or more pollution-intensive manufacturingsectors—are mostly found.4 High-pollution-intensive sectors represent almost 30 percent oftotal manufacturing output. While non-Statemanufacturing output was 27 percent of totalmanufacturing output in 2004, 37 percent ofoutput from high-pollution-intensive sectors camefrom non-State manufacturing. State- and for-eign-invested manufacturing growth is relatively“pollution neutral” since State manufacturing isparticularly high in the chemical and basic metalssectors and foreign-invested manufacturing ishighest in the chemical and tanning and leather



0%

10%

20%

30%

40%

50%

60%

70%

RRDregion

SE region HANOI HCMC BinhDuong

Ba RiaVungTau

HaiPhong

Dong Nai

Low income High income

F I G U R E 4 . 1 2 Industrial Employment in Urban Vietnam

Source: Derived from VNHS 2002.

Industrial OutputPer Capita 2004

Industrial OutputPer Urban Capita 2004

Provincial Output Growth1995-2004

M A P S 4 . 7 Provincial Industrial Output

product sectors. Non-State manufacturing dom-inates in the fabricated metal products and rubberand plastic sectors (Figure 4.13). These high-pollution-intensive non-State sectors grew in therange of 25 to 30 percent per year, nearly twiceas fast as low-pollution-intensive non-state sectors(Figure 4.14).

Among the non-state sector industries, thecraft industry has witnessed phenomenal growth,especially in the Red River Delta region, where60 percent of the craft industry is located. Accord-ing to MoST/INEST (2005), there are more than840 craft villages in the RRD region. The concen-tration of the craft industry is particularly high



0

4000

8000

12000

16000

20000

Chemicals;Chemicalproducts

Tanning;Leatherproducts

Fabricatedmetal products

Rubber; Plasticproducts

Basic metals Paper; Paperproducts

Coke; Refinedpetroleumproducts

Recycling

State Noe-Sate Foreign invested

F I G U R E 4 . 1 3 High-Pollution-Intensive Manufacturing Output, 2004

Billion Dong in 1994 constant prices

0%

5%

10%

15%

20%

25%

Total State Non-state Foreign invested

Low pollution intensive sectors High pollution intensive sectors

F I G U R E 4 . 1 4 Manufacturing Output Growth, 1995–2004

in Ha Tay and Thai Binh provinces. A quarterof the rural communes in the RRD have at leastone craft village, with some communes havingmore than five.

Rural industries were in the past limited tomines. Mines are located in a mix of poorer andless-poor provinces. They are likely to draw poorerworkers, and their concentration creates highpotential environmental impact on those livingnearby. There are 16,000 hectares formally allo-cated to mining in Vietnam. The area is con-centrated in 11 provinces. In these provinces,0.1 percent of total land area, often concentratedin a few districts, is allocated to mining. Many ofthese provinces are in the Cau River basin.

Industrial pollution in Vietnam is shifting froman urban-only issue to a mixed urban-rural issue.The government has made significant progress inrelocating industrial enterprises away from manyof the major city areas to industrial zones. Thishas alleviated some of the pollution in and nearresidential areas. Some enterprises in the citieshave also installed pollution control technology.These improvements are not complete; numeroussmall- and medium-scale enterprises and remain-ing large enterprises continue to pollute the urban

environment in Vietnam. However, pollutionfrom industrial zones is now increasingly impact-ing the nearby environment, including waterresources, paddy fields, and aquaculture farms.

A major focus of Vietnam’s expanding indus-trialdevelopmentand itspossiblepollution impacts(also on different socio-economic groups) hasbeen the expansion of craft village (CV) industries,now tentatively accounting for almost 20 percentof Vietnam’s Industrial Output Value (KEI 2005).More than 55 percent of Vietnam’s CVs arelocated in the Red River Delta in North Vietnam,with a particular high concentration in Ha Tayand other provinces like Thai Bihn and NamDihn (figures 4.15a and 4.15b). Initial mappingshows that the same pattern appears for the mostpolluted CV industrial enterprises with highestconcentrations in areas located North West andSouth & South East of Hanoi (figures 4.15b). Acritical pattern is that CV industries are creatingpollution patterns in rural communities that werespared earlier from industrial pollution.

There was no difference in poverty incidencein rural communes with and without craft villagesin the RRD in 1999. Average incidence wasremarkably close to the national average. More



0

50

100

150

200

250

300

Ha Tay ThaiBinh

NamDinh

HaiDuong

BacNinh

Hanoi HungYen

VinhPhuc

HaNam

NinhBinh

HaiPhong

F I G U R E 4 . 1 5 Craft Villages in the Red River Delta

Source: From data provided by INEST, Hanoi Polytechnic University, Vietnam.

recently, poverty incidence may have declinedfaster in communes with craft villages, but nosystematic data is available to confirm this sup-position. The craft industry typically consists ofhousehold enterprises. On average, about 60 per-cent of households in the craft villages are engagedin the industry in the RRD, and in a quarter ofthe villages more than 80 percent of the house-holds are engaged. Households engaged in thecraft industry usually have a substantially higherincome than the agricultural households in theirvillages and in the RRD in general. As the house-hold enterprises grow, they hire workers who aregenerally poor people. This is most pronouncedin Ha Tay and Thai Binh provinces, the provinceswith the highest number of craft industries. Thirtypercent of low-income workers are engaged inindustry in these provinces, while only 13 to17 percent in the high-income group are engagedin these activities (Figure 4.16).

Communes with lower poverty incidence in1999 tend to have craft industry with higher pol-lution intensity (Figure 4.17). Craft villages engagein a variety of production activities that range fromnon-polluting to highly polluting, some with longand important cultural traditions and some witha recent commercial and industrial nature. Activ-ities include highly skilled embroidery, weaving,pottery work, and ceramics to food processing,construction materials, and waste recycling, suchas plastic and metals. While the causes of thisrelationship are not clear, it could suggest a trade-off between income generation and environ-mental quality in relation to the development ofcraft industry.

Analysis of the Vietnam Health Survey revealssubstantial inequality in health conditions inurban areas. While it is not possible to establisha relationship between health effects and air pol-lution in urban areas from this survey, this findingdoes suggest that controlling air pollution fromindustry and other sources could provide substan-tial health benefits for low-income households.

The burden of ill health in urban areas, whichmostly affects children to the elderly, is clearly

linked to poverty. Air pollution causes cardio-pulmonary mortality in adults, acute respiratorymortality in children, and respiratory illness inchildren and adults, resulting in hospitalizations,lost work days, and lost school attendance). It isthe elderly and young children that are most atrisk of health effects from air pollution.5 In urbanchildren under 5 years old, nearly 35 percent of allcases of acute respiratory illness (ARI) occur in the20 percent lowest income households. In contrast,only 12 percent of the cases are in the highestincome households (Figure 4.18). In the popu-lation over five years of age, the prevalence of ARIis two times higher in low-income than in high-income households.6 People from low-incomeurban households also die younger. The averageage at death was 60 years in the three lowestincome groups, compared to 66 years in the twohighest income groups (Figure 4.18). Numbersof children are disproportionately high in lower-income households. About 30 percent of childrenin urban areas in Vietnam live in the 20 percentlowest income households.

Visual observation of areas of residence of thepoor revealed the degree of higher exposure ofthe poor to air pollution. The poor in urban areastend to live in cheaper but more polluted andenvironmentally unsafe areas. The urban pooralso are disproportionately affected by occupa-tional health risks. They are over-represented inindustries with high occupational health risk andexposure to pollution, particularly in small-scaleindustries with no pollution control technologyand worker protection. In a lead recycling craftvillage, for example, workers operate open fur-naces to melt batteries. In a tanning enterprise,workers are in direct contact with wastewater. Inaddition to these pollution and health issues inurban areas, evidence is mounting of serious healtheffects from craft industry pollution in rural andperi-urban areas (Box 4.8).

Craft industries generate both air and waterpollution. Craft industry wastewater pollutionconstitutes a health risk to poor households inparticular. Untreated wastewater from craft indus-





MA

PS

4.8

Cra

ft V

illag

e C

on

cen

trat

ion

s an

d P

ollu

tio

n H

ots

po

ts in

th

e R

ed R

iver

Del

ta

Sou

rce:

Map

s g

ener

ated

by

com

bin

ing

INES

T C

V d

atab

ase

and

MO

NR

E’s

clas

sifi

cati

on

of

po

llute

d e

nte

rpri

ses.

tries increases the risk of contamination of dugwells. The poor are less likely to own craft enter-prises. They are often migrants including migrantsfrom minority ethnic groups. The owners areoften also exposed to pollution but not to directoccupational health risks, and they have moreaccess to alternative drinking water sources andto health services.

Policy implications

Environmental regulatory improvements canachieve a “win-win” poverty and environmentimpact if efforts are directed toward reducing theimpacts of industrial pollution on the health oflow-income groups. Environmental health inrelation to industrial pollution relates more topolicies for low-income groups than to povertyreduction. There is a broad need for reform inthe regulatory framework and for parallel enforce-ment of this framework in the workplace and inresidential areas.

At present, the challenges that Vietnam is fac-ing in the context of craft industry developmentare similar to the challenges that China and

Korea have faced in the development of SMEs(Box 4.9).

Impact of craft industries in and around theRed River Delta on the health of the poor requiresurgent attention. Uncontrolled development ofsmall and medium industrial enterprises wouldresult in a drastic increase in pollution. Preventingthis uncontrolled development will require a mixof measures. First, a national regulatory frame-work for craft industries needs to be in place and



0%

5%

10%

15%

20%

25%

30%

35%

Ha Tay Thai Binh Nam Dinh Hai Duong Bac Ninh RRD region

Low income High income

F I G U R E 4 . 1 6 Industrial Employment in Rural Vietnam

Source: Derived from VNHS 2002.

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

low povertyincidence

medium povertyincidence

high povertyincidence

F I G U R E 4 . 1 7 Craft Village Pollution Intensity and Poverty Incidence

Source: Derived from data prepared by Tai, Thang et al (2006). Villages are classified low=1, medium=2, and high=3pollution intensity based on type of craft production.Industrial Pollution and Health



B O X 4 . 7 Craft Village Studies in Vietnam

Three comprehensive studies of the craft industry have been carried out in recent years. MARD/JICA(2005) identified over 2,000 craft industry villages, defined as having at least 20 percent of house-holds produce the same craft or at least 20 percent of village income from craft products. MoST/HPU(INEST) (2005) identified 1,450 craft villages, including 840 in the Red River Delta, defined as villageswith craft production constituting at least 50 percent of village income or having 30 percent ofhouseholds engaged in craft production. MONRE/VEPA (2003) focused on 366 craft villages in sixprovinces in Nhue-Day River basin. The studies by MoST and MONRE concentrated on the envi-ronmental aspects and workers’ occupational health risks in craft villages. KEI, Korea (2005) alsoreviewed environmental conditions in craft villages and explored potential environmental management solutions to the problems.

Percent of ARI Cases (Children under 5 years) Average Age at Death (last 3 years)

0%

5%

10%

15%

20%

25%

30%

35%

40%

1 2 3 4 5

Living Standard Quintiles

565758596061626364656667

1-3 4-5

Living Standard Quintiles

F I G U R E 4 . 1 8 Burden of Ill Health in Urban Areas

Source: Calculated from VNHS 2002.

B O X 4 . 8 Main Findings by the Vietnam Background Study

– Industrial pollution in many cases impacts all people, poor and non-poor alike but the poorhave fewer mechanisms to cope with the impacts from pollution.In Chi Dao commune, in a lead battery recycling craft village, both the poor, non-craft house-holds, and wealthier craft households were impacted significantly by the different types of pol-lution. The craft households (there were only 25 in the 500-household village) live and worknearby heavily polluted water. The wastewater, which has been measured to hold 15 times themaximum lead-content standard, ends up in ponds or in the ground near the craft households.The poor are usually hired by the craft households to do the melting and collecting of the recy-cled lead at the furnaces, and thus are directly exposed to the toxic fumes. The lead content inthe air in Chi Dao village is estimated to be about 4,600 times higher than the standard.

In Du Dai, a food processing craft village in Thai Binh Province, water pollution caused bylivestock production and noodle steaming facilities affect the poor and non-poor indiscrimi-nately. The poor however cannot afford health care treatment or deeper wells. Many of theworkers are ethnic minority people.

