eotN.323K Cambodia Report No.35213-KH Report No. 35213-KH Cambodia Halving Poverty by 2015? Poverty Assessment 2006 Public Disclosure Authorized Authorized Disclosure Disclosure Public Public

February 7, 2006

East Asia and the Pacific Region Public Disclosure Authorized Authorized Disclosure Disclosure Public Public Halving Poverty by 2015? Public Disclosure Authorized Authorized Disclosure Disclosure Public Public

Document of the World Bank Public Disclosure Authorized Authorized Disclosure Disclosure Public Public

Abbreviations and acronyms

ACLEDA Association of Cambodian Local EMIS Education Management Information Economic Development Agencies System AFSC American Friends Service Committee EPI Expanded Program of Immunization ASEAN Association of South-East Asian Nations ESSP Education Sector Strategic Plan ATC Agreement on Textiles and Clothing EU European Union BAT British-American Tobacco EWMI East-West Management Institute CIS Commune / Sangkat FDI Foreign Direct Investment CAS 1) Country Assistance Strategy GDCC Government-Donor Coordination 2) Centre for Advanced Studies Committee CCA (UN) Common Country Assessment GDP Gross Domestic Product ccc Cooperation Committee for Cambodia GNI Gross National Income CCLS Cambodia Child Labor Survey (2001) GSP Generalized System of Preferences CDC Council for the Development of HSSP Health Sector Strategic Plan Cambodia IDRC International Development Research CDHS Cambodia Demographic and Health Centre Survey IFAPER Integrated Fiduciary Assessment and CDRI Cambodian Development Resource Public Expenditure Review Institute IF1 International Financial Institute CE Christian Era ILO International Labor Organization CEDAW Convention on the Elimination of all IMR Infant mortality rate forms of Discrimination Against Women JFF'R Japan Fund for Poverty Reduction CEO Chief Executive Officer KR Khmer Rouge CG Consultative Group LEB Life expectancy at birth CIPS Cambodia Inter-Censal Population M&E Monitoring and Evaluation Survey MAFF Ministry of Agriculture, Forestry and Child Labor Survey (or Cambodia Child CLS Fisheries Labor Survey) MBPI Merit-based pay initiative CMDG Cambodian Millennium Development MCH Mother and Child Health Goals Millennium Development Goals CoM Council of Ministers MDG Ministry of Economy and Finance CPR 1) Common Property Resources; MEF 2) Contraceptive Prevalence Rate MFA Multi-Fiber Arrangement CPUE Catch per unit of effort MFN Most Favored Nation CSD Center for Social Development MMR Maternal mortality rate CSES Cambodia Socio-Economic Survey MOEYS Ministry of Education, Youth and Sport cso Civil Society Organization MOH Ministry of Health CSR Corporate Social Responsibility MOI Ministry of Interior DAC Development Assistance Committee MOP Ministry of Planning DALYS Disability-Adjusted Life Years MOPS Moving out of poverty study DFID Department for International MOT Ministry of Tourism Development MOWA Ministry of Women's Affairs DHS Demographic and Health Survey (or MOWRAh4 Ministry of Water Resources and CDHS) Meteorology DK Democratic Kampuchea MPA Minimum Package of Activities DPSIC Department of Planning, Statistics and MRD Ministry of Rural Development International Cooperation MTEF Medium-Term Expenditure Framework DPT Diphtheria, Pertussis and Tetanus NB C National Bank of Cambodia EC European Community NCD Non-communicable diseases EF Equity Fund NGO Non-Governmental Organization EFA Education For All NIS National Institute of Statistics EIC Economic Institute of Cambodia NPRS National Poverty Reduction Strategy PSU Primary Sampling Unit NSDP National Strategic Development Plan RGC Royal Government of Cambodia NTFP Non-Timber Forest Product SARS Severe Acute Respiratory Syndrome O&M Operations and Maintenance SESC Socio-Economic Survey of Cambodia ODA Official Development Assistance SME Small and Medium Enterprises OECD Organization for Economic Cooperation SMP Statistical Master Plan and Development SWAP Sector-Wide Approach OOPS Out-of-pocket spending SWIM Sector-Wide Management p.a. per annum TA Technical Assistance p.c. per capita TCM Textile and Clothing Machinery PAP Priority Action Plan TWG Technical Working Group PETS Public Expenditure Tracking Survey TWG-PPR Technical Working Group-Planning and PFM Public Financial Management Poverty Reduction PFMRP PFM Reform Program USMR Under-five mortality rate PIP Priority Investment Plan UNDAF United Nations Development Assistance PPA Participatory Poverty Assessment Framework United Nations Transitional Authority in PRK People's Republic of Kampuchea (1979- UNTAC 1989) Cambodia PRS Poverty Reduction Strategy PRSP Poverty Reduction Strategy Paper Acknowledgements

This report has received valuable guidance from Dr. Aun Porn Moniroth (Chairman of SNEC and Secretary of State of the Ministry of Economy and Finance), Dr. Hang Chuon Naron (Secretary General of the Ministry of Economy and Finance), H.E. Vongsey Vissoth (Deputy Secretary General of the Ministry of Economy and Finance), H.E. Chan Tong Yves (Secretary of State of the Ministry of Agriculture, Forestry and Fishery), H.E. Sok Chenda (Secretary General of the CDC), H.E. Chhieng Yanara (Deputy Secretary General of the CDC), Heang Siekly (Deputy Director General, Ministry of Planning), Ly Sina (Division Director, Department of Forestry, Ministry of Agriculture, Foretry and Fisheries), Nhean Sochetra (Director of Gender Equality Department, Ministry of Women’s Affairs), Chhuon Chham (Deputy Secretary-General, Council for Administrative Reform) and Uch Sarom (Consultant, Ministry of Women’s Affairs / UNDP).

Earlier drafts were peer-reviewed by Louise Cord and Kathy Lindert (both World Bank), Shyam Bajpai (ADB), Hong Sokheang (UNDP), Helen Appleton (DFID), Sok Hach and Lim Visa1 (EIC), and Brett Ballard and Larry Strange (CDRI). We gratefully acknowledge inputs and contributions including suggestions from Yusuf Ahmad, Enrique Aldaz-Carroll, Anne Christensen, Elisabet Dahlberg (Stockholm School of Economics), Dourng Kakada (EIC), Ignas Dummer (Danida), Eiichiro Hayashi (JBIC), John McAndrew (ADI), Tomoki Fujii, the staff of UNFPA, Didier Grebert (AFD), Roger Hencke (CAS), Jolanda Jonkhart (EC), Long Piseth (Danida), Caroline McCausland (Concern Worldwide), Roneen Marcoux (NEP), Karla Quizon (IFC / MPDF), Neak Samsen (EIC), John Nelmes (IMF), Nou Keosothea (EIC), Michael Stephen (Mekong Bank), Thomas Markussen, Vijay Rao (WHO), Mao Moni Ratana (ADB), Ram Saravanamuttu (WFP), Sofia Sebastian, Carol Stickler (CCC), Emi Uchida, Hirotsugu Uchida and Cheryl Urashima.

The report was prepared under the overall coordination and guidance of Indermit Gill, Sector Manager, and Nisha Agrawal, Country Manager. The Executive Summary was written by Tim Conway and Chorching Goh; Chapter 1 by Tim Conway; Chapters 2 and 3 by Chorching Goh; Chapter 4 by Eric Sidgwick; Chapter 5 by Chorching Goh with Chris Jackson (Rural Development) and Chan Sophal; and Chapter 6 by Tim Conway and Toomas Palu (Human Development) with Andy Ragatz (Consultant); Chapter 7 by Tim Conway; and Chapter 8 by Tim Conway and Mia Hyun. The data annexes were prepared by Chea Huot.

Inputs were provided by many others in the World Bank team, including Daniel Adler (SD), Beng Simeth (HD), Luis Benveniste (HD), Amitava Biswas (Consultant), Ximena del Carpio (HD), Klaus Deininger (DEC), Gershon Feder (DEC), Mona Haddad (PREM), Minna Hahn (HD), Huot Chea (PREM), Stephen Mink (RD), Nil Vanna (Social Development), Steven Noakes (Consult,ant), Nat Pinnoi (EAS), Steven Schonberger (RD), Dean Spinanger (Consultant), and Rob Taliercio (PREM). Overall guidance was also provided by Ian Porter (Country Director, Cambodia) and Kazi Mahbub-Al-Matin (Lead Economist, Cambodia). The 2004 national poverty estimates on which this report draws heavily were prepared by Jim Knowles (Consultant), in a technical report which will be published separately. Bou Saroeun (Communication Specialist) coordinated the editing and production of the Khmer version of the Report. Assistance on the preparation of the report was also provided by Iv Ek Nimnuon and Zelena Jagdeo.

The key sources of data for this report were the Cambodia Socio-Economic Survey (CSES) 2004, conducted by the National Institute of Statistics (NIS); the Moving Out of Poverty Study (MOPS) and Tonle Sap Participatory Poverty Assessment (PPA), both conducted by the Cambodian Development Resource Institute (CDRI); and four District-level case studies of local differences in growth and poverty reduction, conducted by the Economic Institute of Cambodia (EIC). The CDRI MOPS study and the EIC District-level economic appraisals were both commissioned by the World Bank as inputs to this Poverty Assessment. It is intended that both will be published by their respective institutions as separate research papers in 2006. The CDRI Tonle Sap PPA was commissioned by ADB: our thanks go to both CDRI and ADB for permission to use early findings from this study in this report. TABLE OF CONTENTS

EXECUTIVE SUMMARY

Overview ...... i Structure and sources ...... ii Key facts ...... 111 Diagnosis ...... vi Binding constraints ...... x Talung stock and looking forward ...... xiii Policy priorities ...... xvi

CHAPTER 1: Country Context

Summary ...... 1 Geography and opportunity ...... 2 The legacy of history ...... 4 Subject matter and sources ...... 12 Summary ...... 15

CHAPTER 2: Poverty trends. 1994-2004

Summary ...... 17 Tracking poverty over time ...... 18 Key findings ...... 21 Moving in and out of poverty: panel data analysis ...... 30

CHAPTER 3: The nature of poverty in Cambodia. 2004

Summary ...... 35 Household consumption ...... 36 Poverty in 2004 ...... 37 Household poverty profile ...... 41 Urban poverty: lack of secured housing and insecurity of income ...... 47 Inequality is high ...... 49 Vulnerability ...... 51

CHAPTER 4: Sources of economic growth

Summary ...... 55 Retrospective on economic growth...... 56 Agriculture ...... 58 Industry ...... -64 Binding constraints to agricultural and agri-business growth ...... -72 Second-order constraint to agri-business growth: high cost of capital ...... 76 CHAPTER 5: Raising agricultural incomes

Summary ...... 77 Agricultural sector ...... 78 Crop yields are among the lowest in the region...... 79 The natural resource base ...... SO Fishery sector ...... 81 Forestry sector ...... 83 Land: the primary asset of the poor ...... 84 Factors to enhance agricultural productivity ...... 88

CHAPTER 6: Education. health and wealth

Summary ...... 93 Education, skills and growth ...... 94 Health and wealth ...... 110

CHAPTER 7: Building an effective and responsive state

Summary ...... 131 Building state institutions and public sector capacity ...... 132 Voice, participation and social accountability ...... 147

CHAPTER 8: New partnerships for poverty reduction

Summary ...... 155 Will Cambodia meet its poverty reduction targets?...... 157 The NSDP and CMDGs...... 159 Supporting the NSDP: aid effectiveness in Cambodia ...... 161 Aid and poverty reduction, 1993-2004 ...... 163 Monitoring and evaluating progress ...... 171

REFERENCES ...... 175

ANNEX 1: Technical Appendix ...... 185

ANNEX 2: Regression Results in Chapter 5 ...... 194 ANNEX 3: Tables ...... 199 List of Boxes ... Box 1: The difficulties in establishing poverty trends in Cambodia ...... 111 Box 2: Different parts of the countryside have had very different experiences of growth ...... vi Box 3: Landlessness amidst fenced-off land in Phnom Sruoch...... ix Box 4: “Up to the sky”: without water control infrastructure, farmers are critically dependent on rainfall ...... x Box 5: Education and the poverty trap...... xii Box 6: A health shock is often more serious than a harvest failure ...... xii Box 7: Corruption keeps the poor poor ...... xiii Box 8: The National Strategic Development Plan (NSDP), 2006-2010: key commitments ...... xvi

Box 1.1: Angkor, tourism and growth ...... 4 Box 1.2: Peach has opened up opportunities and created the possibility of a better life...... 9 Box 1.3: Poverty reduction in the post-conflict societies: lessons form international experience ...... 11 Box 1.4: Comparing poverty over time and across space: basic concepts in using the CSES data ...... 13 Box 1.5: Household dynamics and perspectives of the poor: the Moving Out of Poverty Study (MOPS) ...... 15

Box 2.1: Comparing mangos and bananas: measuring poverty in Cambodia, 1994-2004 ...... 19 Box 2.2: Measuring poverty: headcount, gap and severity ...... 21 Box 2.3: Backward-projection of national poverty estimates for 1994...... 22 Box 2.4: Perceptions of inequality amongst rural youths ...... 29

Box 3.1: Choosing a poverty line: national and dollar-a-day poverty lines compared ...... 38 Box 3.2 Gender and Poverty in Cambodia ...... 42 Box 3.3: Ethnic minorities and poverty ...... 43 Box 3.3: The urban poor: small, but undercounted and growing? ...... 48 .

Box 4.1: Internal and cross-border labor migration ...... 62 Box 4.2: Contract farming - fragrant rice and tobacco ...... 65 Box 4.3: Garment factory jobs and remittances: the view from the factories and the villages ...... -67 Box 4.4: Tourism and the poor ...... 71

Box 5.1: The success story of Malai...... 88 Box 5.2: Tracking the frequency and severity of land conflicts ...... 89 Box 5.3: Prevalent landlessness amidst fenced-off unused arable land in Phnom Sruoch ...... 89 Box 5.4: Income fluctuates with the amount of rainfall ...... 90 Box 5.5: A health shock is often more serious than a harvest failure ...... 92

Box 6.1 : Peace. more schools and changes to fee policy have increased participation in primary education ...... 95 Box 6.2: Parents worry about the quality of teaching ...... 101 Box 6.3: The priority action program ...... 102 Box 6.4: Supplementary tutoring mirrors public schooling ...... 103 Box 6.5: Villagers perceive links between education and opportunity ...... 104 Box 6.6: Explaining enrolment and retention: findings from multivariate analysis ...... 108 Box 6.7: Scholarships help girls from poor families stay in school ...... 109 Box 6.8 Access to health and “new diseases” ...... 111 Box 6.9 Contracting-out to NGOs for healthcare: ...... 128

Box 7.1: Cambodia continues to rank low in the eyes of investors ...... 134 Box 7.2: From salary supplements to pay reform ...... 139 Box 7.3: The PFM Reform Program offers a realistic agenda for aligning spending to policy ...... 139 Box 7.4 Early estimates of Cambodia’s oil and gas resources ...... 140 Box 7.5: Popular accounts emphasize the ubiquity and severity of corruption .. 142 Box 7.6: Crime and violence are real-and perhaps growing-problems ...... 145 Box 7.7: Poverty and domestic violence are mutually reinforcing ...... 146 Box 7.8: Qualitative research suggests a decline in domestic violence in some villages ...... 147 Box 7.9: Coordinating NGO activities remains a major challenge ...... 150 Box 7.10:Drug use is a serious, and growing, problem for Cambodian youth .... 151 Box 7.11 Rural Cambodians are adopting new forms of collective action to defend their interests ...... 152

Box 8.1 : Poverty Reduction Strategies: Ownership. participation and poverty outcomes ...... 160 Box 8.2: The National Strategic Development Plan (NSDP). 2006-2010: key commitments ...... 162 Box 8.3: Ownership, alignment and harmonization: basic concepts ...... 163 Box 8.4: Improving outcomes or institution-building? ...... 164 Box 8.5: Cambodia’s Harmonization and Alignment Action Plan ...... 167 Box 8.6: Harmonization and alignment: some practical examples ...... 169 Box 8.7: Towards mutual accountability: lessons from Tanzania and Mozambique ...... 170 Box 8.8: Summary requirements of a national poverty M&E system ...... 172 Lists of Figures

Figure 1: Over the last decade, poverty in Cambodia has fallen by an Estimated 12 percentage points i ...... Figure 2: Household consumption has risen 32 percent in real terms ...... 111 Figure 3: The poorest fifth of the population own more assets and have better access to electricity than ten years ago ...... iv Figure 4: The poorest 20 percent also live in better houses than in the past .... iv Figure 5: Cambodia has been remarkably effective in slowing and reversing the spread of HIV ...... v Figure 6: Industry’s contribution to GDP has doubled -- but it still employs only 8 percent of the workforce ...... vii Figure 7: Yields of rice-Cambodia’s food staple-are very low compared to neighboring countries ...... viii Figure 8: Cambodia has yet to achieve the sustained high rates of agricultural growth associated with long-term economic ... development...... viii Figure 9: Common property resources-fishing and forestry-are very important to the poorest quintile, but only marginally important for the richest quintile ...... x Figure 10: Higher agricultural growth would deliver the property reduction goal faster ...... xiv Figure 11: Small farms are a better economic proposition than large farms .. xvii

Figure 1.1: Cambodia benefits from access to the sea, rich freshwater fisheries, and economic growth in neighboring countries ...... 3 Figure 1.2: Landscape and natural resources shapes spatial variations in Cambodia’s economic and social development.,...... 14

Figure 2.1 Between 1994 and 2004, poverty fell by an estimated twelve percent...... 18 Figure 2.2: Poverty-and food poverty-has fallen in both urban and rural areas ...... 23 Figure 2.3: The depth of poverty has declined in both the towns and the countryside ...... 23 Figure 2.4: Per capita consumption increased significantly over the last ten years ...... 24 Figure 2.5: Consumption has risen in rural areas-but more slowly than in urban areas ...... 24 Figure 2.6: Households-including poor households-can now afford to spend more on non-food items ...... -25 Figure 2.7: The poorest fifth of the population own more assets and have better access to electricity than in 1994...... 26 Figure 2.8: The poorest 20 percent also live in better houses than in the past ....26 Figure 2.9: Net primary enrolment has improved for all groups., ...... 27 Figure 2.10: Over the last ten years. per capita consumption amongst the poorest fifth has risen by only 8 percent-compared to 45 percent amongst the richest fifth of the population ...... 27 Figure 2.11: Poverty has fallen much faster in Phnom Penh and the coast ...... 28 Figure 2.12: Inequality rose in rural areas. and in the country as a whole ...... 29 Figure 2.13: Household dynamics and the changing composition of income. 2001-2004 ...... 34

Figure 3.1 : Rural Plateau and rural Tonle Sap regions had the lowest average of consumption...... 36 Figure 3.2: Per capita consumption in upland areas is only 75 percent of the national average ...... 36 Figure 3.3: Standards of living are lowest in the MountainPlateau and Tonle Sap regions ...... 37 Figure 3.4: The majority of the poor are found in the Plains and Tonle Sap Provinces ...... 39 Figure 3.5: Poverty rates are highest in Kampong Speu. Siem Reap and Kampong Thom provinces ...... 40 Figure 3.6: Households headed by men and women in the prime of thief life (aged 30-50 years) made up the largest proportion of the poor and experience the highest rates of poverty ...... 40 Figure 3.7: There is no statistical difference in the CSES data between the incidence of poverty in male-and female-headed households ...41 Figure 3.8 : Larger household tend to be poorer because of higher dependency ratios ...... 44 Figure 3.9: Households headed by domestic workers and farmers are much more likely to be poor ...... 45 Figure 3.10: The poorest infrastructure is to be found in Provinces in the Mountaidplateau region and Siem Reap Province ...... 46 Figure 3.1 1: The poor are acutely deprived of energy. water and sanitation ...... 46 Figure 3.12: The poorest and next poorest quintiles are less likely to have secure land titles ...... 47 Figure 3.13: Inequality in Cambodia is high. particularly in urban areas ...... 49 Figure 3.14: Inequality in Cambodia is high compared to its regional neighbors ...... 50 Figure 3.15: Distribution of household consumption and vulnerability ...... 52

Figure 4.1 : Contribution to real GDP growth...... 56 Figure 4.2: Sub-sector contribution to agricultural sector growth ...... 58 Figure 4.3: Land is being brought into agricultural production, mainly through conversion form forest- ‘000 hectares ...... 60 Figure 4.4: Growth diagnostics ...... 73

Figure 5.1: Agriculture and access to natural resources is critical for the poor ..78 Figure 5.2: Revenue from timber and non-timber products ...... 84 Figure 5.3: Small farms are a better economic proposition than large farms .....87

Figure 6.1: Public expenditure on education-particularly on recurrent costs-has risen dramatically ...... 96 Figure 6.2: Primary education in Cambodia compares poorly with the rest of the region ...... 97 Figure 6.3: Pronounced inequalities still exist in education outcomes ...... 98 Figure 6.4: Children from poor families are much more likely to obtain a primary education in the south-east and north-west ...... 99 Figure 6.5: Net primary enrolment has improved for all ...... 100 Figure 6.6: Those with a good education are significantly less likely to be poor ...... 105 Figure 6.7: Average wages and earnings increase with each stage of education completed ...... 106 Figure 6.8: Cambodia has achieved remarkable success in reversing the spread of HIV ...... 112 Figure 6.9: Malnutrition rates in Cambodia are extremely high ...... 113 Figure 6.10: Poor and uneducated women have more children than other women ...... 117 Figure 6.11: The level of a mother’s education is strongly and positively associated with the children’s immunization ...... 118 Figure 6.12: Smolung is overwhelmingly a male habit, and strongly negatively influenced by education ...... 120 Figure 6.13: Compared to the rich, the poor spend proportionately more on tobacco and less on medical care ...... 120 Figure 6.14: Frequency (% individuals) and size (mean riels per capita shown by size of circle) of out of pocket spending on health care in the last 4 weeks, by consumption quintile, 2004 ...... 123 Figure 6.15: The 10 percent of the population with bad or very bad health account for 64% of total household spending on healthcare ...... 124 Figure 6.16: Inadequate drug supply. cost of services and distance to facilities are seen as the key problems with health services ...... 125 Figure 6.17: For the poor. distance to health facilities may be as important as the cost of care ...... 126

Figure 7.1 : Governance issues-corruption and bureaucracy-are seen as the most important obstacles to private sector development ...... 135 Figure 7.2: Corruption is most pronounced in the courts-but corruption in education has the greatest direct impact ...... 143 Figure 7.3: Channels for citizen demand may be either direct or indirect ...... 148

Figure 8.1: Aid dependence in Cambodia: declining slowly but still high ...... 158 Figure 8.2: Status of harmonization and alignment in 2004: selected indicators ...... 166 List of Tables

Table 1: Profile of the poor in 2004 ...... v Table 2: The rate of poverty reduction has been good-but a number of other countries have done better ...... xiii

Table 2.1: Updated food poverty line, nonfood allowances, and total poverty lines ...... 21 Table 2.2: Transition matrix of panel households...... 3 1 Table 2.3: Changes in per capita consumption and in positions within the income distribution of panel households, 2001-2004 ...... 32

Table 3.1: Poverty headcounts are highest in the rural areas of the Tonle Sap and MountainPlateau regions, which together account for almost half of all poor ...... 39

Table 4.1: Sources of Growth by main economic activity, 1994-2004 ...... 57 Table 4.2: Agricultural sector performance during early stages of development .59 Table 4.3: Compared to its neighbors, Cambodia has very little irrigated arable land) ...... 61 Table 4.4: Most crop yields did not improve over the last decade, with the exception of maize ...... 63 Table 4.5: Growth in domestic tourists is increasingly important in recent years ...... 70

Table 5.1: Food production and consumption for rural households by quintile ...79 Table 5.2: Agriculture productivity for Cambodia and other countries, 2003 ...... 79 Table 5.3: Rice constitutes the bulk of farm households’ agricultural revenue in Cambodia ...... 80 Table 5.4: Rice still nominates crop production in Cambodia ...... 81 Table 5.5: Comparison of fishing and non-fishing households by income and consumption ...... 83 Table 5.6: Estimated effects of road infrastructure and human capital on agricultural productivity ...... 91

Table 6.1: There have been significant improvements in education infrastructure ...... 97 Table 6.2: Child labor rates in Cambodia are amongst the highest in Southeast Asia ...... 103 Table 6.3: The children of educated mothers are much less likely to be malnourished...... lo7 Table 6.4: Despite gains, Cambodia has some of the lag behind health outcomes in the region ...... 110 Table 6.5: Communicable diseases are the main cause of illness for all groups-although non-communicable illnesses are becoming significant for the rich...... 111 Table 6.6: Mean rates (%) of moderate and severe malnutrition among children under 5 by p.c. consumption quintile, 2004 ...... 114 Table 6.7: The poor are much less likely to seek healthcare when ill ...... 121 Table 6.8: Outside Phnom Penh and the Tonle Sap provinces, wealth makes a big difference to healthcare utilization ...... 122 Table 6.9: Individuals in the richest quintile are five times more likely to use hospital services than individuals from the poorest two quintiles ...... 122

Table 7.1: The business climate in Cambodia is far less attractive than the regional average ...... 134 Table 8.1: The rate of poverty reduction has been good but a number of other countries have done better...... 157

subsistence-oriented economy to one based Overview-~ on the market and open to international The latest (2004) household survey finds that 35 percent of Cambodians live below trade. the national poverty line, down from an On the basis of this performance, estimated 47 percent a decade earlier Cambodia has made significant progress (although the two surveys are not directly towards the first of the Cambodian comparable (Figure 1 and Box 1). Over Millennium Development Goals (CMDGs), these ten years, Cambodia has consolidated which commits the Royal Government to peace and achieved economic growth of halve, between 1993 and 2015, the about 7 percent per annum. Non-income proportion of people living below the indicators such as ownership of consumer national poverty line. Starting from a base durables, housing quality and schooling of very low output and very high poverty indicate that the living standards of the at the start of the 1990s, peace and population, including the poor, have economic liberalization have made improved between 1994 and 2004. These possible a decade of rapid growth and gains reflect the economic and social relatively rapid poverty reduction. The development that has been made possible Governments elected since 1993 deserve since the early 1990s, when Cambodia credit for securing and (with occasional embarked on a three-fold transition from brief setbacks) maintaining peace, for civil war to peace, from one-party to multi- some critical investments in roads and party politics, and from an isolated and services, and for over-seeing the process

Figure 1 : Over the last decade, poverty in Cambodia has fallen by an estimated 12 percentage points

60

50 47 - 43 v8 4- 5 40 0 TI8 30 r p 20 > 0 Q 10

0 Cambodia A Cambodia B: Rural Urban (excluding Phnorn Penh entire country(NE 199314 Phnom Penh) 1993/4 estimate ampling frame based on backward within geographically comparable parts of the country projection) (1993/4 sampling frame)

Source: SESC 199314, CSES 2004. 11 Cambodia - Poverty Assessment 2006 by which Cambodia has attracted agriculture for both growth and poverty international investment and established reduction; (iii)equitable access to common high rates of export-led economic growth. property resources as a critical source of Policy choices on trade and infrastructure income and security for the rural poor; (iv) have undoubtedly helped significantly. increased investments in productivity- enhancing infrastructure; and (v) improved Although all segments of society benefited human development and human capital, over the last decade, the rise in living achieved through the pro-poor delivery of standards has been more pronounced in basic services in education and health. urban areas and amongst the richest quintile. The extreme poor (the bottom 20 percent who fall below the food poverty Structure and sources line) have experienced significantly slower This Executive Summary is structured in growth in real consumption than the five parts. Firstly, it lays out the key facts “normal” poor. While poverty in rural on the nature of poverty in 2004 and-for areas fell by 22 percent, cumulatively, over the first time in Cambodia-on poverty the decade between surveys, it fell by 60 trends (Box 1). This section notes success percent in Phnom Penh and 44 percent in in reducing poverty but also draws other urban areas. In 2004, 91 percent of attention to the growing challenge of poor Cambodians lived in rural areas. widening inequalities. Secondly, it Poverty rates are highest in remote rural provides a diagnosis of the relationships areas with limited access to roads, markets between growth, institutions and poverty, and basic services. highlighting the need to focus on the rural economy. Thirdly, it analyzes the binding The Government’s new five-year National constraints to shared growth, noting in Strategic Development Plan 2006-2010 particular the role of property rights, rural (NSDP) recognizes the need to address infrastructure and human capital. Fourthly, rural development and makes improving it takes stock by comparing Cambodia’s the lives and livelihoods of the rural poor a performance in reducing poverty to that of top priority. In keeping with the NSDP, other countries; laying out alternative this report reasons that improving the lives poverty projections for the next ten years; and livelihoods of the rural poor should be and describing the Government’s new five- a top government priority; and that year development plan. Finally, it provides accelerating poverty reduction in a brief discussion of broad policy priorities Cambodia is largely about helping to raise for public action to promote growth, agricultural productivity and incomes. service delivery and poverty reduction. The engines of growth that have driven improvements since the early 1990s The key sources of information on which (namely garment exports, tourism and this poverty assessment is based include construction) now need to be the nationally-representative 2004 complemented with other sources of Cambodia Socio-Economic Survey growth that are more rural, more broad- (CSES) and a set of studies by the based, and more pro-poor. To accelerate Cambodia Development Resource Institute poverty reduction by improving (CDRI) and Economic Institute of productivity and profitability in the rural Cambodia (EIC), which mix quantitative economy will require: (i)secure property and qualitative analysis to address rights to private land, particularly for particular aspects of poverty analysis. smallholders; (ii)emphasis on small holder ... Executive Summaw 111

Key facts Box 1 : The difficulties in establishing poverty trends in Cambodia As living standards have improved, the A major contribution of this report is to present, poverty rate has fallen for the first time, comparable estimates of household consumption and poverty between Between the first household survey in different years, and so establish the nature of 1993/94, and the latest survey in 2004, real trends over time. Official poverty estimates per capita household consumption (the during the 1990s were not comparable for value used to measure living standards and several reasons. Firstly, each set of poverty estimates was based on a different define the poverty line) rose by 32 percent. geographical sampling frame. Due to security Food shares in total consumption fell, as problems the 1993-94 Socio-Economic Survey households can now afford other needs of Cambodia (from which the first poverty (Figure 2). The increase in average per estimates were derived) covered only 56 capita consumption has been experienced percent of rural villages and 65 percent of the rural population; areas in which the Khmer amongst the poorest quintile (that is, the Rouge were still active were not surveyed. In poorest twenty percent of the population) subsequent surveys, these more remote and- as well as the richest quintile. as has been confirmed from the 2004 survey- poorer areas were progressively included in the sampling frames. The 2004 Cambodia Figure 2: Household consumption Socio-Economic Survey (CSES) is the first to has risen 32 percent in real terms include all Cambodian villages in its sampling $ 3,000 frame. While it should in theory have been m possible to at least compare changes between surveys within the areas surveyed in 1993194, k this was not possible because the description of the 1993194 sampling frame was lost. With the rediscovery of this sampling frame in 2005, it has become possible to analyze change within the geographically-comparable sub- sample. Secondly, differences in the design and management of the surveys in different years, and different approaches taken to the calculation of the poverty line, created further problems for those wishing to identify changes 0 0 in poverty between surveys. c Q, Given all these differences, it was simply not v) 0 valid to compare estimates from different Ia years, or to conclude on the basis of this that 1993/94 2004 Cambodia experienced little or no poverty reduction during the 1990s. The rediscovery 1 0 Non-food consumption of the I993194 sampling frame, combined with Food consumption 1 a basically common structure to the consumption modules of the questionnaires, now makes it possible to compare data from Source: SESC 1993194, CSES 2004. 2004 with 1993194. NB: Comparison is of household consumption within the geographically-comparable sampling frame of the 1993/94 survey. iv Cambodia - Poverty Assessment 2006

As a result, the poverty headcount (the Figure 3: The poorest fifth of the proportion living under the poverty line) population own more assets has fallen by over a quarter. In the area and have better access to surveyed in 1993194 in which direct electricity than ten years ago comparison is possible with the same area in 2004, the headcount fell from 39 percent to 28 percent. Backward projections estimates that the poverty rate in the country as a whole in 1993/94 was around 47 percent; in 2004, the all-Cambodia poverty headcount was down to 35 percent . Radio Television Electric light from city The decline in poverty has been grid, widespread generator or battery Importantly, poverty reduction has not Figure 4: The poorest 20 percent also live been confined to any particular area. in better houses than in the past Poverty fell by between a fifth (in rural areas) and a half (in Phnom Penh), and in - Primary roofing material all but one of the five agro-ecological regions of the country. (Trends in the fifth i7 ] 74% e region cannot be determined with c v) confidence.) Broadly speaking, women as g! 0 well as men have benefited from these 8 a, trends, with significant numbers finding 5 .I- employment in the new garment industry 0 and improvements in primary education 8 now closing the gender gap in literacy Thatch Galvanized Tile amongst younger age groups. iron or aluminum The fall in consumption poverty is corroborated by other indicators Source: SESC 1993194, CSES 2004. Non-income indicators of well-being such as ownership of consumer durables, remarkable success has been achieved as housing quality and schooling confirm the government commitment and strengthened picture of rising living standards, amongst institutions and policies have been backed the population in general and the poor in by effective coordination of external particular (Figures 3 and 4). assistance (Figure 5). Human development indicators also suggest steady progress. Between 1997 and 2004, net enrollment and completion rates at the primary level increased by over 10 percentage points among children from the poorest fifth of the population. In cases such as HIV/AIDS control, Executive Summary V

Figure 5: Cambodia has been remarkably secure land tenure, remoteness from effective in slowing and reversing the markets and services, lack of spread of HIV productive assets, low levels of education, and high dependency ratios - Estimated HIV prevalence (% of adults aged 15-49) (Table l). - The rural population began with the lowest average real consumption in 1993/94 and experienced the least growth during the last decade. As a result, the countryside has the lowest levels of per capita consumption and the highest incidence of poverty, and

Cambodia today is thus essentially -one Source: NCHADS 2005. of rural livelihoods. Rapid economic change is also changing social relationships, with some positive and some negative effects. Despite progress, poverty remains Women employed in garment factories widespread and multidimensional often lead subsistence-level lives as they Although the poverty rate has fallen by struggle to remit much of their wages to around one percent each year for the last their families; and domestic violence has decade, a third of the population still lives been and remains a very serious problem. below the poverty line and one in five live below the food poverty line. The poor face Inequality has risen a number of interlocking and mutually- Comparing the 1993/94 and 2004 surveys reinforcing problems, including lack of suggests that even as the poor have

Table 1 : Profile of the poor in 2004 Quintile Poorest Next Middle Next Richest poorest richest 1 Owned agricultural land is secured by 15.6 21.6 24.5 25.3 28.6 a title (“h) 1 Distance to nearest all-weather road 5.2 3.7 3.3 3.1 1.9 (km) 1 Distance to permanent market (km) 10.8 9.6 8.1 7.1 4.2

1 Households with water pump (“5) 3 8 9 12 13

1 Plots with access to irrigation in dry 6 10 12 13 12 season (“A) 1 Dependency burden (elderly and 96.9 84.5 76.6 64.9 54.3 children: 100 economically active adults) 1 Literate adults, age 15+ (“h) 29.3 38.3 43.2 51.1 60.8

1 Education (average school grades 2.8 3.5 3.9 4.6 6.3 completed by adults) Source: CSES 2004. vi Cambodia - Poverty Assessment 2006

Box 2: Different parts of the countryside have had very different experiences of growth and poverty reduction The Economic Institute of Cambodia (EIC) conducted case studies of different local paths to poverty reduction deriving from different sources of growth in four rural Districts in four Provinces. Puok District in Siem Reap Province has received some benefits from the boom in tourism in the Provincial town, which has created employment in construction, transportation, petty trade and handicrafts. However, production-side constraints (poor soils, no money for inputs such as fertilizer, lack of water control infrastructure or good roads) have prevented most local farmers from responding to increased demand for food crops from hotels and restaurants. Malai District in Bantey Meanchey Province is a good example of successful broad-based agricultural growth. With the integration of this former Khmer Rouge zone into the national Government, land was distributed to local households. Relatively rich soil brought into production through conversion of forests to farmland, combined with good access to export markets (Malai is located on the Thai border) has led to a dramatic increase in production of cash crops and rising incomes for most households. Phnom Sruoch District in Kampong Speu Province-one of Cambodia’s poorest Provinces - provides a contrast to Malai. In Sruoch, the end of conflict enabled the poor to clear previously insecure forested land. However, poor quality soils and exposure to a variety of shocks have meant that many have not been able to retain their land, but have sold it and moved on to clear new land and start again. There has been a steady accumulation of land in the hands of the rich, many living in Phnom Penh, while the situation of the local poor has changed only marginally, and may now be declining as further forest clearance becomes impossible. Mesang District in Prey Veng Province is typical of much of the densely populated south- eastern plains where some 40 percent of the poor live. In the face of population increase and declining land availability, the rural poor in Mesang are encountering the limits of traditional livelihood strategies based on rice cultivation and common property resources, and are increasingly relying on sending family members to Phnom Penh to find employment in the garment industry, construction, or transportation.

experienced a discernible improvement in start widen only at later stages of living standards, inequality has increased development, when levels of average significantly. The Gini coefficient income and consumption were higher (and provides a commonly-used summary poverty headcounts lower). measure of inequality, in which a value of Part of the growth in inequality has been zero signifies perfect equality in the due to widening differences between urban distribution of income or consumption, and rural areas. Part of it, however, has while a value of one indicates perfect been due to growing differences within the inequality. In those parts ofCambodia that rural population, with considerable were surveyed in 1993/94, the Gini for real variation in the rate and distribution of per capita consumption has risen from 0.35 growth in different localities (Box 2). to 0.40. Inequality in all of Cambodia in 2004 was 0.42, making Cambodia one of the more unequal countries in the region. Diagnosis Inequality in Cambodia is somewhat Cambodia’s fundamental geographical unusual in that most countries in South- endowments-including its location in the east and East Asia saw levels of inequality fastest growing region in the world-give Executive Summaw vii it a significant advantage over Figure 6: Industry’s contribution to GDP has doubled-- many landlocked, mountainous but it still employs only 8 percent of the or resource-poor low income workforce countries. Rather than climate or 1 topography, Cambodia’ s 46% Agriculture 0 contemporary poverty is largely 71 % a legacy of over twenty years of political conflict, which resulted in destruction of physical Services infrastructure, decimation of 21 % human capital and distortion of social, economic and political 0 Industry 29% institutions. I Just as poverty in Cambodia has I its origins in conflict, so poverty 100 50 0 50 100 share of labour reduction over the last decade share of GDP force (%) (“10 1 has its origin in the restoration of peace. The fall in poverty has been driven by a combination of Source: MEF, NIS. high economic growth (made possible by the end of conflict and the in the economy, the industrial sector still opening of Cambodia to foreign employed only 8 percent of the labor force investment and trade), macroeconomic in 2004 (Figure 6). stability and improved access to services. The rate of poverty reduction could have Meanwhile agriculture, central to the been even higher, however, if more of the livelihood strategies of the poor, has grown growth had been in the agricultural sector slowly and erratically, with several years and in rural areas. in which production has fallen due to droughts andor extreme floods. Cambodia remains dependent on Agricultural productivity in Cambodia urban engines of growth remains very low compared to other The patterns of poverty reduction are what countries in the region. Yields of paddy would be expected, given where and how rice-which is central to the food security growth has occurred. Between 1993 and and livelihoods of rural households and 2004, economic growth averaged 7.1 accounted for 88 percent of the total area percent per annum. The main engines of under crops harvested in 2004-are the growth over the last decade-export- lowest in the region (Figure 7). oriented garment manufacturing and tourism-were both urban-focused and Urban-rural linkages are weak., . have generated only limited urban-rural Both garments and tourism, Cambodia’s linkages. Between 1993 and 2004, the main engines of growth, have been and share of the industrial sector in the still are concentrated in Phnom Penh and economy more than doubled to 29 percent Siem Reap respectively, while the and the share of services remained at economic linkages between urban centers around one-third. However, despite its and rural areas have been limited. phenomenal growth and significant share Probably the most important link has been viii Cambodia - Poverty Assessment 2006

Figure 7: Yields of rice-Cambodia's food staple-are negative growth since 1999. very low compared to neighboring countries Given that the growth rate of labor supply in agriculture was Cambodia 2.7 percent a year, labor Thailand productivity in agriculture grew by a lackluster 0.5 percent per Malaysia annum. Lao PDR Agricultural growth in Cambodia China has not only been low relative to Indonesia growth in the industrial and service sectors, it has also been Vietnam unimpressive relative to its Yield (kg / ha.) in 2003 neighbors during comparable stages of development. Many Source: FAO, World Development Indicators. East Asian countries registered growth rates in agriculture of five percent per annum or more over a decade in the form of remittances. However, only or longer (Figure 8). Various measures of 13 percent of rural households received Cambodia's agricultural productivity, such remittances from family members working as crop yields, value-added per unit of in cities and the value of these remittances labor or profit per hectare, have been and represented on average less than 10 percent still are the lowest among East, Southeast, of total household consumption. While the and South Asian countries. The problems phenomenal growth of the garment sector of low and variable agricultural has benefited many rural households, these productivity and profitability are have not necessarily been poor households, exacerbated by increasing landlessness and and very rarely the poorest. Lack of declining availability and accessibility of education and, in many cases, the need to common property resources. pay informal recruitment fees to secure one of these jobs has meant that poor households in remote areas have Figure 8: Cambodia has yet to achieve the sustained not been the major beneficiaries. high rates of agricultural growth associated with long-term economic development ,..and the agricultural sector is large, but slow-growing Cambodia (1993-2003) While the share in the economy of Vietnam (1990-2000) the agricultural sector-inclusive Indonesia (1970-1 980) of crops, livestock, forestry and Lao P DR (1990-2000) fishery-fell from 46 to 31 percent Thailand (1965-1975) between 1993 and 2004, it Malaysia (1970-1978) continues to support more than 70 China (1980-1990) percent of the labor force. Agriculture grew by an average of 0 2 4 6 8 3.3 percent per annum over the last Average growth of agriculture mlue-added (% p.a.) decade, including three years of Source: World Development Indicators. Executive Summary ix

~ ~~ ~~ Landlessness is rising.. . Box 3: Landlessness amidst fenced-off The proportion of rural households lacking land in Phnom Sruoch land for cultivation has risen from 13 According to focus group participants in percent in 1997 to 16 percent in 1999 and Traeng Trayeung, Ou, and Mohasaing 20 percent in 2004. This rise in communes, landlessness in the area was between 30 and 40 percent. Available landlessness is relatively rapid given that arable land was estimated at about 3,540 distribution was more-or-less equal hectares. Of this, however, villagers (relative to household labor) when land cultivated less than 300 hectares while the was formally allocated to households in rest was unused but inaccessible or 1989. Although not all landless unavailable for rent. Rich families from Phnom Penh held large tracts of land for households are poor (some have profitable speculative purposes; this land is not non-agricultural sources of income), many rented out because of fear that lessees will are, as with few other rural livelihood claim ownership rights, since most of the opportunities have to rely upon low and land is held without proper land titles. variable income from wage labor. Source: EIC 2006 forthcoming. There are a number of forces underlying the rise in landlessness. A significant proportion of the landless (for example, ...and access to common property those who returned from refugee camps in resources is declining for the poor 1993-94 and newly formed households Access to common property resources such headed by young married couples) have as forests and fishing waters has always never owned land. There is a natural been a critical element in the normal dynamic away from equal distribution. livelihoods of the rural poor in Population growth leads to smaller plot Cambodia-and one that became more sizes in densely populated areas, and a important during lean times (e.g. poor proportion of households, as in any harvests), when intensified use of the economy, sell land, either as a deliberate commons provided a safety net for poor strategy to raise money to invest in non- households. Amongst the poor, a quarter agricultural activities, or less positively to depended on only fishery and forestry meet critical consumption needs (e.g. products for over half of their income in treatment for a serious illness). Finally, a 2004. On average, fishery and forestry small but politically important component products accounted for 25 per cent of of total landlessness is due to land household income amongst the poor grabbing. (Figure 9). By contrast, forestry and The problem is exacerbated by the inter- fishing accounted for just 8 percent of the connected problems of unclear property income of the wealthy. Own-harvested rights (as much as 80 percent of rural fishery and forestry products account, households that owned land were without respectively, for 20 and 10 percent of total land titles in 2004) and-this ambiguous consumption of the poor. legal status of ownership Numerous studies of rural livelihoods paint notwithstanding-the increasing value of a consistent picture of declining access to land in many parts of the country. The these resources, especially for the poor. result is an emerging phenomenon of Part of this decline is due to the increasing increasing landlessness juxtaposed with the privatization of once open-access existence of uncultivated land (Box 3). resources; part of it is due to a real decline X Cambodia - Poverty Assessment 2006

Figure 9: Common property resources-particularly fishing and forestry-are very important to the poorest quintile, but only marginally important for the richest quintile

Richest

Next richest

Middle

Next poorest

Poorest I I %of household income 1 H Cultivation 0 Livestock Fishing 0 Forestry 0 Non-agricltural OWage mother 1

Source: CSES 2004. in resource availability, as population low capacity, lack of transparency, and increase and commercial interests have inappropriate and arbitrarily-applied driven unsustainable over-exploitation. regulations. These phenomena contribute to a low level of official accountability and Bin ding constrain ts thus a low level of incentives to achieve results. They also create opportunities for Below-potential agricultural and rural corruption. This increases the costs and growth can be explained primarily in terms decreases the competitiveness of of poorly-defined property rights; businesses operating in Cambodia, inadequate and degraded infrastructure for water control and transport (Box 4), and a Box 4: "Up to the sky": without water combination oflow levels ofhuman capital control infrastructure, farmers and high levels of household-level are critically dependent on rainfall vulnerability (Box 6). Underpinning much of this is a more fundamental problem of In both high-growth areas such as Malai or low state capacity. Although improving, lagging Districts such as Mesang or Phnom Sruoch, lack of irrigation facilities is weak formal institutions do not yet provide cited as the main constraint for growth. a sound basis for the equitable or Farmers rely only on rainwater because sustainable management of natural the irrigation network is non-existent. In resources or pro-poor service delivery. Malai, yields of soy bean and rice fell dramatically because of lack of rainfall in The challenges facing the Government in 2005. In Phnom Sruoch, most have managing rural development serve as a suffered in recent years due to consecutive microcosm of broader national droughts. Farmers and officials in Mesang acknowledged that irrigation facilities are development challenges. In the economy costly but insisted that without them, the as a whole, investment that might consequences are even more costly. accelerate and diversify economic growth Source: EIC 2006, forthcoming. to the benefit of the poor is constrained by I Executive Summaw xi deterring investment and constraining the higher on irrigated land than on un- expansion of small enterprises. irrigated land for the same levels of other variables (e.g. land type and labor, capital Poorly defined property rights to land and fertilizer inputs). Land with access to are hurting income growth irrigation facilities during the dry season 111-defined property rights to private and has on average 15 percent greater rental public resources appear from the CSES value and 10 percent higher sale value than survey data to constitute a critical binding land without such access. Analysis of constraint on agricultural productivity and farm-level productivity also brings out the growth. Although in many cases there is importance of other rural infrastructure common local acceptance of ownership such as roads. Farms with access to and the renting and sale of land proceeds markets in their villages have 26 percent without the need for titles, increasingly, as higher crop yields than those without. land acquires greater commercial value Farms closer to roads also have and population pressure increases, lack of significantly higher yields; reducing the title is a problem. An “owner” who fears distance from a village to an all-weather- that his or her title might be challenged road by one kilometer raises productivity may be reluctant to invest in improving it by about 30,000 Riel per hectare. or planting higher-value permanent crops (e.g. fruit trees) for a long-term return. He Poor health and education constrain or she will also hesitate to rent out land rural productivity and accumulation which they themselves may not be able to The level of human capital in Cambodia is use (for fear that the tenant later claims particularly low. Illiteracy, morbidity and ownership), retarding the development of a malnutrition are prevalent. As expected, rental market which could ameliorate the the poor have rates of educational problem of landlessness. attainment and malnutrition that are worse The ubiquity and severity of the problems than the (bad) national averages. that arise from lack of legal title is still Qualitative research reveals that the poor subject to debate. However, econometric are very conscious of the value of analyses of data from the CSES 2004 education in improving the employment reveals that secure land tenure, in the form and income prospects of their children- of a formal land title certificate, and brings out how the imperative of short- significantly raises crop yields (by 65 term household survival prevents them percent), sale and rental value of land (by from providing their children with this 38 and 57 percent respectively), and schooling, so perpetuating poverty from household consumption (by 24 percent). one generation to the next (Box 5). Lack of infrastructure increases Analysis based on the 2004 CSES suggests production risk and marketing costs that education significantly enhances productivity, farm profitability and The proportion of arable land that is household per capita consumption in rural irrigated in Cambodia is, at 7 percent in Cambodia. An additional year of 2002, the lowest in all of South-east Asia, schooling for the household head enhances and less than half that of the next-lowest productivity by on average 40,000 riels per country (Lao PDR, with 19 percent). hectare. Analysis based on the CSES 2004 demonstrates that crop yield is 10 percent xii Cambodia - Poverty Assessment 2006

money; and, all too frequently, various Box 5: Education and the poverty trap forms of corruption. Because we are poor, our children quit school at an early age or after only one or two years in order to help us earn a living. Box 6: A health shock is often more Unfortunately, they cannot go as far as the serious than a harvest failure rich do in obtaining skills to earn a living. Results from surveys in two villages south of Because we are trapped in illiteracy, we have Phnom Penh suggest that while crop failures poor knowledge and are without ideas, and illnesses are both devastating remaining short-sighted and powerless. experiences, entailing similar magnitudes of Focus group discussion, Kompong Thom. economic damage (averaging a few hundred Moving out of Poverty Study (MOPS). thousand Riels), households find it harder to cope with illness. The negative consequences Source: CDRl2006a, forthcoming. from health shocks are more damaging to both immediate and long-run livelihood because they require an immediate lump-sum of money Illness implies a simultaneous reduction in for urgent treatment. Because most households do not have sufficient savings (and household labor supply and income and an rural credit markets do not operate well), increase in expenditures. In the context of households are often forced to resort to high cost, low quality health care, ill health distress sale of productive assets (including is a major factor in explaining the fall of land) and/or enter long-term debt, reducing non-poor households into poverty, and their future income streams and increasing their non-consumption expendit ures , poor households into destitution. A respectively. This broadly confirms the number of studies have highlighted the role findings of earlier Oxfam studies which found of ill health (and the costs of medical that half of all distress sales, or around 40 treatment) in causing or perpetuating percent of cases of once-landowning families poverty, particularly as a trigger for losing land, involved health crises. distress land sales (Box 6). Sources: Yagura2005; Biddulph 2004; Ballard and So 2004. Governance issues-low capacity and unresponsive and ineffective state institutions-underpin many of these Improving service delivery will ultimately problems. require fundamental reforms of public Strong macroeconomic performance over financial management (PFM) and public the last decade only partially disguises administration. The Government has some deep-seated problems with economic developed a four-stage program for PFM governance. While OECD quotas and reform. This is an extremely positive step, safeguards have boosted Cambodia’s the implementation of which is critically comparative advantage in garment exports important for long-term poverty reduction. and Angkor has attracted tourists, At one level, corruption is only a symptom businesses in other sectors struggle against of the weaknesses noted above. However, many obstacles. The most important of the scale of the problem in the eyes of these relate to costs arising from the citizens, businesses and donors is such that operation of state institutions and policies. it merits attention in its own right. These include unproductive regulatory Corruption keeps people poor; they have to requirements; insecure property rights; pay (or pay more) for goods and services. inconsistent application of the law; Benefits intended for them are diverted, untransparent public sector contract and they are unfairly disadvantaged and processes that often result in bad value for Executive Summaw xiii

dispossessed in conflicts with other groups Unlike many countries which have who can afford to pay more for official emerged from civil war, it has avoided decisions that favor their cause (Box 7). slipping back into conflict. However, there Services are under-funded as revenue are also a number of low-income countries collection is foregone; and investment is that have achieved significantly higher per deterred and employment suppressed. The annum rates of poverty reduction (Table cumulative effect of corruption in 2). These include a number of countries constraining economic development is that share Cambodia’s history of emerging immeasurable but undoubtedly very high. from conflict during the 1990s (e.g. Mozambique); or, indeed, which still Box 7: Corruption keeps the poor poor suffer from conflict in parts of their Students have to bribe the examiner during territory (e.g. Uganda). This is in one examination for looking at other students’ sense encouraging, as it suggests that, answer sheets or hidden documents. given moderately favorable external Policemen or military police are the circumstances and good government persons who are running the illegal gasoline trade. When we arrest thieves policies, it should be possible to not only that steal our fishing nets, and send them sustain but increase the rate of poverty to the police station ... only about two reduction in the future. minutes later the thieves are released because of money and power. The poor Projecting trends for the next decade are the victim and always wrong, even though they are right, because the poor do Figure 10 uses assumptions about the links not have money and power. between sectoral growth rates and the MOPS focus group discussion, Kampot. incomes of the poor to draw two Source: CDRl 2006a, forthcoming. alternative projections for poverty reduction over the next ten years. It seems that if Cambodia continues with the same Taking stock and looking overall rate (7.1 percent) and sectoral forward pattern of growth that was seen over the last decade, poverty will fall, but at a Assessing the past decade modest rate, and Cambodia will not Cambodia’s performance in reducing achieve its target of halving poverty by poverty is respectable, but not exceptional. 2015. If however the overall rate of GDP

Table 2: The rate of poverty reduction has been good-but a number of other countries have done better Country Period % under poverty line Average annual Start End in start in end rate of poverty year year year year reduction Nicaragua 1993 2001 50 56 0.5 Cambodia 1993/4 2004 47 35 I.I Lao PDR 1993 2003 46 33 1.3 Mozambique 1997 2003 69 54 2.5 Uganda 1992 2000 56 35 2.6 Vietnam 1993 2002 58 29 3.2 Tajikistan 1999 2003 75 57 4.5

Source: World Bank 2003; CSES 2004; Lao PDR 2003; Commission for Africa 2005; Poverty Task Force 2003; World Bank 2005. xiv Cambodia - Poverty Assessment 2006

Figure 10: Current growth rates will not halve poverty by 2015-unless more of that growth comes from agriculture

Baseline forecast: BvStats projection based on 1993-2004 growth (7% total, same sectoral corrposition) 35%

Btimted annual Best fit of rates 1993-2004 PovStats (interpolated using annual poverty PovStats, based on estimates Forecast at same annual growth rates) 1993-2004 (7% p.a.) total growth, but with 4% p.a. agricuitural + CMDG baseline (1993/4) and target (2015) growth

1990 1995 2000 2005 201 0 2015

Source: World Bank projections (backward and forward), using SESC 1993/94 and CSES 2004 data and PovStats. growth is kept constant but within this and education sectors under Ministry-led agriculture grows at 4 percent per annum, sector-wide strategies have begun to Cambodia will be likely to achieve and deliver results. surpass the poverty reduction target. The Government has now embarked on an ambitious but feasible program of Is the last decade a good guide to the public financial management (PFM) next? reform, tackling some of the binding The baseline projection in Figure 10 is constraints on improving service delivery. simply a linear forward projection of the Similarly, government commitment to best-fit line for the poverty trend in the last reform the institutions and procedures for decade. Obviously, changing parameters trade facilitation offers the potential for over the coming years will make it more or more diversified and inclusive growth. less likely that these projections, based on past growth performance and growth- Cambodian NGOs are becoming poverty linkages, will be realized and that increasingly capable and sophisticated as the 2015 development targets reached. contributors to national development, both as partners in service delivery and Assessing the underlying trends and as institutions for channeling ideas to potential change factors that might nudge government policy makers. poverty reduction above or below the straight-line, projections is complex. Donors are now beginning to tackle the Several trends appear to be in Cambodia’s problems of uncoordinated and donor- favor: driven approaches that characterized a significant proportion of the aid 9 In recent years sustained efforts to bring delivered over the last decade. together all the stakeholders in the health Executive Summary xv

Finally, the Government now has a new growth is enhanced. If however markets medium-term development plan-the are unable to respond to this increased National Strategic Development Plan for labor supply, youth unemployment and 2006-2010-which lays out an integrated, attendant poverty are likely to become strategic approach to national serious issues. development. A final factor that could also affect the At the same time, there are a number of pace of poverty reduction over the next ten factors that might threaten Cambodia’s years is if, as now seems increasingly ability to accelerate the 1-to-1.5 percent likely, large reserves of oil and natural gas per annum rate of poverty reduction. are confirmed to exist in Cambodian territorial waters. Managed well, this On their own, neither the garment nor resource has the potential to dramatically the tourism sectors will be enough to improve the level of state spending on absorb the more than 200,000 young economic and social development. people who are going to be entering the However, international experience labor market every year over the next suggests that in countries with low state five years as a result of the second capacity and relatively weak and baby boom in the early 1990s. unaccountable public institutions, such Many non-poor households remain petrochemical wealth may equally well highly vulnerable with levels of per result in a “resource curse” that actually capita consumption only a short retards development and poverty reduction. distance above the poverty line. As If the reserves are confirmed, carefully such, they remain exposed to a wide designed institutional mechanisms that range of household-, community- and promote transparency and accountability national-level shocks which could will be critical if they are to help rather easily push many back into poverty than hinder Cambodia’s escape from the within a relatively short period. ranks of poor nations. Declining access to common property resources poses a particular source of The National Strategic Development 2006-2010 vulnerability and a serious threat to the Plan, possibilities of the poor to work their In recent years international donors, way out of poverty (or to make it learning from the successes of countries through difficult times without falling such as Uganda, have advocated that low- more deeply into poverty). income countries formulate a comprehensive medium-term strategy to Two final factors could also affect the pace identify priority policies for poverty of poverty reduction over the next ten reduction. Over the course of 2005, the years. Firstly, Cambodia is going through Government has prepared just such a a demographic transition. Large numbers strategy, called the National Strategic born fifteen to twenty-five years ago are Development Plan (NSDP) 2006-2010. passing through the labor market at the This provides a framework that will be same time as the fertility rate is falling, refined, operationalized and reviewed resulting in lower dependency ratios. If through annual consultations, and which this generation has access to productive will direct public spending (both own employment and business opportunities, revenue and international aid) towards the potential for accelerated economic priority goals defined in terms of xvi Cambodia - Povertv Assessment 2006

Cambodia’s Millennium Development ~ ~~ Goals (CMDGs). Box 8 summarizes the Box 8: The National Strategic Development key policy commitments in the NSDP. Plan (NSDP), 2006-2010: key commitments Policy priorities o Anti-corruption measures; legal and judicial To sustain and accelerate the trend in reforms; administrative reform; poverty reduction seen between 1994 and decentralisation and deconcentration; and 2004 requires a combination of actions. In military reform. order to build on this positive trend, the 2. Environment for the implementation of first priority is to maintain a high rate of the Rectangular Strategy o Ensure peace, political stability and social economic growth and increase the share of order through elections; strengthen this growth that goes to poor households. development partnerships; sustain a favorable macroeconomic and financial Essentially, this requires removing the environment; further promote economic constraints that currently hold back the integration into the region and the world; productivity and profitability of address poverty, ensuring that all smallholder agriculture, while strategies focus on poverty reduction. simultaneously investing in sound economic governance, physical a infrastructure and skills in order to sustain comprehensive Agriculture and Water Resources Strategy; improve agricultural and diversify job and opportunity-creating productivity and diversification; reform land growth in the industrial and service administration and management, fisheries sectors. and forestry reform; invest in environmental conservation and rural A third engine of growth is needed To increase the share of growth that accrues to the poor, Cambodia has to maintain the momentum ofits two existing infrastructure; improve management of water resources and irrigation; develop engines of growth and develop a third, energy and power grids; manage future oil more rural, more pro-poor engine of and gas resources and revenues; develop growth. International experience information and comm gY reinforces the obvious conclusion that growth in the agricultural sector will have o Carry out the RGC’s Twelve Point Plan the greatest poverty reduction impact, both and the recommendations in the direct and indirect, over the short-, medium Investment Climate Survey; promote SMEs, trade and tourism, rural credit; and long-term. Steps to facilitate create jobs and ensure improved working improvement in agricultural productivity conditions; establish social safety nets for and growth are prerequisites for poverty the disadvantaged. reduction.

o Enhance the quality of education; improve Agriculture health outcomes through the Health Action Analysis of CSES 2004 data provides Plan; foster gender equity; implement strong proof in support of earlier, small- population policy to decrease fertility and promote birth spacing. scale studies that have found small plots to be more economically efficient; that is, small farms have higher productivity (yield Executive Summarv xvii

Figure 11 : Small farms are a better smallholder rather than plantation model of economic proposition than ownership and production. The second is large farms to reinforce the case for systematic land titling, with first priority given to areas in A. Small farms have higher outputs which smaller, poorer farmers predominate. per hectare.. . (y axis plots log of output value / ha.) The Government’s provision of *D reasonably-sized plots of land (with essential complementary services) to landless households under the “social land concession” program has the potential to improve food security, productivity and incomes for poor rural households. If small family holdings are, as the CSES and sub-national surveys agree, more productive than large holdings, the case for land redistribution is not merely “social” (improving the distribution of growth) but clearly “economic” (in terms of higher D rates of aggregate output expansion).