– In the major cities, the poor live in locations with higher exposure to industrial pollution and higher risk to health damage and disasterIn District 6 in Ho Chi Minh City, an inner-city district with the highest poverty incidence (9.19 per-cent) of the districts studied, an open drainage canal that flows through the district is the causeof serious environmental pollution throughout the district. The poorest people, who live in slumareas along the canal, are most severely affected from the exposure to this pollution source.

Source: Tai, Nguyen Van and others (2006).

needs to be enforced. Second, mechanisms tomodernize production facilities and use improvedtechnology are required. Attention to the impacton low-income groups should be incorporated atall stages. While wastewater pollution control fromcraft industry is a priority, immediate attention isalso needed in neighboring communes and areaswith high reliance on shallow dug wells. Waterquality monitoring of wells should be expanded,and programs implemented to provide local peo-ple, especially the poor, with safe water supply.

Pesticide Use and Poverty

Pesticides are the third and last specific researcharea in linkages between poverty and environ-

mental health. Pesticides relate to the agricul-tural sector and may cause diffuse pollution buttheir linkage to poverty is the potential environ-mental health risks that they generate.

Pesticide use in Vietnam started as early as the1950s. Pesticide consumption grew significantlyduring the 1990s, increasing from 20,000 tonsin 1991 to more than 40,000 tons in ___ (FAO2004). These increases followed crop intensi-fication and increased cultivation of crops thatrequired higher pesticide applications, such asdry-season paddy or fruits and vegetables. In addi-tion to the intensification of pesticide use, there isevidence of improper pesticide use and its impacton health. Hospital admission records in Vietnam,for example, trace nearly 11 percent of all poi-



B O X 4 . 9 Relevant Experiences of Environmental Pollution of SMEs in China and Korea

Since the late 1980s, the main engine of China’s industrial development has been industrial small-and medium-enterprises (SMEs), often referred to as Township and Village Industrial Enterprises(TVIEs). In China, TVIE’s share of gross industrial output value increased from about 25 percent in1990 to 58 percent in 1998, while industrial output value itself increased 20-fold between 1985 and2004. From 1989 to 1995, TVIEs accounted for almost the entire increase in industrial-based waterpollution loads. TVIE-generated particulate emissions (industrial dust) tripled and TVIE-generatedsolid waste volumes more than doubled. While industrial state-owned enterprises (SOEs) were sub-ject to a set of government pollution control policies from the 1980s, the TVIEs were not subject tosuch regulations. When the pollution loads and their environmental impacts were realized throughthe nationwide survey of more than 1 million TVIEs in 1995 and 1996, the government decided in1997 to close down about 70,000 small enterprises. As a result, pollution loads were driven down in1998 and 1999. However, China is still struggling with the aftermath of the earlier “laissez-faire”TVIE policies. Some of the critical lessons are:– It was particularly in the period when the TVIE contribution expanded that major pollution

increases took place.– The regulation to control pollution from TVIEs should have been in place at the outset of this

expansion, not at its end.– Technical, economic, and organizational means to control pollution from TVIEs should have

been developed during the first phases of its expansion, not after.Korea followed a similar path in industrial development and environmental pollution of SMEs.SMEs grew rapidly in the late 1970s and the early 1980s as government support was redirectedaway from the large conglomerates. The majority of SMEs were concentrated on high pollutingindustries such as textiles, dying, leather, foundries, metal plating, paper manufacturing, andcement manufacturing. SME-generated pollution loads were estimated to account for more than80 percent of the entire increase in industrial pollution loads from the mid-1970s to the early1980s. The 1963 Pollution Prevention Law failed to contain industrial pollution due to a socialatmosphere that placed higher priority on economic growth over environmental protection. Fol-lowing increasing pollution due to the proliferation of SMEs, the1977 Environment PreservationAct introduced comprehensive environmental standards and total pollution load management,mainly targeting SMEs. Government’s financial support to SME investments in pollution preven-tion facilities proved to be the most efficient policy. From 1983, the Korean Government began to provide long-term low interest loans and tax incentives for environmental investments and thedownward spiral of industrial pollution was gradually reversed.

sonings to pesticide misuse, with approximately840 pesticide poisonings in 53 provinces in 1999(Poison Control Center 2000). Available studies,however, do not provide any analysis of the bur-den of misuse shared by the poor.

At the national level, the National Health Sur-vey reveals a clear pattern of higher pesticide useamong the rural poor than among the non-poor(Table 4.6). There is also a pronounced patternin the application techniques. Of those farmerswho use pesticides, the poor use protective equip-ment during pesticide spraying much less fre-quently than the non-poor.

The lowlands do not show higher pesticide usethan the uplands. First, the use of pesticides isprevalent in the whole country (Map 4.2). Second,the upland regions of the Northeast and the Cen-tral Highlands, which specialize in high-valueperennial crops, have several poor provinces withthe highest prevalence of pesticide use in thecountry.

The government has started an extension pro-gram of good agricultural practices that includeoptimizing the use of pesticides. Just as in othercountries in the region, this shift is necessary toadjust to international market quality standards,and is also welcome in terms of improved envi-ronmental management. If this shift is to be suc-cessful, however, policies have to ensure that thepoor and non-poor farmers, who are important

pesticide users, have access to appropriate agri-cultural inputs and techniques.

Eighty-six percent of the surveyed house-holds report spraying pesticides every year. Pes-ticide use is highest—92 percent—among thepoorer 20 percent of households, and lowest—75 percent—among the better-off 20 percent ofhouseholds. The Vietnam Health Survey data forthe Mekong River Delta is overall lower, with58 percent of rural households using pesticides,but the linkage between income group and pes-ticide use is similar.7

Poor farmers in the survey use on average lessthan half of the total reported annual pesticideuse among the non-poor households. This is notdue to a more limited range of products used.There is no difference in the use of insecticides,fungicides, and herbicides among the poor andthe non-poor. Crop intensity is similar. Eighty-nine percent of poor farmers in the sample and60 percent of the non-poor cultivate less than1.5 ha of land. Rice is cropped three times a year.Application times over one year are 4.5 times forthe poor, which is slightly less than the 5-timeaverage for the non-poor. Average quantities perapplication, therefore, make the difference.

There is a striking difference in toxicity betweenpesticides used by poor households in the surveyand pesticides used by the non-poor. Whenquantities were weighted based on their toxicity,it appears that the poor use a higher amount of“risk-weighted” pesticides (Figure 4.19). Slightlymore than half the total quantity of pesticidesused is, however, only moderately toxic based onthe WHO classification (category II). None ofthe pesticides that the poor or the non-poor reportusing are extremely hazardous (category Ia) andonly 0.5 percent of the total quantity is from thesecond-highest category of toxicity (category Ib).

The poor in the survey use masks and otherprotective measures less often; this is statisticallysignificant at the 10 percent level. Around 42 per-cent of the poor do not use a mask and are, there-fore, likely to use no protection other than a shirtand trousers, whereas only 36 percent of the non-



T A B L E 4 . 6 Share of Rural Households Using Pesticides

HH using

HH using

protective pquipment

Quintile pesticides Always Sometimes Never

1 64% 57% 18% 25%2 73% 64% 17% 18%3 68% 67% 16% 17%4 52% 70% 14% 16%5 35% 72% 11% 17%Total 61% 65% 16% 19%



M A P S 4 . 9 Paddy Area Distribution in Vietnam

B O X 4 . 1 0 The Pesticide Use Survey in the Mekong River Delta

Why is pesticide use more prevalent among poor farmers? Is the use of hazardous pesticides alsomore prevalent among them? Do these facts relate to different awareness levels and behavior patterns about risks related to pesticide use? And do the poor differ from the non-poor in accessto information on risks related to pesticide use and to protective measures? These are the ques-tions that a survey in the Mekong River Delta undertook to answer. The study was also seeking toanalyze whether health impairments attributable to pesticide use were disproportionately higheramong poor farmers.

The study surveyed 603 rice producers in 10 districts about their cropping systems, pesticide useand practices, application precautions, averting behavior, and health effects.

The Mekong River Delta, the largest region of intensive rice production in Vietnam, was selectedbecause it offered the advantage of a simple and intensive production system. This avoided discussingdifferences in pesticide use among a large range of crops.

Households with per capita income below 1.2 Million VND/year (13 percent of the sample) aredefined as poor and compared with other households defined as non-poor.

The full report of the national team is available on CD-Rom (Dasgupta et al 2005, Nguyen 2005)

poor do not use a mask. The difference is morepronounced for protective equipment with lessfrequent use, such as gloves and glasses. Whenasked why, close to one-third of poor farmerssaid that masks were not available, compared toonly one-fourth of the non-poor. Only 34 per-cent of the poor stated that they believed masks

were uncomfortable, compared to 41 percent ofthe non-poor.

Poor farmers also use alternative plant pro-tection methods less frequently. Only 9 percentof the poor state that they used integrated pestmanagement (IPM) technologies, a number sig-nificantly lower than the 16 percent of non-poorfarmers who cited the method. Similarly, 79 per-cent of the poor farmers—versus 68 percent of thenon-poor—stated that they did not use any pestcontrol method other than pesticides.

By contrast, the poor appear to be at least asaware of risks from pesticide use and to have goodaccess to information. Both the poor and thenon-poor in the survey make clear statementsabout their perceived risk from pesticide use. Only3 percent of the poor and 1 percent of the non-poor did not answer the question. 30 percent ofthe poor stated that using pesticides generated ahigh risk, a proportion almost 50 percent higherthan for the non-poor (Table 4.7).

Poor farmers have a fair level of access to infor-mation. Only 5 percent of the poor, or 2 percent-



0.0

1.0

2.0

3.0

4.0

5.0

6.0

Application amount ***(kg)

Risk-weighted amount(per 100 kg)

No. of applications **

Poor (n=79) Non-poor (n=524)

F I G U R E 4 . 1 9 Application Amount,Risk-Weighted Amount(kg), and Number of Applications

***Statistically significant at 1 percent level of significance

**Statistically significant at 5 percent level of significance

0

10

20

30

40

50

60

70

80

90

100

Shirt Trousers Mask Hat Gloves Glasses Shoes

Per

cen

tag

e


F I G U R E 4 . 2 0 Use of Protective Equipments

T A B L E 4 . 7 Perceived Risk of Pesticides(above) and Access to Information (below)

Source Poor Non-poor

Public media 38.8 40.9Pesticide companies 17.8 18.5Other 14.4 12.8Agricultural extensionnists 17.7 11.6Pesticide retailer 5.5 9.1No information available 5.5 7.1NGO 0.2 0.1Total 100.0 100.0

Perceived Risk Poor Non-poor

High risk 30 21Medium risk 25 26Small risk 28 31No risk 14 20Don’t know 3 1Total 100 100

Sample size (n=) 79 524

age points less than the non-poor, state that noinformation is available on pesticides. All farmersidentify the public media as their primary sourceof information on pesticides. Pesticide compa-nies and the staff of the local agricultural exten-sion offices rank second as information sources.The level of education probably plays an impor-tant role in accessing information. Only 1 percentof the poor in the survey have no primary edu-cation, and more than 45 percent have studiedbeyond primary school. Availability of informa-tion from agricultural extension agents is alsoimportant. These are a main source of informa-tion for 18 percent of the non-poor, a proportion50 percent higher than for the non-poor. Morethan 60 percent of farmers in the survey are awareof pollution from pesticides through water, and27 percent state that wildlife could be impacted.

The usefulness of accessible information, there-fore, appears to be a critical factor that triggersmore cautious behavior in relation to pesticideuse. The poor appear to have less access to train-ing and to information that allows real choices.Only 29 percent of the poor compared to 34 per-cent of the non-poor in the survey say that theyhave participated in formal training in safe han-dling of pesticides. While this difference is notsignificant, the difference in access to IPM train-ing is. IPM is an appropriate method for all farm-ers, regardless of their living standards, but while60 percent of farmers in the two upper quintileshave received IPM training, only 37 percent ofthose in the lowest quintiles have participated insuch training.

The usefulness of accessible information, there-fore, appears to be a critical factor that triggersmore cautious behavior in relation to pesticideuse. The poor appear to have less access to train-ing and to information that allows real choices.Only 29 percent of the poor compared to 34 per-cent of the non-poor in the survey say that theyhave participated in formal training in safe han-dling of pesticides. While this difference is notsignificant, the difference in access to IPM train-ing is. IPM is an appropriate method for all farm-

ers, regardless of their living standards, but while60 percent of farmers in the two upper quintileshave received IPM training, only 37 percent ofthose in the lowest quintiles have participated insuch training.

Suppliers also play an important role in pes-ticide use. Seventy-six percent of farmers in thefirst quintile purchase pesticides on credit, com-pared to 54 percent in the highest quintile. Thepoor thus depend on credit providers for productselection. The non-poor use pesticide retailers asinformation sources, but may be in a better posi-tion to compare. Only 21 percent of farmers in thelowest quintile regularly change the products theybuy; 32 to 42 percent of those in the upper twoquintiles do. Poor farmers state more often thatthe important factor in their choice of pesticidesis price, not technical effectiveness or safety.

The poor may also be unable to avoid pesticidetoxicity when they are hired for pesticide spraying.This situation is not uncommon, as indicated bythe substantial proportion of rice growers in thehighest quintile of the survey who do not reportspraying pesticides (25 percent). Many of themprobably hire other farmers to spray pesticideson their fields; these are likely to be poorer indi-viduals or even landless people. Rural poverty isoften related to landlessness in the Mekong RiverDelta.8

Potential differences in health impacts are dif-ficult to identify. The tests carried out in this studymay only have an indicative value, but they indi-cate substantial exposure to pesticides. Around60 percent of farmers experienced skin irritations,headaches, dizziness, eye irritations, shortness ofbreath, and other acute short-term health effectsafter spraying pesticides (Figure 4.21).

There is no clear distinction in how the poorand the non-poor are affected, but there is highawareness of the issue among both the poor andthe non-poor. When asked whether they believedthat these symptoms were related to pesticide use,81 percent of poor and 82 percent of non-poorfarmers responded affirmatively. Different levelsbetween the poor and non-poor are visible in some



cases, but may not be significant. Blood testsindicated significant exposure to organophoshatesor carbamates among 42 percent of the poorversus 32 percent of the non-poor.9Thirty-eightpercent of the poor versus 31 percent of the non-poor reacted positively to patch-skin tests forcontact dermatitis, indicating existing exposureto pesticides. In subsequent specific tests of reac-tion to three commonly used pesticides, only 15 to25 percent of the farmers tested positive, andthere was no clear distinction between poor andnon-poor.

Policy implications

There are three main reasons why the differencesobserved among income groups in relation topesticide use, although small in absolute terms,deserve attention in terms of policy making. First,the large range of variables that show a correlationwith income groups provides a solid body of evi-dence of a poverty and environment linkage inpesticide use. Second, many of the differencesobserved are probably on an increasing trend.As the non-poor become increasingly able to avoidthe use of more toxic pesticides, they may con-tinue to be directly exposed to these pesticides.Third, the poor and the non-poor alike are subject

to potential health impacts in an environmentwhere paddy fields are predominant. Waterbornediffuse pollution can become a substantial prob-lem for all those who live in this environment.

The national policy to reduce the use of toxicpesticides is a sound policy with an overall positiveimpact both in terms of poverty and environment.This policy can, however, gain from careful incor-poration of problems that are specific to the poor.Poor farmers tend to participate less in agriculturalextension unless a specific effort is made to reachout to them. The relatively higher levels of educa-tion and access to information create an opportu-nity for information programs on pesticide use.These messages can help them change attitudesand practices when faced with a pesticide issue.

4.2.3. Poverty and Environment inRegional Programs: The CauRiver Basin



0

10

20

30

40

50

60

70

Skinirritation

Headache Dizziness Eyeirritation

Otherproblems

Shortness ofbreath

Fever Vomiting Convulsions Diarrhea

Per

cent

age


F I G U R E 4 . 2 1 Self-Reported Health Impairments AfterUsing Pesticides (% of farmers)

Key findingsIn contrast with classical environmental aware-ness messages, poverty and environment link-age areas cover not only forestry but also lackaccess to safe water and sanitation and impactfrom the mining sector in the upstream. Land-scape degradation is an issue in addition toindustrial pollution in the downstream. Local-ized environmental impacts are at least asimportant as impacts from one section of theriver basin on lower section one.

Social differentiation within communities isincreasing rapidly and is accompanying indus-trial pollution. This creates a need for publicintervention to go beyond pollution remedia-tion. Social differentiation with the upstreamand downstream section is extremely high andis generating a need for balance within the riverbasin in the allocation of investment resources.

The village level is a key level for targetingspecific investments, while the province is akey level for environmental awareness andstakeholder coordination.

The beautiful landscapes and clean waters ofthe Cau River in northeast Vietnam are describedin a well-known traditional song in Vietnam.The river basin is today one of the three highestnational priorities in terms of environmental man-agement. The Cau River originates in Bac KanProvince. It then flows through Thai NguyenCity, a seriously polluted industrial zone in thecountry (Vietnam Environmental Monitor 2003).Bac Ninh Province in the Red River Delta is oneof the downstream provinces and is home to ahigh proportion of the 200 craft industry villagesin the river basin. The Cau River basin is amongthe most important areas in Vietnam in terms ofpopulation concerned with water resource man-agement issues. An interim steering committeewas set up initially to coordinate water resourcemanagement in the river basin. A 13-year masterplan to protect the Cau River’s environment wasreleased in 1998. Water resources and pollutionhave been analyzed, and a substantial number ofengineering solutions have been incorporated in

the provincial development plans. Implementa-tion of the master plan, however, faces severebudget limitations. The Cau River basin wasproposed in 2004 as a candidate to pilot an inte-grated sustainable development program underVietnam’s national Agenda 21.

Poverty and environment in the upstreamcau river basin

Forestry in Bac Kan is an important part of landuse and is clearly household-based, but the sec-tor provides limited income. Close to two thirdsof total land area within the river basin in theprovince and in the districts surveyed has beenzoned as production forest land. Most forest isnatural forest zoned for production due to a sparseforest cover. Satellite images show that activedeforestation took place in the 1990s within theriver basin, and that the limited reforestationoperations have mostly taken place in more acces-sible areas (Castella and Dang 2002). Most of the



B O X 4 . 1 1 Comprehensive Analysis of Poverty and Environment Linkages in the Cau River Basin

How do various poverty and environment issues interact within the limited geographical domainof the river basin? Can these interactions serve as a basis to prioritize investments in the existingmaster plan and in the proposed sustainable development program? And what can be learnedfrom these interactions to design environmental awareness activities within the framework of the national Agenda 21 program? A comparative analysis was carried out in two provinces in theCau river basin to answer these questions.

A 700-household survey in six communes was combined with a methodology of rural appraisalwith focus group discussions, direct observation, and interviews in Bac Kan Province, the upstreamprovince, and Bac Ninh Province, one of the downstream provinces with a high concentration ofcraft industries. Actual household incomes have been recorded, which allows for analysis by quintile.

The survey was initially designed to analyze natural resource use in the upstream river basinsurvey, but the magnitude of mining and drinking water issues led to include these as well.

The establishment of a multisectoral commune-level database was envisioned but turned out toraise substantial difficulties10. Instead, we use district-level data from the PEN II land use databaseand the national health survey.

The master plan is designed to benefit the 54 districts of the 6 provinces that are members ofthe river basin committee11. The analysis takes into account this level as well as (1) the 30 riverbasin districts defined as affected by the Cau River in the master plan;12 (2) the impacted districts,or those districts plus the 12 districts of Hai Duong Province; and (4) the 16 riverside districts thatdirectly create or receive pollution impacts along the main riverbed (Map 1). Area and populationcovered each vary from 1 to 3.7 between definitions (4) and (1).

The full report of the national team is available on CD Rom (Socio-economic developmentcenter 2005)

area is in a stage of natural regeneration. Ninety-five percent of households have been allocatedforest land. Land that remains under a state forestenterprise provides fuelwood and benefits fromprotection contracts to the households.

The four districts located within the physicalriver basin are substantially less poor than thethree more remote northern districts locatedoutside of the river basin. The Dao people, anethnic group generally living in disadvantagedconditions compared to the main ethnic groupin the province, the Tay, account for a high pro-portion of the population in all districts but ahigher one in the northern districts. Environ-mental stakes differ as well. Forest in the northerndistricts is mostly zoned for environmental pro-tection and includes the Bac Be National Park.Slopes are markedly steeper than within the riverbasin. The impact of rotational agriculture, apractice that the Dao have retained in all districtsto complement their small areas of paddy fields,is likely to be higher than within the river basin.Erosion within the river basin is generated bycauses broader than rotational agriculture andmining is an important cause.

The mining sector is present in three of BacKan’s four river basin districts. Its environmentalimpact on soil and water resources appears to besubstantial. Cho Don district plays a crucial rolein the water balance of the whole river basin,because of its location at the source of the CauRiver. Cho Don is also the district with the largestproportion of land allocated to mining uses—2 percent of total land area13.

A combined issue of water quantity and waterquality is stated by households as one of theirmain problems and affects the poor more thanthe non-poor who can afford alternative sources.Close to 85 percent of households mention watershortages during the dry season. Twenty percentof households in the water source commune haveshifted from dug wells to gravity water pipes,because of declining water quality in the dug wells.This desire to shift to a more modern type ofwater supply can be related to the higher incomes

available from local employment in mines. Averageincomes recorded in that commune are 45 per-cent above the other two communes. However,social disparity also increases with mining andother non-agricultural activities. The income ofthe richest 20 percent of the surveyed householdsis more than 9.5 times the income of the poor-est 20 percent in the river source commune withthe large mine. It is only 2.9 times greater in thefully agricultural and forestry communes. Poorhouseholds are often not in a position to avoidwater pollution in dug wells.

Both large and small mines generate pollutionissues. In one village within one of the three com-munes surveyed, a boom in illegal gold mining inthe early 1990s attracted several thousand infor-mal workers. The activity has been banned since,but households express high concern over waterquality. They claim that they have lost domesticanimals, an important household asset for house-holds in poor communities, due to water toxicity.

Social differentiation is equally visible in accessto forest resources. Poorer households have beenallocated 3.6 ha of forest land on average, signif-icantly lower than the average of 5 ha for theentire sample of households. About 10 percent ofhouseholds have been allocated more than 10 haand up to 67 ha of forest land. Poorer householdsare also allocated forest that is more distant.14

Allocation of distant forest land means feweropportunities for households to use this land togrow timber or cash trees. The number of largedomestic animals, another indicator of access toforest resources since large animals in the areamostly feed in the forests, is also clearly linked toincome groups.

Differences in poverty levels between villagesare visible. The average income in higher-incomevillages is 1.2 to 1.5 times the income of thepoorer villages within each of three communessurveyed. These differences have a clear ethnicdimension. Two of the villages surveyed are fromthe Dao ethnic group in a mixed Tay, Dao, andKinh commune, while the other villages belongto fully Tay communes with only a few Kinh





M A P S 4 . 1 0 Cau River Basin

Upper left: boundaries of river basin. Upper right: population change. Lower left: ethnic minority people in total population. Lowerright: poverty incidence to replace Dao ethnic group.

households. The two Dao villages rank secondand third poorest among the 11 villages surveyed.Paddy accounts for 100 percent of agriculturalland in all Tay villages (except the two higher-income villages, where the land may have moreprofitable uses). It accounts for only 15–20 per-cent of agricultural land in the two Dao villages.The development of tea and other perennial cropson deforested forest land is low −1.3 ha oneaverage - in all villages due to limited forestrybudgets available for the provision of seedlings.The Dao villages have planted the smallest areas.

Ethnicity sometimes relates to poverty andnatural resource use in opposite directions. House-holds in the two Dao villages have been allocated6 and 11 ha of forest land on average, significantlymore than in the seven higher-income villages,where forest land allocated to households is only4 ha on average. This is due to larger forested areasin the remote locations where the Dao villagesare established. All Tay villages are increasingtheir area of paddy field by opening new plots. OneDao village is engaged in the same trend, whilethe other is not. Paddy intensification through ashift from one to two crops a year is practiced inthe two Dao villages and in the higher-incomeTay villages, but not in the other Tay villages.