Farm size (hectares, log scale) The Government will also need to tackle production-side constraints. The factors B. ...and higher crop income per hectare underlying low productivity and low (y axis plots crop income / ha.) profitability include poor infrastructure and low stocks of human capital. The problem of inadequate economic infrastructure is 1’ particularly acute in remote rural areas, I!I where the rates of poverty are highest in large part because access to infrastructure and basic services is limited. The poor in these areas have to travel significantly longer distances to get to markets, roads, schools, and healthcare facilities. While the benefits of rural infrastructure are evident, analyses of the costs of infrastructure have yet to be carried out with sufficient rigor. In summary, the main channels for the Farm size (hectares, log scale) government to tackle low agricultural productivity and profitability are: (i) per hectare) and profitability (income per accelerated, transparent, decentralized and hectare) than do large holdings (Figure 11). low-cost procedures for land titling; (ii) There are two main policy implications to clearer and more equitable rules for access emerge from this finding. The first is that to common property resources; and (iii)an agricultural policies should incorporate a elimination of excessive and inappropriate strong general presumption in favor of a regulation, and the corruption associated xviii Cambodia - Poverty Assessment 2006

with it, that characterizes state has played a role in a number of successes intervention in private sector activity. in Cambodia. However, much has also been badly designed or implemented, Human development resulting in inefficiency and contributing In recent years, allocations of public to the governance problems of the expenditures to health and education have Cambodian state. risen dramatically, helping to implement Meeting the poverty reduction targets laid and further refine significant reforms and out in the CMDGs thus requires improving capacity improvements in these sectors. both national capacity and governance and These investments in human capital should aid effectiveness. This will require help to expand the opportunities available progress on the three fronts of ownership, to the next generation. alignment and harmonization. Policies to improve capabilities and Ownership implies that aid is directed to conserve and replenish human capital activities determined by the Government, . . include: (i)consolidating the gains made in rather than driven by the donors. By increasing primary enrolment, while outlining policies and spending priorities addressing the bottleneck in upper primary for the next five years, the NSDP creates a schooling and improving the quality of framework to which donors can and should teaching; (ii)designing and implementing align their aid programs. Donors now need universal adult literacy programs to to support the NSDP by harmonizing their improve the skills and productivity of the aid (e.g. cooperating amongst themselves current economically active generation; to standardize management systems and (iii)improving the quality and reducing the share analysis) and aligning it to cost of public health services to reduce the government priorities and systems (for direct and indirect effects on the poor of example, through the Public Investment ill-health and out-of-pocket health Program or sector-wide approaches). spending; and (iv) developing While progress towards harmonization and multisectoral programs to tackle persistent alignment has been made in recent years, public health problems such as severe much more is needed. malnutrition and maternal mortality, as Finally, the Government, donors and civil well as acting early to promote behavior society need to establish effective change to deal with emerging health partnerships for monitoring and evaluating problems such as smoking. progress in the implementation and More effective partnerships for poverty outcomes of the NSDP. Once again, reduction progress has been made, and the Statistical Master Plan prepared by the National Achieving poverty reduction will require Institute of Statistics (NIS) provides the further improvements in the relevance and outline of both the specific exercises and effectiveness of government policies and systems for the collection and use of socio- spending. Given the importance of economic data, and the capacity-building external development assistance in improvements required for sustaining this Cambodia, the quality of government- work. Further refining and then financing donor relationships and aid effectiveness is the monitoring and evaluation framework of critical importance to improvements in outlined in the NSDP will be essential for the pro-poor allocation and efficiency of joint monitoring of progress in poverty total spending. Over the last decade, aid reduction. I,Country context Summary In terms of geographical endowments, Cambodia enjoys relatively favorable conditions for development. Soils are not particularly good and the hydrological regime is complex, but Cambodia has a deep-water port, plenty of flat cultivable terrain, productive freshwater fisheries, and a considerable stock of timber. It also has great potential for tourism. Most importantly, it is located in a fast-growing region with land borders that allow it access to markets in Thailand and Vietnam. Cambodia’s contemporary poverty is thus due primarily to almost three decades of conflict, Civil war began in 1970 and ended in victory for the Khmer Rouge, who between 1975 and 1979 killed perhaps one in five Cambodians. Throughout the 1980s Cambodia recovered slowly in a context of “low intensity” conflict and international isolation. Conflict has destroyed infrastructure, decimated the nation’s human capital, and weakened or distorted social, economic and political institutions. From the late 1980s, Cambodia embarked on a transition from war to peace, from one- party to multi-party politics, and from an isolated, subsistence-oriented economy to one based on market principles and integrated into international trade and capital flows. It is now fourteen years since this transition was formalized in the 1991 Paris Peace Accords: Cambodia today is thus in a transition from a post-conflict society to a “normal” developing country. The effects of conflict are still seen, particularly in institutions that are oriented primarily to sustaining the state and less to responding to citizens’ demands; but, after more than a decade of international integration, investment and growth, the current generation of leaders and citizens have started to move beyond the legacy of war to define new policies and livelihoods oftheir own. Although Cambodia is unusual in this post-conflict (and “post-post-conflict”) experience, it is not unique; the problems it faces, and the possible solutions it might adopt, can be compared to those in other low-income, post-conflict states. In the first decade since the Paris Peace Accords, the Government has achieved success in consolidating peace and promoting economic growth. People recognize that peace has made an immeasurable difference to their lives, improving physical security and opening up far greater livelihood opportunities. The Government has actively pursued integration into regional and global structures such as ASEAN and WTO. Since the late 1990s it has also judiciously managed the opportunity presented by the Multifibre Arrangemeni (MFA). It is important to see progress in this first decade as in large part due to a historically- unusual ‘peace dividend’: peace and economic opening have made possible rapid growtt and relatively rapid poverty reduction from a very low base. Policy choices on trade and infrastructure have expanded on this window of opportunity, but primarily in terms oi aggregate economic growth, with less attention to the distribution of that growth. As Cambodia enters its second post-conflict decade, it faces new challenges and rising expectations. A new, more complex and equally fundamental set of institutional anc policy changes will be called for. 2 Cambodia - Poverty Assessment 2006

he scale and nature of poverty in The topography of the country also puts it contemporary Cambodia, and the an advantage compared to many processes by which poverty mountainous states (e.g. Lao PDR, Nepal, T Bhutan, or much of Vietnam). Highlands increases or falls over time, cannot adequately be understood in isolation of on the northern, north-eastern and south- the country’s recent history. This first western borders have low potential chapter attempts to provide the background productivity, but the majority of Cambodia within which Cambodia’s current poverty is relatively flat, so that farming and status, recent trends, and future prospects investments in communications can be set in context and assessed. infrastructure do not require expensive adaptations to the landscape: land can be Geography and opportunity worked and roads built without needing to overcome slope. The Tonle Sap lake in the In terms of fundamental geographical middle of the country gives a large endowments, Cambodia has much going proportion of Cambodian villages for it. Unlike many least-developed relatively good access to water transport nations (for example, Lao PDR, Nepal or down to Phnom Penh and eventually Uganda), Cambodia is not landlocked, or southern Vietnam. The lake also supports isolated from its neighbors by terrain: the a unique ecosystem which has-at least to deepwater port at Sihanoukville, Phnom date-provided one ofthe most productive Penh’s position on the Mekong river, and freshwater fisheries in the world, and a the long lowland borders it shares with at critical source of protein for Cambodian least two of its three neighbors means that diets. there are no major geographical barriers to Any list of the inherited factors that international trade (Figure 1.1)’. Indeed, Cambodia has in its favor must include the the land borders between Cambodia and archaeological legacy of the temples at the rapidly-growing economies ofThailand Angkor and the massive potential for and Vietnam creates considerable potential tourism that this cultural capital represents for demand for Cambodian goods and (Box 1.1). Tourism, driven primarily by services. Looking beyond its immediate Angkor, has been a major contributor to neighbors, Cambodia is located in one of economic growth in Cambodia over the the fastest-growing regions in the world. last decade. However, the impact of Although the Asian financial crisis of the tourism on poverty has yet to reach its late 1990s put a dent in growth rates, potential: it is notable that the province in Cambodia certainly benefits from a which Angkor is located, Siem Reap, was “neighborhood effect”, with the potential in 2004 the third poorest in the country. to attract investment from and sell to Obtaining sustained economic benefits markets in the region. from Angkor will require a more strategic approach to development of the tourism ’ Topographically, Cambodia can be described as a sector. tilted saucer. Forested highlands, especially to the north-east and south-west, drain inward to Finally, it now seems increasingly likely wide plains and a large, shallow central lake (the that within a relatively few years Tonle Sap) and then, following the “tilt”, to the Cambodia will enjoy access to a significant south-east as the Tonle Sap River, which joins revenue stream from offshore oil and gas the Mekong River and flows on through the south-eastem plains and on into Vietnam’s reserves in Cambodian territorial waters. Mekong delta. Provisional estimates from the explored Country context 3

Figure 1.I : Cambodia benefits from access to the sea, rich freshwater fisheries, and economic growth in neighboring countries

0 50 100 THAILAND

Provincial boundary I* lntemationalboundaty

portions of one block suggest that this area development and significantly improve the contains 400-500 million barrels of oil and prospects for a rapid fall in povertf. 2-3 trillion cubic feet of gas. Further This is not to say that Cambodia suffers no exploration in this and other blocks is geographical disadvantages. Cambodia’s likely to raise the estimates. Although it is soils are (with some notable exceptions) of impossible to forecast revenue flows to any generally rather low quality. Cambodia degree of precision given both the also has a complex hydrology, in which in uncertainty about the size of the reserves the wet season the build up of water at the and about the future price of oil, it is quite confluence of the Mekong and Tonle Sap possible that full-scale commercial rivers results in the Tonle Sap reversing its production could bring in revenues of $2 flow back into the lake, which expands to billion per annum (Vallely, Dapice, and ten times its dry season area and twice its Wilkinson 2006). Managed well, this dry season depth in widespread flooding. could provide an immense boost to Cambodia’s economic and social There are however fears that, as in many other countries (Nigeria being the most obvious and extreme example), the discovery of oil and gas wealth might actually retard Cambodia’s development if it is not managed well. Chapter 7 addresses this concern. 4 Cambodia - Poverty Assessment 2006

Box 1.1: Angkor, tourism and growth telling that in 1969, Cambodia had a higher The roots of Khmer civilization lie in the per capita income and received more "lndianisation" of the people who lived on tourists than Thailand (Financial Times, the Mekong and Tonle Sap floodplains. From the first to the eighth centuries CE 1995; Dixon 1995). The reversal of this village-based society was drawn fortunes since then is clearly due to the under centralized authority, flourishing onset, in the late 1960~~of what was to into the Angkorian empire. Khmer power turn into almost three decades of civil extended over much of mainland south- conflict. east Asia between the gth and 13th centuries, during which time a collection of superb Hindu and Buddhist temples The legacy of history and funerary monuments were built, primarily at Angkor in north-east Post-colonial Cambodia: war, Cambodia. Drawing on Indian traditions, Angkorian art and architecture is revolution and recovery nonetheless distinctively Khmer, with Cambodia gained independence from the several different styles apparent over French in 1954 under King Sihanouk, who different eras. abdicated in 1955 to campaign in elections. With the return of peace, there has been Many older Cambodians recall this period a steady rise in tourists visiting Angkor as a golden age. By the mid-1960s, and staying in the capital town of the province (Siem Reap) in which it is however, growing economic problems, located. This flow has now hit over 1 corruption, and lack of Government million arrivals per annum. Although tolerance for democratic opposition led a Cambodia receives considerable frustrated urban educated class to consider benefits from this tourism as some jobs radical political alternatives (either left- or have been created for the poor (particularly in hotel construction) and right-wing). The Cambodian Government demand for some goods has been also found it hard to keep Cambodia increased, the links from tourism insulated from the war in Vietnam and backward or forward into other sectors of from pressures from both sides in that the national economy (e.g. through conflict to support their cause. tourist-driven demand for food production) are currently well below Eventually, the balancing act failed. In potential, and the benefits specifically to 1970, Sihanouk was overthrown in a coup the poor could be considerably greater than at present. led by General Lon Nol. Cambodia rapidly descended into a bloody full-blown civil war. In exile, Sihanouk formed a "United Front" with the communists that had become known by the disparaging While these extensive floods are part ofthe label of "les Khmer Rouge'' (KR), who normal annual cycle and are critical to rice began to attract broad popular support for cultivation and the Tonle Sap ecology, in the first time. the absence of water control infrastructure an unusually high flood can cause Within the united front, however, the KR widespread damage to crops, houses and consolidated power, purging Sihanoukists livelihoods. and moderates. With Khmer Rouge victory in April 1975, a clique of On balance, however, there is no extremists driven by rabid nationalism and immutable aspect of location, topography commitment to an uncompromising or climate that holds Cambodia back from agrarian communism came to control the achieving its development goals. It is County context 5 revolutionary state of Democratic poverty reduction (it was cut off from non- Kampuchea (DK). Towns were emptied COMECON trade and from most of the and religion and money abolished. international aid system), it did manage to Although the severity of the revolution stabilize the famine situation prevailing in varied between regions, life everywhere 1979 and made slow but steady gains in was characterized by unremitting hard reconstructing schooling and basic labor, hunger, and the threat of execution3. healthcare. The population was organized It is estimated that between 1975 and 1979 into “solidarity groups” (born samakz). around two million Cambodians-more This low-level collectivization ensured than one in five-died in this “auto- equal access to the scarce male labor and genocide”. Most died of starvation, draft animals available in 1979-80, but as exhaustion and disease, but perhaps the economy recovered it came to be seen 200,000 were systematically executed. as a hindrance to growth. Spontaneous de- Officers and officials of the Lon No1 collectivization, formalized through regime were killed in 1975; in the later liberalization and land distribution in 1989, years many of those executed were improved output and was welcomed by themselves KR, killed as the revolution most, although vulnerable groups lost turned on itself. important forms of social security (Frings 1993). The dominant, Pol Pot faction was driven by, among other things, a desire to reclaim Cambodia’s threefold transition: the Mekong delta, lost to Vietnam since the 17th century. Murderous raids on peace, democracy and market southern Vietnam killed thousands of economics civilians and eventually prompted a In 1989 Vietnam withdrew its troops. The Vietnamese invasion in December 1978. conflict dragged to a military stalemate and Democratic Kampuchea was replaced by was eventually resolved in a political the People’s Republic of Kampuchea solution in the Paris Peace Accords of (PRK). Composed largely of ex-KR who October 1991. These made provisions for a had fled the purges, the PRK ruled UN peace-keeping operation of throughout the 1980s. As outside powers unprecedented cost and scope to supervise once again took sides, Cambodia reverted the transition from one-party state and civil to a second, more low-intensity albeit still war to a peacehl multi-party democracy. massively debilitating civil war, between The United Nations Transitional Authority the PRK government on the one hand and in Cambodia (UNTAC) was assigned to an alliance of anti-PRK guerilla groups supervise disarmament and operating out of refugee camps on the Thai demobilization; organize repatriation and border on the other. This conflict reintegration of the 350,000 refugees on continued throughout the 1980s. the Thai border; coordinate international While the Government was unable to reconstruction aid; and ensure conditions achieve much in terms of growth or for free and fair elections among the civil- war-factions-turned-political-parties. The international financial institutions resumed For different perspectives, see Burgler 1990; relations with Cambodia and the number of Chandler 1991; Vickery 1984; Kieman 1996. On regional variations, see Kiernan 1996; Vickery 1984; Bamett 1983. 6 Cambodia - Poverty Assessment 2006

aid agencies and international NGOs remain somewhat underdeveloped. The working in the country mushroomed4. deeply damaging failure ofpolitical parties to agree on the formation of a new Cambodia has since 1993 enjoyed over a coalition Government for a year after the decade of high average economic growth, 2003 elections suggests that economic and driven largely by construction, tourism social development may continue to suffer and, since the late 1990s, a very rapid periodic setbacks resulting as a by-product emergence of a garment sector which benefited from preferential arrangements of political rivalries. On the positive side, Commune/Sangkat elections in 2002 under the Multifibre Arrangement (MFA) created elected local governments and quotas agreed for clothing produced nationwide for the first time, with the under ILO-accredited labor standards promise of gradual gains in participation (Chapter 4). However, political instability and accountability (Hasselskog et a1 2000; persisted beyond the introduction of multi- and Henke 2005)5. Overall, party democracy, with tensions between Ninh however, the Royal Government continues the two parties in the first coalition to be dogged by low levels of capacity, Government eventually culminating in resources and accountability. It remains armed conflict in 1997. This return to too often unresponsive or ineffective in political violence, coinciding with the meeting its impact of the Asian financial crisis in the needs of citizens, and particularly that third of its citizens who Thailand and the withdrawal of the IMF in live below the poverty line (a theme reaction to continued weaknesses in the explored in Chapter 7). management of forestry revenues, created a short-lived period of economic In summary, Cambodia has been contraction in the last years ofthe 1990s. undergoing a threefold transition since the Economic recovery since then has however early 1990s (Hughes 2003). Firstly, it has been relatively rapid. Progress was moved from civil war to peace. Secondly, it has undergone a fundamental transition reinforced in 1999 by the final collapse of from one-party to multi-party politics. the Khmer Rouge insurgency. Although Finally, it has moved from an isolated, the agricultural sector has continued to lag, with severe floods or droughts in three of low-growth , state-managed and subsistence-oriented economy in which the last four years exacerbating a generally everyone was largely equal in poverty, to a low rate of improvement in productivity or market-based economy that is open to incomes (Chapter 5), there has been an expansion of off-farm opportunities in international flows of capital, goods and Phnom Penh or across the border in labor, in which growth has lifted a substantial number over the poverty line, Thailand, prompting the emergence of but in which inequalities are growing very migrant labor and remittances as important aspects ofrural livelihoods and rural socio- rapidly and a significant number are not sharing in the benefits of growth and some economic change. are suffering a decline in their standard of Political institutions and traditions of living. This transition is more fully peaceful political competition however completed in some respects than others,

For alternative assessments of the impact of Cambodia’s administrative structure is comprised UNTAC on Cambodia, see UN 1995; Utting (ed.) of 23 Provinces and Municipalities, below which 1994; Findlay 1995; Metzl 1995; and Frieson are Districts, Communes (or their urban 1996. equivalent, the Sangkats), and villages. Country context 7 and has given rise to a far-reaching but still ensuing DK period; ifthey survived, they imperfectly understood set of social and were perhaps also disproportionately likely economic changes. afterwards to be able to make their way into exile and establish new lives abroad. Defining “post-conflict” As a result Cambodia lost many ofits most This protracted period of political violence able individuals, and has had to reconstruct has had a profound negative impact on a Cambodian state and economy with a Cambodia’s development which continues very small number of doctors, teachers, to be felt to this day. Three decades of economists, agronomists or engineers. It internecine conflict, exacerbated by Cold has been hard to replace this lost War rivalries, not only claimed millions of generation. For the slightly less than four lives but also undermined many of the years of Khmer Rouge rule, school-based foundations for national development. education was completely dismantled, and while the PRK subsequently invested Destruction of infrastructure considerable effort in rebuilding basic education, it had very limited resources The legacy of conflict is seen most with which to do so. obviously in the depleted physical and human capital of the nation. Investments Finally, many of those who survived the in infrastructure-roads; bridges, water 1970s carry a legacy of health problems- control, hospitals, schools-were physical and/or psychological-which can destroyed and have had to be rebuilt. never be adequately addressed, and which Inadequate transport infrastructure imposes limits their ability to earn a living. While costs and delays on travel, raising the costs war and landmine injuries make up a of marketing goods or obtaining inputs, progressively smaller part of the total and limiting access to facility-based health health burden as time goes by, they are and education services. Water control often closely related to extreme poverty. structures are underdeveloped, leaving a Severe malnutrition under DK also great many people exposed to flooding or continues to have an effect, as mothers drought (or, frequently, both in the same whose growth was stunted as a result of year). And although a significant number inadequate consumption of food are more ofschools and clinics have been built since likely to give birth to low birth-weight 1993, there are still too few classrooms and babies, perpetuating disadvantage from hospital beds for the needs of the one generation to the next. population. Weakened social capital and Depleted human capital unresponsive state institutions Cambodia did not only lose the physical As important as the impact on physical and hardware for economic activity and service human capital, the endemic violence ofthe delivery: it also lost the human capital 1970s and the more constrained but necessary for development-that is, the nonetheless severe violence of the 1980s educated, healthy individuals who are best deeply affected the nature of Cambodian able to add value in business and provide social and economic arrangements, the professional skills in public service. Those institutions of the state, and the framing of who in 1975 were the best educated were relationships between the state and the disproportionately likely to die during the society and economy it governs. Many aspects of this institutional legacy of 8 Cambodia - Povertv Assessment 2006 conflict have been deeply detrimental to problems in a simple or direct way to this the prospects for long-term development. history of conflict. Cambodians, both leaders and ordinary citizens, may have Although impossible to measure, it would inherited a post-conflict society, but ten be remarkable if exposure to extremes of years on they have made choices and taken violence, deprivation, surveillance and DK actions that have collectively started to efforts to destroy all existing and shape their own lives and their own potentially alternative identities- trajectories of social and economic change. including the family-did not have a Some 60 percent of the population is aged profound impact on the ability of survivors 24 years or younger, born after the fall of to trust, plan ahead for the long-term, and the Khmer Rouge period, and have no act collectively. The institutional memory of the DK period. Fully 50 underpinning for an efficient modern percent were born after 1985, and so have market-based economy-ease of market little if any memory of the PRK era or the entry, parity in access to information on national struggles within which the current supply and demand, transparent and stable political parties have their origins6. property rights, equality of treatment for all economic actors under commercial law The influence ofpreceding decades of civil and regulation-are yet to be fully war thus continues, but this influence is established. increasingly subtle, and seen mainly in the nature of social and political institutions This in turn reflects the fact that after the and the relationships between them. As fall of DK a “normal” government had to the era of conflict fades gradually into be reconstructed in the context of incipient collective memory, Cambodian citizens famine, massive destruction, a lack of any acknowledge and appreciate the difference resource base, and civil war. The state that that peace has made to their lives (Box resulted was inevitably concerned 1.2). primarily with survival and less with accountability or inclusion. To a certain However, they are also increasingly extent, this legacy continues to shape state looking forward, concerned with other structures and the terms of the relationship issues to do with the economic between citizens and government officials opportunities that they and their children to this day. These issues are explored in face, the quality of services they receive more depth in Chapter 7. from the state, and the perception that inequalities are widening and becoming From peace to poverty reduction more entrenched. The legacy of conflict is thus undoubtedly still present throughout all aspects of contemporary Cambodian life. However, six years after the final remnants of the Khmer Rouge were brought into the fold- and twelve years since the political settlement which brought all the other parties in the civil war into a system of democratic rather than armed competition-it becomes progressively harder to attribute all Cambodia’s Revised population data from Neupert 2005. Country context 9

Box 1.2: Peace has opened up opportunities and created the possibility of a better life In the MOPS research project (Box 1.5), participants consistently emphasized how the end to armed political conflict lifted the burden of fear and opened up livelihood opportunities. This theme emerged in all villages and in discussions with local leaders, those who had prospered and those who got poorer, and in groups of young men and young women alike. However, the importance of peace emerged most strongly in villages in which the conflict ended more recently (in some, the Khmer Rouge were still active until the late 1990s); and in some villages amongst young men in particular (perhaps reflecting the fact that throughout the 1980s young men were conscripted to fight for the PRK).

Before Pol Pot troops defected to the Government, most villagers dared not sleep in their houses; most of the time they slept on the ground close to a safety hole. Whenever Khmer Rouge or government troops came to the village, they took our chickens and ducks or burned our houses if we were not willing to give our belongings to them. In 1993, we could not move freely because we had to ask for a permit from local authorities. By contrast, in 1998 we got more freedom, because we could go everywhere and our community was secure after the integration of the Khmer Rouge. Dang Kda village - Kompo Before the Khmer Rouge integration in 1998, this community was unsafe because this area was a stronghold of the Khmer Rouge. At that time, the situation was not stable. People had to be well prepared and ready to escape from the bombing. They felt unsafe, so never cared about building or improving their houses ... That‘s why security always came first to our mind and we had less time to think about our economic life. At that time I cared only about my pots and dishes rather than other useless things when each time serious fighting took place in this community. I never thought about earning money and how to improve my family’s standard of living because I had no time to think about it. The only one thing I had to do was to learn how to escape from bombing. Kanchor villag This village was free from armed conflict after the defection of all Khmer Rouge troops in 1998. Having security significantly helps improve people’s livelihoods in this area. Source: CDRl 2006 (forthcoming).

As the Royal Government seeks to move outcomes. Citizens want to see progress in the perception and reality of Cambodia living standards and in the quality of from that of a post-conflict country to that public services they receive from the of a “normal” developing country, its Government. Cambodians can and do citizens too are making this transition. As have reasonable expectations of progress, part of this transition, their expectations and do feel that they can hold their elected are rising. These expectations apply to leaders to account, regardless of the fact issues of process (there is increasing that Cambodia is still dealing with a post- awareness of, and importance attached to conflict legacy. These same higher concepts of individual rights, the rule of standards will also be increasingly law, and the accountability of the expected of other powerful actors affecting Government and donors to use resources their lives-including donors (Chapter 8). honestly and in the public interest); and to 10 Cambodia - Poverty Assessment 2006

Assessing poverty reduction in post- society in setting and/or enforcing pro- conflict contexts poor policy commitments. Cambodia is of course unusual but not Managing peace and development: unique in this post-conflict experience. economic and social change in post- While the significance of decades of war conflict Cambodia and revolution cannot be under-estimated, international experience suggests positive From 1993, the Government has actively examples of other countries that have only managed the re-insertion of Cambodia into recently emerged from similar decades- regional and global relationships (for long civil conflict (e.g. Mozambique) or example, through pursuit of ASEAN and which even now continue to suffer from then WTO membership) following two ongoing conflict in a significant part of decades of isolation. This is having their territory (e.g. Uganda), but which complex effects on the Cambodian have demonstrated considerable success in economy and Cambodian society, which allocating available resources (both the Royal Government is seeking to domestic revenue and aid) to achieve understand and manage in order to direct notable successes in growth and human the path of national recovery and development. These cases are summarized development. in Box 1.3. Attempting to distil lessons In terms of economic change, international from these cases-which it should be investment and trade is creating new noted, involves very different political opportunities for employment or earnings, systems-a number of common themes either within Cambodia or as migrant seem to emerge: laborers in neighboring Thailand. At the The importance of sustained same time, the shift to a market-based commitment to poverty reduction on economy has imparted a commercial value behalf ofpolitical leaders. to land and natural resources (notably forestry and fisheries) that they did not . The translation of this commitment have before. Without transparent or into effective public actions through accountable institutions to manage these the integration of planning and public resources, there is increasing accumulation financial management processes: in the hands of the wealthy. This o within the framework of a broad- concentration of resources is often at the based, nationally-owned, poverty- direct expense of local poor communities focused medium-term development who rely upon open access to forests and strategy; waters to meet their own household needs or to earn an income. Furthermore, the o designed on the basis of policy- new owners too often acquire these newly- relevant socio-economic analysis of the privatized resources at significant discount causes ofpoverty and poverty reduction; (meaning that the Royal Government does o devised and kept on track through not receive the revenue it should expect to mechanisms (broad consultations with receive, and which it needs to fund NGOs in Mozambique, local investments in development); and manage accountability mechanisms and them with a focus on generating a short- structured participatory poverty term, one-off gain, rather than a assessments in Uganda) that were sustainable and more widely-spread flow deliberately designed to involve civil ofbenefits over the long term. Country context 11

Box 1.3: Poverty reduction in post-conflict societies: lessons from international I experience

1 Uganda Following independence in 1962, Uganda has suffered long periods of impoverishing instability and civil war. In 1986, Yoweri Museveni’s National Resistance Army took power and brought peace to most of Uganda, although insurgencies continue to the present. The ‘No-Party’ Movement system strives to be all-inclusive and consensus-based, severely limiting the activities of political parties that Museveni claims were associated with the escalation of conflicts in the past. Within this political system, Museveni and the Movement have pursued poverty reduction as a political project to reduce destabilizing inequalities and promote national unity. Poverty reduction as a political goal has in turn become a national technocratic project within the Ministry of Finance, Planning and Economic Development (MFPED), which has translated the political vision into instruments that are credible in the eyes of key outside stakeholders. The centerpiece of this strategic framework has been the Poverty Eradication Action Plan (PEAP), supported by a medium-term expenditure framework (MTEF), sector strategies and working groups; the Poverty Action Fund, which protects poverty-focused priority expenditures from budget shortfalls; and the Participatory Poverty Assessment Project, which has provided crucial grassroots feedback on the effectiveness of policies, enabling problems to be picked up and addressed. Given its weakness in 1986, MFPED has developed notable capacity, forging alliances with others to reach its objectives. A recent study of public expenditure and poverty reduction concluded ‘We are aware of no other country that has achieved such a dramatic pro-poor change in spending patterns in such a short period’. Real GDP growth has driven poverty down from 56 percent in 1992 to 44 percent in 1997 and 35 percent in 2000. The success of the PEAP was the main source of inspiration for the Poverty Reduction Strategy Paper (PRSP) approach advocated by the IFls since 1999 as the centerpiece of effective aid partnerships.

Mozambique Mozambique’s long-running civil war, fanned by outside powers, began in 1975 and killed almost a million people before peace was achieved in 1992 and the first multiparty elections were held in 1994. Since then, strong economic growth (12 percent p.a. over the 1990s) was sustained in large part by a serious effort, backed by donors, to improve governance and policy. The Government resolved in particular to improve the business climate, resulting in a doubling of private investment as a share of GDP between 1998 and 2002. From 1996, determined efforts were made to improve revenue collection by tackling customs fraud and evasion and implementing trade facilitation measures. Goods are now cleared 40 times faster than before the reforms. In the first two years of customs reform, imports decreased by 0.2 percent but customs revenue increased by 38.4 percent. Increased revenue was translated into higher and more efficient social sector spending. High levels of aid under competent government ownership also appear to have played an important role: Mozambique’s poverty-focused medium-term development strategy, based on policy-oriented research and broad participation of civil society, pre-dated the introduction of the PRSP approach at the global level. Donors were persuaded of the strategic vision of Mozambique’s leaders and the sincerity of their commitment to reforms and so agreed to align the bulk of their aid though sector-wide approaches. Critically, high growth rates were sustained even when the aid ratio fell sharply, suggesting that aid had been used judiciously to build a foundation for sustained economic expansion. The combination of growth and improved service delivery has resulted in broad gains in living standards. Between 1994 and 2002, under-five mortality declined from 277 to 135 deaths per 1000 live births; between 1999 and 2003, the number of students in primary and secondary education increased by 43 percent. All of these improvements were reflected in the poverty rate, which fell 14 percent over the course of six years from 69 percent in 1997 to 54 percent in 2003.

Sources: Piron and Norton 2004; Foster and Mijumbi 2002; IMF 2003; World Bank 2005; Commission for Africa 2005; Diogo 2005; Falck, Landfald and Rebelo 2001. 12 Cambodia - Poverty Assessment 2006

Social relationships, perceptions and to 2004, as we now possess comparable values are also changing, with implications national household socio-economic data for poverty reduction. Practices ofvillage- derived from the 1993/94 Socio-Economic level cooperation or mutual exchange in Survey of Cambodia (SESC) and the 2004 labor or raising livestock are now giving Cambodia Socio-Economic Survey way to fully commercialized arrangements: (CSES) for these two years7. In particular, the switch to waged labor may benefit the we possess comparable, recall-based data poor in most years, but it may also increase on per capita household consumption for vulnerability. At the same time, the both survey years. This allows us to track increasing reliance upon short-or long-term changes in poverty rates and the poverty labor migration as a strategy to survive and profile between these dates’. sometimes thrive is having profound The critical findings from CSES 2004 and effects on the resilience of family and comparison between this and SESC 1994 community structures-but also starting to are summarized in the next two chapters. give rise to new forms of mutual support Because oflimitations in the 1993/94 data, (AFSC, UNICEF and IDRC 2000). the CSES data must be used in slightly Over the last decade, increasing exposure different ways to examine trends in to new political concepts (for example, consumption, living standards and poverty; individual rights and the accountability of and the situation in 2004. Box 1.4 state actors to the electorate) and models of provides some details on the structure of social organization (e.g. NGOs and unions) CSES data and analysis. It also explains is changing the way Cambodians think the various levels of geographical about their relationships with Government. disaggregation (Zone, Region and More widespread education, greater ease Province) that are used throughout this of travel along better roads with more report (and particularly in Chapters 2 security, and increased circulation of and 3). domestic and imported television, music and news media are all increasing people’s awareness of the world beyond their ’ For convenience, the SESC will be referred to immediate community. In the process, henceforth as SESC 1994 rather than “SESC these new social and cultural influences are 1993194”; similarly the CSES will be referred to in shorthand as CSES 2004, even though in raising expectations and increasing the reality it spanned 15 months from late 2003 degree to which people are conscious of through early 2005. and concerned about growing inequalities It is worth noting that the CSES 2004 collected between rich and poor. household consumption data in two different ways, firstly through a recall-based module (as in previous surveys in Cambodia) and secondly Subject matter and sources through a new, diary-based module. This report has opted to use recall-based data as the basis of The focus of this report will be on the consumption and poverty estimates, in large part precisely because it enables comparison with period since the threefold transition-from 1993/94 poverty estimates, which the diary data war to peace, from one-party to multi-party does not. The reasons for this decision are politics, and from economic dirigisme and discussed in more detail in Appendix 1. isolation to markets and integration- Appendix 1 also explains the status of this Poverty Assessment document vis-a-vis other began at the close of the 1980s. Most of recent or forthcoming documents that also use the analysis in the following chapters will CSES 2004 data to present poverty estimates for focus specifically on the period from 1994 Cambodia-and which in some cases opt to use the diary-based consumption values. Country context 13

Box 1.4: Comparing poverty over time and across space: basic concepts in using the CSES data Throughout this report, the data from CSES 2004 has been analyzed in a variety of ways along a variety of dimensions. This Box attempts to summarize these to help the reader follow the analysis in the later chapters. Further details are provided in Appendix 1. P This report wil include both analysis of trend (change between the 1993/94 survey and the 2004 survey) and the analysis of the poverty profile at one point in time (essentially, a snapshot of the current situation). For these two different types of analysis, data from the 2004 sample are used in different ways. When describing standards of living and measures of poverty in 2004, analysis will be based on data from the full set of households in the 2004 sample. However, when we talk about trend-that is, change in values (e.g. poverty rates)-the figures that are presented for 2004 will generally be calculated from a sub-sample of the 2004 data. This sub-set includes only those households that fall within the same geographical area as used in the sampling frame used in 1993/94, when only 56 percent of rural villages could be sampled. In other words, to compare like with like, comparisons between 1993/94 and 2004 will calculate figures for 2004 using only a spatially- defined sub-set of the full 2004 sample. (There will be some exceptions to this principle: we attempt to calculate some all-of-Cambodia estimates of the key poverty indicators for 1993/94 using backward projection, as described in Chapter 2.) The map in Appendix 1 shows the areas covered in the geographically comparable sample frame. e Geographically, each part of the country-and each Primary Sampling Unit (PSU) within the CSES sampling frame-is classified as either urban or rural in character. Many sections of this report start by providing a broad-brush comparison between three Zones namely: (i) Phnom Penh; (ii) Other Urban (i.e. covering other urban centers such as Provincial capitals); and (iii) Rural. We can further disaggregate in terms of five agro-ecological Regions, namely Phnom Penh, Plains, Tonle Sap, Coastal and Plateau/Mountain. These group together Provinces that have broadly similar features in terms of accessibility, terrain, economic activity, etc.

Region ProvincelAggregate of Provinces Phnom Penh Phnom Penh

Plains Kompong Cham; Kandai; Prey Veng;~ Svay_” Riengy “ TakeoI I ~-“ Tonle Sap eay Mean Battambang; Kompong Thom; S?em Reap;

pong Chhnang / Pursat_xxx X” _” ~~ x ”XX

mpot OtherCoastai”~~ ~ -_I (Sihanoukville, Kep, - -” --_ Piatea u/ Kompong Speu; Other Plateau / Mountainw(Kratie,Mondul Kiri, Preah Mountain Vihear, Ratanak Kiri, Stung Treng, Oddar Meanchey, and Pailin) Finally, for those Provinces with large populations (and therefore large numbers of households in the CSES 2004 sample), it is possible to calculate statistically significant estimates for Provinces. For the smaller Provinces, the numbers of households surveyed in the CSES are too few for Province-level analysis to be statistically significant. In one case, merging the samples from two neighboring Provinces produces statistically valid estimates that apply to both provinces but cannot be broken down further (thus the estimates for a “PursatlKompong Chhnang” sub-sample). The remaining sparsely-populated Provinces for which it is not possible to provide statistically valid Province-level samples are often presented in tables that address Provincial variations under a residual “Other” category (e.g. “Other Coastal”). The table above summarizes the levels of disaggregation that are possible. Figure 1.2 shows on a map which Provinces fall into each agro- ecological zone. 14 Cambodia -Poverty Assessment 2006

Figure 1.2: Landscape and natural resources shape spatial variations in Cambodia’s economic and social development I 0 50 100

Figure 1.2 illustrates which Provinces are Development Resource Institute (CDRI). classified as belonging to each of the five Box 1.5 provides some background on this agro-ecological regions that are described research which, as with the CSES, will be in Box 1.5 and used in analysis. referred to throughout this report’. Analysis of the data from CSES 1993/94 Other studies drawn on in this assessment and 2004 is complemented with a wide include surveys of garment sector workers; range of other sources of information on a Public Expenditure Tracking Survey the current status and change in living (PETS) in education; studies of economic standards and poverty over the last ten subsectors or aspects of livelihoods; and years. The most important ofthese are the preliminary findings from Participatory 2000 Cambodia Demographic and Health Poverty Assessments (PPAs) conducted in Survey (CDHS) and the 2004 Cambodia the Tonle Sap Provinces. A full set of Inter-Censal Population Survey (CIPS). In references is provided at the back of the addition to these nationally-representative report. surveys, the report draws on several small- scale contextual studies, most notably the While designed primarily as an input to Moving out ofPoverty Study (MOPS) that national debate on poverty, the MOPS research has been conducted by the Cambodia also serves as a case study for a multi-country World Bank research project ofthe same name. County context 15

~ ~ ~~~~~ Box 1.5: Household dynamics and perspectives of the poor: the Moving out of Poverty Study (MOPS) Over the course of 2004-2005, the Cambodian Development Resource Institute (CDRI) has carried out mixed methods research (that is, research which purposively combines quantitative and qualitative analysis) into village- and household-level wealth and poverty dynamics. This has involved revisiting just under 1,000 households in nine villages (see table) which were previously surveyed by CDRl for an earlier study in 2001.

Village, by region Province Basic criteria for selection Tonle Sap region Andong Trach Battambang Wet season rice; labor migration Krasaing Battambang Wet season rice; resettled returnees Khsach Chiros Kompong Thom Floating rice; Tonle Sap fishing Plains region Prek Kmeng Kandal Dry season rice; substantial fishing Babaong Prey Veng Substantial dry season rice Plateau /mountain region Kanch or Kratie Dry season rice; forest resources Dang Kdar Kompong Thom Wet season rice; forest resources Trapeang Prey Kompong Speu Wet season rice; wage labor Coastal region Kompong Thnaot Kampot Wet season rice; fishing; salt production

This panel dataset (that is, a dataset that tracks the same households between repeated surveys) provides a valuable complement to the random sample survey of successive CSES rounds. In particular, it provides insights into how many households have moved out of poverty over a three year period-and, conversely, how many households which were not poor in 2001 have fallen into poverty by 2004, and how many have not seen their status (poor or non-poor) change. With this information, it becomes possible to estimate what proportion of the poor at any one time are trapped in chronic poverty, relative to the transitory poor who may cycle in and out of poverty in response to life cycle events or household shocks. The qualitative component of the MOPS research, meanwhile, has used semi-structured interviews with groups and individuals in the nine villages, to obtain their views on how their livelihoods have changed, and what factors, at the household, village or national levels, help to explain these changes. Preliminary analysis and findings from MOPS are reported in this document: the full research project will be presented as one or more CDRl Working Papers in 2006.

Summary “peace dividend”-that is, a rapid initial rise from a very low base, as the Securing peace, political stability and restoration of peace and international cooperative international relationships recognition made growth and poverty have been priority goals of the three reduction possible again after two decades Governments elected since 1993 lo.Peace of conflict-that has been complemented has now been achieved. To a significant and extended through a “policy dividend”. extent, the gains in growth and poverty The Government deserves credit for reduction over the last decade represent a securing and (with occasional brief setbacks) maintaining peace. It also lo See for example RGC 1998 (the “Triangle deserves credit for leveraging additional Strategy”); RGC 2004 (the “Rectangular growth benefits from the peace dividend Strategy”). 16 Cambodia - Poverty Assessment 2006 through the active pursuit of regional and poverty. The report attempts to map out a international integration (by joining number of possible scenarios with respect ASEAN and WTO); through an effective, to how the incidence, nature and causes of managed response to the opportunities poverty might change over the next five offered by preferential access to north years-and how this might be affected by American and European markets for alternative policy choices. garments; and through overseeing investments in critical infrastructure development, such as national roads, that are starting to connect the Cambodian space-economy internally and to its neighbors. As described in the next chapter, the net effect of these changes has been a reduction in the incidence of poverty in the order of 25 percent (down by 10-15 percentage points to 35 percent in 2004). This has carried the country a significant distance towards meeting the first of the Cambodia Millennium Development Goals (CMDGs), which commits the Government to reducing poverty to half its 1993 value by 2015 (Ministry ofPlanning 2003,2005). However, a deliberate and over-arching focus on reducing poverty has by and large not been very pronounced in policy formulation; what there has been has been somewhat diluted in policy implementation. To sustain or increase the rates of poverty reduction seen over the last decade, Cambodia will need to adopt a set of new, more complex, and more deliberately pro-poor policies if it is to sustain or accelerate recent progress in growth and poverty reduction. The challenge over the coming years is thus to consolidate the political transition and to build upon it with a concerted and coordinated strategic effort for sustainable poverty reduction. This report seeks to provide a concise but comprehensive overview of what has happened to poverty in Cambodia since 1993, and why, in order to support the Cambodian Government and Cambodian civil society in formulating a strategic response to the problem of 2. Poverty trends, 1994-2004

Summary

Within those central parts of Cambodia that were surveyed in 1993/94 and again in 2004 (“the geographically comparable sampling frame”), the poverty headcount fell from 39 percent to 28 percent over the ten-year period. However, the peripheral and upland parts of the country that were not surveyed in 1993/94 were (and still are) poorer. It is possible to estimate the all-Cambodia poverty rate in 1993/94 based upon the trends in the geographically comparable sampling frame and what is known about population and poverty in the peripheral areas. Using this information, it is estimated that the poverty headcount in Cambodia as a whole fell from around 47 percent in 1994 (back-projected) to 35 percent (measured) in 2004. Given the doubts about some of the parameters and the imprecision of the resulting estimate, it is safer to state that the poverty rate in Cambodia in 1994 was somewhere between 45 and 50 percent. It is thus estimated that between 1994 and 2004, poverty in Cambodia has fallen by between 10 to 15 percent. Within the comparable geographical sampling frame, per capita household consumption-the measure of wellbeing that is used in this report and which is used to specify the poverty line-has risen 32 percent in real terms (Le., allowing for inflation). The improvement in per capita consumption and corresponding fall in poverty rate has been observed in both rural and urban areas, and in all but one ofthe four rural agro- ecological regions. (In the fifth region-the Mountaifllateau region-it is not possible to identify trends because the sub-sample for this region in 1994 was too small to yield statistically significant estimates.) Non-monetary indicators ofwell-being corroborate the picture ofrising average living standards and falling poverty rates. The quality of housing materials, human development indicators, access to modern energy sources and ownership of consumer durables all suggest marked gains, including for the poor. While poverty reduction has been experienced across the population, the rates of improvement have not been uniform. The standards of living of the rural population-which started with the lowest average level of real consumption in 1994-grew at the slowest rate. Inequality has increased most notably within the rural population. Thus, while the poorest fifth ofthe rural population have made progress, this is relatively minor in absolute terms and they continue to fall krther behind in relative terms, as the living standards ofother groups improve at a much faster rate. 18 Cambodia - Poverty Assessment 2006

r available household survey Figure 2.1 : Between 1994 and 2004, poverty now makes it clear that there fell by 10-15 percentage points been a significant fall in m 1 47% poverty over the decade since Cambodia began to emerge from conflict and embark on a “normal” development path. As economic activity has picked up, consumption of goods and services has risen, and the proportion of the population who live in households with per capita consumption values below the poverty line 199314 2004 has fallen. Note: Poverty in 1994 estimated by backward The best estimate of the scale of this shift projection. suggests that the poverty headcount-that Source: SESC 1994, CSES 2004, World Bank is, the proportion of the population living staff projections. below the consumption poverty line has not possible to measure poverty trends fallen from 47 percent in 1994 to 35 because the poverty estimates derived from percent in 2004 (Figure 2.1). Allowing for these surveys were not comparable, for a some doubt about the exact value of the variety ofreasons. headcount in 1994, it might be more defensible to say that poverty in 1994 Firstly and most importantly, none of the could credibly have been anywhere surveys covered the whole of the country between 45 and 50 percent. On this basis, and by extension the whole population. In we can say that poverty has fallen by theory, it should nonetheless have been between 10 and 15 percent over the last ten possible to obtain partial insights into years-or by about 1 percentage point a trends in the area ofthe country covered in year. This refutes the general impression the first survey by comparing poverty of growth with little or no impact on measures obtained from different survey poverty--an impression that took hold dates within a common geographical when poverty estimates from surveys over sampling frame. In other words, by the course of the 1990s were compared, defining a sub-set ofthe sample covered in despite the fact that they were in fact a later, more geographically extensive simply not comparable. survey that was drawn only from the same geographical sample frame as the first Tracking poverty over time survey-that is, by defining and applying a comparable (1 993/94) geographical For the first time, trend analysis sampling frame to the data collected in is now possible subsequent surveys-it should have been possible to obtain reliable estimates of A number of large household sample change within this area of the country. surveys have been carried out in Cambodia Unfortunately, it appeared until this year since the early 1990s (Box 2.1). These that the definition of the sampling frame have, to varying degrees, provided insights used for the first survey (the 1994 SESC) into patterns of consumption and, by had been permanently lost, making extension, the proportion of the population comparison with a baseline, even for just that had consumption values below the part ofthe country, impossible. poverty line. Until now, however, it was Poverty trends 19

Box 2.1 : Comparing mangos and bananas: measuring poverty in Cambodia, 1994-2004 The National Institute of Statistics (NIS) of the Royal Government has, with assistance from donors, conducted a number of national sample surveys of consumption and poverty amongst Cambodian households since the Paris Peace Agreement. These are:

1 The 1993/94 Socio-Economic Survey of Cambodia (SESC) - henceforth referred to for convenience simply as SESC 1994.

1 SESC 1996.

1 The Cambodia Socio-Economic Survey (CSES) 1997.

1 CSES 1999.

1 CSES 2004 (actually conducted over 15 months from late 2003 through early 2005; however, unless otherwise noted values from this survey will be derived from the data collected in the 12 calendar months of 2004 only, hence this will be referred to as “CSES 2004” and not “CSES 2003-05”). However, due to security problems at the time, the 1993/94 Socio Economic Survey of Cambodia was only able to cover 56 percent of rural villages and 65 percent of the rural population. Of “other urban” areas, only 84 percent of sampling units were covered. The areas not surveyed in 1993/94 were those where the Khmer Rouge were still active and remote parts of several provinces where accessibility was extremely difficult. Broadly speaking then, the areas excluded from the 1993/94 CSES were the remoter, forested highlands along the south-western, eastern and northern borders of the country, where fewer people lived but where levels of poverty were presumably, considerably higher than in the lowland core provinces that were covered in the survey. In subsequent surveys, the coverage of the country in the sampling frames increased, including progressively more of these poorer and more remote rural areas. This expanding sampling frame, combined with differences in survey design and implementation, meant that the poverty rates derived from the surveys in the 1990s were simply not comparable, and could not be used to construct time series data for poverty. The 2004 Cambodia Socio-Economic Survey (CSES) was the first Cambodian household survey to be based on a sampling frame that includes all villages (which served as the Primary Sampling Units) in the entire country, drawing on the country’s first Population Census (conducted in 1998). Note that in this chapter we only utilize information from the first survey (the 1994 SESC) but not any of the other surveys conducted in the 1990s. There are several reasons for this.

1 The 1996 SESC collected much useful information, but its household consumption module was too limited (data on food consumption was collected as a single category, namely the total consumed during the last week). As a result, household consumption was seriously under- estimated (and poverty over-estimated) compared to, for example, SESC 1994. 1 Data collection for the 1997 CSES took place in only one round (June-July 1997). This period unfortunately coincided with deepening political crisis that was eventually to lead to political violence between the partners in the coalition government. There is good reason to believe the households were cutting back their consumption in anticipation of political turmoil; the low consumption values and high poverty rates recorded in CSES 1997 do not seem to be representative of the actual standard of living in the first half of 1997, given the progress observed in other indicators of wellbeing. For this reason, we have decided not to include the 1997 data. 1 Finally, the data collected in the 1999 CSES show irreconcilably large differences between the two rounds of the survey. The consensus is that standards of training and supervision for the first round were not adequate, resulting in under-recorded consumption and exaggerated poverty; the levels of household consumption recorded in the second round (after much- improved training and supervision had been put in place) were much higher, and the poverty estimates thus much lower. After much discussion among experts and the Government, it has been agreed that the consumption data in the 1999 CSES is not reliable and does not support comparison of poverty rates with other surveys. 20 Cambodia -Poverty Assessment 2006

Secondly, changes in survey design and (measured in terms of expenditure- implementation from one survey to the money spent-or expenditure equivalent) next meant that data (particularly the is necessary for a decent standard of living. critical consumption data that is used to Technically speaking, Cambodia’s baseline construct poverty lines and calculate poverty line consists of a single national poverty rates) was collected in different reference food bundle and three regional ways. As a result, findings from one non-food allowances. The baseline values survey could not easily or credibly be are in 1993/94 Riel and refer to daily per compared with those from another, even if capita levels of food and non-food it were possible to compare the same consumption. geographical sample between surveys. The food poverty line is based on the In mid-2005, a hard copy of the sampling estimated cost of a basket of food that frame of the first survey (the 1994 SECS) would provide a subsistence-level dietary was rediscovered in NIS. This, combined intake of 2,100 calories per day. This is with a good match between the based on the quantities of different foods construction of the consumption module in consumed by persons in the third quintile the 1994 SESC and the 2004 CSES, has of the per capita consumption distribution made it possible, for the first time, to (as this was the first quintile in the 1993/94 compare household consumption and SESC that met the 2,100 calorie poverty within the same geographical areas minimum). Someone who consumes less between 1994 and 2004. than the food poverty line is not receiving the minimum amount of food (calories) This chapter therefore provides for the first necessary to maintain their health, and in time information on trends in living theory has no ability to meet non-food standards (real per capita household consumption needs for clothing, shelter, consumption) and poverty rates between medical care, and so on. 1994 and 2004. It does this firstly by comparing change within a geographically In reality, even the very poor need to comparable subset of the 2004 CSES with consume a minimum of non-food goods baseline data from the first Cambodian and services. To calculate the overall household survey, the 1993/94 Socio- poverty line, non-food allowances are Economic Survey ofCambodia (SESC). derived from the non-food consumption of Cambodians whose total per capita Having provided the dynamic analysis of household consumption is just equal to the trends in this chapter, the next chapter will food poverty line; that is, the non-food then describe in more detail the key allowance is the amount of consumption features of poverty in Cambodia in 2004, that food-poor individuals divert from food using the full CSES 2004 sample (i.e. data consumption to non-food consumption. collected in 2004 from households within the 1994 geographically comparable The non-food allowances, added to the sample frame and data from households food poverty line, yield the overall poverty outside this frame). line. Table 2.1 presents updated food poverty lines, non-food allowances and Setting the poverty line total poverty lines for Cambodia. Setting a poverty line is an attempt to Annex 1 discusses in detail the procedures specify what level of individual used in setting poverty lines for analysis of , consumption of goods and services SESC/CSES data, including such issues as Povertv trends 21

Table 2.1: Updated food poverty line, non- Box 2.2: Measuring poverty: headcount, food allowances, and total gap and severity poverty lines Standard international practice for measuring poverty involves the use of three Region 1993194 2004 mathematically interrelated indices of Food poverty lines (current Riel per poverty, first presented by Foster, Greer and capita per day) Thorbecke (1984). These poverty Phnom Penh 1,185 1,782 measures-known as FGT measures-are: Other Urban 996 1,568 . The poverty headcount index (Po), also sometimes referred to as the poverty Rural 882 1.389 incidence or simply the poverty rate. This Non-food allowances (current Riel per is the most commonly used measure and capita per day) the easiest to understand. Very simply, it Phnom Penh 393 569 expresses the percentage of the population with per capita consumption below the Other Urban 269 384 poverty line; Rural 236 364 1 the poverty gap index (P,), which is the Total poverty lines (current Riel per average difference in the total population capita per day) between individual’s recorded per capita ~~ Phnom Penh 1,578 2,351 consumption and the poverty line, with a zero value assigned to those above the Other Urban 1,265 1,952 poverty line; and Rural 1,118 1,753 ~~ ~ 1 the poverty severity index (P2). This is Source: SESC 1994; CSES 2004. calculated by first squaring individual poverty gaps before averaging them. This gives greater weight to larger poverty gaps, and provides insights into the distribution of how allowances have been updated to very low consumption values (i.e. extreme allow for inflation. poverty).

Measuring poverty Once a poverty line has been defined, the Key findings distribution of per capita household consumption is compared to this line to The poverty rate has fallen in generate a set of different measures of both rural and urban areas poverty. Box 2.2 describes in basic terms Within the geographically comparable the three commonly-used poverty sample frame, the poverty headcount (that measures that will be referred to is, the percentage of the population living throughout this document. under the poverty line) has fallen from 39 percent in 1994 to 28 percent in 2004. When this rate of known poverty reduction is combined with estimates of the distribution of the total population between areas within and outside the 1994 sampling frame, we estimate that poverty in the whole of Cambodia has fallen from 47 percent to 35 percent between 1994 and 2004 (Box 2.3). 22 Cambodia - Povertv Assessment 2006

Box 2.3: Backward-projection of national poverty estimates for 1994 Although the CSES 2004 was the first survey that sampled the entire country, it is possible to use existing information and postulate some simple assumptions to estimate 1994 poverty measures for the excluded areas, and in turn backward-project poverty measures for the entire country in 1994. This backward projection is useful because it provides a best-informed guess of the poverty situation for the entire country in 1994. Whatever the estimates may be, this enables a discussion about which hypotheses can be ruled out (e.g., stagnant, modest, or rapid poverty reduction), and whether the estimates are plausible given the situations then and now. The key requirements for this exercise are: (i) the population shares of areas included in and excluded from the SESC 1994 sampling frame in both 1994 and 2004; and (ii) the poverty measures of included areas in years 1994 and 2004 and of areas outside the comparable sample frame in 2004. Combining the 1994 and 2004 sampling frames and the 1993-2004 population projections calculated from the 1998 Population Census (Neupert, 2005), we estimated: (i) the share of population in the areas excluded from the SESC 19934 sampling frame in 1994 and in 2004; and (ii) the shares of population included in the SESC 1994 sampling frame in 1994 and in 2004. From the SESC 1994 and the CSES 2004, we could estimate the poverty incidence for: (i) included areas in the SESC 1993/94’s sampling frame in 1993 and in 2004 and (ii) excluded areas from the SESC 1993/94’s sampling frame in 2004. It is possible to construct a few scenarios for poverty reduction rates in excluded areas from the SESC 1994 sampling frame. The main text presents just one scenario, in which the rate of poverty reduction in excluded areas is assumed to be the same as that in included areas, within the same strata.

Scenario 1: Best informed poverty estimates for all Cambodia in 1993194 Population Poverty Poverty Gap Poverty shares Headcount (%) (%) Severity (%)

1 Phnom Penh 0.0758 11.4 3.0 1.2

1 “included” areas in Other Urban 0.0567 36.6 9.6 3.6 1 “included” areas in Rural 0.4670 43.1 10.0 3.3 1 “excluded” areas in Other Urban 0.0414 63.5 14.5 4.9 1 “excluded” areas in Rural 0.3591 59.1 16.7 6.2 1 Cambodia 47.0 12.0 4.3

This is the best-informed assumption because there is no other hard data (such as provincial GDP or agricultural revenue) to suggest otherwise. Nonetheless, other scenarios of poverty reduction are also plausible; we present two below. In the first, poverty is assumed to have been stagnant (Le., no reduction) in the excluded areas; and in the second, poverty in excluded areas (rural and urban) fell at the same rate as in the urban areas that were included in the 1994 sampling frame,

Two alternative scenarios of poverty estimates for the entire Cambodia in 1993194 Poverty Poverty Gap Poverty Assumptions/Scenarios Headcount (%) Severity (%) (%) (1) Poverty is stagnant (no reduction) in all excluded areas 41.1 10.5 3.8 (2) Poverty in rural and urban excluded areas fell at the rate of urban included areas 55.2 14.0 4.9

Source: SESC 1994, CSES 2004, sampling frames for SESC 1994 and CSES 2004, Neupert (2005). Poverty trends 23

A poverty headcount of 47 percent for Figure 2.2: Poverty--and food poverty- Cambodia in 1994, at per capita GDP of have fallen in both urban $21 1 in constant 2000 US$, does not seem and rural areas unreasonable. In neighboring Lao PDR, - Percentage of population under poverty the poverty headcount in 1993 was 46 line/food poverty line, 1994 and 2004 percent, at per capita GDP of $247 in (geographically comparable sample) constant 2000 US$ (World Bank, 2005); while in Vietnam, the poverty headcount A. Poverty line was estimated at 56 percent in 1993, at per (food plus non-food) capita GDP of $265 in constant 2000 US$ 50% (World Bank, 1999). In other words, a 143% poverty headcount of between 45 and 50 percent in 1994 seems credible in the context of Cambodia’s GDP at that time. Looking only at trends within the 25% comparable sampling frame (i.e. at trends in measured as opposed to back-projected poverty rates), it is clear that poverty rates have fallen in both urban and rural areas 0% (Figure 2.2). As would be expected, the Rural Other Phnom proportion of the population living under urban Penh the lower, more severe food poverty line has also fallen, at about the same rate B. Food poverty line (slightly slower in urban areas, marginally faster in rural areas).

Figure 2.3: The depth of poverty has declined in both urban and rural areas - Poverty gap (average distance by which consumption of the poor falls short of the poverty line, as % of poverty line), 1994 and 2004 Rural Other Phnom (geographical Iy com parable frame) urban Penh Other Phnom Source: SESC 1994, CSES 2004. Rural urban Penh Cambodia ...as has the depth of poverty Given that both poverty and food poverty headcounts have fallen, it is only to be expected that the depth of poverty has also declined. Figure 2.3 shows change in the poverty gap index, which essentially measures the average amount by which the -9.2 per capita consumption of poor individuals falls below the poverty line. As the living -12 - standards of the poor (as measured by real per capita consumption) have improved, in Source: SESC 1994, CSES 2004. some cases carrying them over the poverty 24 Cambodia - Povertv Assessment 2006

Figure 2.4: Per capita consumption increased significantly over the last ten years E

r(D0 E a0P

F:

0 500 1000, 2000 4000 8000 16000 32000 Per capita Consumption (Riel per day, 1993/94 Phnom Penh prices) Source: SESC 1994, CSES 2004. line and in other cases merely taking them geographically comparable frame, real up towards the line, so the poverty gap has consumption (defined in terms of average been reduced. Poverty is, as expected, per capita household consumption, deepest in the countryside and very measured in 1993/94 riels per day) rose 32 shallow in Phnom Penh. percent (from 2,228 to 2,932 riels per capita per day) in the nation as a whole. Real consumption has increased The fastest rise in average consumption for the vast majority (36 percent) was found in urban areas outside Phnom Penh, while the slowest rise Figure 2.4 presents the cumulative (24 percent) was found in rural areas distribution function of real per capita (Figure 2.5). consumption, based on geographically comparable samples (Le., households Figure 2.5: Consumption has risen in rural areas- from villages that were included in but more slowly than in urban areas the 1993/94 SESC sampling frame) - Average per capita consumption (mean riel per day, as well as the food poverty line and 1993/94 constant prices) the overall poverty line in 1993/94 Phnom Penh prices. It shows that 6,000 1 5,501 real per capita household 5,000 ~ consumption increased for all segments of the population 4,000 - distribution between 1994 and 2004, as seen in a clear shift to the right in the cumulative distribution. One important implication is that poverty incidence would be estimated to have decreased during this 10-year Rural Other Phnom Cambodia period for any conceivable poverty urban Penh line. 1994 02004 Consumption has risen in the 1 countryside as well as the towns, but Source: SESC 1994, CSES 2004. by a lesser amount, from a much lower base. Within the Povertv trends 25

Figure 2.6: Households-including poor households-can now afford to spend more on non-food items - Non-food consumption as a % of total household consumption, 1994 and 2004, by per capita consumption quintile

75% ,

Poorest Next Middle Next Richest Cambodia poorest richest

Source: SESC 1994, CSES 2004.

...leading to increased spending Higher living standards are on non-food items confirmed in other indicators ... Reinforcing the impression of rising living Household surveys in Cambodia have standards and falling poverty, Figure 2.6 collected data on a large number of socio- shows that the share of non-food economic indicators in addition to consumption in total consumption has been household consumption. As with the non- rising for all quintile groups between food share of total consumption, these 1993/94 and 2004. Given that the first socio-economic indicators are very useful priority for household consumption is as a cross-check on the accuracy of the always food, and that the food share of picture of poverty reduction obtained from total consumption is inversely correlated the poverty line measures of headcount, with wealth, an increasing non-food share depth and severity; indeed, they are often in total consumption corroborates the regarded as more reliable than household picture of falling poverty headcounts. This consumption or income as a measure of implies that poor and non-poor households real wealth. They also provide insights can now better afford to meet their non- into non-monetary dimensions of poverty, food needs than was the case in 1994. such as education and health status, which may not be captured in consumption data. Non-income indicators such as ownership ofhousehold durables, housing quality and use of modem sources of energy indicate that living standards of the population in 26 Cambodia - Poverty Assessment 2006 general and of the poor in particular have Figure 2.7: The poorest fifth of the population improved remarkably between 1993/94 own more assets and have better and 2004. Among the poorest fifth of the access to electricity than in 1994 population, ownership of a television increased ten-fold (fi-om 3.4 to 36 percent), while access to generator- or battery- powered lighting increased from 1.6 percent to 28 percent (Figure 2.7). Among this bottom quintile, far more now live in houses roofed with galvanized iron or aluminum, compared to ten years ago, when a large proportion of houses were Radio TeleLision Electric light covered in thatch (Figure 2.8). from city grid, The supply ofbasic services has improved generator or considerably. This is starting to emerge in battery some human development outcome indicators. Between 1997 and 2004, net Source: SESC 1994, CSES 2004. enrollment and completion rates at the primary level increased by over ' 10 percentage points among children from the Figure 2.8: The poorest 20 percent also live in better houses than in the past poorest fifth ofthe population (Figure 2.9). - Primary roofing material However, the size of increases in consumption has varied considerably While consumption has grown for all quintiles between 1994 and 2004, this growth has not been experienced evenly, with rural areas growing most slowly (Figure 2.5) and demonstrating continued higher levels of poverty (Figure 2.2). In 1994, real consumption in rural areas was Thatch Galvanized Tile 67 percent ofthat in other urban areas: this iron or fell to 61 percent in 2004. aluminum Similarly, the growth in consumption has not been uniform across consumption Source: SESC 1994, CSES 2004. quintile groups. Figure 2.10 shows clearly richest quintile. a result, the shares of that while real per capita consumption As consumption of the poor in the country's increased significantly in all quintile total consumption fell between 1994 and groups during 1994 through 2004, the 2004. relative gains were inversely related to the initial level of average per capita consumption. Real per capita consumption in the poorest quintile increased by only 8 percent, but rose by 45 percent for the Poverty trends 27

Figure 2.9: Net primary enrolment has improved for all The rate of poverty groups reduction has thus been - Children aged 6-1 1 years in primary school as % of uneven.. . population in these years, by per capita consumption quintile Progress in poverty reduction was also not uniform among the poor themselves. The extreme poor have 100 experienced significantly slower growth in real consumption than the 80 8 “average” poor. Because the SESC U and CSES were not panel surveys 60 (ie., CSES did not revisit the same - i! sample of households as SESC), we 40 Q,E are not able to comment directly on U the transition of the population into 20 and out of poverty. Nevertheless, the cross-sectional information still 0 allows us to conclude that a ’oorest Next Middle Next Richest substantial proportion of the barely poorest richest poor (i.e., those whose consumption falls short of the poverty line by a relatively small amount) had exited poverty, whereas the magnitude of Source: CSES 1997, CLS 2001, CSES 2004. consumption shortfall of the extreme poor, and the proportion of the extreme poor among the entire

Figure 2.10: Over the last ten years, per capita consumption amongst the poorest fifth has risen by only 8 percent-compared to 45 percent amongst the richest fifth of the population - Real average per capita household consumption, 1994 and 2004, within the geographically comparable sample frame

71500 1

5,000 -

2,500

0 Poorest Next Middle Next Richest Cambodia poorest richest

Source: SESC 1994, CSES 2004. 28 Cambodia - Poverty Assessment 2006

Figure 2.11: Poverty has fallen much faster in Phnom Penh and the Coast

- Poverty headcounts (% of population living under the poverty line), 1994 and 2004 (geographically comparable sample), by agro-ecological region. 60% 58%

46% 42% 40% 39% 40% 1

~ I 20% - 11%

0% c T Phnom Coast Plains Tonle Sap Plateau / Cambodia Penh mountain

Source: SESC 1994, CSES 2004. Note that the small size of the comparable sample coverage in the Mountain/Plateau Provinces means that the poverty headcounts are estimated with large standard errors and the increase in headcount is not statistically significant. poor population, was as acute in 2004 as it coefficients for Cambodia confirm that had been in 1994. inequality rose considerably within the geographically comparable areas between The pattern of growth and poverty 1994 and 2004, driven primarily by reduction is also widening the differences increased inequality within the rural areas in poverty rates between different parts of (Figure 2.12). The rate of increase in the country. Poverty has fallen fastest in inequality seen in Cambodia is unusually Phnom Penh and the Coastal region, which fast. Decomposing inequality changes into has now pulled considerably ahead of the within-group and between-group changes, Tonle Sap and Plains regions (Figure we find that 86 percent of total increase in 2.1 1). inequality was attributable to increased ...and inequality has risen inequality within the rural region and only 14 percent to a rise in mral-urban The Gini coefficient provides a useful inequality. summary measure of inequality in the distribution of income or consumption: a value of zero signifies perfect equality in distribution, while a value of 1 indicates perfect inequality. Estimates of Gini Poverty trends 29

Figure 2.12: Inequality rose in rural areas, The impression of rapidly growing and in the country as a whole inequality, both between urban and rural - Gini coefficients (0 perfect equality, areas and within rural areas, is something 1 = perfect inequality) for per capita real that is strongly felt in popular debate, consumption, 1994 and 2004 (comparable including at the village level. Box 2.4 sample frame) summarizes some of the comments from ++ one of the nine fieldwork sites in the 0 00 a MOPS research project. Similar comments were recorded in other sites. Rising inequality is of course inevitable in the transition from a very poor, centrally- planned economy to one based on market principles. As long as inequality increases in parallel to a significant fall in absolute poverty, it could be argued that inequality per se is not a problem. However, an Rural Other Phnom Cambodia urban Penh increasing volume ofinternational research points to the strong connections between 1994 02004 1 high levels of inequality and slow rates of economic growth and poverty reduction Source: SESC 1994, CSES 2004. (see for example World Bank 2005; Naschold 2002). What is interesting about the MOPS accounts is the feeling that

Box 2.4: Perceptions of inequality amongst rural youths If there is a sense of greater security and basic freedoms in rural communities, there is also a sense of growing inequality in the opportunities available to the children of rich and poor families. The rich have enough food to eat while the poor have nothing. The rich are brave and powerful while the poor don’t have any of these things. A son of the rich has very clean and good clothes while I am a poor son and wear dirty clothes (male youth). Since 1998, the well-being of people in this community has been largely increasing. The rich become richer while the poor remain trapped in poverty or become worse off. The reason for the inequality of people in this community is that the better-off have large arable land and are able to hire the poor and destitute to work for them. Theyjust use only their brain and money in order to get high earnings, while we use strong labour and get very little pay in return (female youth). Some noted that inequalities in wealth meant that the rich are able to enjoy freedoms that are largely theoretical for the poor. We have freedom to do what we want, especially searching for jobs both inside and outside the village in order to support our own stomach and other members in the family. But unfortunately the rich has more freedom than the poor because they have money and can entertain whenever they want (young woman from poor household). If we want to join the community’s ceremony at night we should ask for permission from our parents first. But normally we cannot go because of the fact that we don’t have money and we have to work hard in the morning (female youth.) If one wants to have freedom one should have enough rice to eat (female youth).