Poverty and environment linkages in theupstream Cau River basin appear to have a “his-

torical” stratum related to differences in statusbetween villages in each commune and a morerecent stratum related to social differentiationwithin communities. In the first stratum, poorervillages have smaller paddy field areas and con-tinue to farm sloped land, although they are grad-ually reducing this practice through intensificationof the paddy fields. They may access larger areasof forest land but are less likely to derive incomefrom it. Poorer villages have less access to collec-tive water networks and are at risk of significantimpact from point pollution. In the second stra-tum, the poor are more exposed to dug well pol-lution and less opportunities to derive incomefrom their forest land.

These strata are not visible in statistics alone.The income of the surveyed Dao households is74 percent of the survey average, less than theincome gap between households. The richestcommune is also the commune where the incomeof the poorer 20 percent of households is thelowest. Finally, differences between the poor andthe non-poor in land use are not visible whencomparing average variables for income groupsacross the district.

Poverty and environment in thedownstream cau river basin

Bac Ninh Province, although downstream of ThaiNguyen’s heavy industries, has important pollu-tion sources in the province itself. The develop-ment rate of household-based craft industries inBac Ninh Province is staggering. The number ofhouseholds running craft industries in Yen Phong,the main district surveyed in Bac Ninh Province,grew to more than 5,600 in 2004. In this district,20 percent of households run a craft industry.High pollution impact is generated by craft indus-tries, agricultural ventures and large enterprisesand is concentrated in a small number of villageswithin the communes. Paper recycling is one ofthe heavily polluting craft industries. Phong Khevillage, a well-known craft village part of the sur-vey, turned to paper recycling in the 1990s after



57

94

87

71

103

52

0 20 40 60 80 100 120

% environmentalforests

% forestry land

% Dao ethnicgroup

% poorpopulation

Averageelevation

N. districts

Bac Kan Province

Bac Kan river basin

F I G U R E 4 . 2 2 Forestry and Poverty Indicators in theUpstream Province

firecracker production, its former production,was banned. The second surveyed commune hasa pig-raising and alcohol-brewing village. A localcoal power plant is the main pollution source forthe third survey commune, which has retainedagriculture as its main activity. State-owned indus-trial industries in Yen Phong district increasedfrom 5 to 24 over the same 10-year period, andprivate enterprises grew from 23 to 154.

Social differentiation is growing rapidly and issubstantially higher than in Bac Kan. The incomesof the richest 20 percent of households in thepaper recycling village are 54 times the incomesof the poorest 20 percent in the third commune(the commune that has not diversified out ofagriculture). The poor population in craft villagesis mostly comprised of workers coming from out-side the village to work in local household-basedindustries. There are no less than 4,000 workersin Phong Khe village alone. Occupational healthand lack of healthcare system was mentioned asan issue by the respondents.

Unclean drinking water, a contaminated gen-eral water environment, coal gas, and dust arethe environmental issues mentioned as affectingdaily lives. Commune health workers witnesshigher incidence of several diseases in the craftvillages. The poor reported 9.8 sick days in the

last 12 months versus 6.4 days for the non-poor.Both of these figures are higher than the nationalaverage of 5.8 days. The entrepreneurs see pol-lution as an unavoidable cost of poverty reduction.This view is supported by rapid reduction ofpoverty incidence decreases as pollution increases.The proportion of households with incomes below$1 per day is 57 percent in the agricultural com-mune, 33 percent in the alcohol brewing com-mune, and 13 percent in the paper recyclingcommune.

The poor are at least as aware as the non-poorabout environmental risks. They use water sourcesthat are less safe but state concerns about unsafewater more often (Figure 4.23). They assess dugwells as an unsafe water source but are twice morelikely to rely on them than non-poor households.The poor even raise concerns for water qualityfrom drilled wells.

District and commune leaders underline thetechnical and financial limitations of the two pol-lution remediation policies are in place, (1) hotspottreatment for some of the large industrial estab-lishments; and (2) the development of industrialzones to move craft industries out of residentialareas. The coal power plant has been coveredunder the pollution remediation program for largeindustrial sites but local residents believe that the



100

90

100

85,7

68,8

90,6

77,8

87,1

60

65

70

75

80

85

90

95

100

River andponds

Dugwell Rainwater Drilled well

Poor households

Non-poor households

0

10

20

30

40

50

60

70

80

Drilled well Dugwell Rivers andponds

All households

Poor households

F I G U R E 4 . 2 3 Water Sources in Three Downstream Communes. Left: Percentage ofHouseholds Using Water Sources. Right: Percentage of Households Assessing Water Source as Unsafe

impact from the exhaust has remained signifi-cant. An industrial zone was created in the paperrecycling village, but only 60 industrial estab-lishments, or around one-third of the total num-ber, have had the means to assemble the requiredinvestment to settle in that zone. The leaders statedfeeling as though they were left alone to addresslocal pollution issues with limited resources.They underline that the river basin master planassembled evidence of pollution charges manytimes above national standards as early as 1998,and that yet limited changes have taken placesince then.

Linkages between the upstream and downstream river basin

Bac Kan and Bac Ninh are two provinces onlya short distance away from one another thatreflect the extreme upland/lowland dimensionof the poverty-environment nexus in Vietnam.Bac Kan Province was created when the formerThai Nguyen Province was divided into two byzoning out into the section of the province thatspecialized in forestry. Bac Kan has the third low-est population density in the country (57 people/km2.) Bac Kan, with 62 percent of its land areaunder forest cover, is the third most forestedprovince in the country. The province rankedseventh in terms of poverty incidence in 2004, justas it did in 1999. It only ranked 55th in terms of

absolute numbers of poor people in 1999, despitethe fact that 60 percent of its residents live belowthe poverty line. This figure remained high in2004, estimated at about 50 percent.

Bac Ninh had 445 poor people per km2 in1999, the highest density in the country in rela-tion with very high rural population densities,1,064 people/km2 in 2000 or more than 20 timesthe population density in Bac Kan. With theexception of Hanoi municipality, Bac Ninh isthe province with the smallest territory and thehighest population density in Vietnam. It is alsothe third least forested province in Vietnam, withonly 1 percent of land zoned as forest land. GDPper capita in Bac Ninh is 85 percent higher thanGDP in Bac Kan, though the economy is nowgrowing at the same rate of 2.3 to 2.4 percent peryear in the two provinces. Poverty incidence inBac Ninh decreased from 38 percent in 1999 to3 percent in 2004.

All household and community-level indicatorsof poverty are markedly higher in the upstreamdistricts reflecting both lower incomes and loweraccess to services (Table 4.8). There are howevertwo exceptions. Paddy field per capita is howevernot lower in Bac Kan than in Bac Ninh. Area percapita is very small in Bac Ninh in relation to highpopulation density, and progress in openingnew paddy fields to increase the sustainability offarming has been made in Bac Kan. Children in



T A B L E 4 . 8 Poverty and Environment Indicators in the Upstream and Downstream Districts

Cho Moi/ Yen Phong/ % Bac Ninh/District/Province Bac Kan Bac Ninh Bac Kan

Monthly income 1000 VND (VLSS) 191 326 171% Kinh population 21 100 476Agricultural land per capita 1347 545 40% crop production in income 42 8 18% agriculture and forestry in income 62 22 35% households with temporary housing 26 1 5% households without safe water 85 20 23% households without sanitation 79 54 68% communes with doctor 44 100 229

Bac Kan, although overwhelmingly from ethnicminority groups, have equal access to education.In the future, job opportunities outside subsis-tence farming and information in general mightbecome equally accessible across the river basin.

One of the most common environmentalawareness messages in Vietnam says that increas-ing forest cover rates in upstream watershedswill preserve or improve water balance in thedownstream watersheds and prevent soil erosion.Upland farmers in some of the poorest, heavilyforested districts are receiving this message repeat-edly (Tessier 2002). This message is very far fromreality in the Cau River basin. Forest resourcedegradation impacts the local economy in theupstream section before generating potentialimpact in the downstream section. Most of theforest land is now under natural regeneration,but this process will take a long time.

Assessing the relative responsibilities of sub-sistence farming and of commercial forest exploita-tion in forest degradation in the past is much lessof a priority than building alternatives for thecurrent period. Alternatives are in place for farm-ers who are starting to develop tea and otherperennial crops but are more difficult to identifyfor the local government itself. The developmentof the mining sector, an alternative with a dis-proportionate environmental impact, is not anappropriate solution. Forestry and related reha-bilitation needs do not only relate to the upperriver basin. Just as in the traditional song aboutthe Cau River, trees were an important elementof landscapes all along the Cau River.

The downstream river basin suffers from anuneven water balance, but there is no evidence ofthe impact of subsistence activities in the upstreamriver basin. Much of the pollution increase in thedownstream section is reported as being generatedlocally by a combination of heavy industries,unregulated craft village industries, agrochemicals,and organic pollution. Heavy industries in ThaiNguyen not only generate pollutant emissions,but also have high water uptakes that reduce wateravailability for the downstream section of the river

basin. Flood episodes followed by low water flowsdo put downstream provinces at risk and worsenpollution impact but water shortages are also animportant issue in upstream communes. Localstakeholders mention a possible combinationof factors that include disruption from miningoperations at the river basin head and past degra-dation of forest resources, perhaps in combina-tion with climatic trends. This is consistentwith the fact that soil erosion in Bac Kan maybe a more serious issue in sections of theprovince that belong to river basins other thanthe Cau River basin.

Policy implications

The masterplan mostly proposes investments forforestry development in the upstream section aremostly for forestry development, and concernfor environmental issues is limited to the down-stream section of the river basin. By doing so, themasterplan omits the issue of access to safe waterand control of spot pollution in the upstreamsection. It does recognizes the need for plantingtrees and improving landscapes in the downstreamsection. The continuing focus on forestry in theuplands is more dictated by the absence of bud-gets available from national programs.

The environmental improvements identifiedin the initial master plan significantly exceed thefinancial resources available. The master plan bud-get is $1 per person per year, if the six provincesare fully included. The master plan underlined theimportance of more cost-efficient “soft invest-ment” to overcome these important budget con-straints. Improving coordination between agenciesand exchange of information was identified as animportant type of “soft investment” early on. Thefocus of the Agenda 21 program on environmen-tal awareness is a second type of “soft investment.”Messages based on a comparison of representativecommunes—one in the upstream section, one inthe downstream section—were initially planned.The awareness messages would have describedinitial difficulties and progress made through apilot program in these two communes.



The Cau River basin offers an ideal setting forinnovative environmental awareness messages.Messages that would tell the story of a need toincrease forest cover rates to restore water balancein the downstream section can be replaced withup-to-date, realistic messages. The strong con-trast between less-favored forested uplands pop-ulated by minority ethnic groups and heavilypolluted lowlands that also have an importantpoverty issue is an opportunity to develop mes-sages based on two shared issues, the need torestore water quality and landscape degradation.Messages based on shared issues have higher poten-tial to generate attitude change.

In a constrained budget environment hardinvestments are likely to remain focused onselected locations within the physical river basinitself. The provincial level is however an appro-priate level for cross-sectoral coordination andenvironmental awareness and is recognized inthe master plan. Some of the solutions to envi-ronmental degradation within the physical riverbasin are located in province sections that are outof this river basin. Tourism development is anoption worth exploring as an alternative to min-ing in Bac Kan even though the Bac Be nationalpark is outside of the physical river basin. Well-functioning protected areas may be a powerfulawareness-raising factor. They convey the mes-sage that industrial pollution is not a necessaryevil in economic development.

Interventions prioritized as “win-win” povertyand environment solutions are also cost-effectivesolutions that can help alleviate budget constraints.Safe water access ranks first among these cost-efficient solutions. Affordable options to giveaccess to safe water to the poor in the downstreamsection are likely to be less investment-intensiveoptions than full coverage of the population withtap water systems. Access of specific villages tosafe in the uplands can be more cost-efficient thanlarger-scale tree plantations. When safe wateraccess is secured. Households will continue toshift from cropping sloped land to animal pro-duction and perennial crops as tea.

Community capacity building is another cost-efficient intervention with the potential to be a“win-win” solution in the long term for povertyand the environment in the Cau River basin.Capacity building needs are high among localcadres. With improved capacity access of poorercommunities to development opportunities orto environmental improvements can improve.Building community capacity through variousmeans, from training local cadres to formal en-hancement in the regulatory framework, is anindirect environmental intervention, but a directpoverty reduction intervention that can signifi-cantly improve the environment.