Source: MOPS focus group discussions with youths in Andong Trach village, from CDRl 2006a (forthcoming). 30 Cambodia -Poverty Assessment 2006 inequalities are growing not only in terms situation in which substantial numbers of outcomes but also opportunities. There escape poverty, but a significant number of is a perception that the poor, with few previously non-poor also move in the assets, social connections or opportunities opposite direction, falling back into for education, are increasingly trapped in poverty, the balance between these two low-skilled, subsistence-level manual work transfers resulting in the net change in are not able to compete on equal terms headcount. This is important, because with the rich. experience in other countries show that success in poverty reduction typically This theme will be explored in more detail requires a combination of ladders- in the next chapter, which discusses the creating opportunities for the poor to level of inequality in 2004. For the escape poverty-and safety nets, which moment, however, it is simply noted that if prevent the non-poor fiom falling into the benefits of growth over the last ten poverty, or the poor fiom falling into years had been more evenly distributed destitution. The policies and programs that between the rich and the poor, the rate of poverty reduction could have been are needed to address vulnerability are often quite different from those that are significantly greater. The experience of required to help the currently poor; countries (e.g. Vietnam) which have similarly, the chronic or long-term poor created broad-based growth which includes require different forms of assistance to the the poor to a greater degree suggests also transitory, short-term poor. that such patterns of shared growth are more stable (and often higher) than growth patterns (such as that seen in Cambodia) How stable are the categories of which are based on a few enclave sectors poor and non-poor? and in which the benefits of growth accrue To investigate these longitudinal issues, disproportionately to those who are already the Moving out of Poverty Study (MOPS) rich. designed and carried out by CDRI resurveyed in 2004/05 just under 1,000 Moving in and out of poverty: households that had previously been panel data analysis surveyed in 2001. The result is a set of insights into the balance, over three years, When determined by random sample between the chronic poor (those who survey, aggregate change in the poverty remain poor in both years); the stable non- headcount between any two points gives us poor; those who have escaped poverty; and the change between two snapshots, but a those who have fallen into poverty’. poor sense of actual household trajectories over that time period. The change in headcount may disguise considerable 1 Panel studies typically focus on the difference diversity in the actual experiences of between the chronic poor (alternatively called individuals or households. The same the “always poor”) and the transitory poor (or change in overall poverty rate, in other “sometimes poor”) who spend a period in poverty before escaping (and possibly falling words, could be the product of either a back into poverty again later). It is only really steady across-the-board improvement in possible to define the transitory poor when living standards, in which the non-poor there are three time-points under observation, stay out of poverty and a proportion of the as MOPS has completed only two rounds to poor escape poverty; or a more dynamic date (2001 and 2004) we cannot label a transitory poor group yet. Poverty trends 31

Measuring the direction and scale of movement) of households over the three- change in household living standards year period between surveys. The three shaded cells running diagonally Given the difficulty of accurately or down the middle of Table 2.2 represent precisely measuring consumption in multi- households that have not changed their purpose household surveys, there is a status between 2001 and 2004. Whether reasonable concern that analysis based they were very poor, moderately poor or solely upon whether households crossed an well-off in 2001, they remained in the exact poverty line might give a false same group in 2004. Those who were very impression of precision. Instead of such a poor in both rounds are labeled the chronic binary poorhon-poor division, we have poor. This group accounts for about 14 opted to conduct initial analysis of the percent of the MOPS sample. Those household dynamics in the MOPS dataset “static middle” group who began and on a three-fold classification of household ended the period in the “moderately poor” well-being at any one point in time. This category accounted for another 14 percent. approach distinguishes the clearly poor Finally, the “comfortable rich” who (those well below the nominal poverty remained well-off in both 2001 and 2004 line, labeled the “very poor”); the clearly accounted for the single largest group in non-poor (those well above the poverty this transition matrix, with 24 percent of line, labeled the “well-off’); and those the sample. Taken together, these non- falling in a band 20 percent above and movers account for around 51 percent of below the poverty line, labeled for the whole sample, implying a degree of convenience the “moderately poor”. On stability for many rural households, at least the basis of the per capita household over a relatively short, three-year period, expenditure recorded in MOPS, each ofthe and within these three broad wealth 890 households that were surveyed in 2001 groups. and 2004 were classified into one of these The remaining 49 percent, however, three bands for both survey years. Comparing status in 2001 with status in changed their status upwards or 2004 gives us nine groups, shown as nine downwards by either one or two positions. Slightly more households moved up (24.5 cells in Table 2.2, each ofwhich is defined percent, comprised of the three groups in in terms of the movement (or non-

Table 2.2: Transition matrix of panel households I 2004 I Vervpoor Moderatelv Door Well-off

Very poor

2001 Moderately poor

Well-off Source: CDRl 2006a (forthcoming) 32 Cambodia - Poverty Assessment 2006 the top right third of Table 2.2) than subsequent analysis (Table 2.2 and moved down (22.3 percent, allocated Table 2.3). amongst the three groups in the bottom right third ofthe Table). Explaining changes in household Only a few households (around 7 percent) fortunes experienced a dramatic change in status, defined in terms of a movement upwards Livelihood strategies and their dynamics or downwards by two categories (Le. from are complex, so it is not entirely very poor to well-off-labeled “rags to straightforward to identify what correlates riches”; or from well-off to very poor- of change in household status can be taken labeled “from riches to rags”). Given the as cause of that change, rather than an small size of these two groups (each adaptation or coping strategy arising out of containing only 3.5 percent of the nine- it. We therefore present a relatively village sample, or about 30 households); straightforward preliminary description of we combine these two groups with changes in household characteristics at this neighboring groups for the purposes of stage. More detailed analysis will be Table 2.3: Changes in per capita consumption and in positions within the income distribution of panel households, 2001-2004

Group Change percent of Mean p.c. household in households consum ption (riels position per day) 2001 2004

1. Chronic poor 0 14.0 percent 849 1,093

2. Static middle 0 14.3 percent 1,460 1,716

3. Comfortable rich 0 24.4 percent 2,931 3,362

4. Deepening poverty -1 7.1 percent 1,424 1,166

5. Falling into poverty 15.2 percent 2,353 1,642

5a. Rags to riches -2 3.3 percent

5b. Slipping into poverty -1 11.9 percent

6. Escaping poverty 13.3 percent 968 1,960

6a. Climbing out of poverty +I 9.8 percent

6b. Rags to riches +2 3.5 percent

7. Climbing into wealth +I 11.8 percent 1,459 3,296

Source: CDRl2006a (forthcoming). Poverty trends 33 included in the CDRI MOPS synthesis considerable change from 200 1, when study to be published by CDRI in 2006. income from CPR was nearly as important as agriculture (28 percent compared to 33 Initial analysis suggests that rising total percent), and labor contributed only 28 incomes have buttressed the position ofthe percent. non-poor. The chronic poor, meanwhile, appear to be trapped in low-income Those falling into poverty were activities, preventing them from saving or characterized by a fall in income from investing in ways that might lead to an agriculture and CPR, and a rise (in escape from poverty. Those households absolute and proportionate terms) in who fell into poverty, unsurprisingly, income from self-employment and labor. experienced a large drop in income. This reflects that for many, floods or Similarly, those who have seen their droughts over recent years have affected consumption decline (falling into poverty rice farming; many of the MOPS villages and deepening poverty) have seen the have also experienced quite dramatic value of total household assets decline declines in fish catches. These are trends between 2001 and 2004, while all other that are broadly shared with many other groups-including even the chronic rural households in Cambodia. poor-have accumulated a greater total Interestingly, the MOPS fieldwork-which value ofassets. asked specific questions about shocks The relationship between household faced by each household-suggests no dynamics and changes in the composition significant difference in the incidence of of income is extremely interesting. Figure shocks amongst poor and rich groups. 2.12 plots percent change in each source of This implies that while the majority of income for each group. The role of rural households may face a similar (and common property resources in household rather high) risk profile, it is the ability of income declined for almost all groups, richer households to cope with such shocks while income from agriculture fell for without needing to resort to the sale ofkey those who became or remained very poor, productive assets or other desperate and increased for those who were or survival strategies that distinguishes them became well-off, and for those who moved from the poor, rather than the lower upward or remained in the middle, probability of suffering such shocks in the moderately poor group. Income from self- first place. employment, labor and “other” sources have become far more important, with labor appearing to be particularly important in helping households move out of the very poor group, while self- employment helped the moderately poor to move into the ranks ofthe better-off. Thus for the chronic poor, the most important source of income in 2004 (38 percent of the total) came from hiring out their labor, compared to 23 percent from agriculture and 18 percent from common property resources (CPR). This marks a 34 Cambodia - Poverty Assessment 2006

Figure 2.13: Household dynamics and the changing composition of income, 2001- 2004 Agricu Itu re Selfemployment

3 Confortable rich 3 Confortable rich

7 Uihing into wealth 7 Climbing into wealth

6 Escaping poverty 6 Escaping poverty

2 Static mddle 2 Static mddle

1 Chronic poor 1 Chronic poor

4 Deepening poverty 4 Deepening poverty

5 Falling into poverty 5 Falling into poverty

-50% 0% 50% 100% -100% 0% 100% 200% 300%

% change in income (riels p.c. per day), 2001-2004 % change in income (riels p.c per day), 2001-2004

Common property resources Labor

3 Confortable rich 3. Confortable rich

7 Uirrbing into w ealth 7 Climbing into w eakh

6 Escaping poverty 6 Escaping poverty

2 Static mddle 2 Static mddle

1, Chronic poor ""1 - 1. Chronic poor J Total I I 1 D Total 4 Deepening poverty 4. Deepening poverty '

5. Falling into poverty 5. Falling into poverty

-50% 0% 50% 100% -100% 0% 100% 200%

% change in income (riels p.c. per day), 2001-2004 % change in income (riels p.c. per day), 2001-2004

Other (rent and transfers)

3. Confortable rich

7 Climbing into wealth

6 Escaping poverty

2. Static mddle

1. Chronic poor

4 Deepening poverty

5. Falling into poverty r

-150% 0% 150% 300% 450% % change in income (riels p.c. per day), 2001-2004

Source: CDRl2006a (forthcoming). "Total" is percentage change in total income for that mobility group between 2001 and 2004. 3. The nature of poverty in Cambodia, 2004

Summary The poverty headcount for 2004 is estimated at 35 percent; that is, 35 percent of the Cambodian population is estimated to have been living under the national poverty line. One in five Cambodians lived under the food poverty line. Poverty was considerably higher in rural areas (39 percent) than urban areas (5 percent in Phnom Penh and 25 percent in other urban areas). Poverty in Cambodia is overwhelmingly a rural phenomenon. In 2004 about 9 1 percent of the poor lived in rural areas. The rural areas of the Tonle Sap and MountaidPlateau regions experience both the highest poverty headcounts and the most pronounced poverty severity. Poverty headcount measures in rural Tonle Sap and rural MountaidPlateau regions were, respectively, 45 percent and 56 percent. Poverty severity measures in both regions were about twice the national average. Fully 40 percent of nation’s total poor live in the Plains region; the rural poor in the Tonle Sap and MountairdPlateau regions account for about 50 percent. In terms of occupational profile, households with heads engaged in agricultural activities or employed as domestic workers, experience the highest incidence of poverty (averaging over 40 percent) and the worst poverty severity. Households in which the head of household was employed in one ofthese two occupations together account for almost two- thirds of all the poor in Cambodia. The occupational group which enjoys the lowest poverty incidence was that of households headed by an individual working in the public sector. Poor households have higher dependency burden and lack human capital; they tend to be uneducated, unskilled, and unhealthy. The poor and in particular the extreme poor are concentrated in rural remote areas, with limited access to roads, markets, and basic services. They also lack secure land tenure and access to irrigation facilities. Critical problems for the urban poor revolve around security of housing rights and opportunities for gainful income generation or employment. Many in the squatter settlements claim tenancy rights with limited right to some services but lack secure housing tenure and are vulnerable to evictions.

his chapter presents some basic facts geographic ally comparable areas between on household consumption and 1994 and 2004, this chapter presents the poverty, based primarily on the full overall picture of living standards for all of Tnationally-representative 12-month Cambodia in the year 2004. sample from CSES 2004. While the previous chapter describes trends in various measures of poverty for 36 Cambodia - Poverty Assessment 2006

This difference between geographically- expressed in 2004 Phnom Penh prices. The comparable and full sample accounts for urban Phnom Penh population clearly differences in the 2004 values reported for enjoyed the highest real per capita level of the same variable (e.g. average consumption, having an estimated level consumption or poverty incidence) in well over twice the national average Chapter 2 and this Chapter. Thus, for (Figure 3.2). The rural areas of the Tonle example, the poverty rate in rural areas in Sap and of PlateadMountain regions, 2004 is reported in Chapter 2 as 34 percent meanwhile, had the lowest mean levels of (Figure 2.2), but in this chapter as 39 real per capita consumption. These two percent. The former figure is the poverty poorest rural areas account for over one rate amongst the households surveyed in third of the country’s population, i.e., 26 2004 that fell within the 1994 sampling percent in the Tonle Sap region and about frame; it is calculated to allow direct 9 percent in PlateauMountain areas. comparison with the value from this area in the 1994 survey. The higher value for Figure 3.2: Per capita consumption in upland areas 2004 in this chapter reflects the poverty is only 75 percent of the national average rate amongst the rural population as a - Per capita consumption by region indexed on all- whole, not merely that part of it that falls Cambodia average = 1009 2004 within the comparable sampling frame. 111.5 I5O I, Household consumption 100 j

Average levels of per capita 50 consumption vary greatly between 0 regions and between urban and rural -50 areas within regions. Figure 3.1 presents estimates of per capita household -100 J consumption by agro-ecological region, Plateau / Tonle Plains Coastal Phnom disaggregated into the urban and rural mountain Sap Penh areas within each region. The estimates Source: CSES 2004. of mean per capita consumption are

Figure 3.1: Rural Plateau and rural Tonle Sap regions had the lowest average of con sum ption rn 0

Plateau / Tonle Coastal Plains Phnom mountain Sap Penh

Source: CSES 2004. Poverty in Cambodia in 2004 37

Including the urban poor, the Tonle Sap incidence of poverty was highest in rural and Mountaidplateau regions combined areas (39 percent) and considerably lower accounted for 54 percent of the total poor in urban areas (five percent in Phnom Penh population. The low average consumption and 25 percent in other urban centers). values in these two regions are brought out For reference, the 2004 poverty headcount in Figure In particular, Kompong 3.3. according to the international, “dollar-a- Speu and Kompong Thorn had the lowest day” poverty line stood at 18.5 percent. provincial per capita consumption in the Throughout this report, however, poverty country, followed Siem Reap province by will be analyzed, as is standard practice, and Eastern Northern provinces in and the with reference only to the national poverty MountaidPlateau zone. lines described in the previous chapter. Box 3.1 describes the formulation of the Poverty in 2004 dollar-a-day international poverty line, In 2004, 35 percent of the population lived how it compares to the national poverty below the poverty line, and 20 percent line, and why the national poverty line is below the lower food poverty line. The preferable for national-level analysis.

Figure 3.3: Standards of living are lowest in the MountainlPlateau and Tonle Sap regions

0 50 100 Ai

Source: CSES 2004. 38 Cambodia - Poverty Assessment 2006

Box 3.1: Choosing a poverty line: national and dollar-a-day poverty lines compared Over the last decade, the international community has become increasingly interested in tracking poverty trends at a global level, primarily to enable monitoring of progress towards the first of the Millennium Development Goals (MDGs), which commits the world’s governments and international organizations to reducing the percentage of the world’s population living in absolute poverty to half of its 1990 value by 2015. To track global poverty, it is necessary to adopt a standard measure of absolute poverty that is constant across countries and over time. This need for an international reference point gave rise to the adoption of the so-called “dollar-a-day” line, which defines the poverty line in terms of the goods and services that could be consumed for one dollar per person per day in the USA in 1993. The full description of the international poverty line is thus “one dollar per capita per day, 1993 purchasing power parity (PPP)”. Defined in this way, the international poverty line attempts to account for changes in prices over time (i.e., adjust for inflation) and account for differences in prices between countries (Le., adjust for cost-of-living differences between economies). Because in most developing countries basic goods and services (particularly food) are much cheaper than in the USA, the actual value of one dollar a day, 1993 PPP in these countries is much less than one dollar in current prices. In Cambodia, the international (dollar-a-day) poverty line in 2004 was equivalent to 1,382 riels per capita per day (or US$0.34 in 2004 current prices). The “dollar-a-day” international poverty line is thus in real terms lower than the Cambodian national poverty lines described in Table 2.1 and used in this report, which range from 1,753 riels (US$0.43) per capita per day in rural areas to 2,351 riels (USfi0.58) per capita per day in Phnom Penh. As such, the international poverty line results in a much lower estimated poverty headcount for 2004 (18.5 percent) than that derived by reference to the national poverty line (35 percent)-even though both use the same reference dataset (the recall-based data on per capita household consumption derived from the CSES 2004). This disparity between a more inclusive national poverty line and the more severe international poverty line is not unusual. The level of Cambodia’s poverty line is comparable to that of most low-income countries, many of which have national poverty lines that are less than half a dollar a day in current prices, but which are nonetheless higher (and thus result in a higher poverty headcount) than the international, dollar-a-day PPP poverty line. In Vietnam in 1998, for example, the international, “dollar-a-day” poverty rate was estimated at 16.4 percent, while the poverty rate according to the national poverty line of 4,900 dong per person per day ($0.35 in current prices), based on the 1998 Vietnam Living Standards Survey (VLSS), was 37 percent. In Lao PDR, similarly, the international, “dollar-a-day” headcount estimate was 28.1 percent in 2002 and 25.8 percent in 2003. By contrast the 2002/03 poverty rate according to the national poverty line (3,065 kip per person per day, or $0.26 in current prices), measured in the third Lao PDR Economic and Consumption Survey (LECS Ill), was higher at 30.7 percent. The dollar-a-day international poverty line thus provides a much cruder definition of poverty than a well-crafted national poverty line. It serves an invaluable role in enabling rough comparisons between countries and over time, in tracking progress towards the MDGs; but for the purposes of national-level analysis and policy debate, it is far preferable to use a national poverty line constructed following accepted international principles and with reference to a country-specific consumption bundle. This report therefore uses the national rather than the international poverty line in the construction of poverty measures. Sources: Knowles 2005a; Deaton 2001 ; Deaton and Grosh 2000.

The south-westem Plains region has a the Mountains/Plains regions) but, by poverty headcount lower than the national virtue of the population density in these average (32 percent, compared to 43 Provinces, nonetheless contains 40 percent percent in the Tonle Sap and 56 percent in ofthe nation’s poor (Figure 3.4). Povertv in Cambodia in 2004 39

Figure 3.4: The majority of the poor are found in The left panel in Table 3.1 presents the Plains and Tonle Sap Provinces poverty headcount measures for all A. Distribution of population agro-climatic regions, disaggregated by urban and rural sectors, for 2004. The right-hand panel in this table presents the distribution of the poor population by zones and sectors. In terms of poverty severity, the rural Tonle Sap and rural Mountaid Plateau regions once again demonstrated the worst values, Plains Tonle bbuntain Phnom Coast averaging about 1.5 times and twice Sap /plateau Penh the national average, respectively. When the data are disaggregated 6. Distribution of population under the further, there was found to be wide poverty line variation between Provinces. However, the poorest Provinces (in terms of poverty headcount and poverty severity index) were still those in either the Tonle Sap or Mountaifllateau regions. The Provinces with the highest incidences of poverty were Kompong Speu, Rains Tonle Sap Mountain/ horn Coast Siem Reap, and Kompong Thom plateau Penh (Figure 3.9, while those with the deepest poverty were Siem Reap, Kompong Speu, and Kompong Source: CSES 2004, Neupert 2005. Thom.

Table 3.1 : Poverty headcounts are highest in the rural areas of the Tonle Sap and MountainlPlateau regions, which together account for almost half of all poor

Poverty headcount index Population distribution of poverty headcount (% of all poor) Zone Urban Rural Total Urban Rural Total Phnom Penh 1 9 5 0 1 1

Plains 14 33 32 1 39 40

Tonle Sap 28 45 43 4 33 37

Coastal 20 30 27 1 4 6

Plateau/ mountains 33 56 52 1 14 16

Cambodia 18 38 35 8 92 100

Source: CSES 2004. 40 Cambodia - Poverty Assessment 2006

Figure 3.5: Poverty rates are highest in Kompong Speu, Siem Reap and Kompong Thom

0 100 t

Source: CSES 2004.

Figure 3.6: Households headed by men and women in the prime of their life (aged 30-50 years) made up the largest proportion of the poor and experience the highest rates of poverty

50 - 36 38 38

age ~30 30-39 40-49 50-59 60-69 70 + age of head of household

Source: CSES 2004. Poverty in Cambodia in 2004 41

Fieldwork in Cambodia draws out that Household poverty profile there are differences, but that the category of female-headed household is, on its own, The poor are primarily in inadequate as a targeting category as there households headed by prime- are indeed both rich and poor female- aged adults headed households. What is needed is rather to disaggregate between different Figure 3.6 analyzes the incidence of types of FHH, looking also at the poverty by the age and gender of availability of other adult labor and household heads. The data suggests that dependency rations (Vlaar and Ahlers the highest incidence and greatest number 1998; Box 3.2). of the poor were among household heads of prime age (between 30 and 50 years). There was no difference in poverty Individuals in these households made up measures by marital status (84 percent of about 60 percent ofthe entire poor. heads were married and 13 percent were widowed). Nor was there any statistical In terms of the sex of the head of significance in poverty measures by household, the CSES data does not support ethnicity or reported disability of the contention that female-headed household heads (although this is mainly households experience greater poverty than due to the fact that the sample sizes of male-headed households (Figure 3.7). minority ethnic and disabled heads were This is not an unusual occurrence. In rather small). discussion of poverty and many countries in the world national A ethnic minorities is presented in Box 3.3. sample surveys fail to find any difference between male and female-headed A broad conclusion from the discussion of households, despite relatively consistent the poverty profile so far is that the most evidence from small-scale and qualitative effective way for poverty reduction would studies that there aye differences. be to promote broad-based growth, with a particular emphasis upon agricultural Figure 3.7: There is no statistical difference in the CSES data between the growth that would directly benefit the rural incidence of poverty in male- population which contains 91 percent of and female-headed households the nation’s poor.

100 - The poor lack human capital Male head 83 1 0 Female head Years of schooling and literacy of 75 J household heads are strongly related to poverty outcomes. The probability of being in poverty dropped significantly for 50 i 35 34 household heads with some years of schooling, compared to those with none at all. Mean years of schooling among the indigent were 2.75 compared to an average of 5-6 years for the two richest quintile poverty shares in total groups. Chapter 6 discusses the links headcount poor between human development, productivity and consumption poverty. Source: CSES 2004. 42 Cambodia - Poverty Assessment 2006

Box 3.2: Gender and poverty in Cambodia Anthropological and sociological literature has often stressed the “relative equality” between men and women in Cambodia, and Cambodian women have traditionally enjoyed a higher status and greater independence than, for example, women in India or China (Ledgerwood 2005). Marriage does not involve dowry, while bilateral kinship and a customary preference for a newly-married couple to live with the parents of the wife provides a measure of familial support (UNDP 1998). Women participate actively in the economy and the agricultural division of labor is flexible (and becoming more so). In many cases women are responsible for managing the family’s income and expenditure, with major economic decisions taken jointly. While relations between women and men in Cambodia may be described as relatively equal when compared to many other countries, gender inequalities continue to persist in certain aspects of social, economic and political life. Gender inequality has historically been most pronounced with regard to education (schooling for girls was seen as at best unnecessary and often undesirable); political authority (very few women held positions that might allow them to shape public policy), and the double burden that women face in managing domestic tasks, which are almost exclusively the responsibility of women, while also contributing to agricultural and/or income-generating activities. Occupational choices are strongly determined by gender, and there are very few women in management positions in any sector. In practice gender ideals are one of several factors (including age, wealth, reputation and power) that determine relative status. (Thus, for example, a married woman with children will have more freedom to speak and act her mind than would a young unmarried woman.) Gender relations are also, along with much of Cambodian society, in flux. Some aspects of traditional gender relations are changing for the better (although they still require attention). The gender gap in education and literacy is closing (see Chapter 6) and women’s participation in politics, while still marginal, is no longer entirely non-existent. Other aspects of economic and social change, however, have further disempowered women. Because significantly more men than women died during the wars of the 1970s (and to a lesser degree the 1980s), Cambodia has had a very unbalanced adult sex ratio. Marriageable men, being in short supply, found they could often get away with behavior vis-a-vis women that would not have been tolerated in the past. Although the sex ratio has now significantly balanced out, there is a legacy of changed attitudes, which may be one factor contributing to extremely high levels of domestic violence (see Chapter 7). The legacy of war also created female-headed households, both directly and indirectly (as surviving men found they could abandon one woman and still get remarried). While analysis of consumption poverty data does not show female-headed households to be any poorer than those headed by men, locality studies make it clear that certain types of female-headed households do indeed suffer particular economic disadvantages (UNIFEM et al2004). The public health system does not serve women well (see Chapter 6) as seen most strikingly in a high maternal mortality rate (437 deaths per 100,000 live births). Gender relations and values are also central to one of Cambodia’s major public health challenges, that of HIV/AIDS. Traditional tolerance of prostitution has posed a challenge to which the Government has responded well; however, while prevalence rates have come down due to promotion of condom use in commercial and casual sex, it is harder to promote their use by married couples, and husband-to-wife and consequently mother-to-child transmission is becoming increasingly common. Although surveys find that family planning decisions are typically arrived at jointly through discussion, participatory research suggests that women have limited power to negotiate safe sex and condom use in the context of marriage (CDRI 2006c forthcoming).

~ Finally, women continue to be concentrated in low-wage/low-income economic sectors and are paid less than men for the same work. The rapid expansion of the garment sector has been a major benefit to approximately 200,000 women (who comprise over 90 percent of the workforce in this sector), but this is a small percentage of the overall workforce. Female garment workers also face extremely high expectations from their families, remitting a very large proportion of their wages to their families, often at the cost of their own current consumption (ADB 2004; Dahlberg 2006 forthcoming; ADI/CCC 2005). Poverty in Cambodia in 2004 43

Box 3.3: Ethnic minorities and poverty Cambodia is relatively ethnically homogenous in comparison to other south-east Asian states. The majority group, Khmer, makes up some 96 percent of the population. The remaining 4 percent is a mix of groups with very different economic and social characteristics in terms of their relationship to the Khmer majority and mainstream development. The small size of ethnic minority populations relative to the majority makes poverty analysis based on the CSES very hard, as the number of non-Khmer households sampled was small, and for the smaller groups, in particular, poverty rates drawn from the CSES may not be statistically robust. The Cham (or Khmer Islam) account for 2.2 percent of the population, living generally in distinct villages (with a concentration in Kompong Cham and neighboring Provinces) or urban neighborhoods. They demonstrate some differences in dress and occupational specialization, but otherwise live in much the same way as the Khmer majority. Reinforcing this impression, in the CSES 2004 they were found to fall heavily (76 percent) in the three middle quintiles in terms of per capita household consumption. The Chinese (0.2 percent of the CSES sample) are generally wealthier than the national average; 71 percent of those surveyed in the CSES fell into the top two quintiles, and 58 percent in the top quintile alone. The Vietnamese (0.4 percent) are also typically wealthier than average (58 percent in the top two quintiles), but there are also significant pockets of ethnic Vietnamese living in poor conditions (e.g. in fishing communities on the Tonle Sap). The poorest but also the smallest of the ethnic groups in Cambodia are the highlanders (Khmer Loeu, chunchiet, or “highland tribal groups”) living in mainly forested upland areas, concentrated in the north-eastern Provinces of Ratanak Kiri, Mondul Kiri, Kratie and Stung Treng. The 17 main groups, each of which numbers only between a few hundred and 19,000 members, collectively account for some 104,000 persons, or less than 0.1 percent of the national population. To varying degrees, these groups face particular problems in terms of language barriers (which disempower them in relations with Khmer-speaking commercial and administrative interests): (i) extreme remoteness and low population density (and, once again, language barriers and a power imbalance) which makes it hard for them to obtain access to services such as schooling or healthcare; and (ii) a heavy dependence on forests, both for timber and non-timber forest products but also as a supply of land for shifting swidden agriculture. As commercial and state interests- and Khmer settlers from elsewhere in Cambodia-are increasingly attracted to exploiting the economic potential of the upland forests, highland ethnic minorities are at risk of losing access to forests and land. Low levels of literacy in Khmer and a lack of representation in local or national level government means that they find it hard to realize what rights they do have, as does an incomplete legal environment. The 2001 Land Law contains provisions for indigenous minority communities to claim title to their land, either as individual titles or community titles, but sub- decrees defining the legal requirements for recognition of communal ownership have yet to be drafted. These groups therefore tend to be marginalized and poorly placed to participate in policy dialogue and negotiations on development activities that affect their livelihoods. Small surveys among these groups find very high levels of poverty, very low levels of service provision and uptake, and extremely poor health and educational status in terms of illiteracy, prevalence of fever and diarrhea, contact with antenatal care, and child malnutrition. Reaching these small pockets of deep poverty calls for relatively intense investments in carefully-designed programs for service delivery (success has been achieved with dual-language primary education) and local empowerment. For these groups who are unlikely to benefit from the mainstream of national economic growth and poverty reduction, NGOs and UN agencies have played a major role and are likely to continue to do so for the foreseeable future. Sources: ADB 2002; NGO Forum 2004; NGO Forum 2005; Minorities Rights Group 1995; McAndrew and Oeur 2004. 44 Cambodia - Poverty Assessment 2006

This pattern is essentially similar to that Poor households tend to be observed in previous poverty profiles. larger and have higher dependency ratios Even after converting to an adult equivalence scale, average real Figure 3.8 presents poverty headcounts by consumption values fall with household household size. The data indicate that the size. For instance, the average real incidence ofpoverty becomes significantly consumption per equivalent adult for worse for households larger than 5 persons household of three members was 3,505 (which account for half ofthe population) Riel a day, but the average for a household of nine was only 2,583 gela day.

Figure 3 .8: Larger households tend to be poorer because of higher dependency ratios

A. Poverty rates increase with household size

1 2 3 4 5 6 7 8 9 IO+ number of individuals in household

B. Dependency burden: children and elderly as %of adults in the economically-active age range

Poorest Next Middle Next Richest poorest richest Consumption quintile groups

Source: CSES 2004. Povevtv in Cambodia in 2004 45

These larger households tend to have These patterns are broadly consistent with higher dependency ratios, that is, more the findings ofprevious poverty profiles. young children and elderly individuals to be supported by each adult ofworking age. Poverty rates are highest in The CSES confirms that dependency ratios remote areas are inversely associated with poverty. Access to economic infrastructure (roads and irrigation facilities) and social Households with farmers and infrastructure (basic services including domestic workers as heads tend modern energy sources, water and to be significantly poorer sanitation. and school and healthcare Figure 3.9 presents poverty estimates by facilities) is far more limited for the poor employment status of the head of than for non-poor. The poor and especially household. Poverty headcounts were the very poor (the bottom quintile that live highest among farmers and domestic below the food poverty line) tend to be in workers. Households headed by self- extremely remote and isolated areas where employed farmers and domestic workers their access to any infrastructure or basic accounted for 48 percent and 13 percent, services was limited. The average distance respectively, of the total poor. Households to infrastructure and basic services headed by domestic workers had the increases steadily as one moves from a highest poverty incidence as well in the higher consumption quintile group to a 1994 poverty profiles and account for an lower one. The food poor are especially increasing share of the total population (9 isolated from access to permanent markets percent in 2004). The lowest estimated and healthcare facilities. When the data poverty headcount was among households are disaggregated by provinces, Northern headed by someone in the public sector. provinces in the Mountaid Plateau zones (Le., Ratanakiri, Figure 3.9: Households headed by domestic workers Stung Treng, Preah Viheaer) and and farmers are much more likely to be poor Siem Reap province in the Tonle - Poverty incidence (%) by occupation of head of household Sap zone have the poorest infrastructure measured by Public sector percentage of villages with all- Self-employed, nonfarm 28 weather roads or those with gasoline as an energy source Private wage 9 (Figure 3.10). Other (unspecified) 32 Not in labor force 32 The poor have virtually no UnDaid family worker 32 access fo modern energy Unemployed sources and water and

S elf-em ploy ed , farm sanitation The poor rely heavily on firewood Domestic worker for fuel and kerosene for lighting, 0 20 40 60 while the non-poor have the luxury of connection to gas and Source: CSES 2004. electricity. While there were noticeable gaps between the poor and the non-poor in energy 46 Cambodia - Poverty Assessment 2006

sources, the deprivation in access to clean or public tap and only 3.5 percent of the water and sanitation was particularly acute poorest had access to decent sanitation among the poor. Barely 2 percent of the (Figure 3.1 1). poor households had access to piped water

Figure 3.10: The poorest infrastructure is to be found in Siem Reap Province and in

Figure 3.1 1 : The poor are acutely deprived of energy, water and sanitation

A. The poorest 40% have virtually no B. The poor are heavily reliant upon access to water and sanitation fue Iw ood and kerosene 75% 100%

50%

50% 25% 25%

Y! 0% n- I_ Poorest Next Middle Next Richest Poorest Next Middle Next Richest poorest richest poorest richest

Toilet-water sealed, connected to sewage or septic tan 0 Use firewood for fuel 0 Obtain light from kerosene Poverty in Cambodia in 2004 47

A substantial number of the Figure 3.12: The poorest and next poorest quintiles poorest in rural areas have no are less likely to have secure land titles access to land for cultivation 40% 7 Given that Cambodia is still primarily an agrarian economy with 71 percent I 29% of the labor force in agriculture, and where 91 percent of the poor are in rural areas, the most important assets for these rural poor will be labor and land. Because of limited human capital and lack of skill, the poor and the indigent were less able to make use of their labor to pursue non-farm Poorest Next Middle Next Richest employment opportunities. Thus, land poorest richest becomes the utmost critical asset for a rural poor. Source: CSES 2004. Landlessness-in which the rural landless are defined as households without any access at all to land for cultivation- has increased from 12.6 percent in CSES Urban poverty: lack of secured 1997 to 15.8 percent in CSES 1999 to 19.6 housing and insecurity of percent in CSES 2004. These households income did not own, borrow, or lease any land, and Although Cambodia is still an could not use communal land. In 2004, overwhelmingly rural society, there is about 15 percent of the poorest quintile reason to suspect that the CSES 2004 (in living in rural areas and 13 percent of the common with similar living standards next poorest quintile group were landless. surveys in other countries) may have under-sampled the urban poor, and that ... while those who owned land urban poverty still merits attention (Box had little security of land tenure 3.4). Among those who owned land for A number of surveys on urban poverty, cultivation, the indigent and the poor were mainly of Phnom Penh, have been carried less likely than the non-poor to have a out by UN-Habitat and NGOs. A survey secure land title. Figure 3.12 shows that by the Urban Resource Centre (URC) only 15 percent of the extreme poor held a revealed that, as expected, most residents land title compared to about 29 percent of of poor squatter settlements are employed the richest quintile group. The likelihood in the informal economy as motodop of obtaining a secure land title increases (moped taxi) drivers, construction workers, with income. Secure land rights and domestic workers, or small scale street efficient land administration systems are vendors, while a few work as permanent crucial for agricultural growth and for (or at least not fully casual) employees in facilitating exit from the sector. The the garment industry (mainly female) or benefits of secure property rights are restaurants, or as domestic help (a category widely recognized and well-documented in found in the CSES 2004 to have the international research. highest poverty headcount-49 percent- 48 Cambodia -Poverty Assessment 2006

Box 3.4: The urban poor: small, but contribution to the income of poor undercounted and growing? households, earning around 4,000 riels per Poverty statistics derived from CSES may day as rag pickers, shoe cleaners or under-represent the urban poor for a variety beggars, or remaining at home to look after of reasons. Firstly, many of the poorest in younger siblings while parents are out to urban areas (particularly Phnom Penh) sleep work. on the streets and would have been overlooked in the CSES sampling frame, The problems of the urban poor differ which is based on residences. Most of the somewhat from those of the rural poor. next strata (poor but not destitute) live in Whereas physically healthy people in the informal squatter settlements. While informal settlements existing in 1998 would have countryside can generally obtain building been included in the Census (which provided materials and water relatively easily, in the sampling frame for the CSES 2004), towns these generally need to be these settlements have been growing rapidly purchased; and while richer neighborhoods since then, fueled by migration from rural are connected to municipal services, poorer areas. NGOs estimate that the number of poor households in squatter settlements neighborhoods generally are not, and have doubled between 1997 and 2003. There will to purchase their water (for example) from thus be households within informal private vendors, at a much higher price. settlements which were not included in the Incomes are generally higher than in rural CSES sampling frame. Although not all areas, but so are costs. Many of the urban inhabitants of informal settlements are poor, many are. So while it remains true that poor depend upon casual labor for their poverty in Cambodia is and will remain for income, which is very variable, making it many years overwhelmingly rural, poverty in hard to save. Phnom Penh and other towns is probably somewhat greater than the CSES would If the key daily challenge for the urban suggest. Urban poverty is likely to become poor is the struggle to make enough increasingly important over time as the poor income, the overriding threat to their move to towns to seek alternative livelihoods. livelihoods and wellbeing is the security of The rural-urban shift is especially likely if, as over the last decade, the pattern of growth housing tenure. While the 2001 Land Law remains heavily biased towards urban-based gives existing residents some rights, these manufacturing and services, with agriculture are very limited for those who have settled and the rural economy more broadly lagging on what is classified as State Public Land behind. As landlessness increases and (as opposed to State Private Land) and access to common property resources declines, rural-urban migration can be even those with nominal tenure rights are expected to increase. Given that the vulnerable to effort to displace them in a problems of the urban poor are significantly context of rapidly rising land values, different from those of the rural poor, and Many feel that their lives have improved urban poverty needs to be tackled in the since moving from the countryside to the context of complex urban planning (both economic and spatial), there is a need for city, a move which brought them better specific policies and programs for the urban employment opportunities, better incomes, poor, to a significant degree distinct from and better access to health and education those designed for the rural poor. services. Given this, they have little l interest in the offer of land outside the and to account for a small but growing Municipality. Many in the squatter share ofthe population, and ofthe poor). settlements report having been able to improve their housing since they first As in the countryside, surveys in urban arrived. However, forced eviction (the areas find that children make a significant Municipality has evicted 11,000 families Poverty in Cambodia in 2004 49 from the city centre between 1998 and value to agricultural products and that 2003) can almost instantaneously reduce draw off rural labor, improving per capita them to a level equal to or worse than what land availability and sustaining agricultural they experienced before coming to the city wages and output prices. (UNOHCHR 2004; URC 2004). Crowded settlements on marginal land are Inequality is high also very vulnerable to community-wide It was mentioned in the previous chapter disasters such as floods or fires, which can that inequality within the areas that were reduce thousands of families to destitution surveyed in both 1994 and 2004 has risen overnight. In terms of social cohesion and considerably (from a Gini coefficient of the community basis for collective action, 0.35 to 0.40, when a Gini of 0 represents the urban poor are in some ways better off perfect equality, and a Gini of 1 represents and in some ways worse off than their rural perfect inequality). The figures for equivalents. On the one hand, poor inequality in 2004 are actually more communities have, with support from serious when the full sample is used as the NGOs, organized to demand provision of basis of analysis. In the whole of services or to oppose evictions. On the Cambodia in 2004, the Gini coefficient other hand, the poorest ofthe urban poor- stood at 0.42 (Figure 3.13). This puts it families living off reclamation from towards the higher, more unequal ends of municipal waste dumps, children or the regional scale (Figure 3.14). disabled persons without families living on Inequality in Cambodia is furthermore the streets, commercial sex workers-may unusual in that most of Cambodia’s high- suffer from extremes of social exclusion growth neighbors in South-east and East and vulnerability. Even amongst these Asia saw levels ofinequality start to widen groups, however, there is some early only at later stages of development, when evidence of group formation around levels of average income and consumption workplace-based or service-delivery issues were higher (and poverty headcounts (USG 2004; AFSC et a1 2001). lower). What is notable about Cambodia, One of the major areas in which information is lacking is the nature Figure 3.13: Inequality in Cambodia is high, of economic and social change in particularly in urban areas. secondary urban centers. Although 0.5 - urbanism in Cambodia is very 0.43 0.42 much a primate city model (there are few large cities other than Phnom Penh, which in 1998 was six times the population ofthe next largest Municipality), experiences in other countries suggest that provincial towns may play a crucial role in the development of their rural hinterlands, as nodes for Rural Other Phnom Cambodia the marketing of rural goods, a urban Penh sources of agricultural inputs, and as locales for the emergence of Source: CSES 2004. industries that process and add 50 Cambodia - Povertv Assessment 2006

Figure 3.14: Inequality in Cambodia is high compared to its regional neighbors

Indonesia (2000) 0 34

Vietnam (2002) 0 35

Laos (1 99718) 0 35

Thailand (2002) 0 40

Cambodia (2004) 0.42

Phillipines (2000) I IO46 Malaysia (1997) i IO49 - -- _. ~ 0.0 0.2 Gini 0.4 0.6

Source: CSES 2004; World Bank 2005. in other words, is the coexistence of In recent years, however, increasing significant and growing inequality with a attention to the connections between still-high level of absolute deprivation. economic growth, inequality and poverty Finally, the uncomfortable combination of has challenged or qualified a number of widespread poverty and conspicuous these axioms (Naschold 2002; World Bank inequalities acquires particular social and 2005). It now seems that high levels of political significance in light of a inequality not only (as expected) reduce commonly-held perception (as revealed in the poverty elasticity of growth (that is, the many fora, including but by no means percentage poverty reduction that arises limited to the MOPS fieldwork) that weak from one percent of GDP growth); high governance gives the wealthy unfair levels of inequality may also, it seems, advantages over the poor. negatively affect the rate of economic growth itself. The lessons to be drawn from international experience are open to debate, as findings There are a number of reasons for this. are highly dependent upon the choice of The first is that in imperfect markets, countries, reference periods and analytical inequalities in wealth translate into technique. Until perhaps the last five years unequal economic opportunities, leading to the consensus in international development wasted productive potential and an theory was that: (i)inequality inevitably inefficient allocation ofresources. In other rose with economic development and the words, individuals who may have the transition from “traditional” or socialist intelligence and application to make society to a market-based society; (ii) efficient use of capital, land and education however, changes in inequality, in either may, by nature of their or their parents’ direction, occurred only slowly; (iii) poverty, be unable to access these growing inequality was of little practical resources. significance so long as absolute Secondly and more subtly, entrenched deprivation (Le. poverty) was falling; and economic inequalities are over time (iv) there was little that public policy could strongly associated with political do to control growing inequality, certainly inequalities and impaired institutional not without impeding growth rates and development. In unequal societies, poverty reduction. institutions (political, economic, and Poverty in Cambodia in 2004 51 social) come to serve primarily to evidence suggests that the judicious use of perpetuate inequality by preserving the public action to prevent inequality from advantages of those with power and growing to very high levels is rational wealth. Such institutions distort incentives economic policy: and that, contrary to the in ways that are typically bad for accepted wisdom of earlier decades, there investment, innovation, and judicious risk- are policy tools that can be used to taking-and thus for long-term economic promote equality that do not impede but growth. rather promote economic growth and political stability (Killick 2002). These Interestingly, the economic case for less need not involve the politically difficult rather than more inequality derives very compulsory redistribution of physical strongly from the experience of assets or income. Investing in quality Cambodia’s neighbors. The high-growth universal education that will allow poor east- and south-east Asian tigers were and children to participate equally for jobs and to a significant degree still are opportunities, or in social protection distinguished by relatively low levels of measures (e.g. affordable health care) that inequality compared to, on the one hand, will help moderate risk-avoiding incentives middle-income Latin America (where for poor households, may all help slow the average growth rates have in recent growth of inequalities and increase and decades been lower than east and south- diversify economic activity. east Asia) or, on the other hand, low- income sub-Saharan Africa (where in many countries growth has on average Vu I nera biI ity been low or even negative for much of the last two decades). In trying to explain Estimating levels of vulnerability what made Korea, Taiwan, Thailand, The reduction ofthe poverty headcount by Indonesia or later China and (perhaps most between 10 and 15 percent over the period notably) Vietnam grow rapidly for several 1993/94 to 2004 is a significant decades, a number of analysts have cited achievement. However, a balanced low levels ofinitial inequality (particularly interpretation of this trend requires that a in land but also in average levels of number ofcontextual factors are taken into education and literacy) as a potential account. The first is that, as outlined in contributory factor. Chapter 1, Cambodia has over this period enjoyed some historically exceptional This is not to say either that absolute factors that have helped it reduce poverty. equality is ever feasible, or even that Improvement was relatively easy when equality necessarily should become a starting from a very low base, to which policy objective in its own right. There is, were added the time-bound benefits of first however, a strong case for taking the level quotas and now safeguards for Cambodia’s of inequality seriously as a factor that garment exports. As the fill effect ofthese affects both the rate of poverty reduction one-off boosts to growth wears off, and aggregate economic growth. At the Cambodia will need to find new sources of very least, policy-makers interested in growth and new ways to tackle poverty sustained economic growth and social and reduction. political stability should be concerned to monitor rapidly growing (or even static but The second factor that tempers the high) levels of inequality. A optimistic interpretation of the poverty comprehensive review of the international trend is that while many have escaped 52 Cambodia - Povevty Assessment 2006

Figure 3.15: A large proportion of non-poor Cambodians remain vulnerable to poverty - Distribution of per capita consumption relative to the national poverty line, 2004

4

3.5

3

2.5

1.5

1

0.5

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 household consumption (riels per capita per day)

Source: CSES 2004. poverty, a significant proportion of the small decline in the living standards ofthis non-poor remain vulnerable to risks that group to move them back down below the could push them back into poverty. One poverty line. Nearly 7 percent of way of illustrating this graphically is by households fall within a 10 percent band looking at the distribution of per capita above the poverty line; in other words, if household consumption (Figure 3.15). It the per capita consumption of these can be seen that a significant proportion of households were to decline by just 10 the population is clustered around-either percent, the poverty rate would increase by just below or just above-the poverty line. 7 percent from 35 percent to 42 percent. On the one hand, this is good news, in that it implies that many of the poor require Sources of vulnerability only a small, steady increase in Households in Cambodia face a variety of consumption to gradually lift them over the risks which can, individually or in poverty line. combination, push even relatively wealthy Conversely, however, it implies also that a households into poverty, and poor great many individuals in 2004 recorded households into destitution (World Bank per capita consumption that put them only 2005). Covariant shocks (that is, shocks a short distance above the poverty line. which affect many households at once, and While this means that they are classified as so are likely to overwhelm social coping currently non-poor, it would take only a Poverty in Cambodia in 2004 53 strategies based upon support within . More field- or farmer-specific crop families and communities) include: failure (due to localized flood, drought, or pest damage) or illness or death of Extreme floods and droughts. Floods . livestock. in 2000 affected some 3.4 million people, causing crop failures and Theft or violence, including domestic extensive damage to houses, livestock violence. and property, damaged public goods in . Life cycle events such as weddings, the form of infrastructure, and spread deaths, or births, each of which illness. Droughts in 2002 and 2004 involves some combination of financial also affected millions ofpeople. outlay, lost earning potential, and . Change in international trade affecting increased consumption needs. Some of comparative advantage of Cambodian these are at least partially predictable, goods and services. The garment making it somewhat easier to mitigate sector appears set to survive, although or cope with the impact, but some are probably not expand, following the end not. of MFA, with an extra lease of life Vulnerability to these shocks is from the EU and US safeguards against exacerbated, in some cases increasingly so, Chinese exports. However, these are by: (i)the limited asset base and savings set to expire in 2008, and if not of poor households; (ii) the replaced, then many of those in the underdevelopment of financial markets for garment sector face unemployment. saving, borrowing or insurance; (iii)the Tourism and construction are lack of diversification in many rural vulnerable to fears of SARS, avian flu, households (and communities); (iv) heavy political instability and terrorism. reliance on common property resources as Changes Thai economy or in in the either part of normal livelihood strategies rules on Khmer labor migration would or as safety nets, when access to or make cross-border wage employment productivity of these resources is in inaccessible for many, and less decline; and (v) a lack of rule of law and profitable for others. lack of guaranteed access to justice in . For poor urban squatter communities, conflicts between the poor and wealthier or evictions or fast-spreading more powerful actors. neighborhood fires are important community-wide shocks. Idiosyncratic shocks or individual-specific shocks which typically affect particular households include: Illness, which has been found to be a major factor in leading to land loss. Without a high quality, accessible, affordable public health system, illness, particularly an emergency illness or injury, can prove extremely expensive, both in terms of costs and foregone labor ofthe patient.

4. Sources of economic growth

During 1994-2004, economic growth averaged 7.1 percent per annum. The main engines ofgrowth were garment manufacturing and tourism. Preferential access to the US and EU markets from the mid-1990s drove the growth of garment exports to an average of 44 percent per annum over the whole period, while the inflow of international tourists, attracted by the ancient temples at Angkor, fueled the growth in tourism-related services (especially hotelsh-estaurants, transportation, and communication, and construction), which collectively grew by 10 percent per annum. However, both ofthe main engines of economic growth (Le. garments and tourism) were and remain urban-focused with limited linkages to rural areas. While its share in the economy fell from 46 to 31 percent between 1994 and 2004, Cambodia’s agricultural sector (inclusive of crops, livestock, forestry and fishery), the primary livelihood of the poor, continued to support more than 70 percent of the labor force in 2004. Agriculture grew by an average of 3.4 percent per annum over the decade, including three years ofnegative growth since 1999. Taking into account that the growth rate of labor supply in agriculture was 2.7 percent a year, labor productivity in the sector grew by a lackluster 0.5 percent per annum. The share ofthe industrial sector in the economy more than doubled to 29 percent, while that of services remained at around one-third. Despite their phenomenal growth and significant shares in the economy, the industrial and services sectors employed only 8 percent and 21 percent ofthe labor force, respectively, in 2004. Cambodia’s economy is also extremely dependent upon a few products. In 2005, garments accounted for 80.4 percent oftotal exports by value. There is a pressing need for economic diversification to broaden the foundations ofgrowth. Looking forward, in addition to its existing two main engines ofgrowth, Cambodia needs more pro-poor and rural-focused sources ofgrowth to accelerate poverty reduction. There are strong indicators that the agricultural and agri-business sectors offer considerable potential for higher growth, which would have a strong direct impact on the livelihoods ofthe poor. Unleashing this potential requires easing critical binding constraints. In the agricultural sector, the primary binding constraint is insecure property rights. The lack of secure land tenure inhibits investment because expected private returns to investments in land are not sufficiently appropriable. The second-order constraints are weak infrastructure, especially irrigation, and low human capital, which reduce returns to investment for society as a whole. In the agri-business sub-sector, the primary binding constraint is weak governance. Excessive unofficial fees either prevent profitability or render private returns too uncertain, and because unofficial fees tend to rise exponentially with the scale of operations, they inhibit expansion and growth. Access to and the cost of credit are second-order binding constraints. 56 Cambodia - Poverty Assessment 2006

his chapter presents Figure 4.1 : Economic growth has been driven by growth in a retrospective on the industry and services during the past decade Tthe key sources of i 14 - T 500 economic growth over the last decade, using the newly revised data on the National Accounts of Cambodia 1993-2004, published by the National Institute of Statistics. It then offers a preliminary analysis of the key binding constraints to future growth of the agricultural sector I -2 J - -200 (including agri-business), in which the vast majority of 0 Agriculture (LHS) the poor live and work’. ’

Retrospective on Source: National Accounts of Cambodia, 7993-2004; National economic growth Institute of Statistics.

Despite the turbulent past and the ensuing annually since2. The main determinants of depletion of physical, human, and social economic growth since the mid-1990s capital, and reflecting the benefits from the have been a relatively stable threefold transition begun in the early macroeconomic environment, including 1990s (Chapter l),economic growth in favorable external conditions and markets, Cambodia over the period 1993-2004 has generally prudent domestic financial been impressive, averaging 7.1 percent per policies, and the creation of critical market annum (Table 4. 1 and Figure 4.1). Key economy institutions. However, in terms of initial factors in this stellar growth physical output, economic growth has performance included: (i)the end of two- derived from a very narrow base and-a-half decades of conflict, which comprising the garment, tourism, and enabled a resurgence in savings and construction sectors, and to a much more investment; and (ii)the formation of a new limited extent, the agricultural sector3. Government with market-oriented More importantly, while both real per institutions and policies, including most capita income and real per capita notably the liberalization oftrade in goods, household consumption (i.e. allowing for services, capital, and labor. Real GDP growth hit a record high of 12.6 percent in 2 1999, and has averaged 6% percent Real growth in 2005 is expected to be around [6%] percent, as the anticipated downturn in garment production following the end of the quota system on December 31, 2004 (see below), has not occurred. A more comprehensive analysis of the origins and determinants of economic growth over the Excluding garments, construction, and tourism, past decade and the key binding constraints to real economic growth would have averaged fiture growth are currently being prepared in only 4.2 percent per year over the period (and an accompanying study on the “Sources of only 2.7 percent if agriculture is also Growth”. excluded). Sources of Economic Growth 57

Table 4.1 : Sources of growth by main economic activity, 1994-2004 Share of GDP Annual percentage Contribution to GDP 11 change 1994 2004 Annual 1994 2004 Annual 1994 2004 Annual average average average 1994- 1994- 1994- 2004 2004 2004 Ag r icu It u re 45.9 30.9 38.8 9.9 -2.0 3.4 49.3 -8.7 20.2 Crops 17.7 15.0 16.8 2.7 -3.4 5.2 5.5 -7.4 11.3 Paddy rice 9.9 7.0 9.2 2.4 -12.3 3.7 2.8 -13.7 4.2 Other 7.8 8.0 7.5 3.1 6.0 7.1 2.8 6.4 7.1 Livestock 7.9 5.0 6.5 -2.9 4.3 1.8 -2.8 2.9 2.2 Fisheries 13.0 8.8 11.5 4.1 -3.3 3.0 6.0 -4.2 6.2 Forestry 7.3 2.0 4.0 87.5 0.2 3.9 40.5 0.1 0.5 Industry 13.6 28.9 20.1 14.2 16.1 15.4 20.0 56.3 41.1 Mining 0.2 0.3 0.2 29.2 9.1 10.8 0.7 0.3 0.3 Manufacturing 8.0 21.8 14.4 9.0 17.4 17.4 7.9 45.3 33.8 Garments 0.9 16.3 7.5 25.1 24.9 43.6 2.0 45.6 28.9 Agri-business 5.2 3.3 4.7 8.2 -3.4 2.7 4.7 -1.6 2.4 Electricity, gas 0.3 0.5 0.4 8.6 4.6 11.9 0.3 0.3 0.7 and water Construction 5.0 6.4 5.1 23.4 13.2 11.9 11.2 10.4 6.4 Services 35.4 34.4 36.0 0.6 9.2 6.1 2.6 40.7 31.3 Hotels and 2.6 4.7 3.7 19.5 23.6 14.6 5.0 12.5 6.6 restaurants Other 32.8 29.7 32.3 -0.6 7.3 5.3 -2.4 28.3 24.6

Taxes less 5.4 6.7 5.9 80.8 15.7 16.2 ’ 28.6 12.7 9.2 subsidies GDP 100.0 100.0 100.0 9.2 7.7 7.1 100.0 100.0 100.0 Memorandum items: GDP excl. 99.1 83.7 92.5 9.1 4.9 5.5 98.0 54.4 71.1 garments GDP excl. 91.6 72.6 83.7 8.1 3.2 4.8 81.7 31.6 58.1 garments, hotels and restaurants and construction Source: National Institute of Statistics. I/The contribution to GDP is calculated as the annual percentage change times the share of GDP in the previous period. GDP growth is set equal to 100 each year. inflation) have risen over the period, the effective economy-wide growth-generating rate of poverty reduction has not been economic policies and management by an uniform and inequality within rural regions effective and responsive state. A number and between rural and urban areas has of studies conclude that weak formal increased (Chapter 2). institutions and excessive, unclear and overlapping regulations impose very high Moreover, the impressive performance of costs (formal and informal) on households the two main engines (garments and and firms attempting to do business in tourism), owed more to fortuitous Cambodia. Thus, for example, a 2003 circumstances and narrowly-based survey of 800 Cambodian firms from “enclave-type” development, than to across the range of formal and informal 58 Cambodia - Povertv Assessment 2006 sector found that the share of sales revenue firms had to pay as bribes was twice that in Ag ricu Iture Bangladesh. In its first rating for The pattern of agricultural growth Cambodia, Transparency International in 2005 gave Cambodia a score of 2.3 out of Cambodia’s agricultural sector, inclusive a possible 10 (where 10 is the least corrupt of crops, livestock, forestry and fisheries, country). This places it at 130 out of 159 grew at an average 3.4 percent annually countries ranked, or in the bottom fifth over the period 1994-2004, although the (World Bank 2004; EIC 2005; World Bank volatility of that growth was high. Since and IFC 2006; Vallely et al 2006). In 1999, the sector has gone through three agriculture and the non-agricultural rural separate years of negative growth, affected economy, the challenge is to create trade- by political unrest, the regional financial supporting institutions to overcome market crisis in the late 1990s, severe floods in fragmentation, while avoiding the 2000, and drought in 2004 and 2005. The problems of inappropriate institutions that average annual rate of growth recorded are associated with the investment climate over this period is significantly less than for the urban, formal sector. This theme of the industrial and service sectors of the economic governance is explored in more economy, and agriculture’s share of GDP detail in Chapter 7. fell from 45.9 percent to 30.9 percent. Looking ahead, the fragile foundations for However, the proportion of the labor force economic growth in Cambodia remain having a primary occupation in agriculture vulnerable to exogenous tremors stemming remained at over 70 percent. from lower global growth, higher oil Though there was no significant change in prices, the accelerated recurrence of the overall structure of the agricultural flooding and droughts in recent years, and sector over the period, there were the potential for an avian flu pandemic to differences in sub-sector contributions to severely curtail tourism-generated growth. agricultural growth (Figure 4.2). On In 2005, garments accounted for 80.4 percent Figure 4.2: Sub-sector contribution to agricultural sector of Cambodia’s total growth exports, making Cambodia B 51 one of the countries in the p 4- n world most heavily L 8 3J dependent upon a single (A export-dependent, with the -E 2- 2 r: potential vulnerability that ’ 1- this entails. In these s 2 0- circumstances, sustaining 9 i: growth and making it more *3 -1- pro-poor, especially in rural U areas where the poor predominantly live, poses I:: significant institutional and policy challenges for the Cambodian Government. Source: National Accounts of Cambodia, 1993-2004, National Institute of Statistics. Sources of Economic Growth 59

Table 4.2: Cambodia's agricultural performance has been lackluster compared to neighboring countries in their early stages of development

Cambodia China Lao Vietnam Indonesia Malaysia Thailand PDR Comparison period '93-'03 '80-'90 '90'0 '90-'00 '70-'80 '70-'78 '65-'75 (respective early 0 stage of development) Average growth 3.3 6.2 4.6 4.3 4.5 5.0 5.0 rate of agriculture value-added (% p.a.) during comparison period Per capita GDP-(constant USD) at beginning of 266 180 227 227 245 1103 405 comparison period at end of 314 364 326 397 397 1648 614 comparison period Source: World Development Indicators. average, crops contributed the most (56 weather roads; weak human capital; and percent), while the contributions from the lack of access to, or high cost of, fisheries (30 percent) and livestock (11 capital. percent) were more modest4. The Within the livestock sub-sector, poultry contribution from forestry has been and swine production have each grown at slightly positive over the whole period just over 2 percent per annum, slightly (just over 2 percent), but significantly higher than the rate of large ruminant negative since 1999. production (1.7 percent). In value terms, Within the crops sub-sector, rice paddy has poultry is still the smallest of these three remained by far the predominant activity; livestock activities, and an outbreak of accounting for about half of gross value avian influenza is unlikely to exert a large added from crops, with horticultural crops negative impact on overall growth of the combined being the next most important sub-sector, although a pandemic could category at 10 percent of crop value. No exert a very negative impact on tourism. other crop accounts for more than 4 Total real value added from fisheries is percent of total crops, despite upswings in slightly higher than that from rice maize and soybean production since the production. Within fisheries, the inland late 1990s, which are responding to the capture component accounts for over healthy regional growth of livestock three-fourths ofthe value of production but sectors, analyzed in more detail below, As is the slowest growing, followed in size by the key binding constraints to higher yields marine capture (13 percent), which has and diversification of crops are: insecurity expanded rapidly, then aquaculture (9 of land tenure; poor irrigation and lack of percent). Most fish catch is consumed other critical infrastructure, including all- domestically and exports-between US$40-50 million per annum from 1999- Relative contribution to overall agricultural 2003, mostly inland to Thailand-are growth normalized to equal 100 each year (for example, the contribution of crops in 2003 is almost all unprocessed. Official statistics (3.6/4.0)*100=90 percent). tend to undercount the total fish catch, and 60 Cambodia - Povertv Assessment 2006 a significant proportion of marine fish Main determinants of agricultural catch is sold on the open sea and is growth therefore not captured in official statistics. In the last decade, the contribution of There are indications that the inland agriculture to overall economic growth has capture from the Tonle Sap basin is come largely through accumulation of reaching the limits of exploitation for factors of production-land and labor-as certain types of fish, and notwithstanding part of extensive growth of activity, with some improvements, conflicts over only modest improvement in productivity ownership rights also continue to inhibit from very low levels. Rice production is higher growth. still the overwhelmingly predominant crop, Though the forestry sector grew at an but some diversification and regional average rate of 3.9 percent per annum specialization may be emerging as farmers during the past decade, this performance is take advantage of ago-ecological heavily skewed by the exceptional growth characteristics of different agricultural in 1994 and 1997. Since then, the sub- systems and ofmarket opportunities. Total sector has contracted sharply, by an investment (public and private) in the average of 7 percent every year. The main agricultural sector over the last decade has reasons for this contraction include the been anemic overall. moratorium on commercial logging, Land is being brought into agricultural use, deforestation, and poor management of mostly via conversion from forest land, concessions. which is averaging about 1 percent per Not only has agricultural growth in annum (Figure 4.3). The largest increase in Cambodia been low relative to its absolute area was for cereals, although this industrial and service sectors, it has also was also the lowest proportionate increase, been unimpressive relative to its neighbors while the fastest growth in proportional during comparable stages of development terms was for oilseeds, pulses and roots (Table 4.2). Many East Asian countries' and tubers. Vegetables had the lowest agriculture sectors registered an annual growth in both absolute and proportionate growth rate of 5 percent or more over a terms ofall the major crop groups. decade or longer. For example, Thailand In Cambodia, only 7 percent of arable land and Malaysia during the late 1960s through is irrigated, well below the 20-30 percent 1970s, and China during the 1980s and range in most neighboring countries (Table beyond. 4.3).

Figure 4.3: Land is being brought into agricultural production, mainly through conversion from forest - '000 hectares

4,000 3,000 ...... 2,000 ...... - 0 !.."...... ? 78 1980 82 84 86 88 1990 92 94 96 98 2000 02 - Arable and permanent crops ...... Crops

Source: FAOSTAT Sources of Economic Growth 61

Table 4.3: Compared to its neighbors, Cambodia has very little irrigated arable land

Irrigation land Arable land Irrigation land (‘000 ha) share, % 1990 2002 1990 2002 1990 2002 Cambodia 0.2 0.3 3.7 3.7 6 7 China 48.0 54.9 123.7 142.6 39 39 Indonesia 4.4 4.8 20.3 20.5 22 23 Lao PDR 0.1 0.2 0.8 0.9 17 19 Malaysia 0.3 0.4 1.7 1.8 20 20 Myanmar 1.o 2.0 9.6 9.9 11 20 Philippines 1.6 1.6 5.5 5.7 28 27 Thailand 4.2 5.0 17.5 15.9 24 31 Viet Nam 2.9 3.0 5.3 6.7 54 45 Source: FAOSTAT

Cambodia has only added about 20,000 although some labor also migrates abroad, hectares of irrigation infrastructure over mostly for seasonal or temporary the 1990~~and no large scale irrigation employment (Box 4.1). schemes have been undertaken since 1974 Productivity gains. The agricultural sector (although allocations to irrigation in the grew very little through increases in last budget have risen). At present, only productivity. The simplest measure is about 10 percent of the rice area, or physical yields of output per unit of land 256,000 hectares (and much less during the andor labor input. While yields of a dry season) are effectively irrigated couple of crops such as rice and maize because of the general dilapidated status, have grown during the past decade, the poor initial design, and inadequate majority ofcrops did not (Table 4.4). management of irrigation infrastructure. Thus, most of agriculture is dependent on Cambodia’s relatively modest gains in the vagaries of rainfall, with attendant yields, overall, are telling in light of other higher risks to the use of purchased inputs neighboring countries’ performances. for small farmers, and reduced capacity to Value added of agricultural output undertake crop activities during the dry expressed in terms of labor and land season. productivity for Cambodia were the lowest among Asian countries in 1978 and Labor in agriculture is also growing. The achieved very little progress over the number of working-age adults for whom subsequent two decades to date (year agriculture is the primary sector of activity 1999), leaving it even firther behind. has grown by over a quarter over the last decade, or by about 2.7 percent annually, to 4.9 million in 2003. The high demographic growth rate in Cambodia, and limited absorption of labor into the secondary and tertiary sectors of the economies, leaves the agricultural sector to absorb a growing national labor force (about 250,000 new entrants a year), 62 Cambodia - Poverty Assessment 2006

Box 4.1: Internal and cross-border labor migration has become important for many A mixture of push and pull factors have led a number of farmers to turn to waged farm and off- farm work, both within Cambodia as well as across the border in Thailand. The push factors relate to slow growth or stagnation in the rural economy in much of Cambodia. Increasing landlessness, inadequate irrigation and transport infrastructure, imperfect credit markets, and the pressure of rising input costs not offset by decreasing or stagnant prices for outputs have all made life harder for many rural households. Against this is the pull of seasonal or longer-term casual employment in local towns, Phnom Penh, or Thailand. Both men and women are involved in seasonal waged agricultural work. In the dry season, men usually migrate seasonally to search for jobs as construction workers in Phnom Penh and other towns, while women work in garment factories or in the service industry. Remittances from male migrants are not as reliable, as garment workers have a more stable monthly income than construction workers who are hired and paid on daily basis. The importance of short- or long-term migrant labor is especially pronounced in border Provinces such as Battambang and Banteay Meanchey, where migration has become an important livelihood coping strategy or an alternative source of income for the poor in many communities in the last four or five years. In some communities in Battambang, about 70-80 percent of the economically active population are working in Thailand. For some households, migration is now the main source of income, while almost all households have some members who sell their labor when not engaged in farming. Because of out-migration, people in this community can earn money to support their families by selling strong, unskilled labor. We worked in Thailand for a month and a half and we brought home around 3,000 baht (approximately US$75) after spending for our daily needs and other expenses such as a work permission letter. MOPS focus group discussion, Battambang. Migrant labor in the border Provinces can be divided into three types: in order of importance, they are farm work at the Thai border, farm work at the Khmer border, and non-farm work deep inland in Thailand. Working on the Thai side of the border is more lucrative, and workers often benefit from free fruit and vegetables and a generally lower cost of living. Migrant work in Thailand has been recently facilitated by the offer of formal work permits by the Thai government. Seasonal waged farm work has become increasingly significant in Malai district in Battambang Province, where it is estimated that about half of the workers migrate seasonally to work either within Malai or in Thailand. Most can find employment for on average 6 months a year, earning approximately US$270, compared to the average rice farmer's annual income of US$lOO. Source: CDRl 2006a (forthcoming); EIC 2006 (forthcoming).

Specialization. The Plain and Tonle Sap example, the production of maize is zones dominate in terms of total crop concentrated in Battambang, largely in production value, and these two zones response to cross-border market links to account for five to eight times the value the Thai livestock economy, while the generated in Coastal or Mountaifllateau production of soya, which is mostly zones. All the zones, with the exception of transported to southern Vietnam for Phnom Penh, get three-quarters or more of processing (mostly small-scale), is largely their total crop value from rice production, concentrated in the Plain province of indicating that crop specialization, has yet Kompong Cham. to exert a significant influence on the Investment. Annual public investment, structure of production. Nonetheless, inclusive of both domestic and donor anecdotal evidence indicates that funds, in the agricultural sector has grown specialization may be emerging. For in recent years but remains very low, at Sources of Economic Growth 63

Table 4.4: Most crop yields did not improve over the last assistance has been largely decade, with the exception of maize disappointing. Actual disbursements of foreign Area Yield (MTIHa) direct investment (FDI) in harvested agriculture, estimated by the 2004 1985 1995 2004 IMF, peaked in 1996 but ('000 ha) subsequently sputtered at around 2-4 percent of GDP Rice Paddy 2,093.0 1.2 1.8 2.0 during 2000-2004. In terms of Maize 91 .o 0.9 1.2 2.8 shares, FDI going to the Sweet Potato 8.5 3.0 4.2 4.0 agricultural sector peaked at Cassava 22.0 2.1 6.6 6.4 21 percent in 2002, but Beans, Dry 35.5 0.5 0.8 0.7 dropped back to 1 and 7 Soybeans 29.9 1.3 1.o 1.3 percent, respectively, in 2003 and 2004. Finally, according Groundnuts in Shell 12.0 0.6 0.8 0.8 to CDC data, only five Coconuts 13.0 4.8 4.8 5.5 agricultural or agri-business Seed Cotton 0.2 1.I 1.5 1.4 investment projects were Mangoes 3.5 10.0 12.9 10.0 approved in the four-year Coffee, Green 0.4 0.5 0.8 0.8 period up until end-2004 (two Pepper, 0.4 3.8 6.3 6.9 FDI and three joint ventures). No domestic investment White/Long/Black projects were approved. The 0.4 1.6 1 1.5 Jute, Yield .o relatively low levels of total Tobacco Leaves 9.7 0.5 0.8 0.8 investment in agriculturally- Natural Rubber 43.5 0.9 0.8 1.I based activities in recent years Sugar Cane 10.0 21.1 27.3 22.0 are a clear indication that the returns to investment are low. Source: FAOSTAT. Institutional capacity: about 1.4 percent of agricultural GDP (or Returns to investments in the agricultural about 0.5 percent of total GDP). The main sector are also reduced by the still-limited vehicles for such investments in the sector provision of non-infrastructure public were the programs of the Ministry of goods by the Government as capacity Agriculture, Forestry, and Fisheries continues to be built in two key areas, (MAFF), the Ministry of Water Resources agricultural technology (generation, and Management (MOWRAM), and the adaptation and dissemination) and public Ministry of Rural Development (MRD). regulatory capacity. Public extension Meanwhile, foreign aid flows have been services have yet to build effective highly skewed away from agriculture, and capacity for reaching farmers. Less than despite the size of the agriculture sector, one percent of farmers have access to any the share of foreign aid going to the form of extension services (Castellanet agricultural sector since 1999 has only 2003). The lack of regulation of input been in the 8-10 percent range. Most of markets, where there are often strong and this funding has been for technical important information asymmetries, makes assistance, often for institutional capacity it difficult for farmers purchasing inputs to building, and the impact of technical know if seed, seedlings, and breeding stock 64 Cambodia - Poverty Assessment 2006 are of good quality, whether fertilizer has between US$300-$3,000 per year with the labeled nutrient content, and whether three provinces accounting for 95 percent pesticide content is as advertised. Falsely of total output: Kompong Speu (70 labeled fertilizer with lower element percent), Svay Rieng (16 percent), and content, seed with poor germination rates, Battambang (9 percent). All three are and marketing of expired or supposedly located along major international trade banned pesticides, are all barriers to higher routes; a powerful indication of the impact productivity. oftrade linkages to output. Industry By all indications and by any measure, agri-business has underperformed over the The industrial sector has been the fastest last decade or so, growing by an average of growing in the past sector Cambodia over only 2.7 percent per year, and its share in decade, at an average of 15.4 percent per GDP has declined from 5.2 percent in 1994 year. Its share in the economy has more to 3.3 percent in 2004.5 Contrary to the than doubled to 28.9 percent in 2004, and experience in many other developing the sector contributed over 41 percent of countries, and for reasons explained below total GDP growth over the period, and now associated with poor governance, the agri- employs about 8 percent ofthe labor force. business sub-sector in Cambodia has not Within the industrial sector, garment yet played a key role in forging and manufacturing and construction were the expanding the links between farming and primary growth engines, contributing on industry, and thereby accelerating the average, 28.9 percent and 6.4 percent, transition to a more modern and diversified respectively, to overall growth. Before economic structure. analyzing the underlying factors for the phenomenal growth of the garments sector, As indicated above, paddy rice production the next section reviews the performance contributes 50-60 percent of total crop of the agri-business sub-sector, which has production by value-added but contributes suffered from the inability of the little to GDP on a per ton basis, due to an agricultural sector to move up the value almost complete lack of processing for chain into processing activities, even as the export, despite a global market of about production ofprimary inputs has increased. 12-15 times Cambodia’s surplus production. Rice exports vary wildly with Agri-business no clear pattern over time as these data are not officially reported, and raw Cambodian Agri-business consists primarily of tens of paddy is often supplied to Thailand and thousands of micro-enterprises, a few Vietnam for processing. hundred small and medium enterprises (SMEs), and only a handful of companies with more than 100 employees. Only about one and one-half percent ofthe labor force is involved in agri-business, with the average micro-enterprise consisting of 2-3 workers. Among the SMEs, rice milling is by far the most common activity, while other grain mills such as bean and potato The National Accounts classify agri-business powders are a distant second. Estimates of activities under “Manufacturing”. They value-added per worker vary hugely comprise food, beverages, and tobacco, and paper and rubber manufacturing. Sources of Economic Growth 65

Partly because of low economies of scale failed to take-off. The most well-known from small and fragmented rice millers, exception is that of Angkor Kasekam, low yields and levels ofmilling efficiency, which operates on the basis of contract and high operating costs and unofficial farming arrangements, (along with British fees, returns to commercial milling of American Tobacco (BAT) in the case of paddy rice are very low. Paddy traders are tobacco). Contract farming arrangements small independent operators, transacting have helped ease some ofthe constraints to with just a few farmers and millers they processing activities (Box 4.2). know well, and widespread milling has

Box 4.2: Contract farming - Fragrant rice and tobacco Fragrant Rice. A variety of fragrant rice, Neang Malis, is cultivated free of chemical pesticides and fertilizers in a carefully selected area in central Cambodia where rich soils provide an optimal growing environment. The quality of fragrant rice largely depends on its planting area in four provinces in central Cambodia concentrating in Kampont Speu (with 80 percent of participating farmers) followed by Kandal, Kampot, and Takeo. On average, contracting farmers are able to harvest approximately 2 tons of rice per hectare, or up to 3 tons per hectare with improved irrigation system. Nearly 80,000 families are contracting with Angkor Rice Company (2004) to grow Neang Malis. Since there are no middlemen, Angkor Rice has control of the entire rice production process. To become a member, each farmer enters into an agreement with Angkor Rice to closely follow the growing protocol. At the start of planting season in July, Angkor Rice distributes high quality Neang Malis seed to contract farmers, who are not allowed to plant their own grains. During the growing season, Angkor Rice staff regularly visit the rice paddies to provide additional instructions and support to the members. Members harvest the crop and bring it to the rice mill, where a quality-control team checks the quality of rice grains. Only rice lots that meet the company’s high standard will be processed at the rice mill so that only the best rice is marketed. Pungent Tobacco. Since launching its Cambodia activities in 1996, British American Tobacco (BAT) has expanded its contract farming of tobacco to some 800 tobacco growers in the Kompong Cham province. Farmers and BAT agree to individual, renewable annual contracts. BAT has provided free tobacco seedlings, assistance growing wood lots for fuel for on-farm drying barns, organized technical assistance through village-stationed instructors, and makes credit available for agricultural investment. Farmers agree to sell a specified quantity of tobacco leaf at pre-set prices, and are free to sell any production above these levels on the market. Assisted by these contractual arrangements and services, BAT’S contract farmers have more than doubled their yields, reduced their costs by almost half, and considerably increase their incomes. A number of these farmers have been able to buy tractors and water pumps. With an improvement in the quality and quantity of tobacco being produced by its contract farmers, BAT Cambodia now sources about three-quarters of the tobacco needs for its local factories from domestic supplies, and has even begun modest exports of semi-processed leaf to Sri Lanka, Singapore and Australia. But the domestic cigarette market is saturated, the company complains of unfair treatment vis-a-vis domestic competitors, and export growth for tobacco leaf would depend on continued improvements in quality and favorable international prices, so no further expansion of the tobacco contract farming model appears on the horizon. Sources: Angkor Kasekam Roongroeung Co. Ltd. Homepage, www.anakorrice.com; Ung, Luyna and Sopheakalyanika Sras, SNEC, February 2005. Economic Institute of Cambodia, Interview with Corporate & Regulatory Affairs Manager of British American Tobacco (BAT), January 2005. 66 Cambodia - Poverty A ssessm en t 2 006

Because of measurement issues, it is all of domestic seed oils are imported impossible to accurately verify the growth despite domestic cultivation of soybeans rate of rubber manufacturing.6 Cambodia and sesame seeds which are informally has about 65,000-70,000 hectares ofrubber sent to Vietnam for processing and re- production near the Vietnamese border, importation. Finally, Kampot province is producing 42,000-45,000 tons of rubber reported to produce high quality organic per year, which contributed US$13-17 pepper, once considered by the French million to GDP each year between 1998 culinary to be the best in the region, but and 2003, almost all in the form ofexports. with only a US$2.2 million crop, it also Tobacco manufacturing leveled off after remains an underdeveloped agri-business 2001 after a 15 percent growth per annum activity. in earlier years.7 In 2004, tobacco manufacturing constituted 9 percent of Garments agri-business and 0.3 percent of GDP in Cambodia’s performance in garments 2004. Wood product manufacturing also production over the last decade was both showed an absolute decline from 1993- unpredicted and remarkable. In constant 2004, reflecting in recent years, the decline prices, total production of garments in logging as part of the effort to preserve increased by an average of 56 percent per Cambodia’s forests. However, there is year between 1994 and 2004.’ Virtually , some doubt as to whether this decline is all garment output was exported to the real or reflects under-reporting of illegal preferential US and European Union (EU) logging. markets. A quarter million people are directly employed in the sector. The vast Beverages constitute one of the smallest majority of the garment factory workers sub-sectors of agri-business. Its share fell (85 percent) are women from rural from 15 percent of agri-business and 24 villages. Most of these workers remit a percent of food processing in 1993, to only significant part of the salaries back to their 6 percent of agri-business and 9 percent of families (Box 4.3). food processing, due to rising input costs, pricing pressure, and competition from According to the 2004 CSES, about 13 imports. Processing of other agricultural percent of rural households received some crops remains small and underdeveloped. remittances from family members working In particular, maize has had strong growth in cities, and the annual average amount of largely serving Thai animal feed markets, remittances (327,000 riel), among but very little is processed for export. receiving rural households, constituted less Some 10,000-15,000 tons of raw cashews than 10 percent of total household flow informally to 80 processing factories consumption. The extent of remittances in Vietnam, the dominant exporter to the among the poor and the indigent was very 160,000 ton annual global market. Almost similar to average rural households.

In 1998, the source for rubber value-add measurements switched from NBC/Customs to MAFF rubber directorate, sharply increasing the volume of rubber counted; at the same time, world rubber prices dropped, reducing per ton value. This makes the numbers of pre- and post-1998 impossible to compare. ~ ’ This growth was not even; there were two 8 Classified as “wearing apparel” in the National years with declines. Accounts. Sources of Economic Growth 67

Box 4.3: Garment factory jobs and remittances: the view from the factories and the villages While the impact of rapid growth in export-oriented garment manufacturing on overall GDP is extremely impressive, it is less clear how deeply or widely this sector-specific growth has affected the poor. Surveys have confirmed that those employed in garment factories-almost all of whom are young women-may be poor but are rarely the very poor. This reflects in part the significant costs (in terms of transport and bribes-200,000-300,000 riel or US$50-US$75-paid to supervisors) that are often required to secure a job. A factor that appears to distinguish migrant- sending families from others in the same villages is their ability to obtain credit to cover these entry costs. (Social networks are also important; those with contacts are less likely to have to pay commissions to secure a job.) The allocation of the benefits primarily to the poor or middle- income groups rather than the poorest also reflects the power of employers, faced with a considerable over-supply of labor, to pick and choose only those with above-average education. Nonetheless, it was notable that in those MOPS and EIC case study communities in which migration was important, having a daughter working in the garment industry was considered to be one of the primary factors in allowing families to move out of poverty (and more important than access to migrant work in other sectors such as construction or agriculture). One MOPS focus group respondent noted: Those families who have daughters are lucky because they can earn some money from their children’s labor... About 70-80 percent of girls aged from 16 to more than 20 have gone to Phnom Penh or other urban areas to work as garment workers or housemaids. They can release some burdens from their family because on the one hand their parents don’t have to worry about spending on them and on the other hand their parents can get remittances from them. For most, the cost of getting a job is considered a worthwhile investment as workers are then able to earn between US30 and US$80 a month. A large amount of this (averaging around US$30 a month) is sent home to the workers’ families. Given that there are approximately 230,000 garment workers and some 2.5 million households in Cambodia, the garment industry has the potential to improve the incomes of up to 9 percent of all households. The key questions are how the money is used by rural households (whether it sustains consumption or can be invested in productive assets) and whether it is sustainable, or for how long. Different studies reach slightly different conclusions, strongly suggesting that the uses and impact of remittances may vary considerably between households and localities. In some cases (generally in areas that were already less poor), the receiving household puts the money to use in productive investments (e.g. buying fertilizer or pigs to fatten) to renovate their homes, dig wells, invest in farm inputs or livestock, purchase motorbikes and Ws, and send their younger children to school. (In one study, it was found that significant amounts of total remittances were used to hire labor, implying at least some multiplier effect within the village economy.) By contrast, in other, poorer areas, particularly those in the south-eastern Plains region that were badly affected by natural disasters, the money simply helped the recipient families sustain current, low levels of immediate consumption, allowing flood- and drought-affected households to purchase rice and other food, settle health expenditures, invest in restarting rice production, and pay off debts. There is also the important question of the degree to which the employees themselves benefit. Expectations of the garment workers are extremely high. They remit a very large proportion of their wages to their families, often at the cost of their own current consumption. Thus, despite earning what is in Cambodian terms a good salary, many live at a subsistence level with consequences for their health and wellbeing. On top of this, their social status is ambiguous. The ADB PPA studies found that rural men reportedly often expressed a preference not to marry a woman who has worked in the garment factories in Phnom Penh, based on general. suspicions of the morality of those who have lived in the city for extended periods (CDRI 2006 forthcoming). Sources: CDRl 2006a (forthcoming); CDRl 2006b (forthcoming); EIC 2005; ADB 2004; Dahlberg 2006; ADllCCC 2005. 68 Cambodia - Poverty Assessment 2006

The phenomenal growth of the garments Greater China.’’ Most of the current sector in Cambodia was largely fortuitous. foreign owners are based outside Just over thirty years ago, under the Cambodia and manage the operations of Multifibre Arrangement (MFA), similar factories across a number of quantitative restrictions (quotas) were countries. Decisions on which factories to imposed by governments of developed use to fill orders are made according to a countries on imports of textiles and variety of factors, including availability of garments from the increasingly quotas, desired quality, manufacturing competitive garment companies located in costs, and delivery lead time. Thus, most developing countries’. As a by-product of outlet garment companies in Cambodia this arrangement, many of the well- have no explicit role in deciding either established garment companies in their production levels or marketing developing countries, mostly owned by strategies. Moreover, production in nationals of Greater China, began shifting Cambodia relies on imported textile and some of their operations to other other raw materials with limited backward developing countries with abundant labor, linkages to a few informal and small-scale including Cambodia in the mid-1990s”. sub-contractors. While it has grown remarkably over the While the recent agreements between the last decade, and managed to secure itself at EU and China and between the US and least for now, a corporate social China to maintain quotas until 2008 will responsibility (CSR) “niche”, the garment offer some measure of protection to industry could be potentially only a Cambodian garment exports in the sputtering engine of growth in the future medium-term, it is not unreasonable to because of its “off-shore outlet” nature of predict that unless Cambodia improves its operations. Such operations, which are competitiveness, especially against China, characterized by low start-up costs, are which has much better transportation and subject to high entry and exit rates as power infrastructure, and produces many producers constantly seek cheaper outlets of the required inputs which enable that can quickly and efficiently produce Chinese factories to quickly meet ever- quality garments. By end-2004, nearly shortening production deadlines, the three-quarters of the owners of factories in garment industry is destined to become a Cambodia were from the Greater China much weaker growth engine for region, South Korea, Malaysia and CambodiaI2. Finally, it is also likely that Singapore, and these owners brought with over time Cambodia’s first-mover them capital, expertise, and considerable advantage in terms of its CSR “niche” will experience. Only seventeen percent of owners were Cambodians, and most of I‘ Measuring Competitiveness and Labor them are in partnership with nationals from Productivity in Cambodia ’s Garment Industry; Table 3.6. A recent WTO paper (Nordas, 2004) simulating the effects of the end of quotas and the associated changes in relative prices found that China will be a big beneficiary and that several small clothing producing countries, Including China, Egypt, Hong Kong, India, including Cambodia, will be significant losers. South Korea, Pakistan, Taiwan, Thailand, and Also, China’s imports of textile and clothing Turkey. machinery (TCM) have expanded sharply over lo Comprising mainland China, Hong Kong, the last few years, and now account for about a Macau, and Taiwan. quarter ofworld TCM imports. Sources of Economic Growth 69 wane as other countries eventually adopt numbers of domestic tourists (Abassi, similar practices. 2005)14.

Construction Tourism Construction is the second most important The “hotels and restaurants” classification industrial sub-sector (after manufacturing). in the National Accounts indicates that In 2004, construction activity represented tourism-related activities have grown by an 6.4 percent of GDP and employed 2.4 average of 14.6 percent per year, and percent of the total employed population accounted for 4.7 percent of GDP in 2004. [as percent of labor force]. Over the last The sub-sector’s contribution to overall decade, construction activity expanded by growth increased from 5.0 to 12.5 percent an average of 11.9 percent per annum, and of GDP (out of loo), while its share in contributed an average 6.4 percent per year total services rose from 6.2 percent in 1993 to overall economic growth (out of 100). to 13.7 percent over the period. growth The of the sub-sector has benefited Tourism is a re-emerging industry in from the expansion of both garment and Cambodia, rather than a new one. In the tourist activities, as well as from the 1950s and 1960s, the southern shorelines Government’s infrastructure projects. The were popular tourist destinations year share oftotal construction coming from the round, but thirty years of civil war tourism industry has declined in recent subsequently destroyed most of the years, signaling possible saturation of the infrastructure connecting tourist tourist market around Angkor Wat. Current destinations and eliminated the required construction activity is focused mainly in social capital and hospitality skills. Phnom Penh and Siem Reap (EIC 2005). Fortunately, most of the country’s prime cultural assets in the form of dozens of Services Khmer temples, including the unique site The share of services in the economy has of Angkor Wat just outside Siem Reap, remained stable over the 10-year period in remained untouched. Together with the the range of 34-38 percent, and currently advent of political stability, the listing of employs about 22 percent of the labor Angkor Wat as a World Heritage Site by force.13 Within the services sector, the UNESCO in December 1992 was followed highest average annual contributions to by considerable international financial and overall GDP growth have come from technical support to rehabilitate the site hotels and restaurants (6.6 percent); and significant initial inflows of tourism- transport and communications (6.5 related FDI. percent); trade (4.9 percent); and other services, including health and education (7.6 percent). Overall, the services sector expanded by 9.2 percent in 2004, significantly higher than in the two previous years, reflecting robust growth in tourism, from both a high rebound in l4 international tourist arrivals following the The Cambodian tourist is a person who has traveled over 10 miles for leisure and has spent regional slump in 2003, and sharply higher money as part of the trip. However, the accuracy of the data is hard to verify and l3 CSES (2004). should be interpreted with caution. 70 Cambodia - Poverty Assessment 2006

Table 4.5: Growth in domestic tourists is increasingly important in recent years -Thousands of tourists and percent annual change ------93 94 1995 96 97 98 99 2000 01 02 03 04 Domestic 266 320 38 5 465 282 422 1,012 1,215 1,464 1,757 1,821 4,251 Yo growth ------20% 20% 21% -39% 50% 140% 20% 20% 20% 4% 133% International 154 216 294 339 284 290 368 466 605 787 701 1,055 % growth ------40% 36% 15% -16% 2% 27% 27% 30% 30% -11% 51% Total 420 536 679 804 566 712 1,379 1,681 2,069 2,543 2,522 5,306 % growth - 28% 27% 18% -30% 26% 94% 22% 23% 23% -1% 110% Sources: Ministry of Tourism and Ministry of the Interior.

As a result, international tourist arrivals Puok district, have been much less have increased steadily since 1993, with ~ignificant'~. average annual increases of 29 percent A second study by CDRI covering eight over the period, excluding 1997 and 2003 villages in six Communes in Siem Reap due to political instability and Province came to broadly similar unfavourable external circumstances conclusions. A significant number of (Table 4.5). Siem Reap and Phnom Penh people in the countryside surrounding are still the two main hubs for tourists, but Siem Reap town were benefiting from the the Southern coast (Sihanoukville, and tourism sector, but the benefits of these Kampot and Kep provinces) and the linkages were distributed unevenly. As Northeast (Rattanakiri and Mondolkiri would be expected, those with education provinces) are receiving a growing number (generally non-poor) obtained the better ofboth domestic and international tourists. jobs (with higher wages and greater The Northeast in particular benefits from employment security). The CDRI study eco-tourism attractions, such as ethnic also found considerable movement out of minorities, rainforests, and wetland agriculture, as people in distant Communes ecosystems, while the border areas of migrated to Siem Reap town and those Poipet and Koh Kong attract gamblers living in areas closer to the town sold land fi-om Thailand. to invest in other, tourist-oriented A recent case study of the effects of the activities. expansion of tourism around Angkor Wat on the neighboring Puok district in Siem Reap province found that tourism-related activities contributed more than a quarter of local incomes in 2004 (Box 4.4); but that much ofthis was due to casual labor in l5 the tourism-related construction sector, While local backward linkages (construction, hotel, restaurant, and food preparation workers) which now shows signs of slowing. The are no doubt important to the communities more permanent direct local effects (Le. around tourism hot-spots, including Angkor non-construction related effects, such as Wat, there is no reason to prefer locally supplies of fresh foods to hotels and supplied labor or any other input over that restaurants) from higher-end tourism- sourced from more distant parts ofthe country. What is important for the country as a whole is related activities around the country, the level of labor productivity and the quality including the effects of Angkor Wat on ofother inputs. Sources of Economic Growth 71

Box 4.4: Tourism and the poor One of the more puzzling aspects of the link between growth and poverty reduction in Cambodia over the last decade concerns Siem Reap Province. The province has received significant tourist revenues and relatively high levels of ODA and private (domestic and foreign) investment, yet remains one of the poorest Provinces in the country. A case study by the Economic Institute of Cambodia (EIC) of Puok District does show increasing incomes, including from tourism, at the District level. While the primary income sources are still based on traditional activities (e.g. cultivation of rice and other crops, raising livestock and fishing), an increasing number of households have obtained significant income from new activities associated with the tourism boom in nearby Siem Reap town. Most of these new jobs are in hotel, guest house and restaurant construction. Thousands of workers have been absorbed from Districts around the town, including Puok District, which has witnessed a seven-fold increase in construction jobs (from 450 to 3,300) between 1999 and 2005. Local officials and residents estimate that approximately 350 people from the District are working in restaurants, earning US40 per month; and 450 people are working in hotels (often as cleaners), earning up to US$60 per month. These hotel and restaurant jobs are considered more secure as they are ‘salary jobs’, while construction offers day rates of about US$1.50 per day, and employment tends to be less secure, both in the short term as well as long term. At some point the hotel construction boom will taper off (this may have already started). Another source of tourism-related employment has been derived from the presence, since 1999, of Artisans D’Angkor, a company which recruits and trains young artisans in traditional and contemporary arts and crafts oriented towards a wealthy tourist market. The company has created jobs for over 720 people, including 320 from Puok District, who earn on average US70 per month. The benefits of the tourism boom are felt unevenly over the District. With poor roads, chances to obtain access to jobs or other opportunities in the town drops off rapidly with distance. Seven communes in the south-eastern part of the district closer to Siem Reap town and the national road that runs to it, containing about 47 percent of the population of the District, provide about 75 percent of the construction workers, 86 percent of restaurant workers and 95 percent of hotel workers. The failure of Siem Reap’s farmers to meet the hotels’ and restaurants’ demand for food has fairly prosaic, production-side problems; small land holdings, and often poor soil, a lack of water control infrastructure, and difficulties getting produce to the town in good condition have all made it hard for local farmers to compete with cheaper imports from Thailand and Viet Nam. Likewise, supplying mulberries for the growing silk industry (Artisans D’Angkor imports more than 90 percent of its silk) is another potential source of farming income. Apart from transport, the main constraint preventing District inhabitants from accessing tourist-related jobs, or at least the better- paying jobs as hotel staff or tourist guides, is the low level of education in the District. A conclusion to draw from this experience is that tourism will not result in major benefits for the poor of Siem Reap province without deliberate strategies in place and a framework for cooperation between the national Government, local authorities, local communities, donors and NGOs. Source: EIC 2005.

Social capital was very important; The conclusion from the EIC and CDRI networks of family and friends would help studies would be that rural Siem Reap has a would-be migrant acquire a job or warn a indeed obtained some benefits (in terms of construction worker of contractors known waged employment and to some extent for not paying wages due for work supplying needs for handicrafts, flowers or completed. Those without such vegetables) for the growth of tourism in connections would often have to pay the Siem Reap town. foreman or supervisor to obtain a job. However, the non-poor have benefited more from the poor; and the gains to the 72 Cambodia - Poverty Assessment 2006 poor are somewhat precarious, as they are remained very low. Agri-business, a primarily in the form of casual labor in the traditional link to a more manufacturing- construction of hotels and restaurants, based economy, has failed to expand which cannot continue to grow despite surplus agricultural output. indefinitely. Both studies agree that the Thus, looking ahead, in addition to greatest potential source of tourism-driven sustaining its existing twin engines of benefit to the rural poor-that is, increased growth, Cambodia also needs a more pro- demand for agricultural produce-has to poor and rural-focused engine of growth to date been limited, in part because of accelerate poverty reduction. The obvious production-side constraints, and in part choices are agriculture-related activities. because of marketing difficulties that are An extensive literature concludes that typical of most of rural Cambodia. With growth in the agricultural sector has the limited information, high marketing costs, greatest poverty reduction impact, and limited potential to make risky compared to growth in industrial or in inputs might investments that improve services sectors. In Cambodia, almost two- production, Siem Reap’s farmers have yet thirds of the heads of poor households in to reap major benefits from tourism. As 2004 report farming as their main elsewhere in the countryside, security of occupation, either self-employed tending tenure, good extension services and to their own land or working for other investments in irrigation and rural roads farms. Over 20 percent of the poor could go a long way towards improving depended only on crop cultivation for over the supply response to increased urban half of their income. Thus, because of the demand16. importance of agriculture, steps to facilitate improvement in agricultural Binding constraints to productivity and expansion of agri- agricultural and agri-business business activities are necessary conditions growth for enduring poverty reduction. In light of The patterns of poverty reduction are not this assessment, the following section unexpected given where and how growth offers an analysis of the key binding occurred. The main reason that the rural constraints to agricultural and agri- poor have made advances but have business growth using a simplified version nonetheless fallen behind is that the twin of the conceptual framework recently engines of growth over the period were developed by Hausman, Rodrik and generally urban-focused with weak urban- Velsaco (2004)17. rural linkages. Meanwhile, agriculture, the primary livelihood of the poor, has l7 A more disaggregated and rigorous analysis of attracted little investment in recent years the binding constraints to growth, including and has experienced relatively slow their relative rankings across major sectors and growth, while its level of productivity has sub-sectors, is being prepared for inclusion in the more comprehensive study on the “Sources of Growth” referred to in footnote 1. While the l6 Improvements in national roads in recent years “binding constraints” conceptual framework, have actually disadvantaged many Siem Reap which is being applied in a sample of other farmers, as it has become easier and cheaper in countries as part of a pilot program being many cases to import produce from Vietnam or conducted by the World Bank on “Growth Thailand-or, indeed, from more productive Diagnostics” in consultation with leading parts of Cambodia such as Kien Svay in academics in the field, is a helpful additional Kandal or Chamkar Leou in Kompong Cham. analytical tool to think through the hierarchy of Sources of Economic Growth 73

Figure 4.4: Growth diagnostics

inteslment and

L.ow return to economic rrtivit?

I I I 4 b d LOWsocial tmrns

externalities

Source: Hausman, Rodrik and Velsaco (2004).

The basic idea underlying the growth under the two broad determinants of diagnostics under HRV framework is to private investment; the returns to economic analyze the key binding constraints to activity and the costs of capital (Figure economic growth (or more precisely, to 4.4). Returns to economic activity are, in higher private investment and turn, a hnction of the social returns to entrepreneurship) as separate decision trees investment (Le. the measure of how much of the benefits from investments are shared by society as a whole) and the private the key constraints to growth, the framework appropriability of those returns (i.e. how does not offer any particular insights as to why much of the benefits from investments are particular constraints are binding. Indeed, constraints are likely to be binding precisely recouped by those who finance the because they are not easily eased or eliminated. investment and bear the inherent risk). The Nevertheless, the identification and ranking of costs of capital are a finction of the the key binding constraints are necessary first international and local financial climate. steps to at least easing them. 74 Cambodia - Poverty Assessment 2006

In turn, low social returns can be caused by 1990~~~.The lack of official title and poor geography, low human capital, and/or insecurity of tenure results in farmers inadequate infrastructure, while weak having a short time horizon for agricultural private appropriability of returns can planning and shunning investments which emanate from: (i)government failures, take longer to pay off, such as land such as weak property rights and leveling for irrigated land, soil corruption at the micro level, and monetary amendments for fertility, and tree crops. and/or fiscal instability at the macro level, Thus, weaknesses in the framework for and/or from (ii)market failures, such as tenure security generally serve as a information and coordination externalities. disincentive to agricultural investment in Cambodia. (More analyses are provided in The essence of the analysis is to determine Chapter 5). While titling ofprivate lands in which of the many constraints to higher areas of long-established settlement is growth are actually binding, in the sense proceeding through systematic as well as that if they were eliminated or eased, sporadic procedures, it is expected to growth would immediately ensue. It is require at least ten years for completion of important to analyze the binding the titling process on the estimated 5 constraints to growth at a disaggregated million private parcels in Cambodia. In level, because they differ in nature and addition, as many as a third of all severity across economic activities. households using land may currently be Primary binding constraint to ineligible for title as the land they occupy agriculture growth: weak private is technically state land, even if, as in the case of indigenous communities, they are appropriability of returns from ill- eligible for recognition of the tenure rights defined property rights under the 2001 Land Law. Concerns over private appropriability can prevent investments being made not because they would not be profitable, but because the risks that these returns would All land titles for private and state property not be appropriated by the investor are too were destroyed under the Khmer Rouge, and most land professionals, judges and community high. Examples of key impediments to leaders either died or went into exile. Even private appropriability of returns include traditional claims to land were lost as a result micro risks related to, inter alia, poor of the massive displacements of populations definition and protection of property rights under the Khmer Rouge and during the continued conflict which lasted until 1998. and contract enforcement, weak and More recently, the allocation of state lands, volatile governance, and uncertainty over following the initial distribution of residential the commitment to maintain the current and agricultural lands to all households in 1989, has reflected political as much as rules ofthe game. economic motivations with large tracts Property rights: In the agricultural sector, officially provided to military units, ostensibly to support demobilization of their troops, and the most binding constraint to growth is unofficially to political and business associates the insecurity of land tenure. Many small to consolidate political hegemony. The farmers work on land essentially acquired consequence of this difficult history is that the through allegiance to local commanders in legal framework for tenure security is being re- developed, registries and other inventories of the peace deals brokered through the late property rights are being reinstituted, and the credibility of judicial and non-judicial mechanisms for dispute resolution and contract enforcement is slowly being strengthened, Sources of Economic Growth 75

Furthermore, most ofthe business interests transport losses for marketing of higher which have obtained official contracts to value, perishable products”. develop large scale agricultural land Human capital: Growth in all sectors of concessions (almost one million hectares the economy will be constrained by the since 1995) have been unable to realize the low human capital stock and this constraint intended investments as less than 5 percent is likely to become more binding as land of the contracted area is under production, security, governance, and all types of in part due to continued conflict over the public services improve. Thus, higher boundaries of their concessions which returns to investment and diversification of were never adequately surveyed and the production base will also require an demarcated to ensure they not overlap did improvement in the quality of the stock of with existing settlements or forests or human capital over time. For example, the protected areas (World Bank 2004). returns to schooling are sizeable and increasing over time (Chapter 6). The Second-order constraints to positive impact on growth of contract agriculture growth: low social farming arrangements highlighted above returns from poor infrastructure would seem to exemplify the benefits of and low human capital stock improved human capital in agriculture. Low social returns to investment can result from an inadequate supply of Primary binding constraint to complementary factors of production, such agri-business growth: as infrastructure, human capital, or poor governance geographical endowments. The prevalence of informal feesmay not rank as first- or second-order constraints in Infrastructure: In the agricultural sector, the agricultural sector, but constitute the the lack of basic infrastructure and primary binding constraint to growth of complementary public services constitute agri-business. Large agri-business second-order constraints to growth, in the companies have found that commercial sense that a variety of such constraints returns to both manufacturers and traders would become increasingly binding as the can be attractive, only if either: (i)non- constraint of land security is eased. In commercial barriers are sufficiently particular, the relatively under-developed reduced to compensate for the high small-scale irrigation infrastructure informal fees; (ii) protection from increases the vulnerability and variability excessive fees is secured in exchange for of agricultural production, as well as the exclusive “right” to operate; or (iii) lowers the overall productivity of the production occurs higher up the value activities undertaken. Also, inadequate chain to cushion against high fees. access to road infrastructure and transport, However, the majority of smaller agri- acts as an impediment to increasing farm business companies do not have such productivity, since it raises the cost of

trade, lowers the farm-gate price of 19 While the benefits of rural infrastructure are production, and increases risks and evident, analyses of the costs of infrastructure have yet to be carried out with sufficient rigor i to determine which parts of the country are best suited to what types of investment in infrastructure, in particular irrigation facilities. More analysis is provided in Chapter 5. 76 Cambodia - Poverty Assessment 2006 options. Even for those agri-business earnings and informal networks of family companies already operating, the high and friends, sometimes at high usury rates, uncertainty and seemingly arbitrary nature while formal enterprises, including of these fees make it impossible to plan exporters, rely on retained earnings business, and therefore impossible to (Development Consulting International accurately finance any expansion. 2003). Productivity-enhancing investments are not made, keeping small processors small With the exception of ACLEDA, a former and larger processors few in number. As micro-finance institution that recently has been well documented in recent transformed into a commercial bank, the studies, informal payments for licensing, formal sector simply does not provide securing access to domestic markets, trade financial services to rural populations. facilitation, and carrying out economic Only a third of the 24 provinces in the activities in general is frequent (82 percent country have commercial bank outlets, of responding firms) and substantial (5 apart from ACLEDA, while another third percent of sales).20 Worst of all, those have no formal bank facilities at all. The charged with law-enforcement-police, areas around Phnom Penh, Sihanoukville customs agents, regulators-are often the and the Tonle Sap are comparatively better same people involved in extracting rents. served, but even where branches do exist, they typically are located in provincial Second-order constraint to agri- capitals with no outreach capacity. The business growth: high cost of formal sector's reluctance to lend in rural capital areas has to do with the higher costs of reaching dispersed and small-scale clients, Local finance: The banking sector is coupled with the difficulty of assessing crowded with many small banks (17 credit risks and the absence of reliable domestic) which can only provide collateral. In the case of agri-business, if expensive short-term financing. Reflecting the governance-related constraints were the fact that these domestic commercial eased, potentially profitable activities banks are still relatively weak, interest rate would still find it difficult to get started spreads have remained high, loan because of a lack of access to capital both maturities short, and the price of capital in from formal domestic financial services, the formal market expensive. Because of and from FDI via partnerships with foreign the high costs of, and/or lack of access to, companies because of the poor investment formal finance, start-up and quasi-formal climate (Chapter 7). enterprises rely mostly on retained

2o Such fees have been estimated at 4-5 percent of sales revenue. See Cambodia: Seizing the Global Opportunity: Investment Climate Assessment and Reform Strategy prepared for the Royal Government of Cambodia by the World Bank Group (August 2004). In the case of agri-business, it should be noted that this range is very close to the average profit margins ofworld-class agri-businesses, such as Thai Union Frozen, one of the largest tuna processing companies in the world. 5. Raising agricultural incomes

Summary As growth has lifted living standards first and foremost in urban centers, poverty in Cambodia is becoming an increasingly rural problem. In 2004, 91 percent of the country’s poor lived in rural areas. Agriculture remains the primary occupation for 72 percent of heads of households, but contributes only 31 per cent to GDP. Growth is erratic, with several years of negative growth over the last decade. Lack of water control infrastructure (flood control and irrigation) creates a high level of production risk for smallholder farmers. Crop yields-including for the food staple, rice-remain amongst the lowest in the region. Rice dominates production in all regions of the country. The poor are the most heavily dependent on agriculture and on the natural resource base. This dependence is declining slowly as non-agricultural activities and wage labor become more important. Nonetheless, it remains high, and for the short- to medium-term progress in poverty reduction will require significant improvements in the productivity and profitability of family-based agriculture, with a managed transition to a more diversified, commodified form ofsmallholder production. Apart from paddy, the “average” rural household depends heavily upon access to local common property resources (fishing waters, forests, and so on). These serve as a source of subsistence and inputs to production for sale in normal years, and are used more intensively as a safety net in years of poor harvests. It seems that much of the improvement seen in rural livelihoods over the last ten years has been at the expense ofrunning down this stock ofnatural capital. The evidence from fish catches suggests that overexploitation has not yet occurred, although locality-based research in MOPS and the Tonle Sap PPAs suggest that the poor certainly enjoy less access than in the past. Forest cover remains high but is in decline, with an increasing number ofcommunities finding it hard to meet their needs for timber and non-timber forest resources. As the Government moves away from the concession system, there is an urgent need to find a better model for managing forests in a more sustainable and pro-poor way. Landlessness in Cambodia has risen rapidly in recent years, albeit from a relatively low base, reflecting a variety of demographic and economic forces. These effects are exacerbated by governance problems in the form of unclear property rights, which results in large tracts remaining idle (in the interests of speculation or out of fear of challenges to ownership) and reluctance of smallholders to invest in land improvement. Better property rights, investments in infrastructure (irrigation and roads) and improvements in human capital and risk management would all help promote agricultural growth and rural poverty reduction. 78 Cambodia - Poverty Assessment 2004

s elaborated in Chapter 2, Figure 5.1: Agriculture and access to natural Cambodia's poverty is resources is critical for the poor .concentrated in rural areas A A. The importance of agriculture to household where 85 percent of the population resides. Although rural poverty has gross income declines steadily with wealth. fallen (by four percentage points) since 100 1993, it remains high at 43 percent. Inequality has increased (the 75 consumption-based Gini coefficient has risen from 0.265 to 0.333) indicating that 50 the benefits of growth have been unequally shared among rural households. Rural poverty remains 25 deeper and more severe than elsewhere. 0 Agricultural sector Poorest Next Middle Next Richest Rural households in Cambodia depend poorest richest on multiple sources of income for their livelihoods, depending upon local agro- 1 mother climatic characteristics. Previous studies 0 Wage estimate the range of activities Non-agricultural undertaken by rural households to earn Forestry income in cash or in kind as between 3.1 and 7.2 (see Helmers et al, 2003). The 0 Livestock majority of rural residents still live in 0 Cultivation traditional ways, primarily cultivating B. For wealthy households, non-agricultural and rice and collecting natural resources wage labor is far more important (gross revenue, from water bodies and forests (see thousand riels, 2004). Figure 5.1). Agriculture-defined here ...... ,"...... ", .... ., ...... " "".l "" ".l." .... "l".._ as including crop and livestock 25,395 production, forestry and fishing- remains the primary occupation for 72 percent of households, yet contributes only 31 percent of GDP'. The importance of off-farm income (remittances, wage labor and non- agricultural self employment) has grown steadily, constituting substantial income shares. Poorest Next Middle Next Richest poorest richest

Source: CSES 2004.

All GDP is 2004 unless otherwise stated. Raising Agricultural incomes 79

Table 5.1: Food production and consumption for rural households by quintile Quintile Quintile Quintile Quintile Quintile 1 2 3 4 5 Purchased food Riels 3,469 4,947 6,206 8,584 13,753 % 64 67 71 77 84 Own produced food Riels 1,974 2,413 2,588 2,594 2,599 % 36 33 29 23 16 Total food consumption (riels) 5,443 7,360 8,795 11,177 16,352 Source: CSES 2004. Sample constrained to rural houses outside Phnom Penh. Including rural households in Phnom Penh leaves results for the bottom quintiles unchanged, but increases the market dependence of the top two quintiles.

Table 5.2: Agriculture productivity in Cambodia In 2004 about 30 percent ofthe poor’s total is the lowest in the region (2003) income was from crop cultivation; 10 percent from livestock rearing; 25 percent Crop yields (kglha.) from common property resources (i.e., Rice Maize Cassava fishery and forestry); and the last third Cambodia 2,150 2,111 6,318 from non-agricultural activities (10 China 3,849 3,485 16,249 percent); wage (possibly farm) Lao PDR 3.31 6 2,333 19,762 employment (20 percent); and “other” Vietnam 4,634 3,225 14,066 sources (e.g., remittances and transfers). Indonesia 4,538 3,252 14,902 As one moves from the poorer to the richer Malaysia 3,178 __ 9,737 end of the distribution, the reliance on Thailand 2,455 3,913 17,552 agriculture (i.e., crop cultivation, livestock, Source: FAO, World Development Indicators and common property resources) declines, while non-agriculture and wage growing by 14 percent over the last five employment becomes important. years3. Despite the improvement, Cambodia still has one of the lowest levels Crop yields are among the of productivity in agriculture among East, lowest in the region and rice Southeast, and South Asia (Table 5.2). still dominates Interviews with fanners and Ministry officials suggest that irrigation remains the Crop production is crucial for Cambodians. major constraint to intensifying production Wet season (essentially rain-fed) paddy of paddy as well as other crops. Poor production is primarily for subsistence technology and extension services also fail consumption. Self-sufficiency is beyond the reach of many rural households. (double the pre-war level) driven largely by Subsistence agriculture dominates among technology (both variety improvements and poorer households (Table 5.1). inputs), improved market access and the liberalization of agricultural prices (Nesbitt, There has been improvement in rice yields 1997). per hectare,2 with wet-season yields Specifically, to account for annual climatic fluctuations, we calculate average yields for Auffiet (2003) reports that yields increased two periods 1997198 - 1999100 and 2003104 - from 1.2 tonnedha in 1980 to 2.0 tonnedha 2004105. 80 Cambodia - Poverty Assessment 2004 to meet the demands of farmers Table 5.3: Rice constitutes the bulk of farm especially cash crop producers households’ agricultural revenue in (Kang and Chan 2003). The Cambodia public extension system has Non-Door Poor Cambodia limited reach and is skewed in Rice revenue 0.84 0.61 0.75 favor of non-poor households in (m Riel) Total agriculture 1.I4 0.73 0.98 more easily accessible areas; revenue NGO-funded extension projects (m Riel) dominate. Rice share in 73.3 83.5 76.2 Cultivated land is expanding total agriculture revenue (%) relatively slowly in the Plains in Average number 1.39 1.25 1.34 contrast to the Tonle Sap region of crops where expansion into the Number of 4300 2582 6882 surrounding forested areas has households taken place at significant rates4. Source: CSES 2004. Some of the fastest expansion in areas with significant existing paddy constituted three-quarters or more of total production is occurring in the far north- agricultural revenue in the provinces west Provinces of the Tonle Sap region in (Table 5.4). Battambang and Banteay Meancheay. In these regions there is also extensive maize There is significant potential for increasing production to meet market demand in production of non-rice crops in Cambodia Thailand and elsewhere. due to the abundance of fertile land and recently cleared forests. This potential is Paddy remains central to household crop already being realized in parts of the production for the poor and non-poor. On country. Various studies pointed out that average, households still concentrate soybean, maize, cassava, mung bean, primarily on paddy production. The cashew, and rubber have the potential for average number of crops that households much greater expansion, but this is grow on land they owned or operated was conditional on additional physical and hovering around one. Poor households institutional improvements (World Bank, depend on paddy for three-quarters oftheir 2004). (gross) annual revenue from agriculture. There is thus still very little crop diversification in rural Cambodia (Table The natural resource base is 5.3). Even at the provincial level, it is important as a source of evident that except in Kampong Cham livelihoods, safety net and (and Phnom Penh), revenue from rice nutrition for the poor Cambodia benefits from unusual natural There is ample evidence from land use maps resource endowments. Almost two-thirds that deforestation is clearing land for of the land area is forested. Cambodia’s agricultural uses. Wet season cultivated land freshwater fishery ranks fourth in terms of under paddy is a reasonable proxy since encroachment is typically motivated by tonnage in the world. Poverty rates remain subsistence needs and is un-irrigated. Between the highest in those areas with the richest 2003104 and 2004105 cultivated land grew by 6 forest resources - and where exploitation ‘percent in the Plains and by 15 percent in the has been most severe. Tonle Sap. Raisina Aaricultural incomes 81

Table 5.4: Rice still dominates crop Coalition Team 2001). The sector is the production in Cambodia major source of income for 45 percent or so households residing in riparian Provinces or grouped Rice share provinces (de Lopez 2002). Rice and fish provinces in total are Cambodian staples; indeed 80 percent agricultural revenue of animal protein intake comes from (%) aquatic resources (de Hortle et a1 2004). Takeo 96.5 Furthermore, fisheries may provide a Prey Veng 94.5 safety net, particularly for the poor, in an Kompong Speu 94 event of shocks. Thus, fish resources are in in in Banteay Meanchey 83.8 important Cambodia many ways, terms of income, nutrition, food security Kampot 83.7 and household risk management. Other Plain/Mountain 83.7 Kompong Chhnang/Pursat 81 .I Previous studies reveal that Cambodian fisheries and people relying on fisheries Battam bang 79.7 are facing changes in their surrounding Svay Rieng 79.4 environment. In terms of resource stock, Kompong Thom 77 there are concerns regarding over Siem Reap 72.4 exploitation of fish and other aquatic Coast 71.9 resources, deforestation and construction Kandal 63.2 of dams that could adversely affect the Phnom Penh 51.9 hydrology, and thus the ecology, of the Kompong Cham 45.5 Lower Mekong River Basin (Hortle, et al. Cambodia 76.2 2004; Sverdrup-Jensen 2002). In terms of people who rely on fisheries, there are Source: CSES. 2004. reports that privatization of what used to be the common fishing grounds (e.g. lake shores) via public auction is adversely The poor in general and the indigent (the affecting the poor by limiting their access bottom consumption quintile) in particular, to these fishing grounds (e.g., de Lopez depend heavily on fishery and forestry 2002; Fisheries Action Coalition Team resources as a vital income source; a safety 2001; World Bank 2005). net in the face of adverse shocks; and an important source of nutrients. Analyses Status of aquatic resources based on the 2004 CSES suggest that 39 percent of rural indigent households In general, there are no clear indications depend on only fishery and forestry that wild aquatic resources are resources for over a quarter of their total overexploited (Sverdrup- Jensen 2002). For income and that 16 percent of these example, according to the statistics indigent rely on fishery and forestry published by the Department of Fisheries income for over halfof their total income, of Cambodia, the catch per unit of effort (CPUE), measured in terms of catch (kg) Fishery sector per boat per annum, has been steady around 1,847 kg (Department of Fisheries two Over million people derive 2001). However, there are changes in the employment from the fishery sector and species composition, especially a decline related activities (Fisheries Action in the abundance of larger and slower 82 Cambodia - Poverty Assessment 2004

growing species. This phenomenon (Sloth et al 2005; CDRI 2006, suggests that for certain species, and areas forthcoming). inhabited by such species, the present Another threat is the impact of exploitation rate may be too high. On the environmental changes in areas other hand, the Tonle Sap bagnet (or dai) surrounding the Mekong River basin. Two fishery for 2004-05 season have recorded main threats are dam construction and the largest catch in the past decade (Hortle, declining forest coverage. Seasonal et al. 2005). flooding of the Mekong River and its The access of the poor to aquatic tributaries (such as Tonle Sap) is critical resources, nonetheless, is facing several for fish reproduction, habitat and threats. One of them is the privatization of migration’. Dams could reduce the size, what used to be a common pool resource. shorten or alter the timing of flooding. In 1987 the Cambodian government Also, deep pools along the Mekong River introduced a two-year concession system and its tributaries, with some reaching of demarcated water areas called fishing more than 80 meters deep, are an important lots, which were awarded to the highest refuge for larger fish during the dry season bidder in public auctions (World Bank (Sverdrup-Jensen 2002). These pools 2005). Although the concession system could be compromised by the dams as included measures to mitigate the impacts well. on households that are excluded from these As indicated, income from fishery is lots, there have been claims that such rules crucial for the rural indigent and the poor. were often not honored by the Income from fishery constituted 10 percent concessionaires and that the boundaries of total household income. As expected, were illegally extended (so-called water- fishery dependent households tend to be grabbing). Consequently, the use of these located in the Tonle Sap and Coastal zones resources by others, notably by the rural where there is greater supply of fishery poor, became limited (de Lopez 2002). A resources. process of reform began in 2000, and in 2002 resulted in the cancellation of Table 5.5 reveals that fishing households commercial fishing lots on about half the were, on average, poorer compared to non- total lot area, with the cancelled lots fishing households. The average total net reverting to community access. This has income of fishing households was 1,621 substantially improved access and reduced thousand riel (US$404), about 55 percent the frequency of conflicts over fishing of the average income of non-fishing rights (and indeed competing agricultural households. Own-harvested fish rights on seasonally-inundated land). consumption was a significant 22 percent However, the impact of the cancellation in total food consumption. Thus, access to has been less than might be expected fishery resources, especially among the because: (i)the cancelled concessions poor, not only constituted an important were of generally lower productivity; and source of livelihood, but also a source of (ii)in several cases communities continue nutrition (particularly). to be excluded from lots which were in theory cancelled, as local authorities Research shows strong positive correlation collaborate with large fishing operators between the maximum water level of the flood and local military units to continue to plain and harvest volume (e.g. Hortle, et al. manage the waters as a private holding (2005); Poulsen, et al. (2002); van Zalinge, et al. (2003)). Raising Agricultural incomes 83

Table 5.5: Comparison of fishing and non- combines profitability and environmental fishing households by income and sustainability, it is also important not to consumDtion exclude the poor from an important Fishing Non- livelihood option and a coping strategy. household fishing More analyses and discussion will be household necessary to decide what may best serve Net total income 1,621,2 2,972,7 (riel) 73 Cambodia’s needs. ($404) Net income from 347,325 Forestry sector fishery (riel) ($86) Forestry occupies 63 percent of the Per capita 1 13,423 country’s area (data from Forest consumption of 2002/03). fish (riel) ($28) clearing has accelerated in recent years, affecting particularly heavily populated Per capita own 37,829 consumption of areas. Forest coverage in Cambodia was fish (riel) ($9) once estimated to be 70 percent in the Fish consumption 1960s and 1970s6. Official statistics in total food 21.8 suggest the sector contributed only 2.1 consumption (%) percent of GDP, although this is likely to Own consumption under-report extensive illegal logging, the in total fish 17.7 consumDtion (%I real value of non-timber forest products in Note: Exchange rate applied here is US$1=4,016 (NTFPs), and ignore the reduction value riel. of remaining forest assets. In the Plateau Sample size is 5,862 for fishing households and region, where most forest resources 6,127 for non-fishing households. remain, 53 percent of rural households Source: CSES 2003-04 remain below the poverty line. Poverty rates may have actually increased in this region7. Regression analyses suggest that access to fishery could be an important coping strategy for households confronted with adverse shocks. Households whose heads were sick and those experiencing natural disasters in the previous five years were significantly more likely to rely on fishery as an additional source of income. Ensuring legal access rights to fishery will be important for the rural poor. 6 The forest coverage was 73 percent in the 1970s. This recent forest cover estimate is in The current fishing lot system, particularly Forestry Administration (2005), Cambodia: with the two-year renewal rule, has created Forest Statistics 2004, (Phnom Penh: May incentives for concessionaires to over- 2005) 7 harvest in an unsustainable manner and See Knowles (2005). Using comparable sampling frames, the headcount rate is illegally expand the lot borders to estimated to have increased fiom 40 percent to maximize benefits during the two-year 58 percent fiom the 1993 CSES to the 2004 concession period. While having a CSES. The small sample size suggests care carefully crafted scheme that encourages should be taken in interpreting these results concessionaires to behave in a manner that literally, although the broad direction ofchange is probably accurate. 84 Cambodia - Poverty Assessment 2004

Forests provide an important Figure 5.2: Revenue from timber and non-timber products source of rural livelihood in (riels, 2004) Cambodia, particularly for 250,000 I--- local communities'. As indicated, income from fishery is crucial for the rural indigent and the poor, accounting for 16 percent of 150,000 their total household income. Access to forestry resources 100,000 is more important for the poor than the non-poor. A 50,000 significant component of forestry production was for 0 own-consumption among the Poorest Second Middle Second Richest poor; the largest share was poorest richest firewood. Besides firewood, non-timber products such as Firewood, sawnwood, charcoal 1 root crops, hits and 10Non-timber forest products 1 vegetables, palm juice and fibrous material are Source: CSES 2003-04. contributed to household cash Note: Non-timber forest products include Rattan, bamboo, palm income and nutrition intake leaves, other fibrous material, palm juice, root crops, fruits and vegetables, herbs, honey, wild animals and birds, and other (Figure 5.2). products. It is commonly perceived that incomes from forestry are identified the fastest rates of forest declining; a fact documented by many degradation occurring in areas surrounding empirical studies' and recent land use (expanding) villages, and alongside new mapping exercises" and reinforced by a road corridors. Conversion of forests to recent study of perceptions of Commune agricultural use is occurring with little heads (Danida, 2005). Perceived loss of regard to the optimum use of land or soil forests was more severe in the quality. Of the 1.3 million hectares of Mountaifllateau regions where forest land cleared between 1997 and households depend on forest resources for 2002, 37 percent was of low soil quality, a higher proportion of their income. The compared to 19 percent of land existing in Independent Forest Sector Review (2004) 1997 (Dummer, 2004).

8 See, for instance, the World Bank- Land: the primary asset of the commissioned Poverty and Environment poor Nexus study and Moving out of Poverty study. See for instance: ABiC (2004a), ABiC The average farm size among the rural (2004b), Chan Sophal and Sarthi Acharya poor in Cambodia is 1.5ha. However, 40 (2003), Global Witness (2002), Helmers, K., percent of the rural population lives off Gibson., J., and Wallgren, P. (2003), - McKenney, Bruce and Prom Tola (2002), less than 0.5ha sufficient to meet only Prom Tola and Bruce McKenney (2004). about half of the per capita (milled) rice 10 See, for instance, the Forest Sector Review, 2004. Raising Agricultural incomes 85 requirement of 165kg/year1’. Population this landlessness is “functional”, as a pressures and hereditary practices proportion of rural households sell land to contribute to further land fragmentation. At invest in (potentially more rewarding) off- the same time, distress land sales, land farm activities. However, the rise in grabbing, and speculative land purchases landlessness is nonetheless of concern are leading to a polarization in the pattern because: (i)it is relatively rapid, given of land ownership between urban that land distribution was more-or-less landowners making speculative equal (relative to household labor) when investments and rural households with land was formally allocated to households small plots. in 1989; and (ii)the economy is not growing sufficiently fast to provide A case study undertaken for this survey poverty-reducing jobs or self-employment (EIC, 2005) reported that in one opportunities for the increasing number of Commune, 80 percent of the land was landless households. Although not all estimated to have changed hands in favor landless households are poor (some have of Phnom Penh residents. The price paid is profitable non-agricultural sources of often low, around US$SO/ha, compared to income), many are, as with few other rural the mean rental value of US$15O/ha and livelihood opportunities, they have to rely much below the value placed on land by upon low and variable income from wage households (Deininger 2005)12. While labor. additional employment opportunities are created through clearance and planning, There are a number of forces underlying wages are low (around US$1.20 a day). the trend towards increasing landlessness. Overall, such land purchases are unlikely A significant proportion of the landless to be ofbenefit to communities. (those who returned from refugee camps in 1993-94 and newly-formed households Rising landlessness.. . headed by a young married couple) have The proportion ofrural households lacking never owned land. There is also a natural land for cultivation has risen from 13 dynamic away from equal distribution: percent in 1997 to 16 percent in 1999 and population growth leads to smaller plot 20 percent in 2004. An additional and sizes in densely populated areas and, as growing number of households are “near mentioned above, as in any economy a landless”, owning only very small plots. proportion of households choose to sell As with inequality, a trend to rising land (either to finance investment in non- landlessness is inevitable given agricultural enterprises, or in response to a Cambodia’s transition from a socialist to a pressing need for cash for current market economy since 1989; and some of consumption, most typically urgent health treatment). A small but increasingly With average yields of paddy of 1.8tonnesiha, important part of total landlessness is due and an average household size of 6.1 people to land grabbing, as more powerful actors (Deininger, 2005)) a household with 0.5ha can displace current, generally poor owner- produce 0.9 tonnes of paddy, equivalent of cultivators in order to acquire valuable 585kg of milled rice (using a conversion factor of0.65) per year, or 96kg per person per year. land (typically those near urban centers or Based on survey results of the 2004 CSES, roads). Deininger (2005) estimates an average willingness to pay for agricultural land of US$1,500/ ha. However, this seems too high for remote, not fertile land. 86 Cambodia - Poverty Assessment 2004

,. .and insecure tenure by a perceived need to commercialize agriculture, in practice, few (only about Land tenure in Cambodia remains insecure 10) are operational. Reviews point out that for most, with consequent disincentive such concessions are often awarded effects for productivity investments and without adequate investigation in the field, access to credit. The benefits of property resulting in overlap with traditional rights, rights are widely recognized and well- leading to conflict and intimidation against documented in the literature. Issuing full which responsible institutions (cadastral titles to the 4.5 million applications commission and courts) are often helpless received after the 1989 land privatization (Kato 1998, Chan, et al, 2002, Leuprecht (86 percent of which are outstanding) 2004). would have significant (positive) impact on farmers with existing land13. A number of concessions appear to be speculative, and/or a mechanism to Analyses, based on the CSES 2004, find circumvent more stringent rules pertaining that secure land tenure, in the form of to forest concessions (Leuprecht 2004)16. certificates or application receipts, It is reported that less than one third of significantly raises crop yields, value of concessionaires have paid the required land, and household con~umption'~. deposit to the Government, or make rent Secure land tenure, i.e., formal land title in payments or pay taxes, suggesting that at the form of a certificate, increases rental least thus far the scope for using value by 57 percent; sale value by 38 concessions to collect additional revenue percent; crop yields by 65 percent; and has been quite limited (McKenny and Tola household consumption by 24 percent. 2002). Suspicion over the motives for The main reason is that well-defined awarding these concessions is fueled by property right improves private the fact that 14 concessions exceeded the appropriability of returns. Owners with 10,000 hectare limit above which a special secure land titles are more willing to invest review is required - yet no review has yet in higher-risk and potentially higher-payoff been implemented. activities, such as planting perennial trees and diversifying into vegetable and cash while there is a strong crops. ... economic case for smallholder Large tracts of farmland remain agriculture idle.. . The Government is now in the process of canceling a number of non-operational Over the past decade, the Government has economic concessions. It is proposed that constituted a system of economic these lands be transferred to landless (and concessions by providing 70-year leases land-poor) households under a system of covering 889,399 hectares of land to 49 social land concessions (established by private c~mpanies'~.Ostensibly motivated

Natural Resources and Environment, l3 Deininger (2005) estimates that holding a full September 2005. title (compared to holding a receipt of title 16 Even though the law mandates that cancelled application) is associated with higher land forest concessions be protected as a natural prices and household welfare. forest area, the two largest agricultural l4 Detailed regression results are in the Annex. concessions (of 315,000 ha and 101,000 ha, l5 List of Land Concession Companies respectively) seem to be effectively used as distributed by the Technical Working Group on forest concessions (McKenny and Tola 2002). Raising Agricultural incomes 87

Decree in 2003). Although the motivations making the case for speedy for this policy are primarily, as the name implementation of the social concession suggest, social or distributional, there is a policy. More broadly, the inverse strong economic rationale too. As in other relationship between farm size and countries, it appears that rural households productivity suggests that there should be a in Cambodia are small but efficient general presumption in favor of secure agricultural producers. According to smallholdings, rather than very large analyses from the CSES 2004, small farms commercial holdings, as the foundation of are more efficient than large farms, the Cambodian agrarian structure. whether measured by crop income, crop Although there may be a case for large yields, profit or output value per hectare commercial farms under certain agro- (Figure 5 -3). This corroborates similar ecological and market access conditions, findings in other, smaller-scale studies (see this will need to be established on a case- Ballard and So 2004). by-case basis. In the meantime, there is a sound economic rationale for policies and Reallocating idle “economic” concession programs that prioritize support to small land to smallholders through “social” farms (e.g. through creating an enabling concessions has a clear rationale in terms environment for the emergence of of improved productive efficiency. intermediary institutions, such as voluntary Increased land utilization compared to the private marketing associations or contract existing economic concessions will boost aggregate economic activity. However, farming arrangements between smallholders and processinghransporting the implications of the findings on small agents, that will allow them to overcome farm productivity go beyond simply

Figure 5.3: Small farms are a better economic proposition than large farms

A. Small farms have higher outputs per B. ...and higher crop income per hectare hectare.. . (y axis plots log of output value / ha.) (y axis plots crop income / ha.)

* 4 s

-io -5 0 5 -10 -5 0 5 Farm size, ha (log) Farm size, ha. (log) Source: CSES 2004 (see regression results in Table A2.4 in the Annex). 88 Cambodia - Poverty Assessment 2004 marketing problems (SNEC et a1 2005). land allocation initiated a boom in non-rice The greater local linkages typically cultivation; the magnitude was so associated with small-scale agriculture significant that Cambodia’s remarkable relative to plantation-type approaches will growth during recent years in maize, also help boost local non-agricultural cassava, and soybeans could be attributed economic growth and result in greater to the Northwest region. poverty elasticity (Ravallion and Datt, 1996; Thorbecke and Jung, 1996). Factors to enhance agricultural Early experience in moving forward on productivity social land concessions illustrates how Based on regression ana lyse^'^ from the difficult it will be to implement even on a CSES 2004, the most important factor limited scale (Dinravy and Groetschel affecting agricultural productivity is secure 2004). Yet progress is needed on both land title. Other factors, for example more secure titling for existing landowners infrastructure and human capital, also and redistribution of land from (redundant) influence productivity to varying extent. A economic concessions to the landless (and, number of other factors such as extension possibly, the land-poor) under the social services and credit access are potentially land concessions program. important but we are not able to investigate An example of the potential growth these factors in depth using the CSES 2004 benefits of the equitable distribution of data. land is that of Malai District in Cambodia’s northwest, in which Well-defined property rights deforested land was allocated to Agricultural productivity and growth will households following the integration of be enhanced with well-defined property this former Khmer Rouge stronghold into rights of both private and public land. the Royal Government (Box 5.1). This Firstly, as much as 80 percent of the rural households that owned land were without Box 5.1 : The success story of Malai land titles in 2004. Second, access to common property resources, Le., forestry Peace resulting from the integration of and fishery, has been increasingly former Khmers Rouges into the insecure. The frequency of reported land Government in 1998-99 led to remarkable conflicts has increased and their severity development of the Malai district. The land time concessions which were provided by the worsened over (Box 5.2). Government to the former Khmer Rouges in International evidence suggests that the exchange for peace played an important social and political costs of land conflicts role in igniting growth. Main factors that can be astronomical. contributed to the development and growth were the relatively large plot (5 hectare) given to each household and good governance in land distribution that resulted in equitable distribution of land by the authorities. As a result, the boom in the district from cash crop production has benefited not only the locals but migrants from other parts of the country that came to work as farm labor.