4.3. POLICY OPTIONS: INTEGRATINGPOVERTY AND ENVIRONMENTIN NATIONAL POLICIES AND PROGRAMS

1. Geographical Targeting of Policiesand Programs

Prioritizing the upland sector or the lowlandurban and industrial sector is not relevant. Thereis a dual need for interventions in the marginalrural areas and in the urban or peri-urban areasthat are in a rapid industrialization process. Thereis, however, a striking contrast between existingpoverty reduction programs in upland rural areasthat are mostly conducive to environmentalimprovements, despite room for improvement,and environmental programs in urban and indus-trial areas that have so far largely missed a pro-poordimension. The former deserve continuation inthe long term, while the latter require innovativeand pro-active approaches. The danger would beto reduce “well-known” interventions in uplandrural areas to concentrate on innovative inter-ventions in the lowlands. Some lowland rural areasalso have important poverty and environmentissues that are unrelated to industrialization. Theseinclude the Southernmost areas of the Mekongfor access to safe water and coastal areas with poorpopulation subject to typhoons.



Upland rural areas with persistently highpoverty rates and pressing natural resource man-agement issues will however deserve specificintervention programs in the long term. The gov-ernment already has a long-standing policy ofspecific support to these upland, ethnic minorityareas and a specific agency, CEMMA. The initialgoal of this policy was centered on security con-cerns in border areas, and economic developmentwas mostly promoted through in-migration. Thisgoal shifted to poverty reduction in the early1990s (Communist Party of Vietnam 1989) andseveral targeted programs are implemented, todaywith support from several international donors.The 135 Program under the Ministry of Planningand Investment, initiated in 1999, is probably thebest-known program for upland areas along withother programs under MOLISA. This programtargets poor and remote communes primarily forsmall infrastructure development. The numberof beneficiary communes totaled 1,875 in 1999,and increased to 2,335 in 2003.

The Northern mountains region stands outfor its very high poverty incidence and povertydepth. Poverty and environment linkages areequally important in the central regions. Regionaldifferences can best be taken into account throughprograms with a regional dimension. And thespecific need for an actively pro-poor continuationof forestry reform in the central regions needs tobe fully recognized by Government and donorsalike.

The Red River Delta and adjoining provinceshas an outstanding issue of rapidly growing pol-lution and degraded river basins such as the Cauriver. These deserve geographically targeted inter-ventions. Interventions covering both uplandsand lowlands and oriented towards soft invest-ments such as environmental awareness can havehigh efficiency compared to pollution remediationalone.

Targeting poor villages within communes isboth a need and a challenge. Poverty and environ-ment issues are combined in some of the villagesin some cases, separate in others. Most impor-

tantly capacity to prioritize village needs andinterventions needs to be enhanced among gov-ernment and at commune level and specific toolsare needed.

Finally, some poverty and environment issueswould be best addressed at a nationwide level.Environmental health risks in relation to pesti-cides, for example, are prevalent throughout thecountry.

2. Priorities and Importance ofMulti-Sectoral Approaches

The policy implications of each detailed PENresearch area together create a framework ofpriorities (Table 4.47). Other priorities thathave been covered under the PEN research includepro-poor disaster preparedness programs, envi-ronmental health issues such as malaria andindoor air pollution, and issues specific to pro-tected areas.

Potential “win-win” impact on poverty reduc-tion and the environment lies more in incorpo-rating better attention to the environment inexisting poverty reduction programs and takinga “pro-poor” approach in existing sector policiesand environmental programs than in innovativeinterventions.

Most priorities identified are multi-sectoralwithin broader rural or urban and industrial devel-opment areas. Interventions needed for balancedand equitable industrial development for exam-ple are multi-sectoral by nature. A poverty andenvironment approach needs to take into accountnot only pollution control and prevention throughrelocation of industries to industrial zones forexample, but also how each solution may affectthe poor differently from the non-poor. Urbandevelopment, including transportation, deter-mines the locations where the poor can afford towork and live. Health services in urban areas canbe pro-poor by focusing on children and theelderly and on occupational health in high-riskoccupations. Social protection measures accessi-ble to the poor are also a necessity. Although the



government is already implementing some multi-sectoral interventions, the agencies that have takenpart in the PEN research or contributed data—MoNRE, MPI, MARD, MOH, CEMMA—canimprove the poverty and environment impact ofthese interventions through better coordinationamong themselves as well as among other stake-holders that include enterprises, local governmentsor the media.

3. Opportunities for Poverty andEnvironment Initiatives

There is an opportunity to give high visibility tobroad poverty and environment issues by makingthem stand-alone initiatives. Four such initiativeshave been identified (Box 4.12).

The specific difficulties of marginal uplandareas that are home to ethnic minority peopledeserve to be addressed by a separate povertyand environment initiative for the Cordillera.Such an initiative could facilitate recognitionthat, despite existing policies, inequitable accessto resources or lack of capacity to compete withother economic operators creates powerful limi-tations in poverty reduction, sound environmen-tal management, or both. Building communitycapacity in areas such as land administrationwould also have potential in such an initiative.An initiative in the region could also focus on bringing pro-poor information that, un-like general mass media information, would be tailored to the practical needs of the localpopulation.



TABLE 4.9 Summary of Program Priorities in Relation to Poverty and Environment Linkages

Poverty &environment linkage area Program/policy priority

Environmentalhealth in rural areas

Environmental health in urbanand industrial areas

Natural resource use in upland areas

Natural resource use and environmental health

Clean water & sanitation: improved services for rural households:a—with no improved water supply or sanitationb—using polluted dug wells.Information/education programs.Community leader information and training.PesticidesStrategic information campaign for users and retailersIndustry regulatory framework and institutional capacity buildingExample: pro-poor national regulatory framework for craft industriesUrban & peri-urban developmentPriority to access to safe water and sanitationAffordable improvements in working & living conditionsCapacity building for land law implementationContinued poor commune program with improvements in:Community capacity buildingInformationLand administrationDevelopment of modern land use plans, with simplified specifications in mar-

ginal areasCommunity leader information and training on legal frameworkLinkage to pro-poor forestry reform in the transition from State to enterprise sectorArea-based integrated programEnvironmental awareness program combined with selected priority investments

Access to clean water and sanitation is such aclear and broad poverty and environment prioritythat it deserves a specific nationwide initiative.The contrasted issues of the marginal upland areasin the Cordillera and of the rural and peri-urbanareas that are increasingly affected by diffusepollution would justify a double initiative. Aspecific clean water access program—insteadof the existing general infrastructure-building135 program—would serve to target those cur-rently without access and would help balanceinvestments between clean water and roads orother much-needed but expensive infrastructure.In the lowlands, affordable, broad-based solutionsto address diffuse pollution need to be introduced.Again, access to appropriate information for thepoor would be an important element of suchinitiatives.

An initiative in the urban and industrial sec-tor would facilitate innovation, a critical elementfor success in these interventions. Innovativeapproaches are needed for issues that are partlyrural, partly peri-urban, and partly urban. Inno-vative technical management systems need to beset to address issues such as diffuse pollution,as well as in terms of institutions to deal with,for example, pollution issues concentrated in vil-lages within communes. Disclosure of informa-tion on pollution levels, wherever possible basedon improved pollution monitoring, can be a use-ful measure to help the poor and the non-poor getrelevant information. Public awareness campaignscan also help promote risk-averting behaviorsamong low-income groups. The proposed Agenda21 program in a degraded river basin is one of

these innovative interventions. A regulatory frame-work for craft industries and other major gaps inthe legal or regulatory framework require urgentaction. Improved capacity for multisectoral urbandevelopment policies is a critical skill. Improvedinstitutional capacity can bring in improved imple-mentation and enforcement of the regulatoryframework, enhanced coordination between stake-holders, and increased attention to innovation.

The Ministry of Planning and Investment(MPI) has expressed interest in Bank assistanceto urgently address industrial pollution issues in“hot spots” particularly in the Cau River Basinas a follow-up to this study (note that this proposalis being shaped at present in line with findings pre-sented in map 4.8). The planned pilot projectwould provide technical assistance to support (i) the government in strengthening monitoringand enforcement of compliance with environ-mental regulations, implementing environmentalinformation disclosure programs, and introduc-ing regulations for craft industries; (ii) industrythrough the preparation of environmental man-agement audits, the introduction of self-monitor-ing and environmental management systems suchas ISO 14,000, adoption of clean technology, andtraining programs for workers’ safety in highhealth risk industries; (iii) the banking sectorthrough the preparation and adoption of envi-ronmental manuals. On the investment side, theproject would pilot pollution management invest-ments including (i) the establishment of central-ized craft industry zones and centralized wastewater treatment systems for highly polluting craftindustries in areas with high pollution risk of



B O X 4 . 1 2 Potential Poverty and Environment Initiatives and Cross-cutting Limiting Factors

There is potential for four poverty and environment initiatives:– A nationwide clean water and sanitation initiative– A poverty and environment initiative for ethnic minorities in the Cordillera (in relation to the

existing partnership for poorest communes)– An area-based comprehensive initiative in degraded territories (already planned for the Cau

River basin) to support innovations in pro-poor environmental health and pollution remediation.

potable water supplies and where the density andsize of craft industries make this economically jus-tified; and (ii) investments, in a cost sharing basis,in cleaner technologies and pollution control inmedium and large scale enterprises in highgrowth-high pollution intensive sectors in enter-prises, which make commitments to introduceenvironmental management systems and to com-ply with environmental regulations. MPI is con-sidering the inclusion of an environmental healthcomponent into the proposal, reflecting healthimpacts polluting industries may have on certainpopulations, for example through the use ofunsafe drinking water sources (see figure 4.23).

Endnotes1. The latest World Bank poverty update for Vietnam

shows that in 2004, 39 percent of the poor were ethnicminorities, a figure that has risen from 20 percent in1993.

2. Since it was too early to examine land titles, theanalysis was done on allocation of land use rights onforest land.

3. Foreign-invested industrial output has no significantassociation with provincial poverty or living standards.This might be influenced by the concentration of for-eign-invested industry in a few provinces, and is notnecessarily inconsistent with the foreign-invested indus-try having important macroeconomic benefits, and thusindirectly contributing to higher living standards.

4. Pollution intensity defined as pollution load per unit ofproduction output.

5. World Health Report 2002. World Health Organi-zation.

6. Estimated from the VNHS 2002.7. The health survey covers all rural households, while this

survey is focused on rice growers.8. The study was focused on rice growers and therefore did

not cover exposure of landless people to pesticide use.9. The blood test used a cholinesterase enzyme technique

measuring active cholinesterase enzymes in erythrocyteand plasma using the Test-Mate-Kit (EQMR-USA).

10. IRRI’s SAM project has established a detailed databasefor Bac Kan Province.

11. Hanoi municipality only has two river basin districtsand is not a full-fledged member; Vin Phuc is a fullmember with six river basin districts, although theCau River des not flow through the province; HaiDuong is a member because it receives downstreamimpacts.

12. The actual number of districts is changing. Three dis-tricts have recently been merged or divided into two. Inall districts,AQ4 at least part of the physical river basinare included. Six of Bac Giang’s 10 districts, which arethe watershed of a lower tributary, are excluded.

13. Mostly under a joint venture between a state-ownedenterprise from Thai Nguyen province and two Asianenterprises.

14. Poorer households in other case studies were found tobe allocated more distant forest land with poorer qual-ity land (Sunderlin and Huynh 2005). The survey onlyrecorded distance to land with protection contracts, buta similar situation is expected for production forest.

ReferencesPEN Case study reports in CD Rom AttachmentDasgupta, Susmita, Craig Meisner, David Wheeler and

Jostein Nygard 2005. Poverty/Pesticide Use in theMekong Delta, Vietnam. Washington, DC: WorldBank.

Nguyen, Huu Dung. 2005. Poverty and Pesticide Use inRice Production in the Mekong Delta, Vietnam. HoChi Minh City: Environment Economic Unit, Univer-sity of Economics.

Socio-Economic Development Center. 2005. “Environ-ment and Poverty Linkages in the Cau River Basin—Vietnam Poverty and Environment Nexus Study.”Hanoi: background paper to the PEN study.