Source: Economic Institute of Cambodia, l7Detailed regression results are in the Annex. 2005. Raising Agricultural incomes 89

Box 5.2: Tracking the frequency and Box 5.3 Prevalent landlessness amidst severity of land conflicts fenced-off unused arable land Alternative sources all suggest that the in Phnom Sruoch frequency and severity of land conflicts in According to focus group participants in Cambodia is rising, even if they paint a Traeng Trayeung, Ou, and Mohasaing slightly different picture on the detail of communes, landlessness in the former trends. Oxfam GB has supported the commune was between 30 and 40 percent maintenance of a database of reported and available arable land was about 3,540 land conflicts, which enables the analysis hectares. However, local villagers of trends up to April 2005. This finds that cultivated less than 300 hectares while the conflicts became slightly less common in remaining was unused but inaccessible or 2001 and 2002. This is probably unavailable for rental. Phnom Penh elites attributable to the fact that in 2001, the held large tracts of land for speculative Ministry of Land Management, Urban purposes, and land is not rented out for Planning and Construction was fear that lessees will claim ownership rights established, and the Land Law was because most of them do not have proper passed. It appears that there were land titles. expectations from would-be contestants and from the institutions responsible, that Source: Economic Institute of Cambodia, the new Law would provide for serious 2005. penalties for those appropriating land. In 2004, however, land disputes rose again by 50 percent, which may be seen to Large tracts of cleared forest land held by reflect the fact that the Land Law had not the urban elites that could have been rented been strictly implemented. out for cultivation are left idle for fear that Newspaper coverage of disputes, lessees will claim ownership (Box 5.3). although an imperfect measure, The opportunity costs of keeping these corroborates the impression of a growing problem. Cases of land conflict reported large tracts of land idle are substantial, in Phnom Penh Post between June 18, whether measured in terms of foregone 2004 and August 11, 2005 suggests that employment, income, or growth. such conflicts became more frequent and more violent. Analyses based on the 2004 CSES demonstrate that secure land tenure on Early findings from research on the World Bank Justice for the Poor project provide private plots raises revenue, yields, some insights into the reasons underlying productivity, rental value and sales value, increased frequency of disputes. In three as well as household consumption". Districts in two Provinces, the research Secure land rights and efficient land found that villagers were: (i) increasingly administration systems are critical for willing to press claims, even against the rich and powerful, reflecting more agricultural growth and for facilitating exit pressure on land and thus more serious from the sector. A survey carried out in consequences if people are disposed; (ii) 2004 in three provinces with relatively increasing spatial integration, with people good market linkages found that there is a in remoter areas now feeling less isolated positive correlation between secure land and vulnerable; and (iii) increased access to outside powers via personal tenure and formal credit, in that secure connections, political parties and NGOs. land tenure provides an incentive for lenders to provide additional credit on Source: Oxfam GB 2005; Phnom Penh Post 14 (5-15), 13 (13-23), 2004-2005. more favorable termslg. Importantly, the

18 Detailed regression results are in the Annex. l9 CDRI (2004). 90 Cambodia - Poverty Assessment 2004 survey found that most credit clients were Box 5.4: Income fluctuates with the women, underlining the need to emphasize amount of rainfall joint titling in the name of husband and wife (70 percent of titles under the current In both high-growth areas such as Malai in Banteay Meanchey Province or lagging systematic titling program). Districts such as Mesang (Prey Veng) or Phnom Sruoch (Kompong Speu), lack of Irrigation facilities irrigation facilities is cited as the main constraint for growth. Farmers rely only on Intensive agriculture will remain of rainwater because the irrigation network is marginal importance until more reliable non-existent. In Malai, yields of soy bean water supply is assured. Even in the Plains, and rice fell dramatically because of lack of where proximity to the Mekong would rainfall in 2005. In Phnom Sruoch, most suggest an advantage, 44 percent of plots have suffered in recent years due to consecutive droughts. Farmers and have no irrigation in either the wet or the officials in Mesang acknowledged that dry season. To illustrate the potential of irrigation facilities are costly but insisted dry-season irrigation, the one-in-four plots that without them, the consequences are in the Plains benefiting from dry-season even more costly. irrigation, supply 86 percent of (national) Source: EIC 2005. dry-season rice production2'. Analysis renewed emphasis on irrigation based on the CSES 2004 show that investments, although effective use of irrigation facilities enhance farm public expenditure remains a challenge. productivity. Crop yield is 10 percent Environmental considerations are also a higher on irrigated land than on un- factor. Many potential areas for large scale irrigated land, given the same levels of irrigation are now in protected or preserved other inputs (e.g., labor, capital, fertilizer, areas. land type, etc.)21. Land with access to irrigation facilities during the dry season has 15 percent greater rental value and 10 Roads and other infrastructure percent higher sale value than land without Agro-climatic conditions and remoteness (Deininger, 2005). Only 3 percent of poor are highly correlated; local topographical households owned the water pumps conditions (particularly rain-fall patterns necessary to distribute irrigation in 2004, and irrigation) limit agricultural potential whereas 12 percent of richer households while simultaneously explaining the lack had water pumps. Qualitative assessments of previous investments in physical based on interviews with farmers in infrastructure. Historically these various regions of the country also reveal cumulative effects have disadvantaged that the lack of irrigation facilities hurts certain communities. their livelihood (Box 5.4). Analyses of farm-level productivity also Current budget allocations of US$14 indicate the importance of other rural million for capital investment under the infrastructure such as access to markets Ministry of Water Resources suggest a and roads22. In particular, farms with access to markets in their villages have 26 percent higher crop yields than those 2o Moreover, dry season paddy is typically higher without. The analysis also finds that farms quality varieties sold to market, compared to with shorter distances to all-weather-roads wet-season paddy which is produced for household consumption. Detailed regression results are in the Annex. 22 Detailed regression results are in the Annex. Raising Agricultural incomes 91

Table 5.6: Estimated effects of road infrastructure and human capital on agricultural productivity Effects at mean level of productivity (7=6.348 thousand rievhectare): One std. dev. increase in x One unit increase in x Absolute change Percent change in Absolute change in Percent change in in productivity productivity productivity (‘000 productivity 8 sd(x) (‘000 riel per ( bsd(x) *I00) riel per hectare) (8’100) hectare) (BY) ( &sdsd(x> 1 Distance to all-weather -0.005 7.712 -235.9 -3.6 -30.6 -0.5 road (km.) Years of schooling of 0.006 3.415 132.8 2.1 38.9 0.6 head

Source: CSES 2004

have significantly higher crop yields. rural Cambodia23. For example, an Controlling for all other factors, decreasing additional year of schooling for the the distance from the village to an all- household’s head will enhance weather-road by one kilometer will productivity by 40,000 Riel per hectare enhance productivity by about 30,000 riel (Table 5.6). On the health front, recent per hectare (Table 5.6). evidence suggested that illnesses tend to While the benefits of rural infrastructure cause more immediate damage to livelihood than crop failures, even though are evident, analyses of the costs of both can be devastating to rural households infrastructure have yet to be carried out (Box with sufficient rigor to determine what 5.5). forms of investment in infrastructure (in particular irrigation) are priorities for different regions in the country. A study currently under preparation by the Technical Working Group on Agriculture (forthcoming in 2006) should help to answer these questions.

Human capital In addition to secure land title and rural infrastructure, human capital such as physical well-being and education is an indispensable complement to enhance productivity. As with findings in the international literature, analysis based on the 2004 CSES suggests that education significantly enhances productivity, farm

profitability and household consumption in 23 Detailed regression results are in the Annex. 92 Cambodia - Poverty Assessment 2004

Box 5.5: A health shock is often more serious than a harvest failure Results from surveys in two villages south of Phnom Penh suggest that while crop failures and illnesses are both devastating experiences, entailing similar magnitudes of economic damage (averaging a few hundred thousand Riels), households find it harder to cope with illness. The negative consequences from health shocks are more damaging to both immediate and long-run livelihood because they require an immediate lump-sum of money for urgent treatment. Because most households do not have sufficient savings (and rural credit markets do not operate well), households are often forced to resort to distress sale of productive assets (including land) and/or enter long-term debt, reducing their future income streams and increasing their non- consumption expenditures, respectively. This broadly confirms the findings of earlier Oxfam studies which found that half of all distress sales, or around 40 percent of cases of once-landowning families losing land, involved health crises. Sources: Yagura 2005; Biddulph 2004; Ballard and So 2004. 6. Education, health and wealth

Summary .Human development is both intrinsic to the definition ofwhat constitutes a good standard of living and instrumental to economic development. Investments in human development are investments in human capital. A healthy, well-educated population is more productive than one which is poor and illiterate. .Cambodia at present records extremely low scores in most human development indicators, including those used to measure progress towards the Cambodian Millennium Development Goals. However, the Ministries of Education and Health are amongst the most advanced in terms of strategic vision, an integrated approach to sector policies and programs, and coherent planning, budgeting and monitoring systems. Improved policies are, increasingly, resulting in improved outcomes. .The education system has made impressive gains. A focus on primary education, backed with rising public spending, has resulted in more schools and trained teachers; lower direct costs to households; and increasing primary enrolment rates (which have risen particularly fast for girls and for children from the bottom quintile). .Child nutrition remains extremely serious and the incidences of preventable diseases are very high, but there are clear gains in both preventative care (e.g. vaccination coverage and antenatal contacts) and curative care. There has been notable success in halting and reversing the spread of HN/AIDS. Indicators suggest that child survival rates are improving, but an accurate measure will only be possible when findings from the 2005 Demographic and Health Survey become available in 2006. .Many challenges remain. On the supply side, per capita public spending is still far below that required for a minimum standard ofbasic service delivery. Particularly with regard to healthcare, Cambodian households have to compensate by spending large proportions of their total income on generally rather low-quality care. High out-of-pocket spending can easily push non-poor households into poverty. .On the demand side, poor households may not be in a position to take up improved services. The main costs of education for poor parents are the opportunity costs of foregone child labor, which remains very important for poor families .In both sectors, sustained increases in public expenditure (although still far short of what is needed) have been supported by the development of a sector strategy, effective sector-wide partnerships, and policy changes. Further improvements in education and health outcomes, for the general population and for the poor, will require policies to tackle both deliverj (quantity and quality) and uptake of services. Finally, improving service delivery and human development outcomes will also involve addressing “governance” issues regarding the transparency and accountability of service provision; community participation in local service delivery; and increased volume. efficiency and effectiveness of public spending, increasingly oriented towards the needs oj the poor. 94 Cambodia - Poverty Assessment 2006

ow levels of human development- Effective government policies to mitigate that is, poor health and limited the impact of preventable disease, protect education-are both cause and the poor from impoverishing costs of L health care, and ensure access to effect of poverty as measured in terms of per capita income or expenditure. A affordable, essential high quality health foundation of basic education and literacy care and education are, in other words, key expands an individual’s economic public goods that increase the stock of opportunities, helping them to acquire human capital available to Cambodia and better jobs with higher and more facilitate long-term economic growth. dependable income and so work There is a strong presumption throughout themselves and their families out of this chapter, based on sound basic poverty. Education can also a powerful be economics, that the Government has to vehicle for improving awareness of health play a key role in the activities that add up issues and the uptake of preventative to the population’s (in particular the poor) health practices. access to an acceptable minimum package Similarly, an individual in good health of health and education services. remains more productive, and has more However, an overall strategy for money available to spend on other improving human development needs to essential goods and services, than an take account of the important role that is individual who is unable to work due to already being played by non-state actors illness, or who has to spend large amounts (particularly in the health sector), where of money on treatment. The economic the private sector is already a very consequences of a disease episode or significant provider of services as well as chronic illness that requires large mobilizing and managing funds. expenditures can push individuals and households into poverty, in particular Education, skills and growth when poorer households are forced into selling productive assets or into debt. Between Cambodian households as between countries, there is strong I11 health also has inter-generational effects correlation between increasing levels of in the context of (inter-linked) high child education or literacy on the one hand and mortality and high fertility. High fertility decreasing levels of poverty on the other. rates reduce the ability of households to It is of course important to be careful in provide adequately for health and interpreting this correlation. It is possible education for children, resulting in a that a better education leads to higher vicious circle of reduced opportunities for incomes, but it may also be simply that future earning potential. increasing wealth enables households to There are also society-wide impacts of ill- educate their children in a way that is not health. HIV/AIDS and even malaria possible for poorer households. epidemics can slow economic growth Intuitively, it seems probable that there are through their impact on labor and labor causal effects at work in both directions. productivity. New public health threats In particular, it is important to remember such as SARS or avian flu can that literacy and education can only significantly affect tourism and the service increase the potential productivity of industries. individuals. If the economy is unable to respond to this increased supply of skilled Education, health and wealth 95 individuals, they will remain underutilized Box 6.1: Peace, more schools and and the economy will continue to perform changes to fee policy have below its potential (Chapter 4). increased participation in primary education The RGC, with support from a number of Participants in the MOPS focus group donors, has over recent years, developed a discussions (FGDs) in all nine villages successful sector strategy which has begun agreed that the availability of schooling, to pay off in terms of improving level and and the level of participation of village equity in educational participation, children in education, was now much better than in 1998. To differing degrees in particularly at primary level’. different villages, this was attributed to: (i) the end of armed conflict; (ii) the Positive trends construction or upgrading of school buildings; (iii) rising standards of living Starting from a low base, the Cambodian which meant that more people could afford education system has made some to send their children to school; and, in impressive gains, a fact that is recognized particular, (iv) the abolition of the by participants in qualitative research (Box enrolment fee in 2000 or 2001. 6.1). The Government’s focus on primary Our young brothers and sisters are education (e.g. school operational budgets luckier than us because now they can for routine expenditures, the abolition of have better access to schools than us, [and] the schools are near our houses. enrolment fees, and school-based learning Kompong Thom remediation programs) has resulted in marked progress, particularly in Parents now are more alert and tend to push their children to school more than enrolment. This progress has been pro- before - and there are good schools, and poor, with enrolment (and completion) education is free. Male youth, FGD, increasing faster amongst the poor than the Kratie. rich, faster amongst girls than boys, and Source: CDRl2006a (forthcoming). faster in rural areas than urban areas. While the inequalities in education described above are still significant and Significant increases in public call for sustained action, it is important to funding for education acknowledge that some of them (notably These improvements have been driven by between boys and girls and between urban a steady rise in allocation of public and rural groups) have clearly been expenditure to education. Education’s narrowing since 1997 (Ragatz 2005, 12- share of the total RGC recurrent budget 14). rose to 18.5 percent in 2003, nearly double the level of around 10 percent in the late 1990s. In the sector as a whole, capital spending has declined over time as recurrent expenditures (wage and non- wage) have risen (Figure 6.1). While ’ Cambodia’s educational structure consists of donor assistance continues to account for a primary education (grades 1-6 for, in theory, major part of total public (i.e. non- ages 6-1 1 years); lower secondary education (grades 7-9); and upper secondary (grades 10- household) education spending, it no 12). Primary and lower secondary education longer exceeds government education are, in principle, compulsory, while upper spending (as it did in 1999). ODA flows secondary and early childhood education (age to education fell considerably in 2001 and 3-5) are not. 96 Cambodia - Poverty Assessment 2006

Figure 6.1: Public expenditure on the poorest 40 percent of households education-particularly on account for 45 percent of all pupils), recurrent costs-has risen government spending results in the balance dramatically ofbenefits accruing to the poor.

There are now more, and beffer, feachers and schools The results of increased public sector spending, made possible and directed by increasingly strong sector policies and institutions, are then seen in a number of 150 intermediate indicators of education supply, access and quality. The number of trained primary school teachers has increased by 12 percent (from 42,000 to 100 47,000) between 1999100 and 2002/03; an increased percentage of the teaching workforce now have u per secondary or 50 graduate qualifications! The quality of 2000 2001 2002 2003 physical infrastructure in the education sector has also improved considerably; a Source: MEF. variety of sources paint a consistent picture of more schools with better have remained largely unchanged since, facilities (see Table 6.1). declining in share relative to spending by the Ministry of Education, Youth and As more girls and children from the Sport (MOEYS), which has been poorest households enter and in some increasing since 1998. cases complete schooling, so the gaps in literacy between sexes and wealth groups This increase in financing for the sector is starting to close in younger age cohorts has made possible progress on the issue of who have more recently left school. teacher salaries. Average teacher salaries Whereas amongst those aged 25-29 years, have risen in a number of steps since 2002, the literacy rate in the richest quintile is 39 most recently by 37 percent (bringing the percent higher than in the poorest quintile, average to US$43.3 per month) between among the cohort aged 15-19, the gap has 2004 and 20052. narrowed to 22 percent. In this youngest These trends in public expenditure on cohort of 15-19 years, the gap in literacy education have also been strongly pro-poor between male and female youths is a mere in nature. Beneficiary incidence analysis 4 percent. brings out that, by prioritizing primary education (within which the children from

However, these across-the-board increases need now to be complemented by efforts to address the salary structure, which remains RGC 2003. The quality of pre-service very flat by international standards, and does training, however, remains low, and the not reward experience or working on priority opportunities for professional development tasks. limited (Ragatz 2005, p.41). Education, health and wealth 97

Table 6.1: There have been significant Current educational status improvements in education infrastructure Despite these positive trends, the current educational status of the a. Coverage 1998 2004 Cambodian population remains very Districts w/out lower N 32 14 low (Figure 6.2 for international secondary school % 17% 8% comparisons). To a significant degree, Districts w/out upper N 81 45 this reflects the legacy of a lost decade secondary school % 44 % 25% in the 197Os, in which all formal b. Facilities: 1999 2004 education ceased and those adults with schools with ... education (including many teachers) Drinkable water 31 % 64% suffered particularly severely from Latrine 34% 64% revolutionary violence under the c. Teachers' perceptions of change in school conditions over last 5 years Khmer Rouge. The collapse of the Primary Secondary Democratic Kampuchea regime was Lower Upper followed by only slow recovery as the education system was rebuilt from Improved 72% 65% 65% scratch under extreme resource Stayed the same 78% 26% 25% constraints from the 1980s onwards. Deteriorated 10% 9% 10% Source: EMIS, CSES 2004 Village Questionnaire. This legacy is reflected in both supply- and demand-side problems and in very Figure 6.2: Primary education in Cambodia low levels of completed schooling and compares poorly with the rest of the literacy amongst the adult population region compared to other countries in the region. By one indicator-that of gross Cambodia primary enrolment, which is the total 2004 enrolment in a specific level of Indonesia education, regardless of age, as a 2002 percentage of the school-age population corresponding to the same Philippines level of education in the given school- 1998 year(s)-the Cambodian education Vietnam system outperforms all its neighbors. 2002 However, this measure, which reflects rapid increase in participation in 0 25 50 75 100 education and (broadly) the ability of Adults with completed primary education the education system to absorb these (% of adults age 25+ who completed numbers, also reflects the prevalence of primary education - country-specific grades) over-age enrolment. By all other measures Cambodia scores below its Net primary enrolment (Students in primary school of national official primary school neighbors, including Lao PDR and age as % of population of this age) Myanmar4.

Primary school participation (Children age 6-1 1 currently in school, regardless of level, ' EdStats data, derived from national EMIS as % of the population aged 6-1 1) (Educational Management Information Systems) data. Indicators from Demographic and Health Surveys (DHStgenerally Source: Last DHS (year of survey noted for each country). 98 Cambodia - Povevtv Assessment 2006

Figure 6.3: Pronounced inequalities still exist in education outcomes

A. Amount of schooling (mean years completed) 10 6.4 5.4 5 3.7

0

Total Female Male I Rural Other Phnom Poorest Richest I urban Penh

bySX i by location by consumption I qui ntile

6. Percentage of population over 25 with no schooling

50 32

Total Female Male Rural Other Phnom ; Poorest Richest urban Penh ' by location ' by consumption quintile

Source: CSES 2004. In one of the most pronounced gender of schooling as those living in the capital. disparities in Cambodia, women have There are significant differences in net markedly less education than men (Figure enrolment rates amongst the poorest two 6.3). This is seen not only in years of quintiles, with most of the remote highland schooling but in literacy rates, in which the Provinces in the north, north-east and average (69.6 percent) masks significant south-west recording much lower rates differences between male (80.3 percent) than the Tonle Sap and Mekong Plains and female (60.3 percent) literacy rates. Provinces (Figure 6.4). Geographical inequalities are also stark; Late entry, repetition and dropout those living in the countryside have completed on average, only half the years rates are falling but still high Late entry is pervasive; only 28 percent of considered more reliable than EMIS, but not new entrants in grade one were of the available for all countries in all years- official entry age (six years old) or confirms the picture. Education, health and wealth 99

Figure 6.4: Children from poor families are much more likely to obtain a primary education in the south-east and north-west - Net primary enrolment (%) of poorest and next poorest quintiles, 2004

Source: CSES 2004. younger5. Comparing average age on underlying dropout and low completion entry from 2001 (the Child Labor Survey rates, however, is the importance of child or CLS ) to 2004 (the CSES) shows no labor to poor households. Although the change, suggesting that this phenomenon dramatic increase in primary enrolment is is to a large degree a structural problem, a major success to be celebrated (Figure rather than merely a transitional feature of 6.5), it appears to be much harder to keep an education system recovering from children in primary school beyond initial conflict. enrolment, and even harder to keep them attending through into secondary Overage enrolment is a problem primarily education. because it seems to be a significant factor in explaining high dropout. Those As expected, primary completion rates are children who start school late are highly by a significant margin higher for urban likely to drop out before they complete areas and the top wealth quintile. The primary education. The primary factor primary completion rate for the poorest quintile is very significantly lower than all other groups-and, unlike the gap between CSES2004. 100 Cambodia -Poverty Assessment 2006

The underlying problems Figure 6.5: Net primary enrolment has improved for all which prevent children (and A. For all wealth groups girls, the rural population, and the poor in particular) from completing education will need to be tackled if Cambodia is to attain the CMDG target of increasing the percentage of universal completion of primary education. The percentage of children needing to repeat grades has fallen considerably between 1997 and 2003 (from 26 Poorest Next Middle Next Richest percent to 10 percent in the poorest richest primary grades). Nonetheless, rates of repetition (particularly in grade one) remain very high, B. For both girls and boys C. For rural and urban second only to Lao PDR in populations the region. 100 Quality of education loo T---7 To get a full picture of the "Vm U 80 Cambodian public education system and its ability to 60 60 provide citizens with the skills they need to secure 40 40 higher-paying jobs and attract foreign investment, it is 20 20 necessary to look beyond simply getting children into 0 0 the classroom and keeping Female Male Rural Urban them there, to address the quality of teaching. Key Source: CSES 1997, CLS 2001, CSES 2004. indicators of this ability to transmit learning to pupils male and female or urban and rural include pupil-teacher ratios and the level completion rates, this gap between wealth of teacher qualifications. In looking into groups is not closing over time (Ragatz the quality of education, it is clear once 2005). again that the education system itself While primary completion rates do not suffers from the nation's short supply of now differ much for boys and girls, a human capital. There are not enough significant difference between the sexes teachers for the needs of a young emerges in lower secondary education. population. Although they have started to Education, health and wealth 101 drop in the last two years, pupil-teacher Box 6.2: Parents worry about the quality ratios are the highest in the region for of teaching primary education (in 2004 far below other Although far more children are in school, countries) and the third-worst at secondary the overall quality of education-and the levels6. availability and pay of teachers-was still a common complaint amongst parents in the The paucity of teachers is felt particularly MOPS villages: badly by the poor (pupil-teacher ratios in I have doubts about the quality of schools in the poorest communes average schooling-for example, my son who is in 79:l compared to 46:l in schools in the grade 6 now cannot even read his name. richest communes). These pupil-teacher I wonder what he has done in school and ratios, furthermore, got worse as success in why the teacher just lets him pass the class. Battambang. recruiting more teachers failed to keep up with success in increasing enrolment, We really appreciate that there are better school facilities now. But we are Since 2002, ratios appear to have concerned with the increasing stabilized in primary levels: they now need malpractices engaged in by the teachers to be brought down significantly. because of their low salary. They normally put as the topics for examination There is also a need to improve not merely topics that have been in their private the quantity of teachers but also their classes, especially after the primary level. qualifications and skills. This, and other Parents have to spend a lot on children’s dimensions of change required to improve education, especially from the secondary further the education and literacy of level. The teachers require our children to attend their private class and pay 400- Cambodian citizens, will require 500 riels per session. Kampot. significant and sustained increase in the Source: CDRl 2006a (forthcoming). level and efficiency of public financing allocated for education. In particular, there is a compelling need to further increase service delivery, there is a compelling teachers’ salaries if Cambodia is to need for government-wide reform of achieve benchmark goals set in the Fast public sector salaries and human resource Track Initiative. Globally, it is management (Chapter 7). recommended that low income countries Finally, while enrolment rates have risen pay teachers around 3.5 times the average (and been followed, with a lag, by rising national income (as measured by per completion rates), hours of classroom capita GDP) in order to ensure quality and the instruction within the Cambodian equitable education7. In Cambodia, by educational system is still far too low to contrast, the starting salary for teachers is enable significant learning. At a maximum below per capita GDP; mid-career salaries of 760 hours per annum, Cambodian average only c. 1.5 times per capita GDP. children receive far below the Education Low salaries-combined with weak for All (EFA) target of 850 hours per human resource management in the annum Tackling this will once again education system-contribute to the low require further increases in the allocation average quality of teaching (Box 6.2). and efficiency of public expenditures to Here as in many other aspects of RGC education.

6 Cambodia data from EMIS; other countries’ data from Edstats. EFA benchmark. 102 Cambodia -Poverty Assessment 2006

Level and allocation of public Box 6.3: The Priority Action Program spending on education The Priority Action Program (PAP) is one As noted above, the pattern of public of a series of attempts to address problems of rigidity, delay and uncertainty arising expenditure on health is broadly pro-poor from over-centralized budget execution. in that it favors primary education, in Introduced first as a pilot in education and which the poor are most prominently health ministries in 2000 and soon represented. However, as the unit costs expanded to MAFF and MRD, PAP has per pupil are significantly higher in sought to ensure that high priority activities obtain their full budget allocation by (i) secondary and post-secondary education removing PAP funds from the discretionary (in which the children of rich families are monthly cash allotment system; (ii) over-represented relative to the children of improving management flexibility by poor families), secondary and tertiary reducing the need for line item detail; (iii) education still capture a very large share of introducing budget management centers in spending agencies to allow them to the education budget in absolute terms, manage PAP funds according to pre- and the children of the rich are still approved plans; (iv) replacing pre-audit receiving more, in per child terms, than the with post-audit; and (v) creating in children of the poor. There is thus still Provincial treasuries dedicated accounts room to improve the poverty focus of that are not subject to discretion over cash release. The evidence for the impact of education sector policies and financing. PAP has been mixed; budget execution There is also a need, as in other sectors of still appears weak, suggesting both that RGC public financial management, to PAP failed fully to solve the problem of cash management and also that the improve the efficiency of disbursement of spending agencies themselves have been public funds. In an effort to tackle slow to disburse. There are also concerns frustrations with budget execution, the that, without greater capacity for post-audit Priority Action Program (PAP) was in the spending ministries, PAP has introduced in Education (and other weakened control and accountability. Over the medium- to long-term, efforts to ministries) from 2000 (Box 6.3). A strengthen PAP need to be framed in recently-completed Public Expenditure terms of incorporating its best features- Tracking Survey (PETS) for the education and eventually PAP itself-within gradual sector found that school operational funds moves towards a single, unified budget channeled through the Priority Action Plan execution * system for non-wage expenditure. (PAP)8 were transferred down as intended. However, the survey was based on official Source: World Bank and ADB 2004. and incomplete receipt data (the PAP system is characterized by low quality The costs and benefits of record-keeping), so it is still not clear how to interpret the findings. PETS-based education analysis of the older Chapter 11 Crucial to improving educational Operations and Maintenance (O&M) outcomes is the calculus that households, budget line found little of this money was rich and poor, make when assessing the spent below the Provincial level, despite returns to education against the costs the intention that it complement PAP involved. funds at the school level.

8 Strictly speaking, PAP 2.1, which is the primary schools’ operating budget. Education, health and wealth 103

Direct costs are coming down.. . Box 6.4: Supplementary private tutoring mirrors public schooling For poor families, the costs of educating their children can be a significant Supplementary tutoring, which operates as a sort of shadow system alongside the deterrent. The direct costs of education mainstream schooling system, consumes (that is, fees, uniform, transport, etc.) have considerable household resources, come down in recent years with the especially in urban areas and for key final abolition of enrolment fees (mentioned grades. One study indicates that in grades 3- often in focus group discussions in the 8 it accounts for approximately 20 percent of total education expenditure in urban areas. MOPS fieldwork). Nonetheless, although By grade 9 it is 38 percent. Whereas in they are now quite low in the early years primary school, tutoring usually covers the of education, they increase rapidly with whole curriculum, in lower secondary school each grade until they account for on it is specialized by subject. In Cambodia, average, nearly a quarter of total non-food much of the tutoring takes place in the students’ own schools and is provided by expenditure for children in grade 7, and 45 their own teachers. The tutoring system that percent of non-food expenditure when (or has evolved can be considered regressive in if) children reach grade 12. that it is a cost that poor households typically cannot afford. Supplementary tutoring is . . .but out-of-pocket costs continue closely linked to the poor working conditions of most teachers. Thus, for policy changes to constitute a burden on, and on tutoring to be effective, they must be barrier to the poor accompanied by changes in teacher remuneration. Although at any given level of schooling poor and rich households spend roughly Source: Ragatz 2005. the same on education as a proportion of their total non-food expenditure, this amounts to a significantly greater absolute Table 6.2: Child labor rates in Cambodia amount that rich parents invest in their are amongst the highest in children’s education. South-east Asia (% of age group) The composition of direct education costs Countrv 1997 2003 changes with age and varies between Malaysia 3 2 wealth groups. Tutoring costs-that is, Vietnam 8 4 costs for supplementary private teaching Philippines 7 4 outside school hours, although typically Thailand 15 10 provided by the same teacher-become Myanmar 24 22 progressively more important at higher Cambodia 24 23 grades, and is a cost that poorer families Lao PDR 26 24 can less frequently afford (either in Source: EdStats. monetary terms or in terms of children’s time free ofhousehold tasks (Box 6.4). Asian nations (Table 6.2), child labor in . . .and the indirect, opportunity costs Cambodia is common and begins at an early age. In both rural and urban settings, remain very high poor families often rely heavily on their More important than direct costs for poor children to help with a variety of tasks households, however, are the indirect (generally household-based) essential to costs. As in other low-income South-east the wellbeing ofthe family. 104 Cambodia - Poverty Assessment 2006

These forms of child labor include both Box 6.5: Villagers recognize links income-earning or subsistence-oriented between education and productive or trading activities, and opportunity household reproductive tasks (caring for Despite the improvements in availability younger siblings, cooking, cleaning, and uptake of schooling seen in recent years, education beyond the primary level carrying water, etc.)-much of which, of remains by and large a possibility only for course, is necessitated in order to free up the children of wealthy families; and some both parents for economic activity, children from poor and very poor families including in some cases long absences are still left out of primary school because from the home in pursuit of income they have to help their parents to earn or to look after the house. Amongst the poorer opportunities elsewhere in Cambodia or in families in the MOPS study villages, there Thailand. The CLS in 2001 found that was a strong sense that education was a children contributed on average 28 percent way out of poverty-but that, equally, oftotal household labor income'. poverty would continue to prevent their children obtaining an education. For many Cambodian children, work and Because we are poor, our children quit school attendance are largely substitutes school at an early age or after only one for one another, with this trade-off or two years in order to help their parents becoming more pronounced as children get earn a living. Unfortunately, they cannot older. The burden of housework and go as far as the rich do in obtaining skills productive work is a particular problem for to earn a living. Because we are trapped rural children, children of families in the in illiteracy, we have poor knowledge.and are without ideas, remaining short- poorest quintile, and girls. From age 12 sighted and powerless. Kampong Thom. onwards, household work is significantly Education is the most important asset of more likely to interfere with girls' people in this world. Being well education than boys. Children who work educated, someone can find a good job more than 14 hours a week-who are more with a high standard of living. With good likely to be poor, rural, and female-start education, s/he can manage to have a to fall behind in grade attainment better life or able to set a clear goal. The educated person never fears facing compared to their peers who only attend financial security since s/he has a school. This difference is insignificant at permanent, good paid job. For example, age ten, but pronounced amongst 17-year a literate woman is able to find a good olds". The high proportion of children job at the garment factory now. Those reported to both attend school and work who have no education cannot find such fortune and become mobile laborers, a suggests that the household need for child more risky life. Female youth group, labor not only contributes to outright Kampong Speu. dropout from school, but will also make it Source: CDRl (2006) forthcoming. hard to increase the currently very short school day. The benefits of education: returns to households and the nation The benefits ofeducation, however, appear to be significant (and are certainly seen as This figure coincides remarkably with the such by parents and young adults (Box estimates of their contribution to household income provided by youths, both male and 6.5). Analysis of the 2001 Labor Force female, in MOPS focus group discussions. Survey suggests that the returns to lo Ragatz 2005, p. 25-6. Education, health and wealth 105

Figure 6.6: Those with a good education are significantly schooling are sizeable and less likely to be poor increasing over time". - Poverty headcount by literacy status and level of education Increasing levels of education completed, 2004 improve the likelihood of having Literate paid employment (further Illiterate classified into self-employment and wage employment) rather TOTAL than unpaid work. This is particularly true of those who Uniwsity complete primary schooling Technlcal/mcatlonal compared to those with no Upper secondary Lower secondary schooling, with the increase in Pnmary likelihood declining with higher Other stages of schooling. Within the None category of paid employment, 0 10 20 % 30 40 50 however, secondary education significantly increases the Source: CSES 2004, analyzed in Knowles 2005a. probability of getting a waged (rather than self-employed) paid job in a way that primary schooling does These findings suggest that, as expected, not. education does expand an individual's economic opportunities and their chances Schooling also has a significant impact in ofescaping or staying out ofpoverty13. terms of the stability of income, as measured by being in permanent rather At a household level then, there would than temporary (seasonal or short-term) appear to be clear benefits to education employment. For employees, completion within the economy as it is currently of primary and lower secondary schooling structured. Analysis suggests that this resulted in the greatest improvement in the reflects a low-level equilibrium: although likelihood of being in permanent rather current demand for skilled and educated than temporary employment. Amongst the workers in the labor market is low, there is self-employed, it is secondary education nonetheless a return to education because that has the greatest effects12. the levels of overall education are correspondingly low. As in previous national surveys, CSES data shows that the likelihood of being One way of looking at this is that even poor drops significantly with completion though the sectors that currently dominate of each successive level of education the formal, waged economy (e.g. garment (Figure 6.6). Conversely, analysis of adult manufacturing) do not require a high level wages (of salaried employees) and earnings (of the self-employed) show 13 Since education precedes work in an increasing average incomes with higher individual's life, it is presumed that education levels of education (Figure 6.7). explains income rather than vice versa. However, one or both variables might be determined by one or more other, unobserved variables: for example, income might be 'I Ridao-Can0 2005, pp. 24-35. shaped primarily by education or wealth of l2 Probit analysis of 2001 Labour Force Survey parents, with own income moving in line with data in Ridao-Can0 2005, pp. 27-30. own education but not causing it. 106 Cambodia - Poverty Assessment 2006

Figure 6.7: Average wages and earnings increase with each stage of education completed A. Wages (employees), 2001

Universlty

Technical & vocational

Upper secondary

Lower secondary

Rimry

No schooling

0% 25% 50% 75% 100%

I

Upper secondary

Lower secondary

Rimry

No schooling

0% 25% 50% 75% 100%

c 49,999r 50,000r-99.999r rn 100,000r-149,999r 8 150,000r-199,999r 200,000r-299.999r 0 300,000r-499,999r > 500,000r

Source: Cambodia Labor Force Survey data, analyzed in Ridao-Can0 2005, pp. 32-3. of education, employers can choose to somewhat weak effect on the Cambodian select only those with, for example, lower labor market. secondary education. If however the Cambodian economy is to It is important to distinguish between a sustain the economic growth of the level static and a dynamic analysis of the and (equally important) the kind needed to relationships between human capital (in drive an economic transition and long-term this case, education and skills) and development, it will need to diversify economic development. The static manufacturing and exports beyond its analysis above suggests that at present current, rather risky dependence on a few employers are broadly satisfied with the largely low-wage growth sectors (Chapter skills available in the labor market. This 4), and find alternative sources of decent is in large part because the structural employment and earnings for the vast factors that normally drive demand for majority currently employed primarily in educated, skilled labor within an low-input, weather-dependent, economy-for example, broad-based trade subsistence-oriented small-holder and foreign investment-currently have a agriculture (Chapter 5). To enable Cambodian workers to expand their Education, health and wealth 107

Table 6.3: The children of educated mothers are much less likely to be malnourished

- Mean rates (%) of moderate and severe malnutrition among children under 5, by mother’s level of schooling, 2004

Height for age Weight for age Weight for height Mother’s (stunting ) (und ewe i g ht) (wasting) schooling Moderate Severe Moderate Severe Moderate Severe None 59.6 38.9 51.6 18.5 12.9 2.8 Primary 53.6 32.6 46.8 16.3 12.2 2.6 Lower secondary 46.7 28.0 34.3 8.6 9.3 2.4 Upper secondary+ 38.9 21 .I 31.4 9.4 10.9 2.5 Total 54.1 33.6 46.0 15.7 12.0 2.7 N 5,738 5,738 5,738 5,738 5,738 5,738

Source: 2004 CSES (15-month sample).

economic opportunities, and Cambodia as not attending, “need to contribute to a nation to attract the domestic and foreign household income” and “need to help with investment required for this economic household chores” emerge as the most transformation, will require an increase in common answers, each of which accounts the stock of educated workers who can for between 14-31 percent of the total supply the needs for higher-wage, greater (depending on the sub-sample used). value-added economic activities. Interestingly, reasons to do with spatial Non-economic returns to education coverage-“no suitable school available/ school too far”-emerged as important but Education also has important inter- (at 9 percent of all who have never generational, inter-sectoral and often attended, and 4 percent of those who societal effects. The level of a mother’s attended but stopped) less important than education in particular appears to be a in the past. The impression is that with the significant determinant of child health. exception of extremely remote and One powerful example is with regard to sparsely populated areas (in which only a child malnutrition, in which the level of very small proportion of the population education of the child’s mother appears to lives), the physical availability of schools be a strong influence upon malnutrition is not the critical i~sue’~. (Table 6.3).

Barriers to access for the poor 14 Other parts of the CSES dataset reinforce These answers can be disaggregated between those children who have never attended school the impression that it is the indirect, and those who have attended in the past but are opportunity costs of education that no longer attending (7.8 percent and 18 percent constitute the most important barrier of those aged 10-18 years, respectively). preventing the poor from obtaining an Amongst the “never attended” group, “Idon’t education and all the advantages that go want to” (25 percent) was by a small margin more common than chores (22 percent), need with it. When those not attending school to contribute income in this group accounted were asked the most important reason for for another 14 percent. 108 Cambodia - Povertv Assessment 2006

Multivariate analysis of the determinants improving the educational opportunities of enrolment and retention reveals more and outcomes for the poor, which will help detailed insights into the interplay of them to better contribute to and benefit factors that influence households’ from national development, requires decisions regarding the schooling of their increased attention to post-primary children (Box 6.6). education. At the same time, the

A low level of education is one of the few Box 6.6: Explaining enrolment and variables that are found to be significantly retention: findings from associated with self-reported disability. multivariate analysis Only 2.3 percent of those who have Demand side, household-level factors completed upper secondary education or Among all children aged 15-17, the higher report one or more disabilities, mother’s level of education is the primary compared to 7.4 percent of those who have determinant of girls’ schooling while the father’s level of education is the primary no schooling. In other words, disabled determinant of boys’ schooling. children have significantly less access to A child’s relationship with the head of the education than do non-disabled ~hildren’~. household may have more impact on Disabilities in vision, hearing or speaking retention than on school entry. appear to be the most important in 0 Poverty and stunting play a key role in preventing children from receiving an determining dropout rates. education. 0 Late school entry is significantly associated with higher dropout rates. Governmen t policies, funding Supply-side, school-level factors and targets 0 The higher the average experience and education levels of teachers, the lower As described above, the improvements the dropout rates seen in education have been achieved Schools with health- and learning-related through a combination of (i)concerted facilities have consistently lower dropout efforts to develop and implement a rates. strategic framework within which The quality of school management, education policies have been refined; and teacher monetary incentives and the (ii) significantly increased public degree of community participation in the expenditure on the education sector. school play a role in attracting children to school early and keeping them in school As MOEYS works with MOP and MEF to Incomplete schools contribute to late formulate the education strategy that will entry and higher dropout and repetition be reflected in the 2006-2010 NSDP, it rates in those grades offered. faces the need to balance a number of Children who live physically near a lower competing considerations. Further secondary school are more likely to stay in primary school. Schools with preschool facilities attached l5 The conclusion is that disability is a barrier to to them have consistently lower dropout education, rather than that less educated people and repetition rates. are more likely to become disabled. This is based on the observation that there is no Contextual factors difference in incidence of disability based on a 0 Schools located in communes with higher mother’s level of education for children under levels of inequality have significantly 15 years old (who were excluded from the higher dropout rates. calculations above). If low education caused disability, this would not be the case. Knowles Source: World Bank 2005a, pp. 50-57. 2005b, p. 4-6. Education, health and wealth 109

Government needs to protect the budgets Box 6.7: Scholarships help girls from for basic education by developing poor families stay in school sustainable financing strategies to support High direct and indirect costs constitute a Education for All goals. While post- significant barrier to post-primary education primary education does now merit more for children from poor families. Girls in attention, it is important to recognize that particular are unlikely to be able to continue into secondary education without the bottleneck is now faced first and special assistance. A scholarship program foremost in the upper years of primary supported by the Japan Fund for Poverty school rather than lower secondary school. Reduction (JFPR) and administered by the It is at this stage that direct and indirect ADB attempts to address this particular costs for households appear to act as a group. In 2004 the program targeted 93 lower secondary schools. Within each significant barrier to the continued school, 45 girls or ethnic minority children attendance ofchildren from poor families. who were starting grade 7 received The poorest quintile in particular faces scholarships of US45 each. An early evaluation of the program suggests that clear cost barriers, as long as they are the JFPR scholarship had a large and forced to rely on their children working to positive effect on the girls’ participation in support their families. Many poor families school, concluding that a girl who was explicitly recognize that pulling children offered a scholarship was 10-30 percent out of school to work solves a short-term more likely to stay in school than was a comparable girl who was not offered a problem of income flow at a high long- scholarship. The result of this program term price. Without this investment in suggests that scholarships can be a useful their education, the children of the poorest tool for ameliorating the economic 20 percent of Cambodians are unlikely to pressures that give rise to inequities in be able to realize their potential to secure education (which in turn perpetuate the inter-generational transmission of poverty). employment at decent wages or earnings which could, in turn, allow them to work Source: Multivariate analysis by Deon their way out of poverty and contribute to Filmer and Norbert Schady, cited in Ragatz 2005. the development of the nation. In the interests of both equity and Parental and community involvement in macroeconomic efficiency, there is thus a running local schools can make an strong case for public investments in invaluable contribution and should be promoting access to education for poor addressed within the ESSP, but will only children. International experience be able to achieve small gains until teacher suggests that preschool or early childhood pay and human resource management education may well merit more significant (including training) are tackled. investment as a way to tackle late entry Although indirect costs are the most and the problems that follow. There is significant barrier to the education of poor also a good case for scholarships for pupils children, direct costs remain important. from poor families, as a mechanism to The current situation in which underpaid compensate poor families for lost income public school teachers teach short hours from child labor and so increase the and supplement their incomes with out-of- probability that their children will remain hours private tutoring is regressive and in school (Box 6.7)-and for re-entry and contributes to the low internal efficiency equivalency programs. of the education system. Addressing this problem in a comprehensive manner will 110 Cambodia - Poverty Assessment 2006 ultimately require tackling the systemic health-related problems that relate to problem of low public sector pay through poverty in Cambodia. whole-of-Government civil service reform.

Table 6.4: Despite gains, Cambodia’s health outcomes lag behind Health and the rest of the region wealth - Selected aggregate health indicators, 2003 (CIPS 2004) some There have been Infant Under-5 Maternal Life expectancy notable improvements mortality mortality mortality at birth in Cambodians’ health rate rate rate female male over the last decade. Cambodia 97 (66) 140 (83) 450 57 (65) 50 (60) Although Cambodia Indonesia 31 41 230 68 65 still has the highest Lao PDR 82 91 650 60 58 prevalence of Myanmar 76 107 360 63 56 HIV/AIDS in the Thailand 23 26 44 73 67 region, it has been Viet Nam 19 23 130 74 68 successful in arresting Sources: WHO and UNICEF. Cambodia estimates derived from and reversing the CIPS shown in brackets. growth ofthe epidemic. Similarly, the tuberculosis (TB) epidemic shows a declining trend. There are also 111 health and the poor in reported improvements the provision of Many of the aggregate health indicators health care, in particular for maternal and for Cambodia have been stagnant. The child health, delivery of the minimum most recent internationally reported package of activities (MPA) and the TB estimates’* for average life expectancy, program. child and infant mortality as well as for While many key MDG indicators (notably maternal mortality date from 2003 and child and maternal mortality rates) have show no improvement over previous until recently remained stagnant or reports. worsened (UN 2005), new evidence However, the 2004 Cambodian Intercensal suggests that the infant mortality rate Population Survey (CIPS) and 2004 CSES under-five (IMR) and mortality rate suggest significant improvements in some (USMR) may have started to decline16. of the aggregate health status indicators. This would need to be verified with other The CIPS reports that the IMR has sources, most notably the 2005 Cambodia declined from 93 deaths per 1,000 births in Demographic and Health Survey (CDHS), 1998 to 66 in 2003; and that U5Mdeclined the results of which will be available in from 124 in 1998 to 97 in 2003. The CIPS late 200617. estimates that average life expectancy at Malnutrition and lack ofaccess to essential birth has increased for men from 52 to 60 health care appear to be the most salient years, and for women from 56 to 65 years between 1998 and 2003, Le., gaining nine l6 Based on data from the Cambodia Inter-Censal Population Survey (CIPS) and the CSES, both conducted in 2004. Internationally reported (WHO, UNICEF) l7 The CIPS and CSES rely on indirect health indicators appear to be extrapolations of measurements compared for direct country and regional trends as no surveys were measurement ofkey MDG indicators by DHS conducted in 2003. Education, health and wealth 111

Table 6.5: Communicable diseases are the main cause of illness for all groups-although non-communicable illnesses are becoming significant for the rich - Reported type of health problem or care need during the past four weeks by per capita consumption quintile, 2004.

h problem or care

Source: CSES 2004.

percent), Communicable diseases are years oflife expectancy over five years. If more prevalent among poorer groups (see years confirmed, these would represent Table 6.5). Epidemiologic transition is very significant changes occurring over a limited and more detectable among the very short period oftime. The reliability of richer parts ofpopulation. these new estimates need to be verified by subsequent demographic studies (e.g. the Box 6.8: Access to health and “new 2006 DHS). diseases” But even with those improvements, In general, the picture of improvements in health services and health status obtained in Cambodia would still rank behind most of the nine MOPS villages was more muted than the countries in the region and with other the equivalent picture of improvements in countries ofsimilar economic standing. education. On the positive side of the ledger, it was acknowledged that access to public Overall morbidity curative services had increased with the construction of health centers and hospitals, In the 2004 CSES, 19 percent of the and many villages reported the introduction of population reported illness during the preventative health programs and significant preceding four weeks. Communicable declines in the number of people affected by diseases dominate all age groups, traditional diseases such as TB, malaria or accounting for 83 percent of the reported diarrhea. However, there were numerous complaints regarding the cost of healthcare, disease burden. Even for the elderly (age which were often perceived to be rising relative 75+), communicable disease accounted to average earnings; and a widespread from 67 percent of reported health perception that “new diseases”-particularly problems. For the age group 0-5 year, the high blood pressure and stomach ulcers- rate was 96 percent. were on the increase (often attributed to increasing consumption of purchased foods Communicable disease also ranks high as and the suspicion that these contained harmful the leading cause of death (37 percent), chemicals from pesticide-intensive agriculture), second only to accidents and injuries (42 Source: CDRl2006a (forthcoming). 112 Cambodia -Poverty Assessment 2006

.The prevalence of reported morbidity is however, that child mortality is strongly positively correlated with socio-economic and negatively correlated with the status. This is not an uncommon finding education of a mother. A child born to a and is usually linked to higher sensitivity mother with secondary and higher to changes in health status, greater education has about 2.5 times better knowledge of and ready access to health chances to survive. The statistical analysis care for the higher income groups. of the CSES data also suggests that child Similarly, the 2004 CSES did not indicate mortality is negatively correlated with any clear overall relationship between access of health services and positively reported disability and consumption correlated with malaria prevalence. poverty. The overall disability rate at the Deaths among children below five years of individual level was 4.7 percent (higher age make up 20 percent of all deaths in the than the 3 percent reported in the 1996 poorest socio-economic quintile, compared SESC and 1.6 percent reported in the 2000 to 13 percent among the richest. This DHS although this may be due to different reflects the “fertility trap”. Significantly methodological approaches). The CSES higher mortality rates among the poor reveals, however, that disability is more create an incentive for more births, which prevalent in rural areas, disease, fever, increase the health risks for the mother and difficulty with deliveries and malnutrition reduce the level of per capita spending on accounting for about 35 percent of total health and education available for each reported disability. child. Another look into disease patterns and HIWAIDS poverty comes from village leaders who under the CSES questionnaire were asked Controlling the spread of the HIV to identify up to three most important epidemic has been a notable success for health problems. Twenty nine percent of the Royal Government (Figure 6.8). This the population in the poorest quintile can be attributed to the law on prevention reside in villages that ranked malaria Figure 6.8: Cambodia has achieved remarkable as the top health issue, compared to 13 success in reversing the spread of HIV percent in the richest quintile. By - Estimated prevalence (%) among adults aged 15-49, contrast, injuries and accidents were 1995-2003. identified as relatively more important health problems in the villages where 47 the rich reside (7 percent versus only 1 2.9 3.0 2.8 2,7 percent in the villages inhabited by the 2.5 2: poorest quintile). 231 2: 2.2 2.1 ,,9 Child and infant mortality do not have a consistent relationship with per capita consumption, although 1-1 estimated mortality is lowest in the 0~ , , richest consumption quintile. It is, 1995 96 97 98 99 2000 01 02 03 however consistently higher in the Mountain and Plateau and Tonle Sap regions and lower in Phnom Penh and Source: NCHADS surveillance data, December 2004, reported in Addendum to UNAIDS 2004. urban areas. The CSES data suggests, Education, health and wealth 113 and care concerning HIV and AIDS. This 2000 DHS. The levels are much higher provides a sound basis for comprehensive than in other low-income countries in the multisectoral interventions. This enabling region (Figure 6.9). policy environment and firm government In the 2000 DHS, the prevalence of commitment has helped to mobilize donors stunting and low weight for age amongst and civil society to help implement anti- children below age 5 was 45 percent, AIDS programs. compared to 54 percent and 46 percent Cambodia is also one of the few countries respectively in CSES 2004 (see Table 6.6). worldwide that has been successful in Cambodia clearly exhibits a pattern of making anti-retroviral treatment available widespread protein energy malnutrition. to AIDS patients and meeting the target set by the global “3 by 5” initiative”. Figure 6.9: Malnutrition rates in Cambodia are extremely high However, there is concern that the A. Low weight for age epidemic is shifting. Husband-to-wife (underweight) infection is now the major mode of 75% - transmission and one third of all new HIV I infections are from mother to child. I Increasing youth risk behavior and illicit drug use among youth and people in labor intensive activities could also become potential drivers of HIV infection. The 2004 CSES reports high overall knowledge of the disease (over 90 percent) Carrbodia LaoFDR Myanmar among the population, but somewhat lower among the youth. The knowledge of B. Low height for age HIV/AIDS and methods of prevention is (stunting) lower among the poor and among women in all consumption quintiles. As anti- retroviral therapy is becoming available in Cambodia, concerted effort on prevention activities should be maintained to counterweigh stabilization of prevalence of HIV/AIDS and potential resurgence of risk behaviors because of risk Cambodia Lao PDR Myanmar compensation.

Nutrition C. Low weight for height (wasting or thinness) Malnutrition remains a serious problem for DSevere UModerate Cambodian children and Cambodia 1 generally. The 2004 CSES data suggests there has been no improvement since the l9 The “3 by 5” initiative sought to ensure that, globally, three million AIDS patients received ARV treatment by 2005. The target for Cambodia was 10,000 patients. Cambodia Lao PDR Myanmar

Source: UNICEF database for Lao PDR, Myanmar, 2001. CSES 2004 for Cambodia 114 Cambodia - Poverty Assessment 2006

Table 6.6: Mean rates (%) of moderate and severe malnutrition among children under 5 by per capita consumption quintile, 2004

Low height for age Low weight for age Low weight for height (stunting) (underweight) (wasting) Qui nti le Moderate Severe Moderate Severe Moderate Severe Poorest 57.8 37.7 53.9 19.7 13.2 2.4 Next poorest 54.6 33.1 46.6 15.7 12.3 2.6 Middle 55.2 34.5 45.7 16.7 12.2 2.0 Next richest 52.3 31.3 42.4 13.6 10.8 3.1 Richest 48.4 29.3 36.5 10.1 10.6 3.3 Total 54.2 33.7 46.1 15.7 12.0 2.6 N 5,905 5,905 5,905 5,905 5,905 5,905 Source: 2004 CSES (15-month sample); Knowles 2005b, p. 30.

Malnutrition indicators climb severely feeding and inappropriate complementary between 6 and 24 months of age, food) and poor public health (a high risk of indicating a combination of un-nutritious diarrhea, TB and other infections that complementary food and childhood prevent nutrition uptake) also play a diseases (acute respiratory infections, significant role. diarrhea and other infections such as TB). Cambodia also exhibits moderate to The 2004 CSES indicates that only 59 potentially severe micro-nutrient percent of individuals in the poorest malnutrition. The 1997 MOH survey quintile reported having had adequate food estimated a gross national goiter rate at 17 supply for every day in the past 12 months percent and it is estimated that about 1.3 compared to 88 percent of the richest million people are at risk of iodine quintile. deficiency disorders. There are some If malnutrition rates are significantly signs of improvement, however, as the higher for the poor, the differences proportion of households reporting use of between wealth quintiles are less iodized salt has increased from 12 percent pronounced than could be expected and in 2000 (DHS data) to 27 percent in 2004 less than in other countries2'. In (CSES data). It should be noted however Cambodia, even the children of the that the DHS estimates were based on salt relatively rich suffer from high rates of that was tested, whereas the CSES malnutrition. estimates are self-reported. There are significant differences between socio- This supports the perception that while economic strata: 12 percent of households inadequate food access (and by extension in the poorest quintile report using iodized low incomes and low and variable salt compared to 50 percent of richest productivity of food crops for household quintile. consumption) is critical, poor infant and child feeding practices (e.g. too early or Vitamin A deficiency is linked to access to too late introduction of complementary vitamin--rich food for children, poor breastfeeding practices, and high

2o prevalence of childhood diseases such as Although malnutrition as a cause of disability is significantly more common amongst the acute respiratory infections, diarrhea and poorest. measles. According to the 2004 CSES, 82 Education, health and wealth 115 percent of children under 5 were given stunted mothers are more likely to give vitamin A. This reported uptake of birth to underweight children. vitamin A is considerably higher than the Several studies have documented a strong 29 percent reported in the 2000 DHS. correlation between adult height and There may be some ambiguity however, as earnings. Malnourished and sick children the 2004 CSES did not distinguish are more likely to reach adulthood with between vitamin A supplements and food reduced height. There is also evidence rich with vitamin A. The 2000 DHS that the health and nutritional status of reported that 76 percent of children children are linked with educational consumed vitamin A rich food. Children attainment, both in terms of cognitive of mothers with at least primary schooling capability and school attendance. The are at least 7 percent more likely to receive effect on adult earnings and productivity is vitamin A than those without schooling. estimated at 10 percent for stunting and 4 In the past, Cambodia has also reported percent for childhood anemia. Estimates high levels of iron deficiency (63 percent of these losses to the economy are among children, 58 percent among estimated at 2-3 percent of GDP in low women, 66 percent among pregnant income countries (ADB 2001). is likely women, 2000 DHS). This a It has been estimated in one recent contributory factor in high maternal death international study that the present value rates in Cambodia (437 per 100,000 births, (using a discount rate of 5 percent) of the DHS 2000). benefits of shifting one child at age 3 from The causes of malnutrition are complex. malnourished to normal nutritional status They range from household food is US US$51421. If this estimate (or insecurity to inadequate care practices to anything close to it) is accepted, the lack of access to essential health services. potential benefits from interventions Malnutrition combined with childhood designed to reduce Cambodia’s currently illnesses can have significant human and very high rates of child malnutrition are economic costs. Malnutrition contributes about US$84 million annually or about to 20-25 percent of global burden of US$6.50 per capita per annum22. Bank disease (World 1993) and the WHO As with other health indicators, child estimates that 16 percent of global malnutrition sharply declines with Disability-Adjusted Life - Years (DALYs) increases in the level of the mother’s a measure of the impact of ill-health on education, in particular between primary labor productivity can be attributed to and secondary level schooling (Table 6.3). malnutrition. Lack of information and knowledge about There is clear evidence that the major the “abnormality” and potential impact of damage caused by malnutrition takes place malnutrition may be a reason why in the womb and in the first two years of life; that this damage is irreversible; that it 21 Behrman, Alderman and Hoddinott 2004 cited in has Knowles 2005. consequences for cognitive and 22 physical capacity (which in turn reduce According to the 2004 CSES, there were about 277,000 children age three in 2004. productivity, slow economic growth and Multiplying this number by the percentage of perpetuate poverty); and that malnutrition children age 3 who are estimated to be passes from generation to generation, since moderately stunted and the estimated potential benefits per malnourished child of US$514 yields the estimate cited in the text. 116 Cambodia - Poverty Assessment 2006 malnutrition is not reported as a major monitoring and evaluating nutrition health concern (malnutrition made up only program process and impact indicator^^^. 0.7 percent of self-reported overall morbidity and 0.2 percent for the age Fertility group 0-4). This lack ofinformation could High fertility rates can lead families with be linked to the low level of education resource constraints to under-invest in the among women, only 18 percent of whom health and education of the next have secondary or higher levels of generation. Lower child mortality rates schooling. combined with lower fertility rates would It is clear that significant effort needs to be allow households to improve their put into improving nutritional status to economic status through investments in improve future opportunities, promote health and education. economic wellbeing, and break the vicious These demographic effects also have an circle of poverty. Responses need to be impact on the economy-wide level. When multi-sectoral, including overall economic child mortality rates decline, followed by and social policies to spur economic declines in fertility rates, overall growth, reduce poverty and mitigate the population growth tends to slow, the impact ofpoverty. As in other countries, it average age of the population begins to appears that educating women is a critical rise and dependency ratios fall, so boosting factor in reducing malnutrition over time. per capita income and economic growth Although health sector interventions on (Bloom and Canning 2001). nutrition are fairly limited, in the short and medium term, effective and accessible The most recent international sources health services and nutrition programs estimate Cambodia’s total fertility rate at 5 could help to mitigate the impact of in 2000 (Figure 6.10). This compares to malnutrition. Education is important and 5.6 in 1990. However, as with other key links need to be sought with the education demographic indicators, the 2004 CIPS sector for school-based interventions (e.g. estimates are significantly lower than school feeding programs or micro-nutrient previously thought. The CIPS puts the initiatives). Effective nutrition programs fertility rate at 3.3 in 2003 and the CSES at are likely to have pro-poor impact given 2.9. Factors underlying high fertility rates the higher prevalence of malnutrition include high child mortality rates, cultural amongst the poorer quintiles. beliefs, lack of schooling of women and limited work opportunities, and lack of An assessment of the sector reveals a essential health services and effective number ofnutrition programs with varying counseling. However, the proportion of scope and scale, but, at present, without married women using modem birth adequate coordination. Most nutrition spacing methods has increased. In 2004, programs are NGO- and donor-led. There 20 percent of married women between is a need for more significant government ages 15-49 accessed contraceptives involvement, strengthening of relevant through the public sector24. There has also institutions and resource allocation, and been significant contraceptive distribution

23 Cambodia Nutrition Investment Program, 2002. 24 Based on MOH administrative data from 2004 based on users of public health services. Education, health and wealth 117 by means of the private sector Figure 6.10: Poor and uneducated women have through donor-funded social more children than other women marketing programs. - Number of children ever born to a woman between This represents significant ages 15-49 improvement since 2000, when 19 percent of women had access to Plateau I Mountain contraceptives, including 13 percent Coastal from public sector25. Tonle Sap Plains There are significant variations in the Phnom Penh use of contraceptives. The CPR of the wealthiest quintile is Rural approximately twice that of the Urban poorest, the highest provincial coverage is about three times that of Richest Next richest the lowest, and the CPR of women Middle with the highest education level is Next poorest about twice as that of women with Poorest no education26. These challenges need to be met by implementing the Upper secondary and aboie Cambodia population policy Lower secondary (announced in 2004) and by Primary mitigating the impact of high No schooling fertility, for example by lowering the 01234 cost of and improving access to education and essential health Source: CSES 2004. services. discussed earlier, protection against Disease prevention, health measles is also an important factor in knowledge and risky behavior reducing malnutrition. It is thus encouraging that there has been immunization coverage some improvement in immunization Vaccination against common coverage in Cambodia. The 2004 CSES communicable diseases is one of the most reports full immunization coverage for 40 cost-effective public health interventions, percent of children aged 12-23 months with significant individual and social before they were 12 months of age and for benefits. The WHO recommends that all 50 percent of children between 13-24 children receive one BCG (against TB), months of age regardless of their age at the three polio and three DPT (diphtheria, time of vaccination (this compares to 30 pertussis, tetanus) and one measles percent and 40 percent respectively in vaccination before age one. As already 2000 DHS).

25 Achieving the MillenniumDevelopment Goals 2005 Update, Royal Government o Cambodia. 26 Cambodia at a Glance - Population, Gender and Reproductive Health, 2005. UNFPA Cambodia. 118 Cambodia - Poverty Assessment 2006

But by all standards, coverage is Figure 6.1 1 : The level of a mother’s education is still low and some of the strongly and positively associated with vaccination coverage may be children’s immunization partial. For example, the MOH - Children between 13-24 months of age never vaccinated, reports that 85 percent ofchildren by selected characteristics below one year received DPT3 vaccine in 2004 compared to 64 Plateau / Mountain percent in 2002. Children in the Coastal 15% poorest consumption quintile Tonle Sap have almost twice the probability Plains ofnever being vaccinated. Phnom Penh Schooling of a child’s mother Rural also makes a difference to the big Urban uptake of immunization services. Twelve percent of the children of Richest mothers with no schooling do not Next richest get vaccinations, compared to 7 Middle percent amongst children of Next poorest mothers with primary education Poorest and 4 percent amongst children of - mothers with secondary Upper secondary and aboE education (Figure 6.1 1). There Lower secondary also appear to be problems with Primary immunization coverage in coastal No schooling regions (which corresponds with 7 low status of health service 0% 10% 20% utilization in the coastal regions). Source: CSES 2004. Use of bed nefs Sleeping under insecticide-impregnated malaria regions in the Mountaifllateau bed nets is a cost-effective measure against regions should be targeted first; here, the malaria and other mosquito-borne reported use of impregnated nets is still diseases. The CSES 2004 reports a high only 14 percent. This finding is a little level of utilization of bed nets by surprising given that the MOH reports that population (94 percent) but only 4 percent 80 percent of villages in endemic malaria of the sample reports using impregnated regions received annual re-treatment and nets. Although richer groups report replacement of bed nets. Limited slightly higher use ofnets, the difference is knowledge by villagers ofwhether the nets not large (96 percent ofthe richest quintile are impregnated or not may be the reason compared to 90 percent of the poorest). for the cited difference. Low utilization of impregnated nets is likely to be linked to lack of availability or Breastfeeding information, as there is no significant Breastfeeding is an important factor in difference in use between different socio- protecting infants from diarrhea and acute economic groups. When scaling up respiratory infections and stimulates the impregnated net use, heavily infested Education, health and wealth 119 development of the immune system. born one year or more before the 2004 Breast milk contains all the important CSES survey were 5 percent less likely to nutrients and antioxidants needed for have been given breast milk as first liquid infant development and protection. to a child. Given that this coincides with According to some studies it also enhances increased efforts under the joint health cognitive development. WHO sector development program between recommends early initiation of MOH and donors that report improved breastfeeding (within 1 hour of birth) access to appropriate antenatal care and exclusive breastfeeding during the first 6 deliveries with qualified attendant in 2004, months of life and continued breastfeeding the health system may have had impact in to two years, accompanied by appropriate improving breastfeeding practices. complementary feeding. As discussed earlier, the Cambodian population has high Tobacco and alcohol consumption levels of malnutrition and childhood As discussed above, the epidemiologic illnesses, and in this context promoting transition (from communicable diseases of appropriate breast-feeding practices matter poverty to non-communicable “lifestyle” a great deal. diseases) is still at an early stage in In 2004, 98 percent of children under three Cambodia and still confined to the richer have been breast fed; 14 percent of infants parts of the population. However, given began breastfeeding within one hour of the time lag between some health birth and 49 percent within one day. This behaviors and their impact on health, it is compares to 11 percent and 24 percent worth looking at smoking and other risk respectively in 2000. But still, only 14 factors as proxies for future outcomes. percent of male infants and 20 percent of Cambodia exhibits high levels of smoking female infants aged 0-6 months were being rates among males (42 percent). These exclusively breastfed on the day of the rates steadily increase with age (up to 70 CSES interview. Infants of mothers with percent among males of 65-74 years of primary education are 11 percent than age). Low smoking prevalence among infants of mothers with no education more women (4.1 percent) help to bring the likely to receive breast milk as their first overall smoking prevalence down to 21 liquid after birth; infants of mothers with percent. On average, a smoker in secondary education are 15 percent more Cambodia smokes 14 cigarettes a day. likely. There is no time series data available to There appears to be an impact of some assess whether the prevalence of smoking recent developments as infants who were is increasing or decreasing. 120 Cambodia - Poverty Assessment 2006

The CSES reveals, however, Figure 6.12: Smoking is overwhelmingly a male habit, that the likelihood of and declines with education smoking is - Smoking prevalence per consumption quintile and level of to per capita COnSUmptiOn schooling as proportion of population and schooling. The poor and men with no schooling have Richest the highest smoking Next richest prevalence rates (Figure Middle 6.12). Although in general, knowledge about the health Next poorest impact of smoking is high Poorest (87 percent of population over 15 years of age is aware), twice as many Upper secondary and abow respondents in the poorest quintile than in the richest Lower secondary quintile think that it is not Primary harmful (7 percent and 3 No schooling percent, respectively). 0% 25% 50% 75% High smoking levels among the poor suggest that as the ' Females demographic and epidemiologic transitions Source: CSES 2004. proceed, the prevalence of non-communicable diseases future but also today, in the form of (NCD) is likely to rise among the poor increased opportunities to spend more on population. In order to prevent some health, education and productive avoidable NCD morbidity and mortality in investments. the future, public policy measures to curb smoking Figure 6.13: Compared to the rich, the poor spend are warranted today. proportionately more on tobacco and less on medical care The poor spend more on - Percentage of household non-food consumption on alcohol, tobacco than on medical tobacco and medical care. CSES 2004. care as a proportion of their non-food consumption. The Richest --I combined share of tobacco and alcohol Next richest I 1 spending among the \ \ poorest quintile is more Middle - than twice that spent on Next poorest medical care (Figure 6.13). Thus, potential Poorest benefits from reduced risky behavior would 0% 5% 10% 15% 20% 25% accrue not only in the 1 WAlcohol UTobacco UMedical care 1 Source: CSES 2004. Education, health and wealth 121

Access to health services and healthcare. Less than 60 percent of the patterns of use poor who are in need of health care use it, compared to 74 percent of the richest Since the mid-l990s, the Royal (Table 6.7). This difference is not that Government has made significant efforts marked in Phnom Penh. The poor appear to improve delivery of basic health to have the least access to health care in services. In recent years, these efforts the coastal zone where only 36 percent of have been framed by the health sector the poorest in need utilize health services, strategy and joint annual health sector as compared to 67 percent of the richest programming with external partners, (Table 6.8). Hospital care is particularly adopted in 2003. This has helped to inaccessible for the poor. The two richest increase attendance at public health centers quintiles use hospital service five times from 0.38 to 0.42 for all age groups and more often than the poorest quintile (Table from 0.54 to 0.74 for children below five 6.9) years. In 2004, 164 health centers were able to provide integrated management of Access to health services depends on childhood illnesses, compared to just 45 in several factors: physical access (distance, 2002. Provision of vital micronutrients to condition ofroads); ability to pay (for care children and pregnant women has also and transport); knowledge and information increased. Between 2002 and 2004, the about availability; personal beliefs and percentage of pregnant women who perceptions of need and quality of health receive at least two antenatal consultations care; and of the functioning of referral from public health services has risen from systems. 29 percent to 47 percent; deliveries attended by a trained public practitioner has increased from 20 percent to 32 percent. All 962 health centers and 70 referral hospitals provide the full DOTS program for TB patients. But huge challenges remain. The poor find it hard to obtain access to quality

Table 6.7: The poor are much less likely to seek healthcare when ill - Percentage of persons reported ill or in need of health care during the past 4 weeks who sought care Quintile By sex By location Total Male Female Urban Rural Poorest 60 58 59 59 59 Next poorest 60 63 63 62 62 Middle 64 64 60 64 64 Next richest 68 70 70 69 69 Richest 74 75 81 72 75 Total 66 67 73 65 66 Source: CSES 2004. 122 Cambodia - Poverty Assessment 2006

Table 6.8: Outside Phnom Penh and the Tonle Sap provinces, wealth makes a big difference to healthcare utilization - Percentage of persons reported ill or in need of health care during the past 4 weeks who sought care Geographical zone Per capita consumption quintile Total Next Next Richest Poorest poorest Middle richest Phnom Penh 88 78 80 86 87 86 Plains zone 54 57 58 69 72 63 Tonle Sap zone 64 67 68 68 67 67 Coastal zone 36 55 61 55 67 57 Plateau/Mountain 59 66 73 69 73 67 zone Total 59 62 64 69 75 66

Source: CSES 2004.

Physical access is a significant issue, traditional healers) are available within 1-4 especially for hospital care. Only nine km. In terms of health programs, village percent of the poorest quintile lives in a leaders judged that only the immunization village with a health center; the mean program is available for all segments of distance to the closest health center for this the population in their village. The quintile is 7.6 km and 16 km to referral Mother and Child Health (MCH) and hospital. This compares to 6 km and 9 km family planning programs, HIV testing and for the richest quintile. In comparison, iodine programs are all more readily private traditional providers (traditional available in communities where rich birth attendant, kru khmer and/or other people reside.

Table 6.9: Individuals in the richest quintile are five times more likely to use hospital services than individuals from the poorest two quintiles - Annual hospital admission rate per 1,000 persons Per catita consumdion auintile Total Poorest Next Middle Next Richest poorest richest by sex Male 28 29 56 88 144 69 Female 28 31 44 92 179 75 by location Urban 39 56 45 63 98 73 Rural 27 28 51 95 197 72 Total 28 30 50 90 162 72 Source: CSES 2004

Secondly, there is the issue of cost true of public health services as well as barriers to access. Health care in private care. Because the public health Cambodia is relatively expensive. This is service is desperately under-funded, Education, health and wealth 123 obtaining service requires high fees at the From the 2004 CSES, it is estimated that point of service. In 2004, the budget out-of-pocket payments (OOPS) for health allocation to health was 1.26 percent of care amount to US$15.48 per capita per GDP or 7.6 percent ofgovernment planned year. However, this estimate is based on expenditures. Allocations have steadily data collected only from individuals who increased over the recent years. However, were reported to have had an illness, injury the execution ofpublic budgets has been a or other health problem during the problem (Chapter 7). In 2004, only 83 preceding four weeks and may therefore percent of the operating budget was be under-estimated. The 2000 DHS disbursed by the end of the calendar year; reported out-of-pocket expenditure at this was even lower at the provincial level US$24 per capita. The size of OOPS (63 percent). This has been accompanied increase with the level of household with unpredictable and delayed flows of consumption, both in absolute terms and in hnds that mostly occurred in the second terms of share of non-food consumption. semester of the calendar year. As a While nearly 19 percent of those in the consequence, per capita public health richest quintile paid for healthcare in the expenditures in 2004 were only US$2.94 four weeks preceding the CSES interview, equivalent (less than in 2003 and 2002, only 12 percent of the poorest paid for US$ 3.16 and US$4.10 re~pectively)~~. care. At 14,857 riels per capita, the average out-of-pocket spending on health This under-funding puts considerable care during the past 4 weeks amongst the pressure on front-line services and the richest quintile was almost 19 times that of institutions have had little option but to the poorest quintile (Figure 6.14). charge user fees for services. In the early and mid-l990s, there was no scope for formalizing these fees (as healthcare was Figure 6.14: Frequency (% individuals) and theoretically free); fees were therefore size (mean riels p.c. shown by informal, untransparent and largely size of circle) of out of pocket arbitrary. User fees in public facilities have spending on health care in the been institutionalized by the Government last 4 weeks, by consumption quintile, 2004 since 1997 with the aim to smooth out public financing fluctuations, improve salaries of front line staff and regulate I Nextrichest, unofficial fees. Although the formalization and standardization of fees, together with exemptions for poor patients is a clear improvement, these fees can often still Richest 14,857- contribute a barrier to access. User fees in Middle. 2,493 r r public facilities and out-of-pocket Poorest, 800 r expenditures for private health care amount to 86 percent of total health expenditures per capita, i.e. six times more than government expenditures.

Per capita consumption quintile 27 It should be noted though that for the referenced years, undisbursed budgets were Source: CSES 2004. carried over to the following year. 124 Cambodia - Poverty Assessment 2006

The same applies to catastrophic health than five times higher than that of people costs (i.e. treatments requiring payments of with no disability. This reflects lack of more than 50 percent of a household’s insurance in Cambodia and makes this part annual non-food consumption). The same of population extremely vulnerable to ill pattern ofOOPS is true whether analysis is health induced poverty (Figure 6.15). in terms of per capita or per household High-out-of-pocket expenditures on health costs. However, the population in rural and health services may throw a household areas appears to have higher OOPS than in into poverty. Households may be forced urban areas across consumption quintiles. to sell or mortgage productive assets (e.g. The burden of OOPS is thus lower among animals or land), or simply cease the poor than the rich. However, this is expenditures on items that can enhance most likely because the poor simply do not household productivity in either the short use health services, and if they do, they or long term (tools, fuel to go to market, use lower cost and lower quality services. children’s education, etc). The 2004 In this context, it is surprising that only 12 CSES reports that borrowing is one of the percent of village leaders representing the coping mechanisms for covering OOPS. poorest communities list the high cost of Illness and injury rank fourth as the most health services as the most important common reasons for taking out a loan (13 health services problem in the CSES percent of all loans). The relative burden (behind lack of medicines and long of health-related indebtedness is higher distance to better quality care). among the poor (1 1 percent of value of all outstanding loans compared to 7.5 percent Figure 6.15: The 10 percent of the population for the richest). In a recent study, illness with bad or very bad health was the number one reason for land sales. account for 64 percent of total Unlike harvest failure, the impact of a household spending on healthcare health crisis requires significant lump sum - Percentage of total out-of-pocket spending by payments. Even when households first reported health status, 2004 borrowed to meet the costs of a

catastrophic health event, they were later 75Qh 1 60% forced to sell land because of high interest 55% - rates (Kenjiro 2005). It is clear from the above that protecting the population from high health care- related out-of-pocket costs should be one of the priorities for health system development. While the poorest should be the first priority, it is also desirable to Very Good Awage Bad Very good bad provide protection against costs over a certain threshold to the non-poor among Source: CSES 2004 the 10 percent of population with more serious health status facing high out-of- The 10 percent of the population with bad pocket costs. This would not necessarily or very bad reported health status account mean abolishing user fees but rather for 64 percent of total OOPS. A similar institutionalizing ways of protecting the pattern is also true for people with poor and seriously ill from the full impact disabilities, whose mean OOPS is more ofuser fees. Education, health and wealth 125

Figure 6.16: Inadequate drug supply, cost of services and distance to facilities are seen as the key problems with health services - Frequency of answers given by village leaders when asked to choose top three problems with health services (public and private) in their village of residence

Not enough medicines, drugs 30%

Health serdces are too expensie

Long distance to better quality care 15%

No physician or qualified medical assistant amilable Lack of beds in hospital, equipment etc.

Staff are unhelpful 8% Poor quality of serdces 6% No secondary nurse or midwife amilable 0 3% Other b2%

Source: CSES 2004 (Village Questionnaire) Poor quality of care is another factor households (and the costs of that care are explaining low utilization of health care. somewhat less important, relatively Although this continues to be a problem, speaking: (Figure 6.17). the trend is positive. The 2004 CSES Matching the CSES data against the asked village leaders in the 900 sampled database of public health facilities and villages whether health services for the 2002 Health Coverage Plan reveals that people in their villages had improved over significant rich-poor differences exist in the past five years. Village leaders overall availability of staff, availability of reported improvement for villages midwives and doctors, staff-to-bed ratios, inhabited by 56 percent of population and and doctor-to-bed ratios in the primary deterioration for only 6 percent of the coverage facility to particular population. population. According to the same survey, For example, only 5 percent of the poorest lack of drugs is the most commonly quintile is served by a primary health perceived problem with health services, center staffed with a doctor, compared to with expense and distance also accounting 25 percent of the richest quintile. for a large proportion of reported health Similarly, the primary hospital for the service problems (Figure 6.16). There is poorest quintile had 13 midwives, considerable consistency regarding the compared to 18 in the richest. These nature of the problems with healthcare. indicators suggest that, in addition to The main difference that can be imputed physical access and financial barriers, the between groups is that, when the answers poor are also disadvantaged in terms ofthe given by village chiefs are assigned to the quality of health care facilities that are households in their villages, it would seem available to them. that distance to quality health care is relatively more important for poor 126 Cambodia - Poverty Assessment 2006

Figure 6.17: For the poor, distance to health population who reported using health facilities may be as important as the care in the cost of care There may be a number of factors that - frequency problem was mentioned in village level deter the poor from seeking health care questionnaire, by household consumption status from public health care providers. A. Health services are too expensive Likely explanations are the still-low 25 public perception of the quality and 20 availability of publicly-provided health services (absence of qualified 15 personnel, formal and informal charges, and lack of drugs) as well as 10 the proximity of private traditional providers. 5 Amongst the poorest quintile, 78 0 percent consulted a private provider Poorest Next Middle Next Richest when they utilized health care. The poorest richest most common private consultations occur in drug shops (33 percent). The B. Long distance to better quality care other important private providers are 20 off-duty or absent-from-duty public n providers (for ex amp le, nurses) , 15 usually located in the villages they 10 serve. Thirteen percent of private consultations occur in the doctor’s or 5 nurse’s home, as villagers find it much more convenient than traveling to a 0 health center; and 18 percent of Poorest Next Middle Next Richest consultations occur in the patient’s poorest richest home by a visiting private health care provider. Source: CSES 2004 (Village Questionnaire). There is evidence, however, from both Public and private provision of CSES 2004 as well as from the health services Public health centers and referral hospitals 28 Focus group discussions in the MOPS and are closer to the poorest segments of PPA research studies revealed a continued preference for private treatment over public- population (7.6 and 16 km, respectively) despite the higher cost of private treatment- than equivalent private facilities (private based on a perception that the skills or clinic 17 km; private hospital 94 km). medicines of the private practitioners were However, publicly-provided health better than those ofthe public facilities. Speed services are used by only 17 percent ofthe of treatment and convenience were also reported as important issues. If a villager is taken ill at night they will seek treatment from the pharmacists who live in the village rather than make the long journey to the nearest health center (CDRI 2006a, forthcoming, 2006b, forthcoming). Education, health and wealth 127

Government reporting that children and expensive high-end services by public women utilize health centers more providers. frequently. The likely explanation is that MCH services are more readily available Engaging civil society as in health centers. partners in service delivery Over the 199Os, much effort went into The Government of Cambodia has improving the rates ofutilization ofpublic experimented with community health services, on the assumption that the participation and innovative public- public sector provided a better quality of private-partnerships to improve the care than traditional healers and modem governance and management of public providers or drug sellers. Private health services. The best-known example providers are perceived to have low levels of these attempts to improve the delivery of medical knowledge, sometimes of public services through partnerships prescribing harmful treatments. This, with non-state actors involves the accompanied with low level of regulation contracting of NGOs to manage health and incentives to over-prescribe, may services at the Operational District level result in treatment that is ineffective, over- (Box 6.9). The Government has also priced, or in some cases actually harmful. expressed a willingness to channel part of Many Cambodians self treat with drugs Government funds through NGO-run purchased from pharmacies or other retail equity funds that help to compensate outlets selling drugs. facilities for the user fees and other health care-related out-of-pocket costs for the Although private providers are not poor. There is emerging evidence that necessarily the cheapest, their these initiatives have helped to increase responsiveness to patients’ preferences and transparency of health facility operations their availability seems to make them (through user fee schedules approved by attractive options. The Government has communities), access to health services for recognized this and has included changing the poor and coverage by essential health the behavior of public health care interventions (e.g. antenatal care, assisted providers as a policy priority. So far, deliveries or vaccination). however, the impact on utilization of public providers has been limited. Since the early 199Os, attempts to ensure community participation and oversight in There is also a question about the the management of health facilities has beneficiary incidence of government moved from pilots to become official expenditures on health. Although the poor policy. In 1997, the first national tend to use public services more than the guidelines on community participation rich, the difference is not large (19 percent were included in the Health Finance vs. 17 percent). As the poor tend to use Charter, which gave Health Centre more health centers and district hospitals Management Committees (HCMCs) the (87 percent of those who used public task of setting user fees and exemption services) when the rich use provincial and policy. In 2003 the MoHestablished national hospitals (53 percent ofthose who (with assistance from a standardized used public services), public expenditures UNICEF) policy which mandates two types of on health are likely to provide more benefit to the rich, who consume more Committees for each health center. The first (the Village Health Support Group) serves primarily as a channel for 128 Cambodia - Poverty Assessment 2006

1 Box 6.9: Contracting-out to NGOs for per Commune to the Health Centre healthcare: public-private Management Committee (HCMC). The partnerships for service delivery HCMC sets objectives for the annual In an experiment to improve management and operational plan; monitors implementation incentives in the delivery of rural health care, and achievements; introduces and the Government, with support from the ADB, selected nine Operational Districts (ODs) in monitors the health finance scheme; and which to try alternative approaches. Two ODs manages the HCbudget29. were contracted out (in which the with contractors-international NGOs-had full A review of experience user fees responsibility for service delivery and full examined (among other aspects) the link to management control, including to hire and community participation and concluded fire). Three Districts were “contracted in” (the positively that the existence of functional contracted NGOs provided management management and feedback committees support and the Government covered appears to correlate with increased usage recurrent costs through normal channels); and four served as control Districts, in which and improved internal management and normal public health service arrangements morale of staff (Wilkinson, Holloway and applied. Surveys in 1997 (before the Fallavier 200 1). experiment began) and again in 2001 showed that while service delivery improved markedly in all Districts (reflecting the impact of national Health sector strategy and reforms of the public health system), the rate challenges ahead of improvement was dramatically greater in the As stated in the MOH strategic document contracted-out Districts. In part, this could be argued to reflect the higher per capita “Pushing Momentum for Health Sector Government/donor/NGO health spending in Development: Annual Operational Plan for contracted-out Districts (US$4.50 compared to 2005,” the Cambodian health sector is US$2.82 in the contracted-in Districts and confronted by the challenges of an US$1.86 in the control Districts); but this economy emerging from a period of increased level of spending was itself to some degree a reflection of improved management conflict, with widespread burden of practices. The NGO managers adopted infectious diseases, and malnutrition that innovative practices (e.g., in one District the poses a risk to the country’s human use of formal user fees, the proceeds of which capital. One may add to this widespread were used in a transparent manner to improve poverty, the threat of new infectious salaries and motivation; performance-based staff contracts; and the decentralization of diseases such as SARS and avian flu, and more highly-trained personnel to Health the likely increase burden of non- Centers). The improvements had a significant communicable diseases in the future. equity impact, with major reductions in the After initial improvements in the mid- average out-of-pocket health expenditure of 199Os, there was a setback in key MCH the poorest half of the population and a corresponding dramatic increase in the indicators in the late 1990s and early utilization of public facilities. With additional 2000s. But more recently there are signs donor funding, the Government has now of improvement that now need to be taken expanded the contracting out approach to further and made sustainable. cover 10 percent of all Cambodians. In the Government’s “Rectangular Sources: Keller and Schwartz 2001; Conway and Crossland 2002; World Bank 2003. Strategy” (2004), health sector strategy forms part of the “growth rectangle” of communication between the health center and beneficiaries on public health issues, 29 but also elects one man and one woman Committees have been much rarer, and less successful, in hospitals. Education, health and wealth 129

capacity building and human resource and/or where high numbers of the poor development. The health strategy is live. Given that out-of-pocket spending described in the Health Sector Strategic mostly impact the 10 percent ofpopulation Plan for 2003-2007, which identified six with poor health who may or may not be key areas for health sector development. among the poorest, the equity hnd These are health services delivery; initiative may also be needed to cover very behavioral change and communication for high out-of-pocket expenditure (above a health service providers; quality certain threshold, or for certain health improvement; human resources conditions with high costs) among the development; health financing; and non-poor to mitigate potential institutional development. These generic impoverishing impact of high out-of- categories need to be harnessed to address pocket spending. priority concerns of the MOH, including The execution of the government budget high infant and child mortality; high rates needs to become more predictable and of malnutrition; high maternal mortality stable. This would require coordination (including death from obstetric trauma and and cooperation with overall reforms of septic abortions); high case fatality rates public financial management (Chapter 7), from HIV/AIDS, malaria and TB; and high led by the Ministry of Finance and total fertility rates and population growth. Economy, as well as overcoming capacity The MOH and donors have committed to constraints within the Ministry of Health. joint annual operational planning and There is currently a disconnect between reviews ofhealth sector performances. the annual operational planning process Evidence presented in this report supports and the budgeting process. Compatibility the strategic directions of the Royal between MOH priorities, key areas of Government of Cambodia. The report also health sector development used in Annual points towards some of the key challenges Operating Planning (AOP) purposes and that need to be overcome. budget categories need to be established. This may require amendments in both First, public expenditures on health need to become more effective and more AOP structure as well as in current budget organization in the MOH and MEF. targeted towards the poor. There is direct evidence that the poor do not have access Moves towards more program budgeting need to be accompanied by a performance to essential health care when they need and monitoring system that would be able to that high out-of-pocket costs constitute an link expenditures and activities to important barrier. There is also indirect outcomes. evidence that the benefits from government health expenditures are The 2005 AOP estimates the cost of the consumed more by the richer segment of public health program to be US$ 7.70 per the population. capita. The Government would be able to Scaling up equity funds is a promising contribute US$3.48 per capita, leaving a initiative to protect the poor against out-of- gap to be supported by international pocket expenditures but achieving partners and donors. Such gaps are likely meaningful population coverage is likely to continue in the short- and medium-term. To accurately predict the financing needs to take time. In this context it is important objectives that the equity funds be established first in for the priority public health locations where poverty rates are highest and ensure financing, Cambodia will need to develop and adhere to a medium-term 130 Cambodia - Poverty Assessment 2006 expenditure framework that would provide underway, adequate resources, and some certainty about available government institutional strengthening. funding in the medium term. Given that only 20 percent of the Second, the delivery of public health population uses publicly-provided health services and programs need to become services, there is clearly scope for more effective. This would require not integrating the role of the private sector in just establishing and equipping new public helping to achieve public health goals. health centers and hospitals. Facilitating The Government has already recognized and rewarding good performance should the need to enact laws and regulations and become part of the institutional culture in to strengthen institutions to provide the health sector. Cambodia has adequate oversight and ensure standards of experimented with contracting NGOs to private providers. But these initiatives can manage and/or take over the provision of go further and include public-private public health service providers, with partnerships in service delivery. In encouraging results. Although the future selected cases, the Government could of scaling-up contracting in Cambodia is become a purchaser of health services unclear, the lessons learned should be rather than the integrated provider it is supported to improve health service now (e.g. through equity funds). Private performance, either through continued providers could also be integrated into NGO involvement, or in internalized referral systems by training them to public service delivery and recognize health conditions and symptoms implementation of equity funds. for which referral to more qualified providers should be considered. Work on Similarly, the ongoing civil service pay facilitating community participation in reform process may provide instruments to local health services delivery should also improve health service delivery. Different be continued. approaches are likely to be needed for motivating health ' administrators on Third, it needs to be acknowledged that the central, provincial and operational district determinants ofhealth outcomes are multi- levels, and for front line staff actually faceted and multi-sectoral. Time and delivering services to the population. The again, this report presents evidence that the need for greatest service improvement is in level of schooling of women and mothers rural areas where distances are longer and is a key factor for improved outcomes in where more ofthe poor reside. fertility, infant and child mortality, nutrition and breastfeeding, and There is also scope for improving the knowledge of HIV/AIDS. In particular, performance of selected health programs. secondary education seems to make a The 2004 CSES indicated that only the significant difference. Similarly, immunization program was evenly improvements in health status are critically available while others were more available dependent on further poverty reduction in villages with relatively wealthier and improved living standards (including populations. In particular, the nutrition nutrition), access to clean drinking water program needs to be strengthened, given and sanitation, and improved knowledge, the importance of the problem and attitudes and behavior with regards to apparent lack of improvement over the last behavioral health risk factors. four years. This would require improved coordination of various initiatives 7. Building an effective and responsive state

Summary Previous chapters discuss the factors underlying improved economic performance and service delivery, and some of the factors holding back further progress. Underlying sector-specific challenges are a number of systemic problems, many of which are fundamentally about “governance”, that is, how the rules, institutions, and systems of the state operate and how the state relates to citizens, civil society and the private sector in terms oftransparency and accountability. Ten years of strong growth performance only partially disguise some acute problems in economic governance. Quotas and safeguards have boosted Cambodia’s comparative advantage in garment exports; Angkor has driven tourism and construction. In other sectors, however, businesses struggle against many obstacles, the most important ofwhich relate to costs imposed by state policies and institutions. If the economy is to become more diversified (Chapter 4) and slow agricultural and rural growth is to be accelerated (Chapter 9, it will be necessary to address insecure property rights; unproductive regulatory requirements; inconsistent application of the law; un-transparent public sector contract processes; and corruption. Significant further improvements in service delivery will require fundamental reforms of public financial management (PFM) and public administration (Chapter 6). The state needs to: (i)raise more revenue; (ii)improve budget execution so that funds are made available to operational units on time and in full; (iii)align public expenditures to policy priorities; and (iv) dramatically improve the pay and management ofthe civil service, The four-stage program for PFM reform is an extremely positive step, the implementation of which is critically important for long-term poverty reduction. Corruption is arguably only a symptom of the weaknesses noted above. However, the scale of the problem in the eyes of citizens, businesses and donors is such that it merits attention in its own right. Corruption keeps people poor; they have to pay (or pay more) for goods and services. Benefits intended for them are diverted, and they are dispossessed in conflicts with other groups. Services are under-funded as revenue collection is foregone, and investment is deterred and employment suppressed. While the end of armed conflict has transformed the opportunities available to the poor (Chapter l),new threats to safety and property are emerging. Many people report a rise in violent youth gangs, driven partly by traffic in illegal drugs. There may, however, be some decline in domestic violence-though the evidence is mixed. Governance problems cannot be fully resolved through top-down reforms internal to the state. Harnessing the inputs of citizens-individually and as formed into groups-would help to create mechanisms for transparency and accountability that can provide a powerful force for reducing corruption and improving efficiency and equity (as seen in the health sector). In the absence ofequitable rule-based systems for allocating resources and resolving disputes, peaceful collective action by citizens is a crucial mechanism for leveraging increased state responsiveness to the needs ofthe Door. 132 Cambodia - Poverty Assessment 2006

revious chapters have examined laying the foundations for development. how poverty appears to have been These rather fragile structures were affected by the rate and pattern of subjected to a new set of pressures at the economic growth and by improvements in end of the 1980~~as external support was access to public services such as health and reduced and the PRK had to manage the education. This chapter attempts to pull transition to a market economy. out of this earlier discussion some over- In the decade since the Paris Peace arching conclusions about the role of the Accords, the newly-democratic Royal Government the development process, in Government ofCambodia (RGC) has, with a specific focus on its role with in assistance from a number of development managing a long-term, sustainable process partners, undertaken a broad range of ofpoverty reduction. public sector reforms. These have helped In addressing this issue of state produce state institutions better suited to effectiveness, the second half of this managing market-driven economic growth chapter looks in particular at the and improving service delivery. relationships between the state and its Fundamentally, however, the RGC remains citizens, both directly and as mediated in many ways a post-conflict state. Human through organized civil society. The and financial resources available to the assumption is that increasing the Government are severely limited. responsiveness of the state to the needs of Institutional arrangements and the its citizens improves its effectiveness in allocation of roles, resources and managing economic and social responsibilities often do not promote development. This responsiveness can transparency, accountability or efficiency. involve improving the channels by which The linkages between problem analysis, problems and preferences might be policy formulation and subsequent channeled from society and business to the implementation are weakly developed. In Government (issues often addressed under particular, public financial management the headings of participation, consultation systems do not effectively link policy or “voice”); and by establishing means to priorities to budget formulation, and hold the Government accountable for using weaknesses in budget execution undermine public resources honestly and effectively policy or results focus in actual allocations. in the public interest. In sum, the Government’s capacity to manage economic growth and deliver basic Building state institutions and services is improving, but remains public sector capacity underdeveloped.

At the end of the Khmer Rouge period, there were no functioning state institutions in Cambodia. The basic structures of the state had to be rebuilt from scratch by the new government of the People’s Republic of Kampuchea (PRK), and for the next ten years this government had to deploy much of its limited resources on countering insurgency, with very little remaining available to invest in reconstruction or Building an effective and responsive state 133

unsustainable open access exploitation and Creating an enabling informal commercial capture and enclosure environment for shared growth ofhigh-value resources, neither ofwhich is to the long-term advantage of the rural “Cainhodiun public institutions are a poor. disadvuntuge to the country, believed [by btisinessmen] to negativelv uflkct the process In this section we attempt to distil from the of developmenl. The perceived serious level of earlier chapters a number of challenges corruption and weak enjbrcement of contracts regarding the role ofthe state in fostering a and laws incite u lack of corlfideiice uniong poverty-reducing pattern of economic business people, investors and the public. ” growth. Good economic governance - EIC 2005, reflecting the findings of a facilitates productive, long-term survey of 100 businesses. investments in Cambodian goods and Ij.vou wunt to see it [corruption], just travel services, and helps to increase the flow of dong ihis road at night and you will ,see the benefits that Cambodians, including poor iirobile check-points on the roads. I fie1 that all such checkpoints are leeches sucking our Cambodians, obtain from these blood,fi.om every corner of our life, We would investments. If the Government deploys waste our lime arid spoil our goods if we would public investments, policies and regulation argue

Bank 2004a; World Bank 2004b; Table 7.1: The business climate in Cambodia is far EIC 2005; World Bank and IFC less attractive than the regional average 2006; Transparency International Business Cambodia Regional OECD 2005). Although the “Doing operation I average average Business” scores for 2005 represent indicator a slight improvement on those in Starting a business 2004, they remain extremely poor. Procedures 10 8 6 Table 7.1 and Box 7.1 summarize (no.) the findings ofthese surveys. Time (days) 86 51 19 cost 276% 42% 7% At present, there is a widespread (% p.c. GNI) perception that government Capital required 81 % 118% 29% institutions and practices generally (% p.c. GNI) reduce, rather than enhance the Dealing with Licenses (2005) competitiveness of businesses Procedures 28 17 14 operating in Cambodia, and impose (no.) cost and time barriers to both Time (days) 247 153 150 domestic and international trade. cost 607% 142% 68% Businesses consistently rate (% p.c. GNI) governance-related factors as more Source: World Bank and IFC 2006; important constraints on business www.doinqbusiness.org. than the (serious) problems with underdeveloped transport infrastructure or Box 7.1: Cambodia continues to rank low very high energy costs (Figure 7.1). in the eyes of investors If economic growth is to be the engine of EIC surveyed the chief executives of 100 sustainable poverty reduction in companies operating in Cambodia as part of Cambodia, tackling improved economic a global study of the business environment in governance would contribute to poverty a large sample of countries. In this survey, reduction in Cambodia in a range of direct the points in Cambodia’s favor included: (i) a and indirect ways: good mobile phone network; (ii) low inflation; (iii) stable exchange rates; (iv) relative . By providing economic actors-from simplicity of the tax code; and (v) a lack of smallhold farmers to foreign discrimination against foreign-owned businesses. Overall, however, Cambodia investors-with security of property was ranked 112 out of 117 in one index and rights. This is directly important to the 109 out of 116 in another. Allowing for the poor who suffer seriously if fact that some reforms enacted in the last six dispossessed from their land. months have improved a few of the Indirectly, it is also important as an indicators (resulting in, for example, a large drop in export fees), the picture is still one of upon influence economic growth and very serious constraints on Cambodia’s change; as lack of secure ownership ability to compete for investment. Asked creates a reluctance to invest in what would make the most difference to productivity-enhancing improvements economic growth, respondents listed (in in land (Chapters 4 and 5). order) public administration reform, legal and judicial reform, and reforms of the financing . For investors, lack of confidence in and banking sector. property rights, together with doubts about the impartiality of treatment Sources: World Bank and IFC 2006; www.doinqbusiness.orq; EIC 2005; under the law, often contributes to a Cambodia Daily 30thSeptember, 2005. Building an effective and responsive state 135

Figure 7.1 : Governance issues-corruption and bureaucracy-are seen as the most important obstacles to private sector development - 100 CEOs were asked to choose from a list of 14 factors the five most critical constraints on business in Cambodia

Cormp t ion lneffi ci ent God. bureaucracy Inadequately educated workforce Inadequate supply of infrastructure Access to financing Policy instability Gownment instability / coups Poor work ethic in national labour Tax regulations Crime and theft Restrictiw labour regulations Tax rates Foreign currency regulations Inflation

% of respondents naming factor as critical Source: EIC 2005 p. A53.

more all-encompassing reluctance to other words, can keep businesses from invest in Cambodiaper se. If investors growing. are reluctant to put money into Economic governance promotes Cambodia, jobs, wages and economic . poverty reduction by ensuring that linkages through demand and supply scarce public expenditure is used to will not develop fast enough to reduce maximum effectiveness for economic poverty. and social development. When The World Bank and EIC studies have contracts are signed with private sector highlighted that complex; overlapping companies for developing and/or and expensive regulation creates room managing infrastructure or public for discretion and rent-seeking, which services, transparency and add to business costs. It may also accountability help ensure that the create a disincentive for businesses to nation is getting the best possible deal. expand, as transition from informal the Progress has been made on these issues to formal sector necessitates an over the last few years. To begin with, exponential increase in costs, which Cambodia’s accession to the WTO in 2004 most small enterprises are not able to will require a comprehensive range of pay. Poor economic governance, in changes in economic policy and practice. This is broadly understood amongst 136 Cambodia - Poverty Assessment 2006

Cambodian policy-makers. The policy served well to control spending and so agenda for private sector development has contribute to macroeconomic and fiscal also taken up the findings of a detailed stability, it fails to deliver resources to study of the investment climate in spending agencies on a reliable basis. A Cambodia (WE3 2004). These were review of PFM in Cambodia in 2003 (the incorporated in the design of a reform Integrated Fiduciary Assessment and program which aims to improve the Public Expenditure Review or IFAPER) regulatory environment, establish credible identified four broad areas in which market-supporting institutions, and government systems needed to be increase transparency and competition in improved in order to protect public finds the administration of public-private from misuse and to ensure that they are partnerships. There are some initial signs used effectively to realize the of success in terms of reduced export Government’s goals (World Bank and fees-and some upward movement ADB 2003)’. between 2004 and 2005 in Cambodia’s The first challenge is to improve domestic scores in the “Doing Business” surveys. resource mobilization. Although an However, as the scores and ranks in the increase in the figure for 2003, most recent surveys show, in 2005 Cambodia’s resource mobilization is, at Cambodia still suffers from a number of 11.7 percent ofGDP in 20052, amongst the serious problems which prevent lowest in the world, constraining the Cambodian firms and workers from ability of the Government to invest in achieving their potential, and which hold economic and social development (and back the expansion and diversification of contributing to the low level of ownership the Cambodian economy as a whole. of development policy, which in the absence of domestic funds is financed to a Ma king Government large degree by ODA). Progress is being a dminis tra tion and spending made; actual tax collections in 2004 effective and pro-poor surpassed the budget target, due in part to Over recent years, increased budget one-off measures but also suggesting some allocations and disbursements to the social success in fighting tax evasion (including sectors have enabled improvements in the smuggling). Actual non-tax revenue delivery of health and education services. collection relative to GDP has deteriorated, However, levels of health and education however, worsening from 2003 to 20053. spending per capita remain extremely low The second challenge for pro-poor PFM is and the flow of funds unpredictable and to improve budget execution, cash heavily back-loaded, making it hard to management andJinancia1 control systems. manage available funds in an efficient or The system of cash-based payments has effective manner (Chapter 6). emerged as a major constraint, and a The Royal Government faces a number of inter-linked challenges in reforming public ’ It should be noted that government officials, financial management (PFM) and public particularly those from MEF, were closely administration systems and practices so involved in the analysis and policy that they are better aligned to government recommendations included in this study. * Projected 2005 data as of January 6, 2006. policy objectives-such as poverty 3 RGC and IMF estimates, December 2005. See reduction. While RGC arrangements have also Oum 2004 for discussion of revenue relative to budget projections. Building an effective and responsive state 137 significant factor in the problem of Furthermore, while allocations to key severely delayed and unpredictable release productive sectors (notably agriculture) of finds. Typically, about one-third of have now started to increase, they remain annual recurrent expenditure is not posted severely under-funded6. Over the medium until the last month of the financial year. term (for example, the period covered by This back loading undermines effective the 2006-2010 National Strategic operational planning and leads to a build- Development Plan), it is unlikely that up of arrears. It also contributes to a lack increased revenue mobilization will be of transparency which increases fiduciary sufficient to increase finding for the risk. relevant economic sector ministries while simultaneously maintaining and further While the four priority sectors of health increasing finds allocated to health and (MOH), education (MOEYS), agriculture education (which may have increased (MAFF) and rural development (MRD) significantly, but still remain far below the have suffered less in recent years, they too minimum level necessary for real are nonetheless still affected4. The effectiveness-see previous chapter). To problems of late and unpredictable release increase the funds available to agricultural are compounded by the centralization of and rural development and rural commitment authority in the Ministry for infrastructure (both social infrastructure, Economy and Finance (MEF). such as water and sanitation systems, and The third aspect of PFM reform that economic infrastructure, such as roads and requires improvement is in the alignment water control systems) will thus require of the formulation and management of further cutting non-priority budget items public expenditure policies to poverty and/or reducing administrative overheads. reduction objectives. Since the late 199Os, To increase the stock of essential physical there has been a very positive shift in capital will require not only cutting non- budget allocations away from defense and critical uses of public expenditure but also security and towards social sectors that an appropriate legal and regulatory serve to improve wellbeing and increase framework within which to administer human capital. The Government has made public-private partnerships. Given that the consistent commitments to increase engagement ofthe private sector on public funding for four priority ministries. There sector construction projects is essential, it have been significant successes: between becomes imperative that these contracts 1998 and 2001, education sector spending are awarded through a transparent process doubled, while health spending nearly that ensures that taxpayer (and donor) tripled. Despite this, however, actual money is used in the most efficient and spending in these sectors still fall short (by cost-effective way to deliver high quality a greater or lesser amount) of the sums allocated in the Budget Law, due to the problems in budget execution mentioned Development), through -1 1.4 percent (Health) and -10 percent (Agriculture) to -4.7 percent already’. (Education). 6 Budget expenditure for MAFF has increased from 3,016 bn riels in 1998 to 5,691 in 2004; See Oum 2004 for discussion of budget for MRD, from 3,783 bn riels to 64,741 bn; execution during the first 11 months of 2004. and for MPWT from 12,248 bn riels to 99,214 The budget deviation index (average of 2002- bn. The Ministry of Water Resources and 04) has varied considerably amongst the four Meteorology (MOWRAM), created in 1999, ministries, from -19 percent (Rural had a budget expenditure of 17,859 bn in 2004. 138 Cambodia -Poverty Assessment 2006 public goods. This is both an issue for allocation (too few service delivery staff in how the Government promotes long-term, remote areas) and management, resulting stable, private sector development (as in very weak linkages between merit (job described above) but also an issue for performance) and salary. public financial management. draft The The Government laid out a framework for law on concessions currently under review increasing the efficiency and effectiveness lays out a good framework for managing of the public sector in the National state-issued contracts private sector with Program for Administrative Reform actors, including public-private (NPAR 2002-2006), and specifically for infrastructure partnerships. public sector pay policy in sub-Decree 98. Rationalizing PFM with poverty reduction Donors have now accepted that they have objectives in mind requires not merely an important role and responsibility in this reallocation of funds between sectors or process of salary reform. The widespread ministries; it also requires increasingly use of ODA-financed salary supplements effective poverty focus in sector policies has created a pattern of complex and and sector planning processes. Substantial inefficient variations in pay rates which progress in this direction has already been pull the best staff towards projects that achieved in education (Chapter 5)7. There donors want to implement, making it much is also considerable room to strengthen the harder for government institutions to links between PFM and (already-existing) manage their human resources on a merit- poverty reduction goals through based, Ministry-wide (let alone geographic targeting of public spending to Government-wide) basis. In recent years spatial concentrations ofpoverty. resolving this issue has been incorporated in the objectives of the aid harmonization The fourth and final challenge is to and alignment agenda (Box 7.2). undertake a comprehensive reform of civil sewice pay and employment policies to The RGC, with support from donor ensure that public sector employees are on partners, is currently engaged in thinking the one hand well motivated and managed, through the elements of improved civil and on the other held accountable for service management and how to combine performing their jobs-and so deliver these elements in a coherent strategic services to citizens in an efficient, approach. Ultimately, this will need to equitable and poverty-reducing manner. incorporate a broad range of At present, the vast majority of public complementary reforms including merit- sector employees receive extremely low based controls on hiring and promotion, salaries; inevitably, they take a second job integrated with the budget process; (often during official working hours) or improving civil service pay, with more turn to bribery to get by8. Low salaries and differentiation to attract and retain senior motivation are compounded by poor managers and staff in priority service delivery ministries; and rationalizing ’ See also WB 2004;~. xii. deployment, so that the sectoral and Public sector employees suffer not merely from geographical spread of public sector very low official wages, but also from partial employment reflects service delivery and late payment of these wages due to the priorities. Some progress is being made- pervasive problem of budget execution. In the increases in average salaries have occurred first 11 months of 2004, salary payments were only 58 percent of what was budgeted for the and the Government has announced plans year. Oum 2004. for further increases-but much remains to Building an effective and responsive state 139

to support its implementation. This reform Box 7.2: From salary supplements to program consists of four stages of pay reform sequenced “platforms” for improving the The Government‘s Action Plan for effectiveness ofPFM (Box 7.3). Harmonization and Alignment of ODA addresses the pressing need to phase out The PFMRP has in turn been broken down ad hoc donor salary supplements and into action plans with regular reports, the redirect these resources towards supporting first ofwhich shows improvement, but also a proper, Government-led strategy for reforming civil service pay and employment. identifies numerous areas in which It is recognized, however, that to manage progress needs to be acceleratedg. A this transition well requires certain number of donors have demonstrated their preconditions and that government support to this program by pooling their institutions vary in the extent to which they support to PFM reform in a Trust Fund for meet these preconditions. There is thus an agreement to adopt a ”two speed” strategy. MEF to use to implement the PFMRP. Each Ministry first assesses the current situation by conducting (through the relevant TWG) an inventory of the rates and practices used by donors in that Ministry. Box 7.3: The PFM Reform Program offers The more advanced ministries will utilize a realistic agenda for aligning the Priority Mission Group (PMG)/Merit- spending to policy Based Pay Initiative (MBPI) approach, which requires that donors phase out salary In December 2004 the Government launched supplements, “pooling” their support to fund the PFMRP. This consists of the following merit-based schemes. In other sectors, four “platforms”: Ministries and institutions that are not yet ready for pooling funds will first rationalize 1 Platform 7: The budget is made credible salary supplementation at the institutional because it delivers resources predictably level, agreeing on a strategy for and reliably to budget managers (which harmonizing and aligning these rates and provides the basis for accountability by practices. One PMG/MBPI scheme is improving budget execution); already in place (in the MEF) and another a Platform 2: Initial improvements are made four Ministries are considering such in internal control and holding managers schemes. The remaining sectors and their accountable (which enables a focus on supporting TWGs are required to begin what is done with resources by providing moves towards the second amroach. better data, effective discipline, and greater internal transparency); be done if the Government is to be able to 1 Platform 3: Improvements are made in linking policy priorities and service delivery meet its targets for service delivery and targets to budget planning and economic and social development implementation (which enables greater outcomes. accountability for performance); and, 1 Platform 4: Accountability and review On all four fronts, the 2003 public processes for both finance and expenditure review found that there had performance are integrated (which would been little progress since the previous such result in greater external transparency and study in 1999. Since 2003, however, the provide a solid basis for deconcentration). Government has developed a detailed Source: MEF 2004. action plan for improving PFM (the Public Financial Management Reform Program or PFMRP) and an organizational structure (a Reform Committee within MEF and a Government-donor PFM Working Group) MEF 2005; World Bank 2005; MEF 2004. 140 Cambodia -Poverty Assessment 2006

Petrochemical revenues and poverty is the most obvious but far from poverty reduction only example. There is now a good possibility that the existence of large-scale reserves of oil and Box 7.4: Early estimates of Cambodia’s natural gas will be confirmed to exist in oil and gas resources Cambodian territorial waters (Box 7.4). This has the potential to dramatically It has yet to be proven that the recent findings improve the level and effectiveness ofstate of oil and natural gas in Cambodian offshore spending on economic and social waters are commercially viable. However, the indications are sufficiently strong that it is now development. Managed well, a major important to start considering the legal, revenue stream from offshore oil and gas institutional and policy arrangements that need could finance infrastructure investments, to be developed to ensure productive use of public sector salaries and services and these revenues, should they be proven. So dramatically increase the rate of poverty far, four exploratory wells drilled in Block A (one of six Blocks in Cambodian waters) have reduction. Such revenues could help all produced oil and gas. Provisional estimates Cambodia escape least-developed country are that this explored portion of Block A status in a much shorter time frame than contains 400-500 million barrels of oil and 2-3 has previously been thought possible. trillion cubic feet of gas; total Cambodian reserves may quite possibly be as high as 2 However, international experience billion barrels of recoverable oil and 10 trillion suggests that such petrochemical wealth cubic feet of gas. Depending upon the world may equally well result in a “resource price of oil, Cambodian reserves may be curse” that actually retards development contributing annual revenues of US$2 billion per annum-several times the current level of and poverty reduction. The main reason is domestic revenue and ODA combined-within that a government that receives all the perhaps five to ten years. money it needs from oil does not feel as accountable to its citizens (taxation Source: Vallely et a/ 2006. becomes irrelevant) and does not have the necessary incentives to do the hard work of fostering broad-based economic growth. Apart from the distorting effect that These problems are more pronounced in “unearned income” from minerals can countries with low state capacity and have on the political economy ofa country, relatively weak formal institutions, and at there are further problems commonly their worst in post-conflict countries in associated with oil wealth. The first is so- which the levels of transparency and called “Dutch disease”. Oil and gas accountability are very low and state- exports enable a country to maintain a society relations have been shaped by a constant exchange rate, while higher public tradition of authoritarianism and spending, fueled by the inflow of extraction. Internationally, the discovery petrochemical revenues, creates domestic of oil has on average reduced national inflation which makes other sectors of the growth rates, not raised them. Nigeria, economy (e.g. agriculture and where an average of $12 billion per annum manufacturing) uncompetitive in export in oil wealth over a thirty year period has and perhaps even domestic markets. fostered pervasive corruption and done Governments often come to depend upon nothing to promote growth or reduce oil revenue, even borrowing heavily against it, with serious consequences when the price of oil falls. Finally, the oil and Building an effective and responsive state 141 gas industry tends to be an enclave and tax collection (Figure 7.2). For all but industry with few linkages to other sectors the last of these institutions, surveys of the host economy (especially if located suggest that corruption has worsened offshore and in a poor economy which between 2000 and 2005. This perception cannot produce the technology- and of worsening corruption may, however, be knowledge-intensive goods and services due to changing values amongst those the industry requires). surveyed, that is, it may reflect a declining willingness amongst respondents to accept There is thus good reason to be cautious corruption as a “fact of life”, rather than about predicting the impact of any future (or as well as) an actual increase in the stream of petrochemical revenue on the incidence of corruption. It is worth noting Cambodian economy and the prospects for that the majority of respondents in the Cambodia’s poor. Carefully designed 2004 survey rejected the notion that institutional mechanisms that promote corruption becomes acceptable because it transparency and accountability in the use is so common, or that a small salary of this national asset will be critical if it is entitles a civil servant to bribes (CSD to help rather than hinder Cambodia’s 2005). escape from the ranks of least-developed nations (Vallely et al 2004; UNDP 2006). Attempts to quantify the prevalence and The potential for a dramatic increase in impact of corruption in terms of direct revenues adds further urgency to the need costs to households suggests, contrary to for the strengthening of public financial common perception, the direct costs of management as described above-and corruption (in terms ofbribes that are paid) sharpens the case for improving falls more heavily on wealthier and urban transparency and accountability in public groups than poor groups. This however life in order to combat corruption. reflects the fact that the rich and urban are more likely to make use of services, which Corruption increases the likelihood that they will The problems of Cambodia’s weak state (have to) pay a bribe. For the poor, institutions, and the resultant low levels of services may simply not be accessible; or, state effectiveness and responsiveness to the cost of the fee required (either the the needs ofthe poor, are felt most directly formal fee or the bribe) means that they are and acutely in the form of corruption. As excluded from obtaining the service, rather than it an in English, corruption in Khmer can cover obtaining at inflated cost. a range of practices and value-judgments, The CSD survey in 2004 suggests that on but it is important to recognize that it is average each household pays US$24.5 per seen as a problem, and a critically annum on bribes-r between 1.4 - 2.2 important one, by citizens, businesses and percent oftotal expenditure, or 5 percent of government policy-makers (see Nissen total income”. The composition of this 2005; CSD 2005, World Bank 1999, total-in terms of average amounts paid to World Bank 2004). Indeed, as Box 7.5 different elements of the state-is makes clear, popular understanding ofwhy summarized in Figure 2, along with the and how different forms of corruption occur is quite sophisticated.

Those institutions seen as most corrupt lo World Bank 2000 found 2.2 percent of income; were the customs service, courts, police CSD 2004 found 1.4 percent of expenditure and 5 percent of reported income. 142 Cambodia - Poverty Assessment 2006

Box 7.5: Popular accounts emphasize the ubiquity and severity of corruption Participants in all nine sites in the MOPS research talked about the impact of corruption upon their livelihoods and their prospects of being able to work their way out of poverty. Many noted the number of ways in which corruption impinged on their daily lives-and how the poor always lose when bribery determines the outcome of state decisions: Students have to bribe the examiner during examination for looking at other students' answer sheets or hidden documents. Policemen or military police are the persons who are running the illegal gasoline trade. When we arrest thieves that steal our fishing nets, and send them to the police station.. . only about two minutes later the thieves are released because of money and power. The poor are the victim and always wrong, even though they are right, because the poor do not have money and power. Kampot Another type of corruption is when the village authorities take money that should be spent for the collective interest and do nothing for the villagers' benefits. Also, corruption means that there are many checkpoints along the road ... that is, many police, customs or people equipped with guns and known to have authority to demand people to pay when carrying goods or even passengers through their areas. Battambang There is also a clear understanding of the structural nature of much corruption, and how those at the bottom pay the cumulative cost for leakage or reduced profit margins at each level of the hierarchy: We learn corruption is committed by all actors involved from the top to the bottom ... Each key person takes bit by bit from the top until at the bottom level there is hardly anything remaining. Kampong Thom The cumulative effect of pervasive corruption may be to disillusion citizens with regard to the Government-or to cause them to lose faith in the ability of political leaders to effectively control corruption. If we have so many cases of corruption, maybe it is not a good government ... We heard that the government announced they would remove all military staff in the forests except at the border areas ... but since then we saw a huge increase in the number of military and related authorities in our forest ... We asked ourselves 'is the society clean and pure?'-and finally could find an answer that it is not, because there are many cases of corruption. Kampong Thom Source: CDRl 2006 (forthcoming). probabilities that a given interaction will benefits that should have reached involve paying a bribe. households but were diverted by authorities; The direct cost of bribes paid is, however, a very incomplete measure ofthe impact of . an increase in the prices paid for corruption on the household economy. In consumption or investment goods; part, it fails to capture the real impact of revenue collection that is forgone from corruption because, as mentioned above, . public finances and therefore an inability to pay the bribe prevents the unavailable to finance service delivery; poor from obtaining access to essential basic services. However, it also reflects . the dispossession ofpoor and/or poorly the effects of corruption that are not connected families from access to measurable in terms of bribes paid at the resources that arises when local household level. Some common forms of authorities sell off as private property non-bribe corruption costs would include, forests or waters that have customarily for example: been accessible as common property resources; or Building an effective and responsive state 143

often time-bound) advantages Figure 7.2: Corruption is most pronounced in the such as MFA quotasor, now, courts-but corruption in education has the greatest direct impact OECD safeguards policies. By imposing unproductive - Probability of paying a bribe and annual total bribes, by institution, 2004 costs on business, corruption reduces the number of jobs 100% $15.00 available and the salaries paid, Likelihood of bribe and reduce the rate at which 0 conditional on the economy grows and a 80% contact B .p B L L diversifies. I) o Average annual cn $10.00 .-c bribe amunt ; The cumulative effect of all these 5 60% 5 manifestations of corruption is .-c i! v) e s extremely to calculate 0 -m hard with 59 40% C c any degree of precision, but is 0 0 $5.00 Lc certainly significant. Although 0 8 E 20% s corruption is arguably merely a Q symptom of more fundamental problems with regard to state 0% $- accountability and effectiveness, it is a critical symptom that must be tackled if Cambodia is to develop. As has been agreed between the Government and its external partners, an anti- corruption law is an essential pre- requisite; but equally important, this law must then be enforced Source: CSD 2005. through credible, independent institutions.

1 the deterrent effect of corruption, in Post-conflict human security terms of the way in which corruption costs effectively reduce the One of the core state functions which competitiveness of Cambodian determine the legitimacy of a government businesses, deterring foreign investors, in the eyes of its citizens, and an essential and creating disincentives for but often marginalized element of successful informal sector enterprises “development” policy, is that of ensuring to formalize, at which point they that citizens enjoy an acceptable level of become liable to a range of safety from crime and violence. cumbersome regulatory requirements Maintaining effective law and order while and attendant requests for payments to ensuring just treatment that respects due “facilitate” approvals). This will act to process and civil rights is an essential part slow economic growth, or at least to of what any government should do. It is, ensure that growth proceeds only in a however, particularly hard to achieve in few sectors in which the cost of post-conflict societies in which corruption can be offset by unique (and institutional capacity and accountability amongst the police, security and justice 144 Cambodia - Poverty Assessment 2006 services is limited; formal legal channels security, and its role as a factor causing often cannot provide equal access to justice and arising from poverty. Domestic for all; small arms are widely available; violence has been recognized as a serious and sustained exposure to violence as a issue since the mid-1990s. Violence means of surviving, resolving disputes and against women remains a major problem in getting ahead has weakened the Cambodia, arguably reflecting both the effectiveness of social sanctions against lower social status of women and a violent behavior. background change in values in which decades ofviolence and trauma have made As mentioned before, Cambodians are the use of violence in variety of contexts manifestly glad that peace has finally been a more normalboth in the sense of “more achieved. During the 1980s (and in some common” and “less shocking”-and places early- to mid-l990s), the conflict subject to less (or less effective) social between the Khmer Rouge and the sanctions (Zimmerman 1995). A recent Government, in which civilians were study on domestic violence for the potentially at risk from both sides, was an Ministry of Women’s Affairs (Indochina unmeasurably serious constraint on Research 2005) found that 64 percent of development. The final resolution of that respondents knew a husband who had been conflict has freed up people to travel, to physically violent toward his wife; 23 choose their own occupations, and to percent of women reported suffering accumulate investments and possessions violence from their husbands (an incidence without the fear that they would be very similar to that recorded in the 2000 destroyed or seized by combatants (Box 2, CDHS). In a separate study on local Chapter 1). dispute resolution practices, both The Government also deserves credit for community respondents and Commune taking concerted action since the late councilors identified domestic conflicts, 1990s to collect and destroy the small arms including domestic violence, as the second that were in widespread circulation. most common form of conflict at the Reducing the prevalence of military- village level (CAS 2005). Similarly, in purpose weapons is not a complete research on access to justice, Commune solution to the problems of crime and authorities estimated that 48 percent of the violence, but it is a major contributor. cases brought to them were related to However, Cambodia still faces serious domestic violence. In Communes where challenges with regard to ensuring human actual case data was collected, domestic security (Box 7.6). Fieldwork in the conflicts (including divorce and domestic MOPS and Tonle Sap PPA studies violence) accounted for between 37 revealed ubiquitous concerns with the percent and 53 percent of total cases emergence of youth gangs; drug dealing (UNDP 2005). and drug use; and on occasion the use of The survey for MOWA also found violence by locally powerfbl actors, significant acceptance of domestic sometimes in collusion with local violence, with around a third of authorities, to exclude local communities respondents agreeing with statements that from newly- and illegally-privatized lands it was “sometimes acceptable”. and resources. Domestic violence-and violence in It is also important not to overlook the general-is a poverty issue for several intra-household dimensions of human reasons. Firstly, a broad definition of Building an effective and responsive state 145

Box 7.6: Crime and violence are serious-and perhaps growing-problems There is some variation in perceptions of trends in crime between research sites in the MOPS research. In two villages, participants felt security was quite good, with only petty crimes, and women could walk safely at night. For most, however, there was considerable concern. Theft and robbery are something new in our village. Before we felt safe to leave our cattle or buffaloes in the fields very far from the village without guarding them for as long as one or two months, but now we cannot do so, and have to accompany our cattle or buffaloes all day. Dong Kda, Kompong Thom. The most commonly cited problems were theft of unattended assets such as fishing nets or cattle, banditry, and, finally, gangs of young men picking fights with other gangs at village celebrations at which there was drinking and dancing. This social phenomenon of rural gangs has emerged only recently (different villages traced its emergence in their locality to no earlier than between 2000 and 2003) but has grown very fast and appears to be a worry throughout much of the countryside. Teenagers in this community are likely to provoke conflict amongst themselves. Even a tiny problem can cause fighting. They don’t use drugs yet-normally they just provoke a conflict during community ceremonies. They carry knives and chains along to be used during fighting. Dong Kda, Kompong Thom. Villagers had a number of ideas, all fairly consistent between the nine communities, about what caused violence and crime. Theft was seen as directly related to poverty; bad harvests in particular had pushed some families into destitution and petty theft was understood (if not tolerated) as one element of their survival strategies. Perhaps the most consistent theme, however, was that the rise of gangs of male youths is associated with increased exposure to “foreign” culture and the glamorization of gangster life, either through television or directly when working as migrant laborers in Thailand. A group of young gangsters has been created since 2000, which was a period when young people went out to search for jobs nearby on the Cambodia-Thai border. They copied this bad behaviour from the people living in Thailand ... Those who think of themselves as young gangsters always provoke conflict with others outside the village, especially during special events such as ceremonies. Andong Trach, Battambang. Sometimes, however, the undesirable external elements blamed for problems may be from no further than Phnom Penh (as in Dong Kda in Kompong Thom, where young village men were afraid to attend local celebrations for fear of becoming embroiled in fights with a group of young men from Phnom Penh who were working locally). The final factor seen to contribute to poor security of property and person was the ineffectiveness of local authorities. People complained that police did not appear to make efforts to prevent crime. If villagers caught a suspected thief or gang brawler and turned them over to the police, they were generally released after they (or fellow gang members) paid a bribe, or the victims would be asked for money to keep the suspect in custody. In particular, it was mentioned that the police go out of their way to avoid having to sanction the children of the rich and powerful. When local authorities did however take action (as occurred in Babaong in Prey Veng), they appear to have been successful. Source: CDRl 2006 (forthcoming). poverty as absence of wellbeing recognizes Secondly, there are strong connections (in multiple dimensions of deprivation, of both directions) between domestic violence which exposure to physical violence is and material poverty (Box 7.7). clearly critical. Participatory research reveals that poverty 146 Cambodia - Poverty Assessment 2006

Box 7.7: Poverty and domestic violence are mutually reinforcing In the MOPS research, domestic violence was perceived to be more common in poor households. This was seen in part because of the stress that poverty puts on relationships, and in part because the moral failings and patterns of behavior (such as drinking and gambling) that account for the poverty of some men and their families, also account for domestic violence. You know, when people are poor they are moody and tend to commit bad deeds easily-go to spend money on unnecessary things like wine or gambling. And when their wife advises them to stop, they just hit her. (poor group, Kompong Thom). It was also felt that conflict and violence within the family causes or perpetuates poverty as the offender often destroys the families’ possessions-or the home itself-in the course of their violent rage. Violence results in poverty, because when they commit violence, they destroy their plates and pots, burn their house and clothes and other people just keep watching them because they can do nothing to help. Dang Kda, Kompong Thom. In the last two or three days there was a conflict between the husband and the wife in a family ... and the husband just burnt out their small hut... Domestic violence results in great loss ofproperty. Female participant, fell into poverty group, Dong Kda, Kg Thom. Apart from the effects on a family’s meager assets, the effect on children’s education was also cited in two villages. Domestic violence was also indirectly but strongly related to poverty in people’s minds because it was often seen to result in divorce or separation, and the downward decline into poverty often experienced by a newly-single woman (who would generally end up looking after the children). There are a lot of things lost when domestic violence happens. Amongst these, children are not able to study because they feel ashamed and also are taunted by others. (Male youth, Krasaing village, Battambang). About ten or twelve years a go, my farther liked drinking wine and always hit my mother when he was drunk. During conflicts, my two siblings and I often ran away ... We started living independently two years ago. Because of those frequent conflicts I did not go to school but now I am the breadwinner, I can send all my younger siblings to school. 22-year old woman, female youth focus group discussion, Dong Kda Kompong Thom. Source: CDRl 2006 (forthcoming). is regarded as a contributory factor in between 5-20 times in the last year domestic violence. The frequency and (Indochina Research 2005). impact of domestic violence is seen as If domestic violence is clearly at greatest amongst the poorest groups. alarmingly high levels, establishing trends Causation also runs in the opposite is somewhat more complex. As with direction; in both the research and MOPS survey-based assessments of corruption, it the Tonle Sap PPAs, respondents see is not clear if the appearance of static (or domestic violence as having clear and rising) incidence reflects actual trends or direct impoverishing impacts on the greater willingness of respondents to households as possessions are destroyed, acknowledge and tackle the issue, or medical costs incurred, and income perhaps, a mixture of both. Cambodian foregone. It also contributes to the communities and local authorities are perpetuation of poverty between increasingly identifying domestic violence generations, as children’s schooling is as a significant conflict at the local level disrupted. Just under a fifth of those who which they are struggling to resolve. reported experiencing violence from their Demand for resources to tackle domestic spouse said their children missed school violence is increasing at the local level, Building an effective and responsive state 147 indicating that domestic violence is no Box 7.8: Qualitative research suggests a decline longer seen as a family affair but an in domestic violence in some villages issue of law and justice. The SEILA In most of the MOPS study villages, participant: Gender Mainstreaming Evaluation (both male and female) reported that the incidence 0: Report found that domestic violence domestic violence had declined significantly in recent years. The improvement was described a: prevention was the number one dramatic-and remaining cases of domestic violence “gender request” made by Commune/ as residual-in six of the villages, and as less Sangkat Councils under their 2004 dramatic but still notable in another two. In the Commune Plans, accounting for remaining village, violence was described as around a quarter of all requests frequent, without comment on whether there was E trend. This improvement was attributed to the impaci (Brereton 2005:47-48). of social and legal awareness campaigns, facilitated Interestingly-and in contrast to the by the spread within the villages of television and general impression of static or perhaps radio (which have carried many messages tackling domestic violence issues). In five villages, mention increasing domestic violence derived was made of human rights training workshops on from surveys-the participatory domestic violence, delivered by NGOs. Participants research in the nine MOPS villages appeared to credit these campaigns with resulted in a fairly consistent considerable effectiveness, to the extent that those impression, from both male and men in the village who persisted in attacking their wives were said to be those who had not attended female respondents, of an improving the trainings. Also important was the perception of situation (Box 7.8). This was seen by stronger laws against domestic violence and of local those questioned as being a result of authorities increasingly willing to intervene (which is education and legal campaigns (which itself a product of public education and legal have, broadly, involved effective awareness campaigns). The situation is still far short from ideal; the emphasis of local police and partnerships between the relevant authorities, as in much of “living law” dispute Government ministry and civil society resolution at the local level in Cambodia, is on organizations). The difference reconciliation and relatively symbolic punishment between surveys and qualitative through a few days confinement in jail, rather than research with respect to the nature of the strict application of the law. Nonetheless, the combination of new laws and active rights education trends in domestic violence is one that appears to have led to an improvement upon the will need further exploration. previous situation, when domestic violence was clearly seen as a family matter in which it was difficult Voice, participation and or impossible for outsiders-both other villagers and local authorities-to intervene. social accounta bi I ity Before, when a husband hit his wife, the wife didn’t dare complain at the police station, but Comparative international history now, after people supporting women’s rights suggests that improvements in have been here, the wife claims her rights at the governance and resulting accelerations police station to have him arrested and warned. in economic and social development Ksach Chiros. have involved both the initiative of Now, we have much less violence in our area far-sighted leaders and reforms than the past five years because some actions against this misconduct have been taken over initiated from within the state, and the same period. For example, the authority organized demand expressed by civil arrests the abuser and detains them at the police society groups, which exerted pressure post for one or two days, so the abuser and for representation of their interests in those people who tend to abuse are afraid and policy-setting. In Cambodia as stop doing wrong. Dang Kda, Kompong Thom elsewhere, enhancing the role of the Source: CDRl (forthcoming) 2006. 148 Cambodia - Poverty Assessment 2006 state as a manager of long-term changes deliver results (e.g. job-creating economic which reduce poverty will require more growth or improved services) by holding than just technical solutions in the form of service providers accountable; or directly, new institutional arrangements and policy by giving citizens and civil society processes within the Government. Citizens organizations the information (e.g. on and organized civil society-including budget allocations for schools or clinics; or business associations-have key roles to on official minimum standards for service play, both as autonomous actors in delivery) and control (e.g. participation in promoting economic and social health centre management committees or development, and as interlocutors with the school boards) that they need to hold Govemment. service providers accountable. Channels for citizens’ participation The evolution of civil society Improving state effectiveness thus requires Relationships between the state and civil a combination of “supply-side” reforms society in post-conflict Cambodia-and (public administration and public financial the possibilities for citizens to channel management reforms, institutional capacity demands along the lines sketched in Figure building, etc.) and “demand-side” 7.3-are complicated by the fact that the developments (increasing the capacity of formal institutions that might provide the Cambodian citizens to participate in policy basis for defining and pursuing an debates and make informed, reasoned and inclusive national identity and effective demands for improved development vision are still somewhat government performance). Citizen underdeveloped in both political and social demand, whether expressed either spheres. Just as the Cambodian state had individually or collectively through an to be rebuilt from scratch after 1979, so too organization, can help to make the Cambodian civil society has had to recover Government and state more accountable from deliberate attempts by the Khmer and in doing so drive improvement in state Rouge to destroy any social institution or effectiveness and development outcomes. form of identity-the family, Buddhism, As in other democracies, Cambodia’s locality or region-that might compete framework of accountability derives from with the party-state for loyalty. Although at a basic level these institutions have been the Constitution and is based on three key relationships: firstly, between citizens and the State; secondly, Figure 7.3: Channels for citizen demand may be between the elected officials and the either direct or indirect public sector employees or voting, contracted private sector companies questions contracts / responsible for delivering essential services to citizens; and, thirdly, between those who deliver services

and the citizens who use them i. ~~ “ ” ” ” . (Figure 7.3). In other words, citizen demand can be channeled either indirectly, as citizens vote for politicians who can participation