Tai, Nguyen Van, Nguyen Trung Thang, Nguyen ThiThuy Duong, Nguyen Duc Minh, and Hoang VanHoa. 2006. “Poverty—Environment Nexus Program,Case study: Poverty and Industrial Pollution in Viet-nam.” Hanoi: background paper to the PEN study.

TECOS (Consultancy Service and Technology DevelopmentCompany for Natural Resources and Environment).2005. Vietnam Poverty and Environment Nexus StudyPhase II. Land Resource—Poverty Study in Vietnam.Hanoi: background paper to the PEN study.

Other referencesCastella, Jean-Christophe and Dang Dinh Quang. 2002.

“Doi Moi in the Mountains: Land Use Changes andFarmers’ Livelihood Strategies in Bac Kan Province,Vietnam.” Hanoi: The Agricultural Publishing House.

Communist Party of Vietnam. 1989. Politburo Decision onGuidelines and Main Policies for Economic and SocialDevelopment in the Mountain Area. Document of theCommunist Party No 22/TQ NW, November 27, 1989.

Curtis, V. and S. Cairncross. 2003. “Effect of WashingHands with Soap on Diarrhea Risk in the Community:A Systematic Review.” Lancet Infectious Diseases 3.

Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. 1991.“Effects of Improved Water Supply and Sanitation on Ascariasis, Diarrhoea, Dracunculiasis, Hookworm



infection, Schistosomiasis, and Trachoma.” WHO Bul-letin 69 (5):AQ2.

Fewtrell, L. and J. Colford Jr. 2004. “Water, Sanitation andHygiene: Interventions and Diarrhea—A systematicreview and meta-analysis.” HNPAQ3 Discussion Paper.Washington, DC: World Bank.

IFPRI, IDS, and Inter-Ministerial Poverty Mapping TaskForce. 2003. “Poverty and inequality in Vietnam: Spa-tial patterns and geographic determinants.” Washing-ton: IFPRI and Brighton: IDS.

Khong, Dien. 2002. “Population and Ethno-Demogra-phy in Vietnam.” Chiang Mai, Thailand: SilkwormBooks.

Korea Environment Institute (KEI), 2005. “Vietnam–Regional Environmental Management for TraditionalVillages,” Final Report, Seoul, Korea.

Sunderlin, William and Huynh Thu Ba. 2005. “PovertyAlleviation and Forests in Vietnam.” Jakarta: CIFOR.

Swinkels, Rob 2004. “Draft Synthesis: Poverty and Envi-ronment.” Hanoi: World Bank.

Tessier, Olivier. 2002. “World Bank Forestry Sector devel-opment Project. Initial Social Screening.” Montpellier,France: Tercia consultants.

Vietnam Ministry of Planning and Investment. 2004. Socio-economic Development of the Poorest Communes:Community Perspectives and Future Prospects. Partner-ship to Assist the Poorest Communes national confer-ence proceedings. November 26, 2004. Hanoi: MPI.

Vietnam Poverty Task Force. 2002. “Ensuring Environ-mental Sustainability. Localization MDGs for PovertyReduction in Vietnam.” Hanoi: Government of Viet-nam, Poverty Task Force.

Vietnam Poverty Update. 2006. Power Point Presentation.Washington, DC: World Bank.

World Bank. 2003. Vietnam Environment Monitor 2003.Washington, DC: World Bank.



SUB-REGIONALPERSPECTIVES

III


INTRODUCTION

The Lower Mekong River flows through Laos,Cambodia, and Vietnam in its lower stretches.Do the three countries form a “subregion” interms of poverty and environment issues?

The Lower Mekong River itself is not a com-mon thread when environment is analyzed fromthe point of view of its linkages to poverty. InCambodia, seasonal floods heavily affect largenumbers of poor. At the same time, seasonalflooding is important for many rural farmers. InLaos, the Mekong floods affect the non-poorgroups that form the vast majority of Mekongplains inhabitants. In Vietnam, floods along thenorthern central coast are related to local topo-graphy and typhoons and not to the MekongRiver, and only the southernmost section of Viet-nam, the Mekong River Delta, belongs to thisriver basin.

This chapter highlights which areas of povertylinkage have high magnitude in only one of thethree countries or are shared among two or threecountries. The practical solutions to these issuesare often country-specific, but broader policyoptions are largely similar. This creates opportu-nities to jointly address poverty and environmentabove the national level.

Subregional environmental initiatives arealready in place, but there is a largely untapped

potential to address poverty jointly with environ-ment in initiatives at that level. The three coun-tries are already part of cooperation bodies, mostnotably ASEAN and the Mekong River Com-mission.1 ASEAN is a driving force behind thecurrent opening of borders and regional roads.2

According to MRC, the international body estab-lished to promote and ensure the sustainablemanagement of water and related resources, thecondition of the common resources “ha[s] impli-cations for how successful people are in raisingtheir standard of living and bringing about relatedsocial improvements” (MRC 2003).

Potential joint initiatives will be a focus topicfor discussion among national stakeholders at thesubregional PEN conference. This draft chapterwill be revised after the conference to fully incorpo-rate the results of this upcoming event.

Shared Features in Poverty andEnvironment Linkages

The PEN research has analyzed the magnitudeof several poverty and environment linkages inVietnam, Laos, and Cambodia. When taking ahorizontal view of poverty and environment,linkages can be ranked as having high, medium,or low magnitude or severity based on informa-tion on spatial distribution of the issue and ofpoverty (Figure 5.1).

Subregional Findings5

�

Poverty and environment linkages of high relevance in more than one country include:3

– Environmental health issues with water supply and sanitation and urban pollution,– Natural resource use issues related to forestry issues (including NTFPs, other resources and the

impact of roads), UXO contamination, and natural disasters.

Environmental Health and PovertyThe analysis of environmental health and pov-erty linkages in each of the three countries hasrevealed three important shared features (Fig-ure 5.1):

• The joint existence of a poverty and environ-ment nexus for clean water and sanitation andof important difficulties in taking it fully intoaccount (in all three countries),

• A combined issue of lack of access of the poorto services and of widespread diffuse waterpollution (in all three countries),

• Low levels of community capacity and infor-mation on environmental health (an issue of special relevance in Vietnam and Laos),and

• Special challenges in urban areas (Vietnamand Cambodia at the present stage).

Access to safe water is spelled out in the nationalpolicy frameworks and is one of the MillenniumDevelopment Goals. Yet the gap between thepoor and the non-poor in accessing clean waterand sanitation, and their lagging behind in theprogress being made, remains a major environ-

S U B R E G I O N A L F I N D I N G S


F I G U R E 5 . 1 Magnitude of Poverty and Environment Linkages in the Three CountriesVN = Vietnam, LA = Laos, KH = Cambodia

mental health issue. Differences between coun-tries are dwarfed by the shared importance of theissue itself (Figure 5.1).

For the subregion as a whole, there is a risk ofdiffuse water pollution of diverse origins. Thishas already materialized in Vietnam. Water pol-lution in upland areas originates not only fromlocal organic sources, but also from pesticides.Water pollution in lowland areas can be a com-bination of organic, agricultural, and industrialpollution. The wide distribution of paddyfields and water bodies in all regions, not onlyin the lowlands, probably plays a significantrole in this problem. The poor and non-poorshare the issue of avoiding pollution when theylive in areas with polluted waters. The poor,even in marginal rural areas, can also be con-fronted with spot pollution. With high andincreasing inequity levels, the poor visibly findit difficult to access the improved services beingcreated.

The PEN research provides evidence thatethnic minority people have less access to cleanwater and sanitation and demonstrates that com-plex demand problems are probably at play. Thepoor are at least as aware as the non-poor ofpollution risks but, as pointed out in the casestudies, communities need to have the capacityand local institutions in place to access servicesand maintain them. Appropriate informationfor individuals is also often missing. The lackof appropriate technology is also a problem,especially in Cambodia’s drought-affected pla-teaus and in the case of Vietnam’s polluted dugwells.

Inequitable access to safe working and liv-ing environments is high in the growing urbanareas in the subregion, as well as in Vietnam’sindustrial areas. Communities do not have for-mal institutional capacity to manage local issues.Information on workers’ rights and residents’rights is weak. Improved management systems,such as urban land use planning, are just start-ing to be put in place in Vietnam through theland law.

Natural Resource Use and Poverty

Regarding linkages between natural resource useand poverty, the analysis derived two importantshared features in the three countries:

• Access of the poor to their resource base appearsto often be a prominent issue compared toappropriate management of resources by indi-vidual households, and

• Communities’ capacity in marginal areas isweak; this stands out as a limiting factor forthe implementation of potential poverty andenvironment policies.

Access of the poor to their local natural resourcebase is restricted in some cases and totally openin others. Issues of restricted access, generally toforestry resources, are insufficiently recognizedin the subregion—in sharp contrast with othercountries in Southeast Asia such as Indonesia(Homes 2002). Stakeholders that compete forresource access include local and migrant house-holds in migration regions, both of which arepossibly equally poor. They also include largeroperators, from small local entrepreneurs tolarge international corporations. The weight ofthe forestry sector in the three countries is a keyelement in this context. Commercial timber is amajor national commodity in Laos and Cam-bodian, and remains an important one in thesouthern two-thirds of Vietnam. This competi-tion is openly discussed in the media in Cambo-dia, but not in Laos and Vietnam. Difficultieswith open paddy fields in UXO-affected areasin Laos are another type of access restriction.Examples of open access include most NTFPs.In Cambodia, most natural resources are underopen access, unless a concession has been set up.When the poor do not have sufficient access toan environmental resource that is critical to theirlivelihood, they generally turn to another, less-restricted resource that can generate negativeenvironmental impact while providing limitedlivelihood improvements.

In situations of open access, the local tradi-tional community management modes may have



allowed sustainable resource management in thepast, but they are often disrupted. Populationgrowth, migration, and the arrival of marketdemand are elements that combine to create dis-ruption. In situations of restricted access, weakercommunities may not be in a position to access theinformation they need to improve the situation.Ethnic minority communities may have retainedstrong management systems in some instances,but they are also more likely to have encountereddisruption. Communities’ capacity is also a criti-cal element in the important and widely discussedissue of balance between the need to build a ruralroad network and forest conservation. The casestudies in Laos have documented communities’desire and confidence in their ability to managetheir own resources in a sustainable manner aftera road is opened, but difficulties are likely to occurunless communities’ legal capacity in resourcemanagement is established and taken into account.

Disruption of community capacity is partly alegacy of war. War legacy is one of the reasonsbehind active rural-to-rural migration in Cam-bodia, and is still visible after 30 years in the mar-ginal areas of Laos and Cambodia. Land usesystems that may appear traditional have oftenalready been through several steps of adaptation.UXO-contaminated areas are among the poorestrural areas in Laos and in Cambodia, while AgentOrange residues, which are believed to continueto generate serious heath impacts, are clearly

located in the poor areas that link southern Laosto southern Vietnam (Lao UXO program 2003).

The management modes of natural resources,from the commercial logging sector to subsistenceagriculture, are in a transition stage in the threecountries. This rapid change further weakens poorcommunities. In Vietnam, state interventions incommercial logging are gradually replaced by cor-porate operators, where national and internationalcorporations are taking over from state forestenterprises. In Cambodia, the concession system isbeing replaced with logging contracts. Farmers’subsistence land use systems are changing. Themost noticeable change relates to traditional rota-tional agriculture (or shifting cultivation, Box 5.1).Rotational agriculture is gradually becoming moreintensive, a change that mostly generates benign oreven positive environmental impacts. Paths ofchange are identical, but stages of progress are con-trasted, Vietnam being the most advanced.

Shared Geographical Features that Shape Poverty and Environment Linkages

Linkages between environmental health or nat-ural resource use and poverty in the three coun-tries occur in a geographical context that displaysthree further and important shared features:

• A dichotomy between lowlands, where urbancenters and industrial growth are located,



B O X 5 . 1 Rotational Agriculture, Poverty, and Environment

Poverty and environment linkages have been repeatedly described as a downward spiral betweenenvironmental degradation and rural poverty. The transition of rotational agriculture in Laos andVietnam is illustrative. As population pressure increases and the desire of more educated householdsfor less strenuous farming tasks rises, fallow periods become shorter until they eventually disappear.

The transition is technically difficult, especially in weed control. Declining yields may occur inthe process. Interaction with government policies or with other stakeholders competing for thesame land resources often makes things worse. When farmers are asked to rapidly reduce theirrotational agriculture area, either fallow cycles are reduced too quickly to allow for adjustment, or subsistence farmers turn to opening other plots of land “behind the hill.”

Soil erosion in upland areas remains an issue, but rotational cultivation is probably not the mainculprit. River basin topography—especially for river basins flowing into Vietnam from the Himalayanplateau—as well as commercial agriculture and non-agricultural activities such as mining all play arole in soil erosion.

and the marginal rural areas that are laggingbehind;

• Legal communities that are composed of sev-eral smaller village communities; and

• The importance of rural-to-rural migration.The upland / lowland structure is present in eachof the three countries. Vietnam and Laos both dis-play a strong contrast between their lowlands andtheir uplands, which by and large form the Anna-mite Cordillera (Figure 5.2). The poor on bothsides of the mountain range account for the largestshare of the rural population and live in a regionwhere national forest resources are located. Mostethnic minority groups in the Cordillera are dis-tributed across the two countries. Yet an impor-tant difference between Vietnam and Laos is thenational distribution of poverty. In Laos, whichhas low population density throughout the coun-try, the upland poor account for a substantial shareof poverty in the country. The highest absolutenumbers of poor in Vietnam reside in the denselypopulated lowlands. Cambodia also has higherpoverty in the plateau and mountain regions,where there is a high concentration of ethnicminorities—but the overall land area and popula-tion within upland areas is much smaller, so thatthe overwhelming majority of the poor reside inthe lowlands, many of them in villages withincommunes on marginal plateau soils. Numbers ofrural poor are also high in the floodable areas.Since 60 percent of Cambodia is under forest, nocorrelation is visible between rural poverty-trapareas and the location of forest resources.

Communes in Cambodia and Vietnam arecomposed of villages that tend to have unequalendowments and therefore contrasting povertylevels. (The commune is the legal institution,while the village is the human community.) Smallvillage communities in Laos are in the process ofbeing merged into larger villages to obtain legalstatus.

Migration is modifying these geographical pat-terns in an important manner. The increasingenvironmental health issues specific to growing

urban centers are not specific to the subregion butare nevertheless an important feature. Migrationinto rural areas, which are often as poor as theregions of origin, is creating pressure and distur-bance on natural resources. In spite of their differ-ent political systems, rural-to-rural migration is acombination of government-sponsored programsand spontaneous migration in each of the threecountries. Population migrates from resource-poor areas to what are perceived to be resource-rich areas.4 Spontaneous migration is especiallyhigh in Cambodia, a combination of the post-warphase, where many still have not found a stablecommunity to settle in, and of traditionally sea-sonal migration patterns. The government is in theprocess of creating “social concessions.” Some of



4Poverty 3incidence 2

1

4Number of 3rural poor 2

1

Importance of 4natural resources 3& poverty issues 2

1

Importance of 4environmental 3health 2& poverty issues 1

Up- Low- Up- Low- Up- Low-land land land land land land

Laos Vietnam Cambodia

F I G U R E 5 . 2 Upland and Lowland in Poverty & Environment Issues in the Three CountriesQualitative assessment of magnitude from 1 = low to 4 = very high

the land will be provided to local residents, andsome may be provided to migrants. In Laos, thevillage consolidation and focal zone policies,which are at the center of many debates among thedonor commodity and have been discussed in thecountry chapter, actively promote rural-to-ruralmigration. Migration from the lowlands to theuplands in Vietnam, a sensitive issue that is oftennot discussed openly, is also a mix of spontaneoustrends and government programs.

Many of the correlations observed betweennatural resource use and poverty in environmen-tally fragile areas retain an unclear link to causal-ity because of these two geographical factors.Does poverty play an important role in resourcedecline or degradation? Or are the poor affectedby resource degradation? Are they at risk of nat-ural disasters that can be largely unrelated to nat-ural resource use? All these remain equally validquestions.

Shared Policy Options

When poverty issues and environment issues areobserved to take place together and to have pos-sible causal links in Laos, Cambodia, or Vietnam,a combination of a small number of limiting fac-tors is generally observed. These factors—accessto resources, information and education, manage-ment systems and community capacity—createtogether a strong obstacle to sustainable devel-opment. While they may be observed in manycountries, what makes them so important in Laos,Vietnam, and Cambodia is the contrast betweentheir visibility and their low-key status in nationalenvironmental policies and donor-supported pro-grams. Each of the ten PEN studies has yieldedstrong evidence that two, three, or even four ofthese factors were at play in poverty and envi-ronment linkages, while at least one was not fullytaken into account in the policy framework:

Despite broad availability of informationthrough the media and improvements in educa-tion levels, the poor continue to face more diffi-culties in coping with and finding solutions to

the environmental problems they face. Althoughthe need for increased community capacity hasachieved high visibility in decentralization andpoverty reduction policies and programs, itsimportance for the environment is often less rec-ognized. There is also a need for better recogni-tion that the lack of appropriate and sustainablemanagement systems, be it in terms of technol-ogy or in terms of institutional arrangements,may continue to limit poverty reduction andenvironmental improvements.

Four important policy implications derivefrom the combination of these limiting factorsand are shared by the three countries:

• “Indirect” policies and programs focusing onimprovements in the four main PEN condi-tions are as needed as direct interventions onenvironmental issues;

• Poverty and environment interventions in thesubregion need to fully take account of chal-lenges in relation to local institutions, espe-cially at community level;

• Appropriate pacing of change is a key elementin policy making; and

• International donors have a role to play toensure that these are fully incorporated innational policies and programs.

Several donor agencies are in the process of sup-porting poverty reduction programs that fullyincorporate environment. While design proce-dures in international programs have generallyled to screen environmentally friendly approachesto poverty reduction and to avoid negative socialimpacts in environmental programs, they nowwant to go one step further. However, directinvestments in environmental improvementsare unlikely to succeed unless improvements inaccess, information, and education, as well ascapacity and management systems are facilitatedat the same time. Indirect interventions on theselimiting factors should also be considered. Per-haps one important value of donor interventionis to facilitate cross-sector interventions. Just as



limiting factors for poverty and environment arenot within a specific sector, cross-sector solutionsare generally needed to generate win-win povertyand environment solutions. Government inter-ventions are often sector interventions. Geo-graphically targeted polices in the uplands ofVietnam and Laos are designed to be cross-sectorinterventions, but they would gain from fullyincorporating issues of environmental health andsustainable natural resource use.

Community capacity is a legal and institutionalissue compounded with social and cultural fea-tures. The two-tiered system of formal and infor-mal communities makes community capacitybuilding an even more complex issue. So doesmigration. Migrating communities have to rebuildcapacity and need time to do so. Targeting poorervillages is an option worth considering in Cambo-

dia and Vietnam, and justifies targeting at thatlevel. However, prioritizing poorer, often environ-mentally fragile villages is no easy task. The com-munes need to build capacity to identify prioritiesand ensure they are taken into account. Migrationcan be so high that it may make geographical tar-geting less relevant. Disruption from migrationcan be so substantial that migration will eventuallyfail to achieve poverty reduction outcomes, or topreserve the environment, or both.

Rapid change is taking place in the subregionin terms of economic growth, access to roads, com-munication technologies and markets. Timingmight be the key to successful policies and inter-ventions. Progress in environmental health, espe-cially in clean water and sanitation, is slow, whilemuch more proactive interventions could takeplace. Conversely, national policy frameworks are



sustainable development

social dimensions

environmental dimensions

growth

poverty reduction

equity

appropriate information access

community capacity

managementsystems

environ-mentalhealth

naturalresource

use

4 conditions for effectivenessof PEN interventions

F I G U R E 5 . 3 The Four Main Conditions for Effectiveness of PEN Interventions in Cambodia, Lao PDR and Vietnam

often designed to hasten change in naturalresource use, while gradual, careful solutions areneeded to ensure success in terms of povertyreduction.

Policy options in environmental health and poverty

• Environmental health policies would be beststructured in the three countries as a dual setof programs, one for marginal areas and onefor urban areas. Marginal rural areas deservean intensified initiative to provide clean waterand sanitation access to the poor. Social dispar-ity in urban areas deserves attention in the threelarger urban areas in Vietnam and Cambodia.Other urban areas will gain from addressingthese problems early on.

• Pro-poor improvements in environmentalhealth in urban and industrial areas requirelegal framework improvements, together withinformation and awareness programs, real ini-tiation of land use regulations, and institutionalreforms giving more capacity to the affectedcommunities. These programs are best struc-tured when they focus both on target residentsand their living neighborhoods and workersand their work environment. A number ofpro-poor urban development programs are inplace and deserve expansion.

• Programs should gain from being joint initia-tives among the agencies building infrastruc-ture, the agencies in charge of sanitationdevelopment, of hygiene information and,more generally, of information (includingthe media), and—in Laos and Vietnam—theagency in charge of ethnic issues. Such jointefforts are required to trigger real change inthe poverty trap areas.

• Information and education would greatlybenefit from careful tailoring to the specifici-ties of ethnic groups. Innovative informationmessages have potential to trigger local inter-est in accessing improved services, to facilitateattitude change toward clean water and sani-tation, and to raise interest in community

management of water systems. Indoor air pol-lution deserves inclusion in the messages. Andcommunity capacity-building obviously needsto include training of local cadres.

Policy options in natural resource use and poverty

Ensuring more equitable access of the poor tothe local resource base or helping them break outof traditional low-return unsustainable activitiesare two policy options that have the potential tobe successful poverty and environment strategiesbut will require long-term efforts in the threecountries. Increasing access needs to go hand-in-hand with the design of new community manage-ment systems. Sustaining these systems whenstakeholder relationships and the market environ-ment are changing quickly will require enhancedcapacity at the community level. Promotingnon-traditional activities requires access to mar-kets and services in the marginal areas and isunlikely to be pro-poor, unless special attentionis paid to giving the poor access to these activi-ties. And non-traditional activities may be moreharmful than subsistence agriculture to the envi-ronment unless new management modes includ-ing appropriate technology are in place.

Geographically targeted programs that incor-porate attention to these issues are already in placein Vietnam and Laos, but there is a need for bet-ter recognition that these marginal areas deservelong-term support policies. It is this long-termdimension, provided it is recognized, that allowspolicies and programs to focus on indirect inter-ventions to manage the environment better andnot on direct environmental interventions.

In this context of a need for sustained supportto the marginal areas in the longer term, threeshorter-term policy options derive from the PENresearch:

• Improvements in access and community capac-ity can best be achieved if policies do not createadditional disruption by promoting rapidchange in local land use systems. National pol-icy frameworks, in sharp contrast with the



recent past, are starting to take this need forslow adjustment into account and they deservesupport in this process. In Lao PDR, the shift-ing cultivation policy is formulated today asstabilization. In Vietnam, allocation of forestland for rotational agriculture has been legal-ized and allocation of forest use rights to com-munities has started. Conversely, migrationpolicies are still viewed as a priority option inthe subregion and have the potential to gener-ate poverty, environmental damage, or both.

• Addressing growing and emerging issues com-pared phasing-out issues will increase the effec-tiveness of interventions. Rotational agricultureis a primary example of a phasing-out issue inthe subregion. Improved management of mar-keted NTFPs under strong market demand is a growing issue. So is the need for increasedtransparency about stakeholders use of naturalresources.

• Innovative awareness programs targeting poorcommunities in environmentally fragile areasare, just as for environmental health, an impor-tant “forgotten” option. Existing programs havelong been sending simple messages relating toforest conservation and to reduction of unsus-tainable slope cultivation. These messages areobsolete, and national governments can bene-fit from introducing modern environmentaleducation, taking into account a broader arrayof issues.

5.2. Potential Subregional Initiatives

Laos, Vietnam, and Cambodia: comparisonand exchange of experience on poverty and environment

Many of the poverty and environment issues areshared between two countries, while the thirdfaces a contrasting situation. Seven fields standout as opportunities for subregional exchanges(Table 5.1). Subregional exchanges focusing onlinkages between poverty and environment largelyremain an innovative approach, especially forrural areas and border areas. Country-to-country

exchange programs tend to relate either to eco-nomic development in general or to the environ-ment alone.