Source: World Bank 2003. Building an effective and responsive state 149 rebuilt, decades of conflict have clearly charity on behalf of a comfortable, resulted in a depleted stock of social established middle class13. and capital (Colletta Cullen 2000)”. These donor and NGO partners have Attempts to mobilize local development fostered the creation or further evolution of initiatives using the language of groups, national NGOs as the framework for the cooperation, solidarity or collective action, emergence of an organized Cambodian must also contend with the negative civil society capable of advocating for poor associations that these words acquired and marginalized groups and delivering under the Khmer Rouge (and, albeit to a services to these groups in an efficient much lesser extent, the PRK). manner (Hughes 2002). To a considerable While there is an active debate about the extent they have succeeded admirably. It current status of, or potential for, is striking that in the MOPS research, community identity and collective action at villagers who were asked if they had the grassroots12, there is an evident received any form of assistance or benefit abundance of non-governmental from actors outside the village cited NGO organizations (NGOs) that operate above projects far more often than they cited the local level. This sector has grown from government service providers. NGOs have nothing since the transition to democracy recorded considerable success in delivering that began a decade ago. Although there both basic goods and services and in are an increasing number of exceptions, imparting knowledge and ideas on topics the national NGOs that have emerged since such as human rights, domestic violence, 1993 have typically been professional and health practices (most notably in service delivery or advocacy organizations regard to HIV/AIDS control). They have which owe their existence more to donor also played an invaluable role in “ground- and INGO support than (as in, for truthing” the Government and donor example, much of Latin America or South understandings of economic and social Asia) to the outgrowth of grassroots change in the country, and sustaining associations based on occupation or policy attention to the needs of the poorest locality, or the expression of organized and most marginalized members of society. However, the growth of the Cambodian NGO sector under the aegis of external funding and technical assistance does raise some relevant strategic issues of 11 “Social capital” is taken here to refer to traditions of cooperation and charity embedded coordination, accountability, legitimacy in trust and collective identity. and sustainability. If Cambodian NGOs 12 It is not entirely academic to note that the have relatively shallow roots in everyday anthropological literature from the 1950s and Cambodian society, there is a danger that 1960s suggest that, even before the start of the despite conscious efforts on their part they civil conflict, rural Cambodia had relatively few forms of village-level organization. On this, and the debate about the current strength 13 INGOs have an unusually important historical of village-level civil society, compare Collins legacy in Cambodia. During the 1980s the 1998; Ashmoneit 1998; World Bank 2005; PRK was denied access to most ODA as it was Biddulph 1996; Ovesen et al 1996; Conway not recognized by the UN. INGOs such as 1999; Daubert 1996; Ebhara 1968; Delvert Oxfam came to play a quasi-donor role, 1958; and Kieman 1982. Aranvind 2005 reconstructing core infrastructure and provides a good overview. providing basic service delivery. 150 Cambodia - Poverty Assessment 2006 may not fully understand the needs and Box 7.9: Coordinating NGO activities remains a major challenge priorities of their constituents (see 0' Leary and Nee 2001). There are also serious At the time of writing, there are estimated to be 308 international and 642 national questions about the timescale required for NGOs operating in Cambodia. In 2004, NGOs to become financially viable NGOs spent just under US$124m (almost without external support, and the potential twice the $66m spent in 1996). Although inefficiencies of so many organizations these NGOs vary greatly in size, skills and competing for finite funds with relatively ambitions, there is an increasing coherence in attempts to link grassroots limited coordination structures (Box 7.9). experience to national debates on policy Grassroots organizations and social and programs. This is however made hard movements are now emerging and by the scale of the effort required to coordinate such large numbers of strengthening. Most notable amongst these organizations. With the beginning of the are the unions which have developed as the explosive growth of the NGO sector in the garment industry has grown. However, early 199Os, a new coordination structure there are also emergent regional and had to evolve. This structure is capped by national federations amongst farmers' the Cooperation Committee for Cambodia (CCC), which seeks to facilitate information groups, and at the village level pagoda exchange, coordination, and (in tandem associations have resumed their historical with the NGO Forum) collective roles and, in some cases, taken on new representation on issues of common functions in local economic and social concern. Below the CCC, 38 informal development. Nonetheless, supra-local networks coordinate between NGOs working in the same sector--Medicam and organizations and movements remain at an EduCam (for health sector and education early stage of de~elopment'~.Donors (and sector NGOs. respectively)--are notably INGOs) need to be very careful how they successful networks that have developed attempt to engage with such movements. good working relations with their respective There is a real danger of "killing them with line ministries and sector donors and participate to a significant extent in debates kindness", or de-legitimizing them by on the direction of sector policy lavishing external funds and advice on development. There are also 15 Provincial them at too early a stage in their evolution. NGO networks which offer the potential for the emergence of a middle level of The growth of civil society since 1993 "intermediary institutions" in Cambodian should thus be seen as what is, in society, improving the flow of information Cambodian terms, a historically novel and ideas between the grassroots and development, driven both by socio- national govern ment. economic change and by increasing Source: Development Weekly 2005; Clarke exposure to global ideas and knowledge. 1998; Mysliwiec 1994; Conway 1999. INGO and donor support to national NGOs has played an important role in helping to political discourse, and in helping to create transmit concepts of civil society, and sustain a measure ofpolitical space for participation, human rights and the emergence of civil society. However, accountability into Cambodian social and donors and INGOs could often have done more to support the "thickening" of civil 14 Historical and socio-political factors aside, the society in ways that put Khmers more in relative paucity of membership organizations is charge of the process, and that rooted in part due to the fact that incomes are so low that it has been very hard to establish standing national NGOs in Cambodian society. The organizations supported by members' relationship between donors/INGOs and contributions. Building an effective and responsive state 151

CNGOs is steadily evolving in this Box 7.10: Drug use is a growing direction, but still has some way to go. problem for Cambodian youth To build on what has been achieved and A variety of factors, including its regional lay out a framework within which location, large youth population and under- resourced and under-trained law government-NGO relations can develop enforcement institutions, foster the along a positive path, Cambodia is now emergence of drug production, use and embarking on the preparation of a law on trafficking in Cambodia. Ready availability NGOs. The World Bank has agreed to and low price has meant that use of illegal provide technical support to the NGO substances is quite high amongst youths; particularly vulnerable youth groups such community in the drafting ofthis law. as those living on the streets in Phnom Finally, those formulating social and Penh. A one-day survey amongst this group conducted in June 2005 by the NGO economic policy need to bear in mind that Mith Samlanh/Friends found almost half not all social capital is pro-poor or pro- were abusing substances ranging from development. While attention rightly glue to heroin. Of those using drugs, 62 focuses on forms of association which percent (or 30 percent of the total sample) benefit the poor either directly as used yarna (methamphetamines); 21 percent (or around 10 percent of the members, or indirectly through positive sample) used heroin (a dramatic rise from social impacts felt outside the group, there 2004, and a cause for concern given its are also “perverse” forms of social capital, potential as a vector for HIV/AIDS in which groups act in ways which benefit transmission). group members but impose costs on society as a whole15. The increasingly out the pro-poor potential of ubiquitous youth gangs described above decentralization. are a good example, as are many other The decentralization process that the examples of organized criminal activity, Government embarked on with Commune/ such as those which support trafficking of Sangkat elections in 2O02-andy drugs (Box 7.10) or humans. As in all potentially, recent moves towards more countries, such criminal organizations are substantive deconcentration of functions to often based in social (kinship or patron- the Province and District level-has client) relationships of loyalty and undoubtedly created new opportunities for protection, and constitute a serious civil society engagement and challenge for Cambodia’s long-term accountability. In some cases, more development. capable and assertive Commune Councils Making decentralization pro-poor are responding to local concerns and acting as a champion or at least intermediary Taking decisions on plans and budget between Commune residents and more allocation down to a local level also has powerful outside interests (Box 7.1 1). considerable potential for increasing Rapid processes of economic and social citizen participation, and the accountability change-processes of modernization-are and effectiveness of state action-although having complex effects on engagements regional and international experience suggests that this is far from guaranteed, between citizens and the state. On the one increasing penetration of market and deliberate efforts are required to bring hand, the relations into remote areas increases the In economists’ terms, this is social capital as a incentives for officials to exercise power membership good with negative externalities. for financial gain. However it also seems 152 Cambodia - Poverty Assessment 2006

Box 7.1 1 : Rural Cambodians are adopting new forms of collective action to defend their interests - A case study of land dispute resolution (Kompong Thom Province)

In World Bank research into actually-existing dispute resolution processes at the local level, an early case study provides interesting insights into what factors seem to enable the poor to obtain justice. The dispute in this case-over land in Kompong Thom Province-revolved around the acquisition, by high-ranking District and Provincial officials, of degraded forestlplantation land on the outskirts of a group of relatively remote lowland rice farming villages. The fact that land in the same area was the subject of a dispute in 2000 and again in 2005 provides an interesting opportunity to study changes in dispute resolution practice over time. Two things stand out in this dispute. Firstly, the Commune experienced significant development between 2000 and 2005. The construction of a new road has put the area within an hour’s drive of the Provincial capital. Significant investment in infrastructure has occurred (e.g., mobile telephones; wired electricity; health center; and schools). While villagers report significant increases in personal security over this time, they also expressed the view that these developments are not really alleviating poverty because the poor are at high risk of shocks (health crises, crop failure, etc.) which push them into the poverty traps of debt or landlessness, despite increasing levels of economic opportunity. Secondly, the way in which the dispossessed villagers dealt with the dispute changed significantly. In 2000. The villagers organized for representatives of the group (an elder and a temple administrator) to speak to the District chief on their behalf. They felt intimidated by the District chief, were unable to push the case forward, and in essence gave up on the claim, explaining that they did not know how they could take any further action. In 2005 when a similar dispute arose over an adjacent piece of land, villagers were more assertive. Initially the villagers confronted the group of District officials who were measuring the land and removed signs which designated the area as state forest. The villagers then organized themselves to collect thumb prints for a petition against the District administration-a petition which they also sent to human rights organizations and the local National Assembly member representing the opposition. Though the dispute had not been finally resolved at the time of writing, the Provincial Governor made an initial intervention on behalf of the villagers and they were allowed back on to the land on a provisional basis. Source: World Bank research on “Justice for the Poor”, 2006, forthcoming. that economic progress, spreading civil society networks and increased access to Information and development information can support the development The free flow of information via of social capital that provides the basis for independent news media-both print and collective action. Recent World Bank electronic-can be powerful tools in research on dispute resolution supports the improving communications between the finding that, though by no means always state and civil society and promoting the case, it is possible for groups of stable, equitable, inclusive development villagers to obtain greater responsiveness (World Bank Institute 2002; Goldin 2005). in the resolution of their grievances if they The presence of independent media build a strong network of relationships facilitates all three accountability with both NGOs and government officials relationships sketched in Figure 7.3. It not at all levels” (Nee and Healy 2003: 102). only keeps politicians in touch with the concerns of the electorate and service providers subject to public critique, but also-by stimulating public debate on the Building an effective and responsive state 153 strengths and failings of state arrangements Finally, an independent media which for service delivery-provides politicians represents a plurality of voices can and policy makers with additional, contribute to the task of building national independent sources of information that unity and political stability. Mass media they need in order to be able to judge and that incorporates uncensored stories from regulate the performance of line Ministries all over the country-and connects the or sub-national administrations. An Kingdom to the wider world-helps independent press, as envisaged in Cambodians to think of themselves as Cambodia’s 1994 Press Law, can, for citizens with common problems, example, help expose corrupt practices, or opportunities, interests and institutions, local abuses, that might otherwise not and encourages people to seek collective, come to. the attention of politicians until peaceful ways of engaging in national they face a frustrated electorate on voting development processes. For many day (Sen 2000). Cambodians, the contrast between the free media of the last decade and the preceding A free media also has a critical role to play decades of isolation and censorship is in making markets work well, by reducing stark; whereas in the past they lived “like a “information asymmetries”-that is, by frog in a well”, over the last ten years ensuring that all economic actors have Cambodians have enjoyed a significantly access to the same information on supply expanded perspective on their own and and demand, can make informed decisions other countries, and seem to value this accordingly, and can compete on an equal basic freedom-and, indeed, to want it footing (World Bank 2001). When only a expanded to encompass the electronic privileged or lucky few have access to media as well as the print media (Phnom critical information on prices, trade Penh Post 2005a, 2005b). While there is policies or government plans for an urgent and ongoing need to improve the infrastructure investment, they obtain an levels of professionalism in much of the advantage that may have nothing to do media, the general point remains--a free with their entrepreneurial or managerial press almost always improves state-society skills. The result is distorted markets and relations, and promotes rather than retards an inefficient allocation of resources. national development, by bringing critical Speculation crowds out productive issues into the open rather than allowing investment and economic activity is them to remain ignored and unaddressed. characterized by a focus on short-term profit maximization at the cost of long- State, civil society and national term, stable, job-creating growth. development The mass media have also played a Cambodians’ average incomes and significant role in raising social awareness As levels of literacy and education rise over and changing behavior on a number of time, it is inevitable that they will ask more issues that are critical for national development. In Cambodia, television, of their Government-and, in all likelihood, will develop the capacity to radio and newspapers have all played an participate in increasingly meaningful important part in spreading messages about HIV/AIDS prevention and other health ways on a more equal footing with service practices, about the immorality and providers and policy-makers. illegality of domestic violence, or about This improvement in the capacity of civil the dangers oftrafficking. society to cooperate for economic and 154 Cambodia - Poverty Assessment 2006 social ends is important for national between the Government and NGO development, just as the development of results in better outcomes than either capacity within state institutions is undertaking a task on its own. The important. There is no one model for how experiences of contracting health these relationships between the state and service provision (Chapter 6), or of civil society should be organized. community groups or NGOs serving as Regional examples include such diverse the administrators of exemptions arrangements as the mass organizations in policies and equity hnds, are good Vietnam, the community development examples. movement in Thailand, or watchdog In other words, if a supportive state can committees in Singapore (Burke and Nil make it much easier for civil society to 2005). The important point is that the develop, so too can actively reaching out to Government and civil society should not civil society, make the task of governing be thought of as alternatives competing in easier and more stable. A strong a zero-sum game. The history of the developmental state and a vibrant, dynamic OECD economies (among others) is of civil society should be seen as state action creating incentives for civil complementary and mutually-reinforcing society organization. For the state, partners in national development, rather meanwhile, having established civil than opposed rivals. society interlocutors has many advantages. Enabling civil society to participate in the policy process allows the state to: obtain regular feedback from the grassroots on the success or failure of its policies-and on how these policies, and the Government itself, are perceived; tap into a broader body of knowledge and ideas when formulating policy. Engaging with independent research institutes that possess specialist analytical skills which might be in short supply within the Government is important if policy decisions are to be informed by evidence. identify and deal with emerging or potential future problems in society, which senior policy-makers might not otherwise pick up on until they have become too serious to resolve easily; mobilize civil society capacity in areas in which they may have comparative advantage (e.g. identifying poor households for targeted programs), or where a division of responsibilities 8. New partnerships for poverty reduction

Summary Whether Cambodia will achieve the poverty reduction goals set out in the National Strategic Development Plan (NSDP) 2006-2010, will depend in large part on whether the Government is able to foster a pattern of economic growth that directs more benefits towards sectors, regions and enterprises that include the poor. This can be illustrated in projections based on alternative growth scenarios. If industry and services were to grow by 10 percent per annum, while agriculture continued to average only 2.5 percent per annum, poverty would fall but at a modest rate, and Cambodia would not achieve its 2015 target of reducing poverty to 24 percent. If however GDP in agriculture achieved 4 percent per annum growth, with net GDP growth held constant at 7 percent per annum, Cambodia would be likely to achieve and surpass the poverty reduction target. Achieving pro-poor growth and improved service delivery will require further improvements in the relevance and effectiveness of policies and spending. Given the scale of ODA in Cambodia, the quality of government-donor relationships and aid effectiveness is of critical importance. Over the last decade, aid has played a role in a number of successes. However, a significant proportion has also been badly designed or implemented, resulting in inefficiency and contributing to the governance problems ofthe Cambodian state. Meeting the poverty reduction targets laid out in the CMDGs thus requires improving both national capacity and governance and aid effectiveness. This will require progress on the three fronts ofownership, alignment and harmonization. Ownership implies that aid is directed to activities determined by the Government, rather than driven by the donors. By outlining policies and spending priorities for the next five years, the NSDP creates a framework to which donors can and should align their aid programs. It lays out poverty reduction goals, aligned to the Cambodian Millennium Development Goals (CMDGs); asserts the role ofthe rolling Public Investment Plan (PIP) in aligning capital expenditures to NSDP priorities; and puts in place arrangements for annual progress reviews and policy adjustments. Donors now need to support the NSDP by harmonizing their aid (e.g. cooperating amongst themselves to standardize management systems and share analysis) and aligning it to government priorities and systems (for example, through the PIP or sector-wide approaches). While progress towards harmonization and alignment has been made in recent years, much more is needed. Finally, the Government, donors and civil society need to establish effective partnerships for monitoring and evaluating progress in the implementation and outcomes ofthe NSDP. Once again, progress has been made, and the Statistical Master Plan provides the outline of both the specific exercises and systems for the collection of socio-economic data, and the capacity-building improvements required for sustaining this work. Further refining and then financing the M&E framework outlined in the NSDP will be essential for ioint monitoring of progress in poverty reduction. 156 Cambodia -Poverty Assessment 2006

he previous chapter attempted to standards that they should have met. identify some of the strategic Heavy reliance on technical assistance challenges that the Royal (TA) and stand-alone projects, while T partially justifiable as a response to Government must tackle to accelerate progress in poverty reduction. We begin capacity and governance problems within this final chapter by mapping out a series government institutions, have too often of scenarios regarding the rate and actually retarded moves to improve those distribution of economic growth, and how institutions over the long term. these different scenarios would result in Thus, if the NSDP is to achieve its goals, it different rates of progress. The next is important that the Government and its section then describes the Government’s external partners identify ways to work new medium-term development strategy together more efficiently and effectively so (the National Strategic Development Plan that ODA programs support the poverty or NSDP), which lays out strategies to reduction policies outlined in the reduce poverty between 2006 and 2010. Government’s NSDP-and in doing so add The primary purpose of the NSDP is to up to more than the sum oftheir parts. The guide policies and resource allocation third section ofthis chapter briefly reviews decisions of the sovereign Royal some of the problems that have affected Government of Cambodia. However, the aid to Cambodia in the last decade, and Cambodian state remains severely under- describes how the increasingly complete resourced and remains heavily dependent institutional and policy framework for aid upon external assistance, whether defined management might address these problems in terms of the size of ODA relative to over the course ofthe NSDP. national GDP (which has averaged around Finally, the chapter addresses the 13 percent for the last ten years); or in requirements for monitoring and relation to government expenditures evaluating progress towards the (which over the period 1999-2004 have Government’s poverty reduction goals. been relatively stable but declining After many years in which donor gradually, with ODA a little under half of assistance to data collection and analysis the budget in recent years). has been somewhat ad hoc and supply- Cambodia is likely to remain heavily driven, the National Institute of Statistics dependent upon flows of official together with the donor community has development assistance (ODA) for several developed a more strategic medium-term years to come. As such, the relevance, framework for statistical capacity-building efficiency and effectiveness of aid become and specific statistical exercises. The final critically important to achieving the targets section describes this framework, and laid out in the NSDP. In the decade since identifies what is needed to improve the the Paris Peace Accord paved the way for availability oftimely, accurate and relevant large-scale aid inflows, Cambodia has information for policy-makers seeking to received substantial quantities of ODA. design and implement anti-poverty While donors can thus take a share of the strategies. credit for some of the improvements in service delivery and outcomes seen over the last decade, it is also widely acknowledged that aid coordination and delivery has often fallen’ below the New Dartnershim for Dovertv reduction 157

Table 8.1: The rate of poverty reduction has been good-but a number of other countries have done better Country Period % under poverty line Average annual Start End in start in end rate of poverty year year year year reduction Nicaragua 1993 2001 50 56 0.5 Cambodia 1993/4 2004 47 35 I.1 Lao PDR 1993 2003 46 33 1.3 Mozambique 1997 2003 69 54 2.5 Uganda 1992 2000 56 35 2.6 Vietnam 1993 2002 58 29 3.2 Tajikistan 1999 2003 75 57 4.5

Source: World Bank 2003; CSES 2004: Lao PDR 2003; Commission for Africa 2005; Poverty Task Force 2003: World Bank 2005 Will Cambodia meet its poverty Mozambique); or, indeed, which still reduction targets? suffer from conflict in parts of their territory (e.g. Uganda). This is in one Assessing the past decade sense encouraging, as it suggests that, given moderately favorable external reduction in poverty by 10 to 15 A circumstances and good government percentage points over the course of a policies, it should be possible to not only decade is a significant achievement. sustain but increase the rate of poverty Accounting for population growth, there reduction in the future. are now some 460,000 fewer poor people in Cambodia than there were ten years 1 Projecting trends for the next ago . decade This is, obviously, a very positive The level and pattern of economic growth development. In international terms, will critically affect the rate at which however, Cambodia’s performance in poverty falls over the coming years, and reducing poverty is respectable but not thus the prospects of Cambodia achieving exceptional. Unlike many countries which the Cambodian Millennium Development have emerged from civil war, it has Goal (CMDG) target of reducing avoided slipping back into conflict. consumption poverty to half its baseline However, there are also a number of low- 1993/94 value by 20152 (i.e. from 47 income countries that have achieved significantly higher per annum rates of poverty reduction (Table 2). These include At present, this 2015 poverty incidence target a number of countries that share is set in both the NSDP and the 2005 CG Update at 19.5 percent. This is based on an Cambodia’s history of emerging from old, straightforward reading of the incidence conflict during the 1990s (e.g. found in the 1993194 survey, which would establish a baseline of 39 percent. If however we attempt to come up with a more realistic ’ Given the difficulties in pinning down the baseline figure for poverty in the whole of 1993194 poverty headcount estimate with Cambodia in 1993194 through backward precision, it may be better to say that there are projection using what we know to have been anywhere between 240,000 and 790,000 fewer the trend in the parts of the country covered by poor now than in 1993194. both the 1993194 and 2004 surveys, we would 158 Cambodia - Poverty Assessment 2006

p.a. agricultural CMDG baseline (199314) and target (2015) 4 growth 0 I I” ,88881

Source: World Bank projections (backward and forward), using SESC 1993/4 and CSES 2004 data and PovStats.

percent to 24 percent). There are various rate of poverty reduction that will occur if ways in which alternative scenarios could growth continues at the same overall rate be formulated. The one we use here is (7 percent per annum) and sectoral relatively simple. Using the CSES 2004 as composition as over the last decade. The the base year, we back-project annual second projection is based upon a scenario poverty headcount estimates for the years in which overall growth remains the same 1994-2004 (Le. estimating poverty rates (7 percent per annum) but the rate of for the years between the 35 percent growth in agriculture rose to 4 percent per recorded in 2004 and the 47 percent annum (with corresponding reductions in estimated for 1993) on the basis of annual the rates of industrial and service sector sectoral rates of GDP growth and the growth). As with the historical annual World Bank’s PovStat Toolkit (see Annex poverty rates between 1994 and 2004, for details of the methodology). The best- these two forward scenarios are estimated fit line for these estimates provides us with using PovStats software. Any changes are a trend line for poverty over the period assumed to come into effect only from 1993-2004. 2007, as it is assumed that the level and Figure 8.1 presents these annual estimates, pattern of growth in 2005 and 2006 will be the same as in preceding years, historical trend line and two forward projections, reflecting different growth As can be seen from Figure 8.1 , Cambodia scenarios. The first projection represents a is unlikely to attain the 2015 poverty linear extension of the trend in poverty reduction under the first, “no change” between 1993 and 2004, representing the scenario. According to this projection, if growth continues at current sectoral rates, estimate the 199314 baseline as 47 percent the poverty rate in 2015 would be closer to (Chapter 2), and the 2015 target therefore at 29 percent than the target of24 percent. 23.5 percent. New partnerships for poverty reduction 159

However, in the second scenario (indicated and, in turn, shape annual planning and by the dark grey line in Figure 8.1) of budget processes. In countries where such slightly accelerated (4 percent per annum) planning systems do not exist or do not agricultural growth, the fall in poverty address poverty, the donor community has between 2007 and 2015 is likely to be since 2000 advocated the development of considerably steeper. Under this scenario, Poverty Reduction Strategies or PRS (Box Cambodia is highly likely to attain its 8.1). MDG poverty target, with a projected The importance of a credible PRS in poverty rate of percent in 201 2 1 5. improving aid effectiveness has been By putting numbers to alternative reinforced in recent years as the global scenarios, these projections help to confirm agenda for aid harmonization and the intuitive case that if poverty reduction alignment has developed momentum. The is to be sustained or accelerated, there March 2005 Paris Declaration on Aid needs to be a concerted effort to raise the Effectiveness identifies the existence of a rate of agricultural growth-and to do so national development strategy as the in such a way that it benefits the rural prerequisite for government ownership. It majority and the rural poor. This does not also commits donors to align their country mean that other sectors can be neglected; programs to partners’ national for example, raising the level of education development strategies (and periodic and mitigating the economic impact of reviews ofprogress against goals set out in high-cost, low-quality health care for these strategies). serious illness will themselves have important positive consequences for rural Medium-term planning in growth. It does however help to make the Cambodia, I 9 93-2 005 case that policies and investments for The Cambodian planning process was economic growth need to focus explicitly rebuilt over the 1980s with very limited on agriculture and the rural economy if resources and on the basis of a centrally- they are to help Cambodia achieve its planned economy. It has evolved since goals for poverty reduction. 1993, with inputs from donors, to adapt to a more complex mixed economy. The NSDP and CMDGs However, it still faces a number of severe To tackle a cross-cutting problem such as constraints growing out of a combination poverty, a government needs to adopt a of: (i) limited financial and human holistic, prioritized medium-term resources; and (ii)institutional and process framework that draws together policies in a arrangements which are not always best broad range of sectors (from suited for the tasks they are assigned (Toda macroeconomic management to enhancing 2001). Cross-government planning and productivity to improving service delivery) alignment of budgets to plans are still with resource allocation and institutional weak3. development strategies. Many governments have institutionalized five- year national development plans as the cornerstone of their policy and planning systems. These form the basis for medium-term public expenditure plans See the previous chapter for discussion of the credibility of the budget. 160 Cambodia - Poverty Assessment 2006

Box 8.1 : Poverty Reduction Strategies: ownership, participation and poverty outcomes The promotion by donors of Poverty Reduction Strategies (or, originally, Poverty Reduction Strategy Papers) emerged from a number of roots. Initially, the proposal was that countries qualifying for World Bank and IMF debt relief would have to produce a PRSP to show that the money would be allocated to poverty reduction, and how. However, the promotion of PRSPs was soon generalized to a// countries in receipt of IF1 concessional financing-and increasingly adopted by other donors as a lynchpin of aid dialogue at the country level. The approach recognized that policies mandated as part of traditional structural adjustment conditionalities were often unfulfilled or unsuccessful because the recipient government lacked commitment to them. Low levels of aid effectiveness were attributed to the lack of “ownership” by the recipient government. By contrast, countries (Uganda was the best known example) that had produced their own comprehensive medium-term strategies for poverty reduction, linked to budget processes and developed with participation from the poor, appeared to have had much more success in sustaining reforms that added up to more than the sum of their parts. A PRS, thus, was intended to help redefine the relationship between a government and its donors. If the government could articulate a coherent and credible medium-plan for orienting policies and public expenditure towards poverty reduction goals, the donors would accept this as the framework for their aid programs. Carried through in sector-wide approaches (SWAPS) and medium-term expenditure framework (MTEF), the PRS would lay the groundwork for: (i) streamlining conditionalities; (ii) reducing the number of donor-driven projects, missions, studies and evaluations, as donors move to budget support, giving up detailed project reporting for a voice in policy and joint annual progress reports; and (iii) improving the predictability of aid flows. Such arrangements would reduce transaction costs and allow for more efficient and strategic allocation of total (domestic plus ODA) revenues. The PRS approach attempted to introduce to national poverty planning a number of innovations. Two of the most important were: (i) an emphasis upon participation; and (ii) an orientation towards final development outcomes. The advocacy of participation was based upon international experience which strongly suggested that poverty reduction policies which involved the poor were likely to be more relevant, realistic and legitimate in the eyes of the intended beneficiaries, and thus more likely to succeed. The emphasis upon outcomes, meanwhile, emerged from new public management theory and attempted to shift the definition of planning goals and targets from outputs (e.g. schools built, midwives trained, increased agricultural production) to ultimate development outcomes (e.g., improved literacy, reduced maternal mortality, rural poverty reduction). While outputs remain critically important and need to be monitored, shifting attention to tracking outcomes can often help tighten the rigor with which policies are formulated, implemented, appraised and revised. From the donor side, thinking on the PRS approach has evolved since it was introduced in 1999. The inherent tension between principles of country ownership and donor conditionality was managed with differing degrees of success in different countries. Reflecting this experience, there is a general preference now for helping countries with an existing medium-term planning tradition to integrate PRS principles (e.g. poverty focus, broad national ownership, participation, results orientation, etc.), rather than, as was sometimes the case, insist on the production of a PRSP in parallel to the “normal” plan.

Although ultimately the responsibility of development of medium-term plans the Government, donors have also suffered from a lack of coordination. contributed to this state of affairs, with Multi-lateral donors have each supported assistance to the Ministry of Planning the development of national strategies or (MOP) taking the form of unit-level strategy-like sets of targets, sometimes in projects that have contributed to a parallel. This resulted in confusion over balkanization of planning functions. Past which plan to follow or which set of efforts by donors to support the targets to work towards, and helps to New partnerships for poverty reduction 161 explain why none of the strategies were umbrella NGOs. Although a positive substantially implemented4. Perhaps development, it has yet to realize its because the plans did not appear to have potential, having met only twice to date. much importance within the Government, The NSDP now provides the Government they exercised very little influence on and donors with an over-arching set of donor programming. The result has been development goals under which sectoral a general loss of confidence in the and sub-national plans can be elaborated. planning process. It is to act as the basis for fbture three-year The Government and donors agreed that rolling Public Investment Plans (PIPS) by the next medium-term plan should address which the Government will steer capital these issues. It was agreed that there investments (both domestically-financed should be just one such plan, which would and ODA-financed) to priority sectors. It provide a set of common targets and is deliberately intended to be a short, monitoring indicators, drawn from the strategic document, which is to be CMDGs. The strategy would be supported monitored and revised on an annual basis by all donors, and used as the reference to reflect progress, new data and emerging point for government-donor negotiations issues. While the NSDP will be finalized on the allocation and modalities of aid. and approved by the National Assembly in Over the course of 2005 a Secretariat December, Box 8.2 provides a summary of based in the MOP has prepared the the priority goals and actions laid out in the National Strategic Development Plan current draft. (NSDP) 2006-2010. The NSDP constitutes Cambodia’s unified medium- Supporting the NSDP: aid term development strategy for the next five effectiveness in Cambodia years. It puts into operation the broad vision laid out in the Rectangular Strategy The production of a comprehensive (announced by the new Government in strategy is a necessary but not sufficient July 2004) and is oriented towards condition for poverty reduction. There is reaching the CMDG poverty reduction also a need for an institutional targets. infrastructure capable of implementing the The Secretariat has discussed their plans plan. Given the significant role of ODA in with and received technical and financial Cambodia, this institutional framework support from a core group of three donors needs to address coordination within and (the World Bank, ADB and the UN between the Government and donors. The system). A Technical Working Group on institutional framework required to Planning and Poverty Reduction (TWG- implement the NSDP must thus ensure PPR) was established to provide partners’ that: (i)Ministries and agencies work input to the NSDP formulation process, together efficiently to implement and including representatives from across the refine the NSDP as a cross-government Government, most donors, and 12 strategy; (ii)donors coordinate between themselves to support NSDP priorities; and 4 Implementation of the last two plans-the (iii)the Government as a whole and donors Second Socio-Economic Development Plan as a collective, find efficient and effective (SEDP 11) and the National Poverty Reduction ways of communicating on strategic issues Strategy. (NPRS) 2003-2005-also suffered to do with NSDP implementation. from the year-long political hiatus following the 2003 election. 162 Cambodia - Poverty Assessment 2006

Box 8.2: The National Strategic Development Plan (NSDP), 2006-2010: key commitments

ption efforts through policy reforms, behavioral change and transparency to combat corruption and instill a culture of service within public administration. o Deepen legal and judicial reforms via institutional strengthening, capacity building, passing basic laws, providing legal aid for the poor, and out-of-court conflict resolution mechanisms. o Implement the National Program of Administration Reform, including rationalizing the civil service through merit-based appointments and promotions, increasing salary levels. o Promote decentralization and deconcentration to strengthen local democracy and participatory development, and improve local level service delivery by building local management capacity and further delegation of responsibilities and devolution of funds. o Reform the armed forces, continue with demobilization.

ral elections for the Senate, National Assembly and Communes. o Strengthen development partnerships with all stakeholders, including private sector, donor community and civil society, and increase aid effectiveness following the Rome and Paris Declarations, and encourage donors to work increasingly through SWAPS and budget support. o Sustain a favorable macro economic and financial environment by ensuring steady GDP growth of 6 percent per annum, containing inflation under 5 percent, broadening the tax base, targeting expenditures to priority sectors, regular budget auditing across government, and transparent management of state assets. Implement the PFM program and the Priority Action Program, and the Financial Sector Blueprint. o Further promote integration into the region and the world by entering into more free trade agreements, and attracting investments. o Address poverty, ensuring that all strategies focus on poverty reduction with the aim to reducing omoting pro-poor targeting of investments. o Formulate and implement a comprehensive Agriculture and Water Resources Strategy. o Improve agricultural productivity and diversification to increase exports and food security. o Carry out reforms in land administration and management, fisheries and forestry reform, o Invest in environmental conservation and rural infrastructure development.

o Improve management of water resources and irrigation. o Develop energy and power grids to reduce cost and increase access to power, through the involvement of the private sector, manage future oil and gas resources and revenues. o Develop information and communication technology.

an and the recommendations in the Investment Climate Survey aimed at strengthening the private sector and attracting investments. o Promote SMEs, trade and tourism, rural credit. o Create jobs and ensure improved working conditions. o Establish social safety nets for the disadvantaged.

ntation of the ESSP. o Improve health outcomes through the Health Action Plan (service delivery, behavioral change, quality improvement, institutional and HR development, health financing). o Foster gender equity via gender mainstreaming, addressing domestic violence and trafficking, increasing women’s access to productive resources, high quality employment, and participation in decision making processes and public administration. o Implement population policy to decrease fertility and promote birth spacing. New partnerships for poverty reduction 163

Aid and poverty reduction, 1993- Box 8.3: Ownership, alignment and 2004 harmonization: basic concepts Over the last decade, Cambodia has 0 wnership. U Iti matel y, major im provemen ts in development cooperation rely on credible received significant quantities of aid. The indications of government ownership of the achievements that this report describes in development agenda. Ownership implies that reducing poverty in its various aspects- the Government understands the problems it higher household consumption, a number faces, has developed a feasible strategy for of improved human development tackling these, and shows sign of real political commitment to implement this strategy. outcomes, etc.-reflects some degree of success on the part of donors in helping to Alignment If persuaded of the adequacy and sincerity of the Government's development support improvements in service delivery strategy, donors should align their country and, to a degree, economic growth. cooperation programs with the Government's However, there is also a widespread lead on policy directions and budget priorities. Where such ownership does not exist, aid acceptance that much of the aid that should be managed in ways that promote the Cambodia has received over the last evolution of ownership-or, at a minimum, do decade has not been as effective as it could not retard its development. A different kind of have been in the fight against poverty. For alignment involves donors commitment to use a variety of tactical and strategic reasons, government administrative and financial systems to manage aid, rather than (as has ODA has not always been well prioritized, conventionally been the case) requiring the coordinated, or delivered in ways which Government to adapt its own processes to a build government capacity. In too many multitude of different donor systems. This cases, short-term aid management systems alignment, however, requires that solutions designed to sidestep the capacity donors (and their taxpayers or shareholders) are satisfied that government systems and and governance problems of the post- procedures meet minimum technical standards conflict Cambodian state (e.g. reliance on of efficiency and financial accountability. technical assistance to fill capacity gaps, or Harmonization. Finally, there is considerable use of stand-alone projects to ensure scope for donor-donor harmonization of their control and avoid the fiduciary risks own procedures by: (i) adopting common involved in working through existing RGC arrangements for planning and delivering aid structures) have often ended up (e.g. greater clarity and commonality regarding perpetuating or exacerbating those the conditions that would lead to increased or decreased giving, or through support to problems over the long term (World Bank SWAPS); (ii) streamlining and rationalizing their 2004). procedures (e.9. through joint, rather than single, donor missions and studies) so as to Over the last few years, there has been reduce the burden on the Government; and (iii) increasing recognition within Cambodia of sharing as widely as possible analytical work the problem of patchy aid effectiveness and information on planned and actual aid and increasing calls for better flows. Donor-donor harmonization can and harmonization and alignment of ODA (see should occur regardless of whether or not they can align their programming and systems with Box 8.3 for definition of terms). The those of the Government. That is, even when volume and design of technical assistance donors cannot align their policies or systems (TA) has been a particular point of with those of the Government (because they contention. The number of foreign TA judge that those policies are misguided or the advisors is very high (estimated at around systems are inadequate to ensure accountable, effective or efficient use of aid 800), as is the proportion of ODA that is resources), donors can take steps amongst spent as TA (circa US$162m in 2001; themselves to harmonize. 164 Cambodia - Poverty Assessment 2006

Box 8.4: Improving outcomes or around 21 percent of combined domestic institution-building? and externally-financed expenditures or more than the Government’s total wage Francis Fukuyama notes that donors often end up “making things worse” by actually bill). There is no doubt that there is an destroying institutional capacity in the important role for TA, which, if managed quest for rapid results. He argues that a correctly, can contribute to long-term donor cannot simultaneously improve capacity development. However, relying service delivery and human development upon advisors to manage projects or outcomes in the short- to medium-term and build capacity over the long term. Until the reform processes can undermine national incentive structure facing donors is ownership and prevent the development of changed, the need to please donor country institutional capacity, especially when taxpayers, politicians, audit authorities and these advisors are weakly integrated into pressure grou ps-and , indeed , developing the mainstream of Ministry processes country partner governments who need to get re-elected or otherwise remain in (Godfrey et a1 2003; Mysliwiec 2004). power-will result in the focus on short- Ultimately, however, the problem is term improvement in outcomes (through broader than just that of TA. The direct donor engagement in service rationale that underlies the excessive and delivery) trumping the longer, slower, more uncritical use of TA-that is, the desire to painful and less easily justified focus on building the capacity of low income states circumvent an underpaid and under- to improve service delivery and outcomes managed civil service in order to achieve themselves. There is overlap between demonstrable improvements in poor Fukuyama’s diagnosis and that of Nancy people’s lives in the short- to medium- Birdsall, who in listing the common “seven term-is the same rationale that underlies deadly sins of development cooperation” includes that of “impatience (with another risky and ultimately counter- institution-building)”. productive aid practice, that of ad hoc donor-paid salary supplements for Sources: Fukuyama 2004; Birdsall 2004. government employees (Box 7.2 in the previous chapter). Both of these deeply nine participants in the DAC Peer Review ambiguous and frequently problematic ofharmonization and alignment. forms ofaid delivery-and a host of others besides-are unfortunately far from unique Structures for Government-donor to Cambodia (Box 8.4). cooperation Progress in addressing aid effectiveness During the first decade of Cambodia’s has been facilitated by the Government’s recovery in the 1980s, coordination was active participation in the work of the not a major problem. Few donors worked OECD Development Assistance in Cambodia and resource constraints were Committee (DAC) to research the state of so pronounced that there was little room aid harmonization and alignment and for debate over government priorities, so formulate global guidelines, notably the sector-level coordination structures worked Rome (2003) and Paris (2005) well5. With the massive rise in the number Declarations on Aid Effectiveness (see of donors and the volume of ODA in the OECD 2003 and OECD-DAC website). Cambodia is one of the fourteen partner As described in the previous chapter, the few countries on the OECD-DAC Working international NGOs operating in Cambodia during the 1980s operated in donor-like ways, Party on Aid Effectiveness and one of the channeling significant hnds to capital investment and basic service delivery. New partnerships for poverty reduction 165 early 1990s, however, these structures health, where multi-donor partnerships proved inadequate. The institutional have evolved gradually in support to framework for RGC-donor-civil society developing sectoral strategies. A Sector- dialogue around development policy has Wide Approach (SWAP) in education and evolved through several stages since then. Sector-Wide Management (SWIM) in health align ODA behind a Ministerial As in other countries, annual Consultative strategy and reduce transaction costs. Group (CG) meetings are now held in the Multi-donor partnerships have also country rather than, as previously, in donor streamlined donor-government relations capitals. These meetings provide a forum and helped to reduce gaps and overlaps in for a strategic review of progress on regard to decentralization and local reforms and discussion of the level of governance (through the Seila program and ODA that donors plan to commit for the the Partnership for Local Government); next year. public financial management (with A number of permanent bodies share the agreement on multi-donor support to a overall mandate for strategic planning and three-stage process of reform, including the integration of development assistance salary reform, in the MEF); and private into RGC policies and plans. MOP is sector development, where donors are responsible for coordinating the production coordinating efforts to support the RGC’s and implementation of the NSDP and 12 Point Plan to Improve the Investment working with the Planning Departments of Climate and Trade Facilitation (2004). In line ministries. However, the Council for the land management sector there is an the Development of Cambodia (CDC)- emerging program-based approach that which seeks to direct capital investment harmonizes donor support to development (including both ODA and private and implementation of key government investment) to RGC priorities-is also hnctions in land resources. assigned responsibility for guiding the However, in other sectors the quality of preparation of national “development sector policy and aid coordination and the vision and strategies” and coordinating effectiveness of total spending have been activities across the Government and much lower. By the early years of this between the Government and donors6. The decade, sector-level arrangements varied Ministry of Economy and Finance (MEF), greatly in strength. In an effort to impose a meanwhile, has the remit for common basic structure, five working macroeconomic, fiscal and monetary groups were formed in 1999 with another policy, and as such primary control upon two (Governance and Partnership) added the allocation of public finances; while in December 2002. Responding to French bilateral cooperation is managed perceived limitations with this model, in through the Ministry ofForeign Affairs. October 2003 the Prime Minister approved Moves towards more coherent ODA that is a proposal to restructure the working better aligned to RGC leadership first groups. Over the course of 2004 the began in earnest at the sectoral level. coordination framework was recast in the Today, the aid reform agenda is most form of 17 sectoral and thematic Technical advanced in respect to education and Working Groups (TWGs). A joint government-donor proposal to strengthen CDC is responsible for collecting and reporting the architecture for aid, endorsed in on the volume and allocation of ODA, most September 2004, laid out two main notably at CG meetings. 166 Cambodia - Poverty Assessment 2006 elements. The first was the 17 restructured coordination mechanisms has helped, the TWGs (since joined by the 18th TWG on performance of TWGs continues to be very Planning and Poverty Reduction), each of mixed. While some work well, others are which was to be chaired by the relevant struggling with limited capacity to manage Government Ministry or Agency and co- their members and the resources that facilitated by one or two lead sector members bring to the table. donors. Each TWG was mandated to prepare joint strategies and action plans for Defining the aid effectiveness the sector, consistent with the Rectangular agenda in 2005 Strategy to mobilize the necessary The latest round of Cambodia's implement the resources to strategy; and to engagement in the DAC-led global debate jointly monitor progress against on aid effectiveness has involved benchmarks set out in the sector strategy as Cambodia's participation in a baseline well as the overarching Joint Monitoring survey of Paris Declaration monitoring Indicators (JMIs) agreed at CG meetings. indicators. Although these indicators are The second element was a new based on self-reported data from a sample Government-Donor Coordination of 16 donors, and involve a degree of Committee (GDCC), mandated to subjective interpretation of the indicator strengthen policy-level coordination, set criteria, they nonetheless provide a useful priorities and resolve issues raised by the benchmark against which future progress TWGs. Quarterly GDCC meetings began can be measured (Figure 8.2). The in 2004. Three NGOs (usually those who indicator for budget support needs to be attend the CG meeting) are invited to interpreted with some caution. A number GDCC meetings, which review TWG of donors (including the ADB and World reports on progress against their action Bank) have provided or continue to plans and the CGJMIs. provide budget or balance of payments

support of , Although the standardization of sector to Cambodia. The value "zero" in 2004 thus reflects a particular year.

Figure 8.2: Status of harmonization and alignment in 2004: selected indicators

"" "." ...... " "..l""."."l"..."."" ...... ". % country analytic studies produced jointly % missions that were joint li % ODA disbursed using common sector arrangements

~ % ODA disbursed as general budget support I 'YO PIUS fully integratea 42%

, Yo ODA using national procurement systems

% total capacity-building TA proided jointly

Source: Completed master questionnaire for OECD-DAC Baseline survey-Paris Declaration on Aid Effectiveness. Notes: Participating donors were Danida, AusAID, EC, CIDA, Sida, UN, DFID, USAID, World Bank, ADB, AFD, French Embassy, Japan, KW, GTZ and DED. New partnerships for poverty reduction 167

However, the fact that none of those Box 8.5: Cambodia’s Harmonization surveyed reported using budget support in and Alignment Action Plan 2004 is still significant, given that several The Cambodia Action Plan on of the other donors surveyed are, in other Harmonization and Alignment was adopted partner country contexts, willing if not in December 2004 and revised during 2005 enthusiastic supporters of budget support. to reflect the recommendations of the Paris That only 11 percent of total disbursed Declaration. The Action Plan establishes ODA amongst the sample donors was the following goals: spent using national procurement systems 1. Strengthening national ownership reinforces the impression that many donors whereby the Government exercises are uncomfortable about the level of effective leadership over its development policies, strategies and coordinated fiduciary entailed using RGC risk in development actions. Thus the Ministry systems. of Planning leads NSDP preparation; However, there are also several indicators CDC leads the process for resource mobilization and aid coordination; and of harmonization (e.g. joint analytical line ministries lead the management of work or missions; TA provided as multi- development assistance. donor arrangements) which remain low, despite the fact that donors could make 2. Promoting alignment. Development partners should base their overall progress to a large degree regardless of support on Cambodia’s development government capacity or governance strategies (e.g. the NSDP and sector, constraints. thematic or issue-based programs), institutions and procedures ( e.g. by Drawing on its experiences of close strengthening RGC’s capacity for public engagement in the Rome and Paris financial management and for planning, processes, Cambodia has translated global managing, and accounting for the results principles into a number of Cambodia- of policies and programs). Development specific policy papers on improving aid partners should increase value for effectiveness, notably the 2004 money in the management of ODA (e.g. Harmonization and Alignment Action Plan by untying aid). (Box 8.5). The Government has also 3. Increasing harmonization. Development increasingly asserted its own priorities for partners’ actions should become ODA, in terms of sectoral allocation (with increasingly harmonized, transparent and collectively effective. This includes more to go to agriculture and economic use of common arrangements and infrastructure, particularly irrigation), simplified procedures; increased geographical targeting (with more to reach complementarity among development rural areas) and modalities (with less to be partners; and incentives for collaborative spent on TA and project aid, and more behavior. delivered through budget support). These have been laid out in the draft NSDP and a a sister document, the National mid-2005 CDC document (the Strategic Operational Guidelines for Development framework for development cooperation Cooperation Grant Assistance, which management) which outlines some guiding provides guidelines for the development of principles for the planning and results-based project cycle management. management of ODA resources on issues These are positive steps towards increased such as the use of loans as opposed to ownership of aid management. However, grants; the role of the CDC; and linkages there is a danger that the debate between to the TWGs. The CDC has also prepared 168 Cambodia - Poverty Assessment 2006 the Government and donors becomes (or term perspective ultimately self-defeating7, remains) unhelpfully polarized. Poor the Government still needs to recognize the governance and poor aid management have legitimate concerns that underpin at least co-evolved over the last decade. some of this behavior. Although some Unblocking this logjam will require donors provide and will continue to simultaneous and complementary reforms provide budget support, and others are by both the Government and its likely to join them over time, there is development partners. In other words, unlikely to be any wholesale shift of significant improvements in aid donors towards budget support until effectiveness will require progress on existing governance problems are governance, and vice versa. addressed and donors have more confidence in the fiduciary arrangements Thus, for example, donors have relied on needed for government execution ofaid. TA and projects in part to circumvent ineffective state implementing agencies There is thus a strong argument that, while and reduce their exposure to the risk of keeping up pressure for budget support as corruption. It is now broadly recognized the desired modality in the long term, the that such attempts to ring-fence donor Government and donors also need to funds within project implementing units define a (complementary) short- to (PIUs) and donor-specific reporting medium-term agenda of intermediate requirements have in turn at best failed to harmonization and alignment reforms that address the systemic problems of can be realistically achieved over the government structures (Le. lack of lifetime of the NSDP. Box 8.6 lists some financial and human resources, positive examples from recent years, while inappropriate institutional design, and low Box 8.7 provides some ideas from good levels of transparency and accountability). practice in other countries where donors Often, they have exacerbated these and the Government have made significant problems. To take the most obvious moves towards putting into operation the example, when the best officials are principle ofmutual accountability. recruited by donors to manage project Building on these developments and the implementation, they are drawn out of the new Government’s strategic framework for mainstream of the Ministry, where they aid suggests the following priorities for should be deployed to lead the improving aid effectiveness: development of core RGC policies and systems. Donor-specific project . Concerted efforts to frame formulation missions, reporting development cooperation in reference requirements, M&E and audit processes, to the NSDP and the PIP that will meanwhile, add to the administrative (hopefully) be used to align capital burden of Ministry under-resourced staff, investments to NSDP priorities. distracting them from the task of policy and institutional reform and routine ’ planning and budgeting exercises. There is a separate argument to be made that project modalities are in any case not However, if the logic of short-term donor- necessarily any more effective at preventing driven enclave models of development corruption. The experience in recent years of cooperation is in a broader and more long- both the World Bank and World Food Program suggests that, at best, projects are far from being a guarantee of protection from embezzlement or misprocurement. New partnerships for poverty reduction 169

Box 8. 6: Harmonization and alignment: government ownership will require some practical examples improvements to the TWGs (e.g. by Joint strategic planning. The ADB, DFID, ensuring that appropriate individuals World Bank and UN system prepared their new are delegated to participate in (2005 onward) country strategies jointly. meetings, creating space for non-state Although this imposed additional transaction actors to participate and so acquire a costs on the four partners, the burden on other better understanding of RGC policies; stakeholders (in terms of, for example, consultation meetings) were reduced and there establishing partnership principles to was some division of labor between the four keep the work of the group focused on parties. The country situation analysis was sector policy issues and not mere shared, resulting in an exchange of knowledge, information-sharing; and embedding a improved analysis, and greater consensus in role for TWGs in annual planning, perspective; and a foundation was laid for better future cooperation between the four budgeting and review cycles, linked to agencies. (for example) GDCC or CG meetings, or the annual NSDP progress review- Alignment of country planning. Within the UN, the Common Country Appraisal-UN which should ideally be brought Development Assistance Framework (CCA- together. UNDAF) process was streamlined and defined A balanced dialogue on technical in reference to the NPRS. . assistance that establishes some basic Joint analytical work. Several strategic studies practical principles for a common (for example, the Cambodia Gender Assessment) have been undertaken by two or approach to capacity-building and the more donors. In some cases these have role of TA. Realistically, there will be provided a broader base for reform programs continued need and demand for TA8. that are subsequently agreed with the Nonetheless, there is considerable Government (e.g. the Integrated Fiduciary room for both reducing the quantity Assessment and Public Expenditure Review, produced by ADB and the World Bank with and improving the quality of TA. inputs from a number of others). There needs to be a commitment by Harmonized procedures and support to the donors, enforced by peer pressure and development of core government systems. counterparts, that the primary purpose The RGC, World Bank and ADB have of TA is to build long-term capacity cooperated to produce a Manual of Standard through skills transfer and institutional Operating Procedures and joint project restructuring (and not gap-filling or financial management system and manual. AFD has subsequently agreed to use these training individuals). Terms of procedures and it is hoped that others will Reference must be agreed with the follow. Government and other sector partners Joint programming. Danida and DFID are through the relevant TWG, and there currently developing a five-year multi-donor needs to be a managed shift from facility for natural resource management and international to national TA. livelihoods promotion with a single management structure. * It is notable that while the NSDP and CDC Strengthening the institutional calls for less TA, Ministries are still often framework for government-donor requesting more TA to help design or dialogue-including, where required, implement particular programs or reform processes (often with good justification). TA completing the missing links. To make certainly needs to be reduced, but it may prove the NSDP effective as the locus for hard to adopt any simple formula for how or how fast to do so. 170 Cambodia - Povevty Assessment 2006

Box 8.7: Towards mutual accountability: lessons from Tanzania and Mozambique Aid dependency is high in much of sub-Saharan Africa, leading to many dysfunctional aspects of Government-donor relationships of the kind seen in Cambodia. In several of these countries, however, significant progress has been made over the last five to ten years. In Tanzania, reform was driven by the onset of economic and political crisis when donors collectively suspended aid in 1994. An independent assessment of government-donor relationships (the 1995 Helleiner Report) dissected the lack of government vision and increasingly rampant corruption that had precipitated this move; but was also scathing with regard to poor donor practices which had significantly contributed to the lack of government ownership. The reformulation of aid relations since 1995 has been profoundly shaped by the Helleiner report. An increasingly mature partnership was codified in the 2002 Tanzania Assistance Strategy (TAS), in which the Government and donors agreed to institutionalize independent monitoring and evaluation reports, to be reviewed jointly as the basis for discussion on next steps to be taken. A jointly-appointed Independent Monitoring Group (IMG) has now reported to the CG meeting twice; the 2005 report notes significant progress, but also observes that TA continues to be the most problematic aid modality. It recommends that the Government of Tanzania formulate a clear TA policy which specifies that the primary role of TA must be capacity building; and calls for project aid to be allowed only when it meets stipulated criteria. In Mozambique the post-conflict government managed to establish and maintain an unusually open and trusting aid relationship which has seen a growing number of donors channel aid as direct budget support. In 2004 these 15 donors (now 16) signed a MOU with the Government that committed both sides to joint review of both Government and donor performance. Sources: ESRF 2005; Killick et al 2005.