Poverty and environment issues mayremain difficult to address through jointsubregional projects or policies

There are two reasons for taking a cautious lookat subregional initiatives. First, there are remain-ing security concerns in the marginal areas thatare also often border areas. These concerns can-not be taken lightly. For example, the southernethnic groups in Laos are neighbors to the Cen-tral Highland groups in Vietnam. The CentralHighlands are regarded in Vietnam as a regionthat has been recently facing unrest.

Second, joint projects may not be feasiblewhen the legal and regulatory frameworks betweentwo countries differ markedly. Transboundaryprotected areas, for example, are an effective andnecessary approach in terms of environment, buttaking poverty reduction into account in theseinitiatives is far from obvious. Integrated protec-tion and development projects have proven to bean appropriate approach on the Vietnamese sideof the Laos-Vietnam border (for example WorldBank 1997, WWF 2005). In Laos, such projectsmight still provide some encouragement forthose who would hope to benefit from a devel-opment intervention to migrate into the pro-tected area. Similarly, in urban areas, Vietnam,Laos, or Cambodia are likely to have very differ-ent answers in terms of urban land use planning,and lessons learned are more likely to be a validfield for exchange of experience than for jointaction.

Subregional exchanges should primarilyinvolve local governments and experts andshould focus on the production of informationand capacity building tools and on monitor-ing. Ethnicity should be taken into accountwhen designing these tools.

Country-to-country exchanges are also alreadytaking place between local governments acrossborders, but they tend to address immediate



issues of illegal activities or to simply promoteeconomic relations. There is potential to expandthe scope of these exchanges to cover priorityissues among the seven subregional poverty andenvironment issues (Box 5.1). Researchers andtechnical assistants often specialize in one of thethree countries. Subregional exchanges wouldallow them to learn from comparative analysisand to facilitate exchanges about lessons learnedbetween the countries.

Researchers and technical assistants engagingin subregional exchanges would be in a positionto develop tools in the form of public informationpackages and media products to support nationalprograms in priority poverty and environment

problem areas. There will be much added value interms of environmental information and educa-tion if these tools take ethnicity into account. Inthe past, communication between ethnic groupsliving in different states has been—and oftenremains today—a sensitive issue. Today, withthe development of IT, appropriate informationon environmental health and natural resourceuse can be best packaged for ethnic minorities atthe subregion level, while information deliverycould remain at the national level and below.

The PEN research has documented howanalysis is limited by the lack of consistent data.Several key poverty and environment indica-tors, such as access to clean water and for forest



T A B L E 5 . 1 Opportunities for Subregional Exchanges

Subregional poverty and Source of relevantenvironment linkage experience Scope of relevance

Environmental Healthand Poverty

Natural Resource Useand Poverty

Cross-sector

Linkage 1. Water Accessfor Ethnic MinorityCommunities

Linkage 2. Diffuse Pol-lution in Wells

Linkage 3. Low-incomegroups in urban areas

Linkage 4. Natural dis-aster preparedness

Linkage 5. Livelihoodimprovement in UXOaffected communities

Linkage 6. Targetedinvestment programsin upland areas

Linkage 7. Environmen-tal awareness pro-grams in poor areas

Infrastructure and formal communitycapacity building inVietnam

Informal communitymanagement systemin Lao PDR

Diffuse Pollution inVietnam

Technical solutions in 3 countries

Existing programs (Asiaurbs and Citynet2001) and projects

Pro-poor flood responsein Cambodia

Pro-poor typhoonresponse in Vietnam

Pilot projects in 3 countries

135 program in Viet-nam and relateddonor-assisted projects

Agenda 21 in VietnamMiscellaneous proj-

ect experience in 3 countries

Primary: Vietnam andLao PDR

Secondary: Cambodia

3 countries

3 countries

Primary: Vietnam andCambodia

Secondary: Lao PDR

3 countries

Primary: LaosSecondary: Cambodia

Primary: Vietnam andLaos

Secondary: Cambodia



Transportation and trade GMS main transport routes Pilot open border points (2005) under

GMS transport Agreement Other border crossingpoints (to be completed)

Examples of flow of

timber, NTFPs, maize

Major urban center (to add) Protected areas

Transboundary

(existing)

Transboundary (planned)

National PAs

nearby main routes

Sources : ADB 2005 ICEM 2004 PEN field observations

Yunnan province China

Myanmar

Thailand

HCMC

Danang

Savanna-khet

Vientiane

Phnom Penh

Bangkok

M A P 5 . 1 Poverty



cover, happen to be MDGs. With a joint effortto use a shared methodology when assemblingthese indicators, the governments and thedonor agencies that support their povertyreduction efforts would have a solid basis formaking investment and policy decisions thatfully incorporate both poverty and environ-

ment. The subregion appears to be an idealspace for the collection and use of a consistentset of poverty and environment indicators. Thedistrict level (except in Cambodia), the com-mune level, and the village level (for villagesamples) have been identified as important tar-geting levels for PEN analysis. A monitoring

B O X 5 . 2 Border Features of Poverty and Environment Linkages

Border areas have a special role to play in subregional initiatives for poverty and environment. Thevast majority of border areas between Vietnam, Laos, and Cambodia are marginal, fragile areas withhigh poverty incidence. Forty-six priority poor districts in Laos—and many of Vietnam’s districts witha high share of communes listed as remote and poor—are upland border districts (Map 5.1).

Border regions had long specialized—perhaps as many marginal areas in the world—in tradethat ranged from informal to illegal, but this is changing as the opening of the border betweenthe two countries is making rapid progress5—a remarkable change compared to recent decades. Ashift from informal to more formal trade is taking place. This shift originates not only from theopening of borders but also from the use of information technology. Increasingly comprehensiveregulatory frameworks in Thailand and China are also encouraging informal traders to turn to for-mal transactions.

Opening borders in poor areas is a coin with two sides for the environment and for poverty.Legal trade can help reduce environmentally adverse illegal activities as local economic operatorsmay find larger formal transactions more profitable than smaller informal ones, provided interna-tional trade does not bypass the local operators of border trade. There is, however, a risk oflarger-scale environmental threats from international trade with limited poverty reduction gains.From a biodiversity perspective, trade in plant and animal products has long been an area of con-cern for environmental groups. This focus rightly deserves to be continued, but legal traderequires as much attention as the persistent illegal activities.

Examples of poverty and environment related to legal or informal trade:

• Logging for export to China and Vietnam increase as stronger environmental regulationscome into operation in these countries.

• Active NTFP exports to China and Vietnam are depleting resources for some species, evenwhen trade is legal. Quantities and prices often remain in the grey area of informal trade,and illegal exports of species with biodiversity value complement legal ones (PEN study onNTFPs in Laos).

• The hill closure policy in China and the development of demand for animal feed in the subre-gion trigger rapid development of hybrid maize cultivation on sloped land in Laos. This oftengenerates forest losses with limited income gains (Prodessa 2002).

• Pesticides are often purchased by farmers directly on border markets in Northern Laos in atotal absence of agricultural extension services (Ducourtieux 2004).

• In Laos, roads built by enterprises outside of road network development plans can lead to for-est depletion with limited improvements in accessibility for local people (PEN study on roads).

Local governments in marginal border areas are confronted with the issue of making improveddecisions in terms of environment and poverty as trade is increasing. The transition stage betweeninformal and formal exchanges creates specific challenges. Increased transactions are taking placewhile local government capacity, be it in controlling environmental impact or in ensuring povertyreduction gains, is limited. Two types of subregional initiatives can support this process:

• Local-government-to-local-government exchanges with a focus on the subregional povertyand environment issues;

• Design of information and training materials targeting local governments.

program collecting MDGs at these levels on aregular basis would be a cost-effective option.

These priorities for exchanges and monitoringfully justify a third phase of the PEN study. PhaseI consisted of an extensive review of available geo-graphically referenced data and an analysis of cor-relations between poverty and environment. PhaseII turned to researching linkages and povertyand environment in terms of policy implications.A third phase could usefully focus on supportingexchanges as well as monitoring poverty and envi-ronment issues and progress.

Endnotes1. Vietnam, Laos, and Cambodia, which together formed

French Indochina until 1954, also take part in the inter-national cultural cooperation forum related to the Frenchlanguage.

2. Vietnam joined ASEAN in 1995, Lao PDR joined in1997, and Cambodia in 1999.

3. Several poverty and environment linkages that have notbeen analyzed in depth in one of the three countries are notincluded in the figure but are however viewed as highly rel-evant. These include indoor air pollution and malaria.

4. The subregion has a centuries-old history of migrationof ethnic groups from the North to the South.

5. Under the Greater Mekong Subregion Cross-borderTransport Agreement that was initiated in 1999, majorroads will soon link major urban areas in the three coun-tries and in Thailand and China (Map 5.1). Full imple-mentation of the agreement is expected for 2007–08. Fivepilot border points are fully open since 2005 (ADB 2005).

ReferencesAsian Development Bank (ADB). 2005. Agreement between

and among the Governments of the Kingdom of Cambodia,the People’s Republic of China, the Lao People’s Democra-tic Republic, the Union of Myanmar, the Kingdom of

Thailand, and the Socialist Republic of Viet Nam for theFacilitation of Cross-Border Transport of Goods and Peo-ple. February 2005.

Asia-Urbs and Citynet. 2001. Cities, Poverty and the Environ-ment. Report and Proceedings from a Regional Workshop,Hanoi, July 30–August 2, 2001. Citynet (Regional Net-work of Local Authorities for the Management ofHuman Settlements) and European Communities(Europeaid Asia-Urbs Program).

Project for the Development of Sayaboury. (PRODESSA)2002. “Improving the Quality and Marketing Channelsof Maize and Other Crop Products in SouthernSayaboury Region.” Report for the French Committeefor Co-operation with Laos. Paris: The French Devel-opment Agency.

Ducourtieux, Olivier. 2004. “Shifting Cultivation andPoverty: a Complex Issue.” In: “Poverty Reduction andShifting Cultivation in the Uplands of Lao PDR, Tech-nologies, Approaches and Methods for ImprovingUpland Livelihoods.” Proceedings of a Workshop heldin Luang Prabang, January 27–30, 2004.

Homes, Derek. 2002. “Indonesia: Where Have All theForests Gone?” Environment and Social DevelopmentEast Asia and Pacific Region Discussion Paper. Wash-ington: The World Bank.

International Center for Environmental Management(ICEM). 2004. “Protected Areas and Development inthe Lower Mekong River Region.” Queensland, Australia:ICEM.

Lao UXO program. 2003. “Map of UXO Impact and Bomb-ing Data 1965–75.” In: “Claiming the future, the impactof UXO and Landmines in Lao PDR.” Herberton, Aus-tralia: Australian Network of the International Campaignto Ban Landmines.Mines Action Group (MAG). 2005. Laos Report 2005.

London: MAG.Mekong River Commission (MRC). 2003 Social Atlas of the

Lower Mekong Basin. Bangkok: MRC.World Bank. 1997. “Vietnam Forest Protection and Rural

development Project.” Staff Appraisal Report. Wash-ington: The World Bank.

World Wide Fund (WWF). 2005. Greater Truong Son Con-servation Action Plan. City: publisher.



POVERTY

ENVIRONMENT

NEXUS

Sustainable approaches to poverty reductionin Cambodia, Lao PDR and Vietnam

Environmental and Social Development UnitEast Asia and Pacific RegionTHE WORLD BANK1818 H Street, N.W.Washington, D.C. 20433Telephone: 1-202-473-1000www.worldbank.org/eapenvironment

The World Bank in Cambodia113 Norodom BoulevardPhnom Penh, CambodiaTelephone: 855-23-217-304www.worldbank.org/cambodia

The World Bank in Lao PDRPatou Xay - Nehru RoadP.O. Box 345 code 01004Vientiane, Lao PDRTelephone: 856-21-414-209www.worldbank.org/lao

The World Bank in Vietnam63 Ly Thai To StreetHanoi, VietnamTelephone: 84-4-934-6600www.worldbank.org/vietnam

THE WORLD BANK

PO

VE

RT

YE

NV

IRO

NM

EN

TN

EX

US

TH

E W

OR

LD

BA

NK

Sustainable approaches to poverty reductionin C

ambodia, Lao PD

R and V

ietnam


Date post:	12-Dec-2018
Category:	Documents
Upload:	buicong
View:	216 times
Download:	0 times

DRAFT CONFERENCE EDITION PE OVERTY NVIRONMENTN...

Documents