Streamlining and harmonization accountable “horizontally” (to the through delegated cooperation Cambodian Government) and arrangements, in which two or more “downward” (to the intended donors that agree on a program of beneficiaries amongst the Cambodian support arrange for one to execute the poor) for the quality oftheir aid. fimds of both parties, with the Engage the full range of Cambodian delegating donor requiring only . stakeholders. This should involve periodic reporting and joint review. more active dissemination of Khmer- More effective oversight of aid language information on country effectiveness through fewer but more programs, projects, cooperation meaningful joint (multi-donor and principles, and evaluations. In a post- donor-government-civil society) conflict society in which the monitoring and evaluation framed in institutions of both the state and civil reference to sector or thematic goals, society need to be strengthened and the and to harmonization and alignment vertical flow of information is very indicators. Lessons learnt should limited, a role for donors may be to contribute to improved practices and provide fora for bringing the accountability not only with regard to Government and civil society together, the donor agency and its home-country where civil society would obtain a constituents (taxpayers, better understanding of government Parliamentarians or shareholders): they policy positions and the reasons for should also provide lessons for RGC them, while policy-makers would policy and institutional reforms, and obtain a better sense of how policies inculcate the principle that donors are and outcomes are seen to be changing New uartnershivs for * vovertv reduction 171

(or not) at the local level. This is a exacerbated by the lack of a link between difficult role to play, but one well performance and budget). To this must be worth the effort. added (iv) the unclear division of responsibilities and lack of communication Coordinated and harmonized donor between MEF and MOP on matters of support for a government-owned monitoring and evaluation of national medium-term strategy for building policy, and uncertainties about the capacity capacity in monitoring and evaluating of MOP (the parent institution for NIS) to progress in NSDP policy fulfill its designated role as clearing house implementation and poverty outcomes: for matching cross-government demand for which is the subject ofthe next section. data and analysis with supply emanating Mon itori n g and eval uat i ng from NIS and other RGC institutions. progress Now that the Government is finalizing a In the past, the development of national new outcome-oriented medium-tern monitoring and evaluation (M&E) systems development strategy in the form of the has suffered from much the same problems NSDP, there is a need for a strong and of poor donor coordination that have been integrated system to collect and analyze described above. With very limited key indicators of progress, and to ensure government funding for either the National that these are fed in effectively to Institute of Statistics or the statistical statistically-literate users in order to shape functions of line Ministries, the collection, decisions on policy development and analysis and use of statistical data was public expenditure allocation (Box 8.8). constrained by severe shortages of Despite the problems described above, financial and human resources. The result many ofthe preconditions are now in place was that statistical development has been for the development of a nationally-owned heavily donor-driven. This was seen most system that will allow all stakeholders to clearly in the fact that major surveys (such review progress and discuss policy as the various rounds of the household changes on the basis of shared and reliable survey) were scheduled and designed on a information. Elements of this emerging largely ad hoc basis. framework for poverty M&E include:

The consequences have been serious. m A new Statistics Law which came into They include: (i)a lack of comparability effect in May 2005 and which defines of key indicators over time (due to principles and institutional changing survey design); (ii)an inefficient, responsibilities for data collection, feast-and-famine pattern of funding for analysis and dissemination. data collection and capacity-building (with large sums of financial and technical . The production of a Statistics Master assistance made available to NIS when Plan (SMP). This lays out a donors decided to fund a survey, followed framework schedule and budget for the by a drop-off between surveys during prioiity surveys and other data which trained staff moved to other jobs collection systems required to track and institutional memory was lost); and NSDP implementation over the next (iii)little demand for or ownership or five years, together with the utilization of the resulting data by investments in training and technology government decision-makers (a problem needed to build the capacity of the national statistical system (which 172 Cambodia - Poverty Assessment 2006

Box 8. 8: Summary requirements of a towards a government-led strategy for national poverty M&E system building the capacity for NSDP M&E. Regular monitoringltracking of: The groundwork for a unified national . Implementation of policy (inputs/ policy and poverty M&E system is thus outputs). Change in intermediate outcomes. now substantially in place. To realize this . Variations in status and trends in potential, a number of actions need to be different groupslregions. taken. First, it is important to clarify the Plus periodic evaluation of: role of MOP in the overall monitoring . Progress towards final outcomes. system. The mechanism for cross- . Reasons for success or failure. government coordination established to . Reasons for differences. prepare the NSDP (the NSDP Secretariat) . Likely sustainability of trends. needs to be institutionalized and By and for: strengthened to ensure that the NSDP Government actually takes root in implementation as a . Civil society . whole-of-government plan that exercises a Donors . significant and positive effect in the To support: allocation of resources between and within Outcome-oriented poverty-focused policy making, planning and resource Ministries. Due to time pressures, the allocation. partnership body that was intended to provide broad civil society, donor and private sector input to NSDP formulation includes the NIS, the statistical units of (the TWG on Planning and Poverty line ministries, and the National Bank Reduction) did not have the opportunity to ofCambodia). perform this function to the full. Agreement on the use of a common Determined effort will be needed to ensure tool (CamInfo) as the “front-end” that it can in 2006 develop a real role in repository ofsocio-economic data from monitoring and supporting NSDP all sources (survey and non-survey), implementation. accessible to all users. Second, there will be a need to The M&E principles and outline differentiate between Ministries in terms of indicator and target framework the current strength of their capacity for included in the draft NSDP. This collecting and using performance data and proposes a substantive annual review thus the forms of M&E capacity-building of progress in NSDP implementation they require. The more advanced and outcomes, to guide refinements of Ministries of Education and Health, for the policies and programs needed to example, should focus on building capacity achieve the 2010 and 2015 CMDG for program budgeting and costing, while targets. This implies a substantial those Ministries which are at an earlier increase in the level of demand for stage of institutional and policy timely, reliable and accurate data on development (such as Agriculture, Rural inputs, outputs and outcomes. Development, or Water Resources) need first to their capacity for articulating Agreement amongst an informal core build policy, including the ability to define group of donors, soon to be expanded, monitorable policy outcomes. Although on ways to coordinate their support the priority for NSDP M&E is national level poverty/human development New partnerships for poverty reduction 173 outcomes, it is important also that management. To ensure national rather programs build in M&E systems with than merely narrow government ownership strong results frameworks (logframes). of NSDP implementation and monitoring, Monitoring at the program level helps with these skills need to be built both within the the task of policy analysis-that is, Government and within independent identifying the contributions towards national academic, research and advocacy ultimate outcomes made by different institutions. policies and programs, and how this Sixth, there is a need to prioritize and impact can be improved-and helps to schedule data collection exercises (and the embed the principle of oversight and social capacity building required for these) accountability. according to the timing of data needs for Third, there is a need to make it easier to NSDP implementation and revision. The obtain access to official statistics (both preparation of the NSDP was somewhat those collected by the NIS and those hampered by the fact that core socio- generated by the management information economic data from the 2004 CSES only systems (MIS) of line ministries). The became available in the middle of 2005. CamInfo software package has been This allowed only six months for this data developed over a number of years as a to be analyzed and interpreted and made user-friendly means of storing and available as an input to the team drafting retrieving the findings of both surveys and the NSDP. The findings of the DHS MIS-generated administrative data. (which, as discussed in Chapter 6, are far However, CamInfo has depended heavily superior to the DHS for a number of upon the support ofUNICEF and UNFPA. critical human development/CMDG NIS needs to build up its capacity to indicators) will not be available until the maintain and upgrade the database and second quarter of 2006, once the NSDP institutionalize its position as the central has been finalized. Ideally, the preparation repository ofdata, available to all users. of the next (2011-2015) Five-Year Plan will have this information-including the Fourth, access to data needs to be updated consumption poverty estimates- complemented by demand from available at the start of the preparation statistically literate users, beginning with period, so that the planners can judge those within the Government who should progress made, identify aspects of be basing policy, program, planning and economic or social development where budget decisions on analysis and evidence. progress is falling behind target, and Ministry officials should take full attempt to find explanations and policy advantage of the training in the use of response to these problems. This will statistics that have been developed by the probably necessitate bringing the full UN and other agencies. (large-sample) CSES and the DHS forward Fifth, these efforts to build basic statistical a year, so that the surveys can be literacy (that is, to build skills in how to conducted in 2008, the data entered, interpret descriptive statistics) need to be cleaned and analyzed by mid-2009, the reinforced with efforts to build national analysis interpreted and policy conclusions capacity in the more technically advanced drawn by end 2009, and this poverty skills of statistical analysis for problem assessment made available to the strategy identification, policy impact assessment drafting team for them to use over the and performance-based public expenditure course of2010. 174 Cambodia - Poverty Assessment 2006

Seventh, there is a need to build links between the collection and analysis of official data on the one hand and civil society and private sector users on the other. This includes institutionalizing efforts to actively disseminate key data to the National Assembly, NGOs, business associations and so on; but it also means establishing dialogue between the official statistical system and initiatives for community-based and independent progress monitoring. Finally, it is necessary to ensure full funding for the development of core statistical capacity and specific key data collection processes as laid out in the SMP, so that the Government (and its partners in civil society, the private sector and the donor community) can be assured of a regular flow of reliable socio-economic data. Such data will be essential for the performance-based policy process anticipated in the NSDP. Timely and reliable progress indicators are essential to improving the partnership between the Government and its development partners. If donors are to align to the NSDP and move towards budget support, it is critical that all parties feel that they have information on policy implementation and outcomes. If they do not have this information, they are likely to continue to impose requirements for donor-specific M&E and reporting (at best, harmonized between donors, at worst specific to each individual donor). While the donors will realistically have to continue to provide the majority of the financing for government data collection, analysis and dissemination, it would be an important contribution to trust-building (and one incorporated in the SMP) if the Government were to allocate a steadily rising sum towards at least the core costs of the NIS and the rest of the statistical system. References Abassi, Zia. 2005. The contribution of Ballard, Brett and So Sovannarith. 2004. tourism to gross domestic product. Cambodia land titling program: baseline Draft. PhnomPenh. IMF. survey project. Final Report: Rural Phase 1. CDRI in collaboration with ActionAid International. 2005. REALAID: MLMUPC. Phnom Penh. July 2004. An Agenda for Making Aid Work. Johannesburg. June 2005. Barnett, Anthony. 1983. “Cambodia: a highly centralized dictatorship.” in David ADB. 2002. Indigenous peoples/ethnic Chandler and Ben Kieman (eds.). 1983. minorities and poverty reduction- Revolution and its aftermath in Cambodia. Environment and Social Kampuchea: eight essays. New Haven: Safeguards Division, Regional and Yale University Press. 212-229. Sustainable Development Department. Manila. June 2002. Biddulph, Robin. 1996. Participatory development in authoritarian societies: AD1 / CCC. 2005. Impact of the garment the case of Village Development industry on rural livelihoods: lessons Committees in two villages in Banteay from Prey Veng garment workers and Meanchey Province, Cambodia. rural households. Analyzing Unpublished Masters of Development Development Issues team and research Administration thesis. National Centre participants. CCC. Phnom Penh. of Development Studies, Australian October 2005. National University. AFSC, UNICEF and IDRC. 2000. Choices Birdsall, Nancy. 2004. Seven deadly sins: from the periphery: vulnerable people’s reflections on donor failings. Center for responses to adversity and change. Global Development Working Paper No. Socio-Cultural Vulnerability and Coping 50. December 2004. Strategies (SCVS) Research Project http ://www .cgdev.org/content/publicatio Report No. 1. Phnom Penh. April 2000. ns/detai1/2737. Amvind, Vinod. 2005. Social relations Brereton, Helen. 2005. Draft Report: and local level power structures in Evaluation of the Seila Gender Cambodia: a review of the literature. Mainstreaming Strategy 2001 -2005 Unpublished discussion paper: World (Phnom Penh, Seila). Bank Cambodia Country Office. Burgler, R.A. 1990. The eyes of the Aschmoneit, Walter 1998 Traditional sew- pineapple: revolutionary intellectuals help associations in Cambodia: how to and terror in Democratic Kampuchea. identi& and cooperate with them. GTZ: Saarbriicken, Fort Lauderdale: Verlag Phnom Penh. breitenbach Publishers. 1990. Ballard, Brett. 2005. “Linking tourism to Burke, Adam and Nil Vanna. 2005. poverty reduction: a Siem Reap case Options for increasing social study.” CDRI. Phnom Penh. Annual accountability in Cambodia. Social Development Review 2004-05. 78- 101. Development Direct. 15 February, 2005. December 1995. 176 Cambodia - Poverty Assessment 2006

Castellanet, C. 2003. A review of Conway, Tim. 1999. Poverty, agricultural extension sewices in participation and programmes: Cambodia. Consultant report for the international aid and rural development World Bank. October 2003. in Cambodia. Unpublished PhD dissertation. St Catharine's College / CDC. 2005. National Operational Department of Geography, University of Guidelinesfor Development Cooperation Cambridge. Grant Assistance. CDC. July 2005 Conway, Tim and Andrea Crossland. CDRI. 2006. Moving out of poverty: 2002. Review of mechanisms to improve Cambodia national synthesis report. equity in access to health care, (Draft). CDRI. Phnom Penh. January Cambodia. DFID Health Systems 2006. Resource Centre. London. July 16th CDRI. 2006a forthcoming. Moving out of 2002. poverty in Cambodia. Phnom Penh. http://www.odi.org.uk/PPPG/publication CDRI. 2006b forthcoming. Participatovy s%5Cpapers_reports%5Cdfid%SCcountr Poverty Assessments (Phnom Penh, iesY05Ccam02.pdf CDRI/ADB) Crosbie, David. 2004. Understanding Chandler, David. 1991. The tragedy Asian Cities: Phnom Penh. URC, of Cambodian histoy: politics, war and Ministry of Land Management and revolution since 1945. New Haven and Urban Planning and Construction. London: Yale University Press. Phnom Penh Christensen, Anne and Thomas CSD. 2005. Corruption and Cambodian Markussen. 2005. A proJile of remittance households: household survey on recipients in Cambodia. Mimeo. World perceptions, attitudes and impact of Bank. everyday forms of corrupt practices in Cambodia. CSD. Phnom Penh. March Colletta, Nat J. and Michelle L. Cullen. 2005. 2000. Violent conflict and the transformation of social capital: lessons Dahlberg, Elisabet. 2005. Insights into from Cambodia, Rwanda, Guatemala, migration and spending patterns based and Somalia. World Bank. Conflict on a small-scale study of garment Prevention and Post-Conflict workers in Phnom Penh. Working Paper Reconstruction Series. Washington, 221. Stockholm School of Economics. D.C. December 2005. Collins, William. 1998. Grassroots civil Daubert, Pierre. 1996. Is there a place for society in Cambodia: a discussion paper collective organization in the rural prepared for a workshop organized by society of Cam,bodia? GRET. Phnom Forum Syd and Diakonia in September Penh. Mimeo. March 1996. Based on 1998. Phnom Penh. November 1998. Pasquier-Desvignes, Rodolphe, Didier Reynaud, Dara Rat Moni Ung, Dider Commission for Africa. 2005. Our Pillot and Phillipe Rousseau. 1995. common interest: report of the Entre mise'rabilisme et populisme: la Commission for Africa. March 2005. voie e'troite du de'velopement au http://www.commissionforafrica,org/eng Cambodge. lish/report/thereportt/english/l 1-03 - 05-cr-report.pdf References 177

Deaton, Angus. 2001. “Counting the Outreach. World Bank Institute. world’s poor: problems and possible September 2005. solutions.” World Bank Research Ebihara, May. 1968. Svay, a Kher village Obsewer. 16 (2). 125-147. Fall 2001 in Cambodia. Unpublished PhD http ://wbro.ox fordj ournals.org/cgi/reprint dissertation. Columbia University. 11 6/211 2 5 .pdf Available from University Microfilms, Deaton, Angus and Margaret Grosh. 2000. Ann Arbor. and “Consumption.” Margaret Grosh EIC. 2005. Cambodia Economic Watch Paul Glewwe (eds). 2000. Designing 2005 - Issue 3 (October 2005). EIC. household survey questionnaires for Phnom Penh. developing countries: lessons from fifteen years of the Living Standards EIC. 2005. Cambodia competitiveness Measurement Study. Washington DC. report 2005-06. EIC with support from World Bank 1 OW. USAID, TAF and WEF. Phnom Penh. September 2005. Delvert, Jean. 1958. 1961. Le paysan Cambodgien. Paris, The Hague: Mouton ESRF. 2005. Enhancing aid relationships and Co. in Tanzania: IMG Report 2005. Economic and Social Research Deninger, Klaus 2005. “Quantifying the Foundation. Draft Final Report of the poverty impacts of land policy Independent Monitoring Group. intervention: The case of Cambodia,” mimeo, the World Bank. Falck, Hans, Khre Landfeld with Pamela Rebelo. 2001. “Chapter 7: the poverty Development Consulting International. reduction strategy process in 2003. Private Sector Assessment for the Mozambique.” ODI. 2001. PRSP Kingdom of Cambodia. TA Report for Institutionalisation Study: final report. the Asian Development Bank, TA No. London. 4030-CAM (December 2003). Financial Times, The. 1995. “Cracks Development Weekly. 2005. Carol behind the prosperous veneer.” Jonathon Strickler: ‘(NGO coordination is very Miller: 27th February 1995. dijjficult”. Phnom Penh. December 5- 11,2005. Findlay, Trevor. 1995. Cambodia: the lessons and legacy of UNTAC. SIPRI DFID. 2005. Review of Asian Research Report No. 9: Oxford Development Bank-DFID- World Bank- University Press: Oxford, New York. UN joint planning process, Cambodia 2002-5. Phnom Penh. August 2005. Foredan, Graham. 2005. Wise Before Their Time: Young People, Gender-Based Dixon, Chris. 1995. “Origins, Violence and Pornigrpahy in Kandal sustainability and lessons from Strung District. (Phnom Penh, World Thailand’s economic growth.” Vision) Contemporary Southeast Asia 17 (1). 38-52. June 1995. Foster, Mick and Peter Mijumbi. 2002. How, when and why does poverty get Diogo, Luisa. 2005. “Post-conflict budget priority? Poverty reduction and Mozambique’s reconstruction: a public expenditure in Uganda. OD1 WP transferable strategy.” Development 163. London 178 Cambodia - Poverty Assessment 2006

Frieson, Kate. 1996. “The Cambodian Implementation. 25 June 2003. Phnom elections of 1993: a case of power to the Penh. people?” in R.H. Taylor (ed.). 1996. The Harrison, Makiko. 2005. Cambodia politics of elections in southeast Asia. Poverty Assessment: Poverty Monitoring Cambridge. Cambridge University Press. System in Cambodia. Draft. 18 August 224-242. 2005. Frings, Viviane. 1993. The failure of Hasselskog, Malin with Krong Chanthou agricultural collectivisation in the and Chim Charya. 2000. Local People’s Republic of Kampuchea (1979- governance in transition: villagers ’ 1989). Centre for Southeast Asian perceptions and Seila j. impact. Studies WP 80. Monash University. CARERE: Phnom Penh. Fujii, Tomoki. 2005. Poor farmers in Hausman, R., D. Rodrik, and A. Velasco, Cambodia: An exploratory study. Mimeo. “Growth Diagnostics”; mimeo. World Bank. September 2004. Available at: Fukuyama, Francis. 2004. State-building: (http ://ksghome.harvard.edu/-drodrikhar governance and world order in the 21Sf celonaoctl9.pdf). century. Ithaca. Cornell University Hendrikson, Dylan. (n.d.) “Cambodia’s Press. security-sector reforms: the limits of a GDCC. 2005. Progress Report of the down-sizing strategy.” Conflict, Security Jooint Technical Working Groups For and Development 1 (1): 67-82. the Period: December 2004 - May 2005. Hubbard, Michael. 2002. Cambodia-a GDCC. country case study. OECD-DAC Studies Godfrey, Martin, Chan Sophal, Toshiyasu on Donor Practices. University of Kato, Long Vou Piseth, Pon Dorina, Tep Birmingham. November 2002. Saravy, Tia Savora and So Sovannarith. http://www.oecd.org/dataoecd/4 111 91809 2002. “Technical assistance and 3042.pdf capacity development in an aid- Hughes, Caroline. 2003. The political dependent economy: the experience of economy of Cambodia’s transition, Cambodia.” World Development. 30 1991-2001. London and New York. (3). 355-373. RoutledgeCurzon. Gorman, Siobhan with Dorina, Pon and IMF. 2003. Staff report for the 2002 Kheng, Sok. 1999. Gender and consultation with Uganda. 29 January Development in Cambodia: An 2003. Overview. CDRI: Phnom Penh. June 1999. Indochina Research. 2005. Violence against women-a baseline survey. Goldin, Ian. 2005. The role of the media Indochina Research for MOWA, with in development. Speech by Ian Goldin, support from GTZ, EWMI and UNIFEM Vice-president, World Bank Group, in CEDAW SE Asia Program. Phnom Penh, October 2005. Keller, Sheryl and J. Brad Schwartz. 2001. Hageman, Robert. 2003. The role offiscal Final evaluation report: contracting for policy in reducing poverty. Presentation health services pilot project (CHSPP) - to the Provincial Workshop on NPRS a component of the Basic Health References 179

Sewices Project. Phnom Penh. Markussen, Thomas. 2005b. A note on November 2001. farm size and agricultural productivity in Cambodia. Mimeo. World Bank. Kiernan, Ben. 1996. The Pol Pot regime: race, power, and genocide in Cambodia McAndrew, John P. and Oeur 11. 2004. under the Khmer Rouge, 1975-79. New Upholding indigenous access to natural Haven and London: Yale University resources in northeast Cambodia. Paper Press. read.at the ADB Regional Workshop on Indigenous Peoples and Communal Land Kiernan, Ben and Boua Chanthou (eds.). Management, Manila, 8-9 December 1982. Peasants and politics in 2004. CCC / ADI. Phnom Penh. Kampuchea, 1942-81. New Haven. Yale University Southeast Asian Studies MEF. 2004. Volume One: the longer term Monograph Series No. 41. PFM Strategic Vision in the policy context and the plan for Stage One Killick, Tony. 2002. Responding to implementation. PFMRP Strengthening inequality. OD1 Inequality Briefing 3 governance in Cambodia through (March 2002). Overseas Development enhanced public financial management. Institute. London. RGC. Phnom Penh. December 2004. http://www.odi.org.uk/PPPG/publication s/briefings/inequality briefinad03 .pdf MEF. 2005. Public Financial Killick, Tony, Carlos N. Castel-Branco and Management Reform Program: summary report for period Is‘February-31S’ July, Richard Gerster. 2005. Perfect 2005. partners? The performance of Public Financial Management Programme Aid Partners in Reform Committee. Mimeo (final draft). Mozambique, 2004. A report to the Metzl, Jamie Frederic. 1995. “The many Programme Aid Partners and faces of UNTAC: a review article.” Government of Mozambique. May Contemporary Southeast Asia 17 (1). 2005. 85-96. Lao PDR. 2003. National growth and Ministry of Interior. 2005. Decision on the poverty eradication strategy. Establishment of Provincial/Municipal Vientianne. CS-Fund Accountability Working Group. http ://www.undplao .org/Ngpes/Lao%20 (Phnom Penh, RGC) PDR%20-%20NGPES%20- Ministry of Planning. 2003. Cambodia %20Main%20Document .pdf Millennium Development Goals Report Ledgerwood Judy. n.d. Cambodian Recent 2003. Prepared with funding from History and contemporary Society: An UNDP and Sida. Phnom Penh. Introductory Course. Women in November 2003. Cambodian Society. Northern Illinois http ://www.un.org.kh/publications/mdg/c University. mdg-en-full.pdf http://www.seasite.niu.edu/khmer/ledger Ministry of Planning. 2005. Achieving the wood/women.htm. Retrieved on: Cambodia Millennium Development December 2,2005. Goals: 2005 update. Prepared with Markussen, Thomas. 2005a. Property funding from UNDP and Sida. Phnom rights and productivity in Cambodian Penh. October 2005. crop agriculture. Mimeo. World Bank. 180 Cambodia - Poverty Assessment 2006

Ministry of Rural Development. 2004. Neupert, Ricardo F. 2005. New ADB Loan 1862-CAM (SF). demographic estimates and updated Northwestern rural Development projections for Cambodia. 17th April Project, Briefing Note No. 4 - Commune 2005. Council Fund Allocation and NGO Forum. 2004. NGO sectoral and Distribution (Phnom Penh, ADB) issues papers on povrty reduction and Minority Rights Group International. development in Cambodia 2003. Phnom 1995. Minorities in Cambodia. London. Penh. March 2004. MOP. 1998. Cambodia Human NGO Forum. 2005. Consultation on the Development Report 1998. (Phnom Draft NSDP in Battambong Province. Penh, RGC) (Phnom Penh, NGO Forum on Cambodia) MOP. 2005. Statistical Master Plan: Draft. NIS 6 August 2005. NGO Forum. 2005. Rethinking poverty reduction to protect and promote the MOWA. 2005. Violence Against Women: rights of indigenous minorities in A baseline survey. MOWA, Phnom Penh. Cambodia. NGO Forum, Phnom Penh. MOWA. 2006 forthcoming. Gender Ninh, Kim and Roger Henke. 2005. Equality in the Civil Service. (Phnom Commune Councils in Cambodia: A Penh) national survey on their functions and Murshid, KA and Phim Runsinarith. 2005. performance, with a special focus on A review ofpoverty and vulnerability in conflict resolution. The Asia Foundation Cambodia (with special reference to the in collaboration with the Center for Tonle Sap Region). CDRI. Phnom Penh. Advanced Studies. May 2005. Mysliwiec, Eva L. 1994. “Cambodia: NGOs Nissen, Christine J. 2005. Living under the in transition.” Peter Utting (ed.). 1994. rule of corruption: an analysis of Mysliwiec 2004 Envisioning a new eveyday forms of corrupt practices in paradigm for development cooperation Cambodia. CSD. Phnom Penh. March in Cambodia. CDRI. February 2004. 2005. OECD. 2003. Harmonising donor practices for effective aid delivery. DAC Naschold, Felix. 2002. why inequality Guidelines and Reference Series: Good matters for poverty. OD1 Inequality Practice Papers. Paris. Briefing 2 (March 2002). Overseas Development Institute. London. OECD 2005 Survey on harmonisation and http://www.odi.orguWPPPG/Dublication alignment: measuring aid haronisation s/briefings/inequality briefings/02.pdf and alignment in 14 partner countries. OECD. Paris. Nelson , Erin and Zimmerman, Cathy. http://www.oecd.org/dataoecd/3 113 713 3 9 1996. Household Survey on Domestic 81948.pdf or see chapter four on Violence in Cambodia (Phnom Penh: Cambodia at Project Against Domestic Violence http://www.oecd.org/dataoecd/4 115 5/33 8 [PADV] and Ministry of Women’s 13377.pdf Affairs). Oum Sothea. 2004. “Cambodia’s national budget implementation in 2004 and References 181

prospects for 2005”. EIC Economic Strengthening governance in Cambodia Review, l(6). October-December 2004. through enhanced public financial 13-15. management. MEF. December 2004. Oversen, Jan, Trankell Ing-Britt and RGC. 2004. Development Cooperation Joakim Ojendal. 1996. When every Report 2002 and 2003: Main Report. household is an island: social CDC. October 2004. organization and power structures in RGC. 2004. The Rectangular Strategy for rural Cambodia. Research Report in growth, employment and equity. Cultural Anthropology No. 15. Department of Cultural Anthropology, RGC. 2005a. Cambodia National Poverty Uppsala University / Sida. Reduction Strategy - progress report (updated 15 July 2005). SNEC, MEF. Oxfam GB. 2005. Briefing background. Phnom Penh. Draft analysis of patterns and trends in land disputes. April 2005. RGC. 2005b. Strategic Framework for Development Cooperation Management. PACT and PADCO. 2005. Fighting RGC. August 2005 corruption in Cambodia: the demand for an international standard anti- RGC. 200%. Discussion Draft: National corruption law. PACT. Phnom Penh. Strategic Development Plan, 2006-201 0 July 2005. (NSDP). 10 October 2005 Phnom Penh Post. 2005a. The roles and Sen, Amartya. 2000. Development as responsibilities of the Khmer media. freedom. Anchor Books. Moeun Chhean Narridh. 14 (23). Sloth, Christian, Heov Kim Sreng and November 19-December 1 2005. Khlok Bottra. 2005. “Natural resources Phnom Penh Post. 2005b. Free press feels and environment: issues, constraints and the pinch. Charles McDermid. 14 (25). challenges.” CDRI. Phnom Penh. December 16-29 2005. Annual Development Review 2004-05. 101-128. December 1995. Piron, Laure-Helhe with Andy Norton. 2004. Politics and the PRSP Approach - SNEC with support from World Bank, Uganda Case Study. OD1 WP 240. AusAID and FAO. 2005. Natonal London. forum on Cambodia ’s agrarian structure http://www,odi.org.uk/publications/work in the context of trade integration, June ingqapers/wp240.pdf 28-9 2005. Proceeding Report. Phnom Penh. Poverty Task Force. 2003. Poverty: Vietnam Development Report 2004. SUPF. 2003. The poor in Phnom Penh. Joint donor report to the Vietnam SUPF. Phnom Penh Consultative Group Meeting. Hanoi. Toda, Hirohito. 200 1. “The planning December 2-3 2003. process in Cambodia.” In ESCAP. 2001. RGC. 2002. The Triangle strategy of the Development planning in a market Second Mandate. economy. Economic and Social Commission for Asia and the Pacific. RGC. 2004. Volume I: the longer-term Least Developed Countries Series No. 6. PFM Strategic Vision in the policy ST/ESCAP/2174. New York and context and the plan for stage I Bangkok. United Nations. implementation. PFMPR - 182 Cambodia - Poverty Assessment 2006

http://www.unescap.org/drpad/publicatio UNIFEM, WB, ADB, UNDP, DFID/UK. n/toc/ldc6-2 174.htm Phnom Penh, April 2004. Uchida, Emi and Hirotsugu Uchida. 2005a. UNOHCHR. 2004. Community timeline Depending on forest resources for and focus group discussion on freedom. livelihood: Cambodia's rural poor. Pilot of MOPS methods in an urban Mimeo. World Bank. context, Tuomnoup Teuk Khan, Chamcannon, Phnom Penh. July 22004. Uchida, Hirotsugu and Emi Uchida. 2005b. Aquatic resources and Cambodian Urashima, Cheryl. 2005. Cambodia households. Mimeo. World Bank. Gender Brie$ (Phnom Penh, MOWA) UN Department of Public Information. Urashima, Cheryl. 2005. Public jinancial 1995. The United Nations and management reform in Cambodia...from Cambodia: 1991-1 995. New York: Blue a gender perspective. Project on Book Series Vol 11. Gender-Sensitive Monitoring, Planning and Budgeting for Implementation of the UNDP. 2005. Access to Justice in Poverty Reduction Strategy. Cambodia with focus on the poor, women and indigenous peoples. (Phnom URC. 2004. Secure tenure and poor Penh) communities in Phnom Penh: Pilot Project 2004. URC. Phnom Penh UNDP. 2005. Human Development Report 2005. International Cooperation at the Urdal, Henrik. 2002. The devil in the Cross Roads: Aid, trade and security in demographics: how youth bulges an unequal world. UNDP. New York. influence the risk of domestic armed conflict, 1950-2000. Paper presented to UNDP. 2006. Review of development the International Studies Association 43'd prospects and options for the Cambodian Annual Convention. New Orleans. 24-7 oil and gas sector. A UNDP-funded March 2003. discussion paper in collaboration with the Cambodian National Petroleum USG. 2005. Nurturing and Enhancing Authority and Harvard's Kennedy Urban Poor Initiatives and Development School. Presented at the Cambodia Processes. USG, Phnom Penh. Economic Forum, 17th January 2006. Utting, Peter (ed.). 1994. Between hope UNDP, OECD, World Bank, ADB, and insecurity: the social consequences African Development Bank, European of the Cambodian peace process. Bank for Reconstruction and UNRISD. Geneva. Development and Inter-American Vallely, Thomas, David Dapice and Ben Development Bank. 2005. Wilkinson. 2006. A SWOT analysis of Harmonisation, Alignment, Results: Cambodia: Executive Summa y. Report on Progress, Challenges and Harvard University Kennedy School of Opportunities. Paris. Government, for UNDP Cambodia. UNIFEM. 2006 forthcoming. Power Vickery, Michael. 1984. Cambodia: 1975- Relations in Contemporay Cambodian 1982. Boston. South End Press. Marriage (Phnom Penh) Village Focus International for NGO UNIFEM, WB, ADB, UNDP and Forum. 2006 forthcoming. PRSP-NPRS DFID/UK. 2004. A Fair Share For Monitoring: Field Research Results. ' Women: Cambodia Gender Assessment. References 183

(Phnom Penh, NGO Forum on climate assessment and reform strategy. Cambodia) Prepared for the Royal Cambodian Government by the World Bank Group Vlaar, Sonja and Ahlers, Rhodante. 1998. (IBRD, IFC, MPDF, PPIAF). August “Gender-blind or Gender-bright 2004. Targeting of Projects in Cambodia”. Edited by Irene Guijt and Meera Kaul World Bank. 2004. Cambodia at the Shah. The Myth of Community. crossroads: strengthening accountability Intermediate Technology Publications. to reduce poverty. Prepared by the World Bank in collaboration with the Warner, Michael. 2006. Oil discoveries: IMF for the Cambodia Consultative blessing or curse. International Group Meeting. Phnom Penh, December experience and lessons learned, 6-7 2004. Presentation at Cambodia Economic Forum, 17th January, 2006. World Bank. 2005. Tajikistan: Poverty http://www.odi.org.ukPPPG/activities/c Assessment Update. Report No: 30853- ountry-level/odpci/msp/InternationalExp TJ, Washington DC, (March). erienceswithResourceBoomsqresentatio World Bank, OECD/DAC. 2005. Baseline n-files/frame.htm Survey Paris Declaration on Aid Wilkinson, David, John Holloway and Effectiveness: Quick Survey for Pierre Fallavier. 2001. The impact of Government and Local Donor Agencies. user fees on access, equity and health World Bank. 2005. Global economic provider practices in Cambodia. Health prospects 2005. Economics Task Force, MoH / WHO Health Sector Reform Phase I11 Project. World Bank. 2005. World Development Phnom Penh. Report 2006: Equity and development. OW/ World Bank. World Bank. 1999. Cambodia Poverty Assessment. Poverty Reduction and World Bank. 2005. Cambodia: social Economic Management Sector Unit. accountability and community November 22,1999. empowerment. Presenation: December 2005. World Bank. 2000. Cambodia County Assistance Strategy. World Bank. World Bank. 2005. Cambodia ’s Public Washington DC. Financial Management Reform: towards a sustainable, high impact approach. World Bank. 2001. Building institutions PEFA Case Study: mimeo (draft under for markets: World Development Report review). 2002. New York. OW. World Bank and ADB. 2003. Cambodia: World Bank. 2003. Nicaragua’s Poverty enhancing service delivey through Assessment - Raising welfare and improved resource allocation and reducing vulnerability. Report No. institutional reform. (Integrated 26128-NI. Washington DC. (December). Fiduciay Assessment and Public World Bank. 2003. Making services work Expenditure Review) World Bank for poor people: World Development Poverty Reduction and Economic Report 2004. New York. OW. Management Sector Unit, East Asia and the Pacific Region; ADB Mekong World Bank. 2004. Cambodia-seizing the global opportunity: investment 184 Cambodia - Poverty Assessment 2006

Department, Cambodia Resident Mission. World Bank and IFC. 2006. Doing business in 2006: creating jobs. Washington DC. World Bank Institute. 2002. The right to tell: the role of mass media in economic development. World Bank. Washington D.C. WHO. 2002. World Report on Violence and Health. (Geneva). Yagura, Kenjiro. 2005. “Why illness cases more serious economic damage than crop failure in rural Cambodia” Development and Change 36(4): 759-783. Zimmerman, Cathy. 1996. Divorce and Domestic Violence in Cambodia (Phnom Penh: PADV). Zimmerman, Cathy. 1995. Plates in a basket will rattle: domestic violence in Cambodia. Asia Foundation and UNDP. Phnom Penh Annex I: Technical Annex Survey Coverage One major problem in preparing a Design of the 2004 CSES comparable series of poverty estimates The 2004 CSES sample was selected over time in Cambodia has been the from 45 strata (24 provinces, urban and changing (increasing) geographical rural)’ in three steps using the 1998 coverage of the various household Population Census as a sampling frame. surveys over time. Coverage was First, 900 villages were selected from initially quite low in the 1993/94 SESC the various strata using systematic due to security reasons, particularly in random sampling (with over-sampling in the Rural areas (Le., only 56% of Rural the urban strata). Second, one Census villages and 65% of Rural households). enumeration area was randomly selected Table 1 presents the regional and from each sample village. Third, 10 (20) province distributions of the comparable households were selected randomly from samples from the three surveys, each urban (rural) sample enumeration weighted by the estimated populations at area, yielding a total sample of 15,000 the time of each survey. households (of which 14,984 were actually interviewed). One thousand households were interviewed in each month of the survey in a randomly selected (and therefore nationally representative) sample of 60 villages. The 2004 CSES is not self-weighting. Two sets of adjusted sample design weights2 are provided, one for use with the calendar year 2004 sample of 12,000 households (of which 1 1,993 households were actually interviewed) and the other for use with the full sample of 15,000 households (of which 14, 984 households were actually interviewed). Unless otherwise noted, all estimates presented in this report are based on the calendar year 2004 sample of 11,993 households actually interviewed and are weighted to be representative of the Cambodian population.

Three of Cambodia’s 24 provinces do not have any rural areas (Le., Kep, Sihanoukville and Pailin). The sample design weights are equal to the inverse of the probability of each household’s selection and are based on the population structure (province, urbadrural, age and sex) of the 1998 Population Census 186 Cambodia - Poverty Assessment 2006

Table 1. Distribution (YO)of comparable samples by province, 1993/94-2004 1993194 1997 2004 Region Phnom Penh 10.74 14.45 13.81 Other Urban 1 1.02 12.29 12.79 Rural 78.25 73.26 73.40 Cambodia 100.00 100.00 100.00 Province Banteay Meanchey 3.61 3.86 2.49 Battambang 4.68 6.07 4.36 Kampong Cham 18.85 16.34 17.03 Kampong Chhnang 2.48 2.64 3.14 Kampong Speu 2.16 1.66 2.41 Kampong Thom 0.92 0.84 0.83 Kampot 2.64 2.15 2.51 Kandal 13.19 13.10 13.85 Koh Kong - - - Kratie - - - Mondul Kiri - - - Phnom Penh 0.74 14.45 13.81 Preah Vihear 0.00 0.00 0.00 Prey Veng 5.32 14.89 13.64 Pursat 3.81 4.12 4.18 Ratanak Kiri - - - Siem Reap 2.96 2.18 3.29 Sihanoukville 0.72 0.86 1.91 Stung Treng - - - Svay Rieng 7.08 6.38 6.60 Takeo 10.84 10.48 9.95 Oddor Meanchey - - - KeP - - - Pailin - - - Cambodia 100.00 100.00 100.00 N 5,578 4,470 7,596 Source: 1993/94 SESC, 1997 CSES, 2004 CSES.

provided in the CSES questionnaires). Household consumption This information is presented below in The detailed consumption categories connection with the description of the used in the 1993/94 SESC serve as the procedures used to estimate consumption reference point for defining the various in each broad category. categories of household consumption presented in Table 2 as these 1993/94 SESC categories were also used to define the various CSES recall categories (as indicated by the capsule descriptions of each category that are Annex 1: Technical Appendix 187

Table 2: Definition of house1 ild consumption based m Source of information Period Number ofdays Household questionnaire, Last 7 days 7 Section 1D: questions 1- 16, 18-20 v.Clothing & footwear Household questionnaire, Last 6 months 182.5 Section 7B, question 1 I 3. Housing & utilities 3.1 Housin I 1

Household questionnaire, Last month 30.4 Section 3, question 28 Estimated rent Household questionnaire, Month 30.4 Section 8, qiestion 6 ~ I Hoteliaccommodation Diary, items 9407, 9415 Calendar month Number ofdays in calendar month Household questionnaire, Last month 30.4 Section 3. auestion 29 3.2 Utilities 3.2.1 Water & sanitation

Household questionnaire, Last month 30.4 Section 3. auestion 17 Sewageiwaste water charges Household question, Last month 30.4 Section 3, question 21 Household question, Last month 30.4 Section 3, question 22 I Household question, Last month 30.4 Section 3, question 24a Gas Household question, Last month Section 3, question 24b Kerosene Household question, Last month Section 3, question 24c Battery Household question, Last month Section 3, question 24f 3.2.2.2 Wood fuels Household question, Last month 30.4 Section 3, question 24d Charcoal Household question, Last month Section 3, question 24e Household question, Last month Section 3, question 24g 1 4. Household furnishings Household questionnaire, Last 6 months 182.5 Section 7B, question 2 Household questionnaire, Past 4 weeks 28 Section 14, question 12 (only for those reported

6. Transportation and Diary, items 7 11 1-808 1 Calendar month Number of days in communications calendar month 7. Recreation Household questionnaire, Last 12 months 365 tSection 7B. auestion 3 188 Cambodia - Povertv Assessment 2006

Source: 2004 CSES questionnaires.

This report presents estimates of The 2004 CSES collected data on household consumption for two of household consumption from 14,984 Cambodia’s multi-objective household households using two distinct surveys, the 1993/94 SESC, and the methodologies, i.e., a set of recall 2004 CSES. Data on household questions similar to those used in the consumption were collected in the 1997 and 1999 CSES and a calendar- 1993/94 SESC for a sample of 5,578 month diary that was completed by households (from a geographically respondents with the assistance of truncated sample) and are generally interviewers who remained in each regarded to be at least reasonably sample village for one full month. The accurate. The 1993/94 SESC household consumption estimates presented in this questionnaire collected recall data for a report are prepared with data collected large number of consumption items (177 during calendar year 2004 and are food items and 266 nonfood items). The prepared for the most part using the 1993/94 SESC was also conducted in recall data. However, diary data were four rounds during a 12-month period used to estimate consumption for two (i.e., from October 1993 through categories (transportation and September 1994). Accordingly, the communications and personal care) estimates of average household because no recall data were collected for consumption it provides are presumably these two consumption ~ategories.~No not affected by seasonal variation in adjustments were made to the reported either consumption quantities or prices. data apart from imputation of an The 1993/94 estimates of household estimated rental value for owner- consumption presented in this report occupied housing for the 5.9% of sample were prepared using the same data file households that did not report one and used by the authors of the 1993/94 for the 2.4% of households that reported Poverty Profile, as reported in that document (Prescott and Pradhan 1997).3 Since the reference period for these items was “the last month” in previous CSES rounds, it was presumably decided that there was no need to The data file is a Stata file with the name collect recall data for the same reference period “reqsma1l.dta”. as the diary. 1 a zero valuee5The consumption data in persons of all ages and both ~exes).~The the 2004 CSES are most comparable to reference food bundle was designed to those in the 1993194 SESC inasmuch as reflect the actual food consumption both surveys collected consumption data patterns of Cambodians who consumed a for at least a 12-month period. However, diet yielding approximately 2,100 the consumption data in the 2004 CSES calories per day. It is based on the are not comparable to the 1993/94 SESC quantities of different foods consumed consumption data in at least one respect, by persons in the middle per capita i.e., recall data were collected for many consumption quintile, as this was the fewer consumption categories in the first quintile that met the 2,100 calorie 2004 CSES than in the 1993/94 SESC minimum. A single reference food (20 food categories versus 177 in the bundle was used for all Cambodians. 1993/94 SESC and 13 nonfood categories versus 266 nonfood categories The 1993/94 baseline food “prices” are in the 1993/94 SESC). in fact the median unit values calculated from household responses on the value Updating the baseline poverty and quantity of individual foods lines for inflation6 consumed in the 1993/94 SESC.8 These Cambodia’s baseline poverty line were the “prices” used to estimate the consists of a single national food poverty 1993/94 cost of the reference food line and three regional nonfood bundle in each region (Le., Phnom Penh, allowances. The baseline values are in Other Urban and Rural).’ The three 1993/94 Riel and refer to daily per capita regional baseline food poverty lines have levels of food and nonfood consumption. been previously updated for the effects Both the food poverty line and the ofinflation in the 1997 and 1999 poverty regional nonfood allowances need to be profiles using a common methodology updated for inflation, Le., they need to (Knowles 1998, MOP 2000). This be converted into average calendar year methodology involves estimating food 2004 prices. This section discusses the procedures used in updating the baseline 7 poverty lines for inflation. Gibson made some minor corrections to the original baseline food poverty lines in 2000 (MOP 2000), and these corrected values are the Updated food poverty line ones used in this report. The food poverty line is based on the 8 Only cash transactions were used in computing estimated cost of a single national the 1993194 unit values because they were considered to be more reliable measures ofactual reference food bundle providing an prices (Prescott and Pradhan 1997). In fact, even average subsistence diet of 2,100 the cash unit values in Phnom Penh are about calories per day (Le., averaged over 11% lower than the corresponding 1994 market prices in the Phnom Penh CPI (Annex 1). However, as noted further along in the text, the By comparison, only 3 households reported a estimates of inflation in this report are not based zero rental value in the 1997 CSES. The reasons on comparisons between prices and unit values. for imputing a nonzero value for such responses 9 Beginning with the 1998 Population Census in the 2004 CSES are discussed in Annex 4. several districts in Phnom Penh were re- Details can be found in Knowles, James “A classified from urban to rural. For surveys new set ofpoverty estimates for Cambodia, from conducted after 1998, therefore, the correct 1993194 to 2004”, August 2005, posted on the regional nomenclature is Phnom Penh, Other World Bank’s Phnom Penh Office website. Urban and Other Rural. 2 Cambodia - Povertv Assessment 2006 price inflation in Phnom Penh using price data from the Phnom Penh CPI and quantity weights from the baseline reference food bundle and using household survey data (Le., village food prices collected in the 1997 and 1999 CSES) to estimate spatial (regional) differences in food prices in a given year (i.e., 1997 or 1999). This two-step procedure is necessary because the 1997 and 1999 CSES did not collect sufficiently detailed consumption data (and no data on consumption quantities) to enable the calculation of unit values. The same two-step methodology is used in this report to estimate inflation in food prices during the period 1993/94 to 2004. One advantage of this methodology is that it avoids direct comparisons between unit values and prices.

Table 3 presents estimates of average annual rates of inflation in food prices for the periods between four household surveys (columns 2-4) and for the entire 10-year period, 1993194-2004 (column 5)." Table 4 presents updated food poverty lines and the food price index using the quantities in the reference food bundle as weights.

loColumn numbering in this report begins with the left-most column in a table (which usually contains labels). Annex 1I. Technical Appendix 191

Table 3: Average annual rates of inflation in food prices (weighted by the reference food bundle), 1993/94-2004 Region 1993194-1997 1997-1999 1999-2004 1993194-2004 Phnom Penh 5.0 12.5 0.5 4.2 Other Urban 3.3 20.1 -0.2 4.6 Rural 2.1 21.2 0.1 4.6

Table 4: Updated food poverty lines (current Riel per capita per day) and food price index (weighted- by the reference food bundle), 1993/94-2004 Region 1993194 1997 1999 2004 Food poverty lines (current Riel per capita per day) Phnom Penh 1,185 1,372 1,737 1,782 Other Urban 996 1,098 1,583 1,568 Rural 882 938 1,379 1,389 Food price index (1993194 Phnom Penh=100) Phnom Penh 100.0 115.8 146.6 150.4 Other Urban 84.1 92.7 133.6 132.3 Rural 74.4 79.2 116.4 117.2

single reference food bundle is used for all regions, the baseline nonfood Updated nonfood allowances allowances vary in their commodity The baseline nonfood allowances are composition. For example, in 1993/94, defined as the estimated value of Phnom Penh households with levels of nonfood consumption of Cambodians per capita nonfood consumption within whose total per capita household 20% of the Phnom Penh nonfood consumption is just equal to the food allowance allocated more of their poverty line. This is a minimal nonfood nonfood consumption to housing and allowance since it represents nonfood utilities (and especially to rent) than consumption that is at the expense of corresponding households in the Other food consumption that could otherwise Urban or Rural regions. This is due to be used to achieve an average daily diet the regression methodology that was with 2,100 calories by consuming the used to derive the 1993/94 baseline reference food bundle.' Although a nonfood allowances (Prescott and Pr adhan) .'

Although persons with total per capita In order to update the different regional consumption below the food poverty line would nonfood allowances for inflation, it is have to sacrifice some food consumption to necessary to have a nonfood price index purchase nonfood items, they would presumably substitute cheaper foods for more expensive foods within the reference food bundle. There is considerable scope for such substitution within l2Although the intention was to obtain different the reference food bundle. For example, although regional nonfood allowances that reflected both items absorbed similar shares of the cost of regional differences in absolute prices, it is clear the reference food bundle, the cost per 100 from the discussion in Annex 2 that the calories of rice was about 20 Riel in 1993194 estimated nonfood allowances also reflect Phnom Penh prices compared to the cost per 100 regional differences in relative foodnonfood calories of 342 Riel for a small mudfish (MOP prices and probably also regional differences in 2000). preferences. 192 Cambodia - Poverty Assessment 2006 similar to the food price index presented Table 5 provides estimates of average in Table 4 above. The main obstacle to annual rates of inflation in nonfood preparing such an index is that there is prices during the period 1993/94 to 2004 very little information on nonfood prices (but excluding the period 1997-1999 outside ofPhnom Penh prior to the 1997 because estimates ofinflation in nonfood CSES, which collected village price data prices are not available for this peri~d).'~ for 42 nonfood items. Apart from Table 6 presents the inflation-adjusted extensive data on nonfood prices in the regional nonfood allowances and (total) Phnom Penh CPI, which dates from poverty lines (Le., the sum of the July-September 1994, data on nonfood updated food poverty lines in Table 4 prices in the other regions are limited to and the updated nonfood allowances) recently discovered data that can be used and the estimated nonfood price index to calculate unit values of clothing and using weights reflecting the estimated tobacco consumption in the 1993/94 commodity composition of nonfood SESC and estimates of rental housing consumption among households with per prices by region that can be obtained capita consumption within 20% of the from hedonic price regressions. The nonfood allowance for the region in estimates of inflation in nonfood prices which they re~ide.'~ prepared in this report are based on these limited data sources and were obtained using a similar methodology to that used to estimate inflation in food prices, i.e., nonfood price data from the Phnom Penh CPI (weighted to reflect the different commodity compositions of each regional nonfood allowance) were used to estimate inflation in Phnom Penh nonfood prices, while survey data were used to estimate spatial differences in nonfood prices in 1993/94 and 1997. Similar methods were also used to estimate inflation in nonfood prices during the period 1997-2004 for which more data are available.

13 No estimates of inflation in nonfood prices were prepared in connection with the 1999 Poverty Profile (MOP 2000). l4No inflation-updated estimates of the nonfood allowances are provided in Table 6 for 1999 because the 1999 Poverty Profile estimated a new set of regional nonfood allowances using a similar methodology to that used to estimate the baseline nonfood all0wances. Annex 1: Technical Appendix 193

Table 5: Average annual rates (YO)of inflation in nonfood prices (weighted by the estimated commodity composition of regional nonfood allowances), 1993/94-2004 Region 1993194-1997 1997-2004 1993194-2004 Phnom Penh 2.9 4.2 3.8 Other Urban 3.4 3.7 3.6 Rural 4.6 4.4 4.4

Table 6. Updated nonfood allowances and total poverty lines (current Riel per capita per day) and nonfood price index (weighted according to the estimated commodity composition of the regional nonfood allowances) Region 1993194 1997 2004 Nonfood allowances (current Riel per capita per day) Phnom Penh 393 428 569 Other Urban 269 297 3 84 Rural 236 270 3 64 Total poverty line (current Riel per capita per day) Phnom Penh 1,578 1,800 2,351 Other Urban 1,265 1,395 1,952 Rural 1,118 1,208 1,753 Nonfood price index (1993194 Phnom Penh prices=100) Phnom Penh 100.0 108.8 144.8 Other Urban 90.4 99.9 129.0 Rural 78.0 89.3 120.4 Note: Total poverty lines obtained by summing updated food poverty lines and updated nonfood allowances. Annex 2: Regression results in Chapter 5

TABLEm.1: DETERMINANTS OF AGRICULTURAL PRODUCTIVITY: BASICMODEL Dependent variable: log output OLS regression IV regression value per hectare Owned with title 0.035 0.730 (1.14) (3.22)** Log of farm size -0.693 -0.699 (28.93)** (28.70)** Log of hh labour 0.152 0.136 (6.25)** (5.48)** Log of spending on other inputs 0.457 0.450 (21.91)** (21.48)** Irrigated 0.099 0.109 (2.88)** (3.05)** Type of land: Dry season land 0.09 0.105 (1.74) (1.98)* Both wet and dry season land 0.366 0.349 (4.15)** (3.88)** Chamkar land 0.179 0.197 (2.88)** (3.18)** Vegetable garden 1.036 1.085 (5.72)** (5.72)** Other types of land 0.155 0.158 (1.10) (1.13) Years of schooling of head 0.007 0.007 (2.22)* (2.33)* Dependency ratio 0.023 -0.046 (1.80) (3.96)** Constant 0.61 8 0.740 (2.40)* (2.74)** Village fixed effects Yes Yes

Observations 14892 14892 R-squared 0.51 Chi-sq overidentification test 7.45 (0.1 14)

Note: Robust, absolute t statistics in parentheses. The omitted category for type of land is wet season land. * significant at 5 percent; ** significant at 1 percent. Property rights variables are instrumented by dummies for mode of acquisition (Le., purchase, gift, inheritance, own clearing of land, marriage, occupation for free, and donation by friends). 195 Cambodia - Poverty Assessment 2006

TABLEA2.2: DETERMINANTS OF AGRICULTURAL PRODUCTIVITY: ALTERNATIVEMODEL SPECIFICATIONS Dependent variable: Log of value ofproduction per hectare All plots operated Plots owned and operated OLS IV OLS IV Owned with title 0.008 0.634 0.006 0.621 (0.30) (3.68)" * (0.23) (3.60)** Log offarm size -0.721 -0.723 -0.717 -0.719 (31.85)** (30.21)** (30.70)** (29.10)** Log ofhh labour 0.154 0.138 0.157 0.141 (5. SO)* * (4.83)** (5.79)** (4.87)** Log ofspending on other inputs 0.497 0.491 0.494 0.486 (25.49)** (23.24)** (24.64)* * (22.28)** Years of schooling ofhead 0.004 0.004 0.005 0.005 (1.47) (1.35) (1.70) (1.61) Dependency ratio 0.042 0.042 0.045 0.045 (2.96)** (2.77)* * (3.13)** (2.92)** Irrigated 0.12 0.132 0.116 0.129 (3.97)** (3.96)* * (3.76)* * (3.81)** Type of land: Dry season land 0.187 0.208 0.196 0.'216 (4.18)** (4.34)* * (4.29)** (4.39)** Both wet and dry season land 0.306 0.184 0.323 0.202 (3 * 5 7)* * (2.19)* (3.25)** (2.1 8)* Chamkar land 0.377 0.385 0.353 0.36 (6.89)* * (6.86)* * (6.34)* * (6.34)" * Vegetable garden 1.175 1.188 1.196 1.208 (6.68)** (6.30)" * (6.67)* * (6.27)** Other types ofland 0.274 0.283 0.24 1 0.254 (2.45)* (2.51)* (2.15)" (2.24)* Village infrastructure Permanent market in village 0.203 0.21 1 0.197 0.205 (3.1 9)* * (2.98)* * (3.14)* * (2.96)** Dist. from village to all-weather road (km.) -0.006 -0.005 -0.006 -0.005 (2.72)* * (1.99)* (2.89)* * (2.13)*

Constant 0.074 0.039 0.105 0.073 (0.3 1) (0.14) (0.43) (0.26) Fixed effects Province Province Province Province

Observations 15,457 15,457 14,813 14,813 R-squared 0.43 0.42

Note: Robust t statistics in parentheses. * significant at 5 percent; ** significant at 1 percent. In the IV-regressions, "owned-with title" is instrumented with mode of acquisition dummies. Only rural areas are included. Annex 2: Regression Results in Chapter 6 196

TABLEA2.3: PROPERTY RIGHTS AND PRODUCTIVITY, ROBUSTNESS CHECKS Measure of nronertv riahtr Any paper Title Observations Farm level 0.279 0.624 8.740 (3.29)** (3.28)** Rice plots only 0.337 0.782 12.066 (3.66)** (3.69)** Non-rice plots only 0.314 0.695 2.462 (1.13) (0.96) Draft animals included 0.391 0.792 11.135 (3.30)** (2.99)** Households using hired labour 0.338 0.684 10.927 (3.18)** (2.98)** Households not usinn- hired labour 0.338 0.734 4.141 (1 52) (154) Note: Robust, absolute t statistics in parentheses. * significant at 5 percent; ** significant at 1 percent. The result suggests that the (IV regression) coefficients on property rights variables are robust to different specifications of the model in Table A2.2. The dependent variable is in all cases log of the value of output per hectare. The property rights variables are in all cases instrumented with different modes of land acquisition (i.e., purchase, gift, inheritance, own clearing of land, marriage, occupation for free, and donation by friends).

TABLEA2.4: Farmsize and productivity Dependent variable: Log of value of output per hectare Households without hired labor Households with hired labor Log of farmsize -0.553 -0.552 -0.700 -0.453 -0.456 -0.677 (17.00)** (16.65)** (21.28)** (16.13)** (16.31)** (20.02)** Log of household size 0.195 0.119 (3.77)** (4.27)** Log of capital input 0.379 0.349 (8.95)** (14.92)** Log of spending on hired labor 0.134 (9.38)** Irrigated 0.14 0.12 0.177 0.134 (2.30)* (1.85) (4.25)** (3.19)** Crop grown last season: Other crops than rice 0.304 0.321 0.33 0.309 (including vegetable) (2.41)* (2.57)* (4.37)** (4.02)** Rice and other crops 0.001 -0.027 0.189 0.148 (0.01) (0.2) (2.04)* -1.65 Perennial trees 0.297 0.357 0.132 0.176 (2.07)* (2.51 )* -1.29 -1.71 NoneAA 0.178 0.662 0.526 0.581 -0.29 -1.08 -1.83 (2.15)* Years of schooling of head 0.006 0.004 -1.02 -1.21 Dependency ratio -0.072 -0.037 (2.82)** (2.60)** Village fixed effects Yes Yes Yes Yes Yes Yes Observations 4432 4432 4222 11295 11295 11040 R-squared 0.48 0.49 0.54 0.43 0.43 0.5 Note: Robust t statistics in parentheses. * significant at 5%; ** significant at 1% AATheyield data refer to the last 12 months. Therefore some plots have registered output even if a crop was not grown in the last season (plots with no output at all in the last 12 months are excluded). The reference category for crops grown last season is rice. Hired labor includes expenses for hired draft power. "Irrigated" is a dummy for whether the plot is irrigated during either wet or dry season, or both.. 197 Cambodia - Poverty Assessment 2006

TABLEA2.5: Determinants of land values (IV estimates) --- plot level Dependent variable (all in US !Wac) Land rental price Land sales price Land title (instrumented) 0.587*** 0.571*** 1.039*** 0.381*** (5.26) (4.45) (9.80) (3.32) Land area (log) -0.505*** -0.522*** -0.549*** -0.521*** (48.27) (53.71) (5 6.50) (61.92) Log land area squared 0.026*** 0.025*** -0.017*** -0.01 1*** (7.99) (8.89) (5.72) (4.69) Land planted with trees 0.665*** 0.559*** 0.598*** 0.581*** (19.20) (17.66) (1 8.85) (21.34) Irrigated in dry season 0.258*** 0.148*** 0.258*** 0.104*** (8.93) (5.46) (9.46) (4.35) Irrigated in wet season 0.134*** 0.080*** 0.158*** 0.132*** (6.82) (4.08) (8.53) (7.58) Chamkar (cash crop) land 0.396*** 0.304*** 0.01 1 -0.015 (16.16) (12.83) (0.49) (0.74) Vegetable land 0.171*** 0.231*** 0.289*** 0.214*** (2.88) (4.3 1) (5.08) (4.48) Dry season land 0.180*** 0.172*** 0.061* -0.1 12*** (5.43) (5.05) (1.96) (3.74) Type of dummy included Province Village Province Village Observations 19,170 19,170 19,654 19,654 R-squared 0.38 0.61 0.32 0.62 Identifyrng instruments are the presence ofan application receipt, dummies for different types of land acquisition, and household characteristics Absolute value oft statisbcs in parentheses * significant at IOpercent; ** significant at 5 percent, *** significant at 1 percent

TABLEA2.6: Welfare impact of land ownership and title (IV regressions) Deaendent variable: Per capita expenditure Land title (instrumented) 0.362*** 0.244** * (5.22) (3.06) Head's education (log) 0.055 *** 0.044*** (8.80) (7.72) Head's education (log) squared 0.006*** 0.005*** (7.55) (6.90) Value ofnon-land assets (log) 0.153*** 0.128*** (5 6.28) (48.36) Female head -o.ooo** -o.ooo*** (2.41) (3.93) Share ofmembers in wage employment 0.424*** 0.298*** (9.38) (7.10) Share of members w. self employment 0.060** -0.028 (2.45) (1.19) Landless dummy 0.177*** -0.001 (5.54) (0.03) Landless * rural dummy -0.109*** -0.007 (4.67) (0.25) No. of household members -0.645*** -0.617*** (50.36) (5 1.5 1) Rural dummy -0.157*** -0.890*** (8.60) (4.76) Type of dummy included Province Village Observations 14978 14978 R-squared 0.53 0.66 Note: Identifying instruments are the presence of an application receipt, dummies for different types ofland acquisition, and household characteristics. Absolute value oft statistics in parentheses: *significant at 5 percent; ** significant at 1 percent Not shown in the Table are log of land area and its higher orders; household head age and higher orders. Annex 2: Regression Results in Chapter 6 198

Explanatory Note about Instrumental Variable (IV) Regressions:

Land title is one of the explanatory variables in the regressions but it is an endogenous variable, and therefore, one needs to use the Instrumental Variable (IV) Approach. What we mean by “an endogenous variable” is that for example, the quality of soil can affect both the productivity (and value) of the plot and the likelihood of having a title. Owners with more fertile land will have greater incentives to acquire title for his land. Since we do not have a good variable to control for quality of soil in an OLS regression, we will purge this endogenous effect of quality of soil by using an IV to identify exogenous variation in land titling. The IV chosen in our case is the mode ofacquisition ofland --- purchase, inheritance, marriage, own-clearance, donation, or government provision. Why is the mode of acquisition a good IV? First, how the plot of land was acquired surely affects the plot’s likelihood of having a title. For example, a purchased plot has a higher probability to come with a title compared to a forest plot cleared by the owner. Second, the mode of acquisition does not and should not affect land productivity (and value). Through the use of IVs, the coefficients on the variable of “land title” estimate the impacts of land title on productivity and land value, purged of the endogenous effects of soil quality. In contrast, OLS coefficients are biased by such endogenous effects from uncontrolled soil quality. Annex 3: Tables

Table A3.1: Population of Cambodia by Gender

Population Male Female Year (Million persons) (Million persons) (Million persons)

1962 5.7 2.9 2.8 1982 6.9 3.1 3.8 1992 9.0 4.1 4.9 1993 9.3 4.3 5.0 1994 9.9 4.7 5.2 1995 10.9 5.2 5.6 1996 11.6 5.6 6.0 1997 11.9 5.8 6.1 1998 12.1 5.9 6.3 1999 12.4 6.0 6.4 2000 12.6 6.1 6.5 2001 12.8 6.2 6.6 2002 13.0 6.3 6.7 2003 13.3 6.4 6.9 2004 13.4 6.5 6.9 2005 13.8 6.7 7.1 2006 14.1 6.8 7.2 2007 14.4 7.0 7.4 2008 14.7 7.1 7.5 2009 15.0 7.3 7.7 2010 15.3 7.4 7.8 201 1 15.6 7.6 8.0

Sources: Statistical Year Book 2005 and CSES 2003/2004, National Institute of Statistics Table A3.2: Projected Population by Age and Sex in 2004, Cambodia Urban and Rural Cambodia Social Economic Survey 2003/2004

Age Total I Urban I Rural Group Males Female Both Sexes I Males Female Both Sexes 1 Males Female Both Sexes

0-4.. . . 777,315 754,097 1,531,412 101,064 97,630 198,694 676,25 1 656,467 1,332,718 5-9.. . . 902,340 877,141 1,779,481 119,888 116,940 236,828 782,452 760,201 1,542,653 10- 14.. . 924,885 892,978 1,817,863 134,346 132,017 266,363 790,539 760,961 1,551,500 15- 19.. . 875,578 829,673 1,705,251 146,290 145,216 291,506 729,288 684,457 1,413,745 20 - 24. . . 717,496 725,826 1,443,322 126,249 119,638 245,887 591,247 606,188 1,197,435 25 - 29. . . 388,058 427,260 815,318 70,609 7 1,798 142,407 3 17,449 355,462 672,911 30 - 34. . . 407,256 444,526 851,782 74,544 72,438 146,982 332,712 372,088 704,800 35 - 39. . . 379,097 422,514 801,611 69,985 69,9 18 139,903 309,112 352,596 661,708 40 - 49. . . 547,902 682,258 1,230,160 105,807 114,915 220,722 442,095 567,343 1,009,438 50 and + 605,697 857,237 1,462,934 84,861 124,209 209,070 520,836 733,028 1,253,864

Total 6,525,624 6,913,510 13,439,134 1,033,643 1,064,719 2,098,362 5,491,981 5,848,791 11,340,772 Source: National Institute of Statistics, Cambodia Socio Economic Survey 2003/2004 -I-

r C

r C

I ! ! ! I 1 ( ( I ! h

08 x0 0 cu cn W 0 Y

.-0

E,C 8 W .-0 0 63

'c 0

-m C .-0 .I-m z Table A3.5: Gross Domestic Product (In Current Prices) and Per Capita GDP (Billions of Riel)

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004p

GDP in billion Riels 6,793 7,093 8,438 9,191 10,130 11,719 13,408 14,089 14,860 15,994 17,311 19,630 Growth Rate in Percent (%) 4.4% 19.0% 8.9% 10.2% 15.7% 14.4% 5.1% 5.5% 7.6% 8.2% 13.4% Per Capita GDP in '000 Riels 717 713 806 833 870 962 1,076 1,110 1,150 1,215 1,290 1,435 Growth Rate in Percent (%) -0.6% 13.0% 3.3% 4.4% 10.6% 11.9% 3.2% 3.6% 5.7% 6.2% 11.2%

GDP in million US $ 2,473 2,760 3,420 3,481 3,387 3,105 3,515 3,651 3,787 4,079 4,355 4,888 Growth Rate in Percent (%) 11.6% 23.9% 1.8% -2.7% -8.3% 13.2% 3.9% 3.7% 7.7% 6.8% 12.2% Per Capita GDP in US $ 261 277 327 316 291 255 282 288 293 310 325 357 Growth Rate in Percent (%) 6.1% 18.1% -3.4% -7.9% -12.4% 10.6% 2.1% 1.7% 5.8% 4.8% 9.8%

Household Final Consumption Expenditure (HFCE) 6,713 6,571 7,812 8,749 8,932 10,887 11,732 12,211 12,351 12,820 13,220 14,922 Per Capita HFCE in '000 Riels 709 660 746 793 767 893 941 962 956 974 986 1,091 Monthly Per Capita HFCE in Riels 49,048 55,011 62,177 66,078 63,948 74,447 78,448 80,198 79,650 81,160 82,126 90,937

Source: National Institute of Statistics, National Account of Cambodia (June 2005)

Table 6: Gross Domestic Product (In Constant 2000 Prices) and Per Capita GDP (Billions of Riel)

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004p

GDP in billion Riels 8,496 9,277 9,883 10,411 10,999 11,545 12,994 14,089 14,863 15,643 16,745 18,032 Growth Rate in Percent (%) 9.2% 6.5% 5.3% 5.7% 5.0% 12.6% 8.4% 5.5% 5.2% 7.0% 7.7% Per Capita GDP in '000 Riels 897 932 944 944 945 947 1,043 1,110 1,150 1,188 1,248 1,319 Growth Rate in Percent (%) 3.9% 1.3% 0.0% 0.1% 0.2% 10.1% 6.4% 3.6% 3.3% 5.1% 5.7%

GDP in million US $ 3,093 3,610 4,006 3,944 3,677 3,059 3,407 3,651 3,788 3,989 4,213 4,490 Growth Rate in Percent (%) 16.7% 11.0% -1.5% -6.8% -16.8% 11.4% 7.2% 3.8% 5.3% 5.6% 6.6% Per Capita GDP in US $ 326 363 383 357 316 251 273 288 293 303 314 328 Growth Rate in Percent (%) 11.3% 5.5% -6.8% -11.5% -20.6% 8.8% 5.5% 1.7% 3.4% 3.6% 4.5%

Household Final Consumption Expenditure (HFCE) 8,087 8,476 9,206 10,020 9,906 10,908 11,639 12,211 12,303 12,397 12,606 13,646 Per Capita HFCE in '000 Riels 854 851 879 908 851 895 934 962 952 942 940 998 Monthly Per Capita HFCE in Riels 71,130 70,956 73,273 75,676 70,917 74,594 77,827 80,198 79,344 78,478 78,312 83,156

Source: National Institute of Statistics, National Account of Cambodia (June 2005) Table A3.7: GDP by Industrial Origin and By Economic Sector in Current Prices (Billions of Riel) ___ In Billions of Riel 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004,

AGRICULTURE, FISHERIES 8 FORESTRY 3,066 3,224 4,029 4,080 4,493 5,212 5,503 5,065 5,046 5,156 5,750 6,097

Crops 1,169 1,313 1,994 1,941 2,056 2,423 2,541 2,328 2,276 2,185 2,773 2,844 Livestock & Poultry 479 436 500 564 575 694 844 757 816 846 855 912 Fisheries 1,133 953 1,034 1,176 1,229 1,471 1,576 1,518 1,506 1,704 1,722 1,924 Forestry & Logging 285 522 500 400 633 623 542 464 448 421 400 417

INDUSTRY 862 975 1,203 1,379 1,663 1,958 2,413 3,078 3,500 4,004 4,562 5,403

Mining 13 16 19 20 20 19 26 34 36 42 47 52 Manufacturing 587 629 771 929 1,181 1.484 1,765 2,255 2,618 2,956 3,361 3,979 Food, Beverages & Tobacco 282 267 346 358 377 444 482 449 455 456 488 513 Textile,Wearing Apparel & Footwear 71 80 123 198 379 584 790 1,297 1,681 1,971 2,294 2,848 Wood, Paper & Publishing 83 118 109 142 167 171 151 132 103 112 105 113 Rubber Manufacturing 24 22 31 42 49 46 56 69 62 74 111 119 Other Manufacturing 128 142 162 189 209 239 286 307 317 343 364 386 Electricity, Gas & Water 24 31 36 44 49 59 57 58 70 86 104 120 Construction 237 299 376 387 41 2 396 565 732 777 921 1,051 1,251

SERVICES 2,686 2,576 2,884 3,311 3,560 4,079 4,741 5,231 5,521 5,931 6,088 7,028

Trade 993 923 1,002 1,142 1,231 1,361 1,476 1,512 1,522 1,562 1,635 1,903 Hotel & Restaurants 155 195 234 250 291 324 419 521 687 836 751 934 Transport & Communications 373 395 441 522 563 635 81 1 930 1,002 1,083 1,163 1,369 Finance 24 27 77 87 98 100 130 175 139 151 154 173 Public Administration 142 229 234 296 305 333 389 377 359 391 406 41 9 Real Estate & Business 622 452 483 537 601 734 765 855 877 901 925 951 Other services 377 355 413 477 471 591 752 861 936 1,008 1,054 1,280

Taxes on Products less Subsidies 195 334 395 484 537 594 865 870 921 1,040 1,065 1,280 Less: Subsidies 2 5 12 15 6 14 12 31 32 51 60 64 Less: Finance Service Charge 15 17 73 63 123 124 114 155 128 137 154 178

GROSS DOMESTIC PRODUCT (GDP) 6,793 7,093 8,438 9,191 10,130 11,719 13,408 14,089 14,860 15,994 17,311 19,630 Source: National Institute of Statistics, National Account - June 2005 p/ - prekrninaw estimates Table A3.8: Share of GDP by Industrial Origin and By Economic Sector in Current Prices (In percent)

(In percent) 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004,

AGRICULTURE, FISHERIES 8 FORESTRY 45.1% 45.5% 47.7% 44.4% 44.4% 44.5% 41.0% 35.9% 34.0% 32.2% 33.2% 31.1%

Crops 17.2% 18.5% 23.6% 21.1% 20.3% 20.7% 19.0% 16.5% 15.3% 13.7% 16.0% 14.5% Livestock & Poultry 7.1% 6.2% 5.9% 6.1% 5.7% 5.9% 6.3% 5.4% 5.5% 5.3% 4.9% 4.6% Fisheries 16.7% 13.4% 12.3% 12.8% 12.1% 12.5% 11.8% 10.8% 10.1% 10.7% 9.9% 9.8% Forestry & Logging 4.2% 7.4% 5.9% 4.3% 6.2% 5.3% 4.0% 3.3% 3.0% 2.6% 2.3% 2.1%

INDUSTRY 12.7% 13.7% 14.3% 15.0% 16.4% 16.7% 18.0% 21.8% 23.6% 25.0% 26.4% 27.5%

Mining 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.3% 0.3% 0.3% Manufacturing 8.6% 8.9% 9.1% 10.1% 11.7% 12.7% 13.2% 16.0% 17.6% 18.5% 19.4% 20.3% Food, Beverages & Tobacco 4.1% 3.8% 4.1% 3.9% 3.7% 3.8% 3.6% 3.2% 3.1% 2.9% 2.8% 2.6% Textile,Wearing Apparel & Footwear 1.O% 1.1% 1.5% 2.2% 3.7% 5.0% 5.9% 9.2% 11.3% 12.3% 13.2% 14.5% Wood, Paper & Publishing 1.2% 1.7% 1.3% 1.5% 1.6% 1.5% 1.1% 0.9% 0.7% 0.7% 0.6% 0.6% Rubber Manufacturing 0.3% 0.3% 0.4% 0.5% 0.5% 0.4% 0.4% 0.5% 0.4% 0.5% 0.6% 0.6% Other Manufacturing 1.9% 2.0% 1.9% 2.1% 2.1% 2.0% 2.1% 2.2% 2.1% 2.1% 2.1% 2.0% Electricity, Gas &Water 0.4% 0.4% 0.4% 0.5% 0.5% 0.5% 0.4% 0.4% 0.5% 0.5% 0.6% 0.6% Construction 3.5% 4.2% 4.5% 4.2% 4.1% 3.4% 4.2% 5.2% 5.2% 5.8% 6.1% 6.4%

SERVICES 39.5% 36.3% 34.2% 36.0% 35.1% 34.8% 35.4% 37.1% 37.2% 37.1% 35.2% 35.8%

Trade 14.6% 13.0% 11.9% 12.4% 12.1% 11.6% 11.0% 10.7% 10.2% 9.8% 9.4% 9.7% Hotel & Restaurants 2.3% 2.8% 2.8% 2.7% 2.9% 2.8% 3.1% 3.7% 4.6% 5.2% 4.3% 4.8% Transport & Communications 5.5% 5.6% 5.2% 5.7% 5.6% 5.4% 6.0% 6.6% 6.7% 6.8% 6.7% 7.0% Finance 0.4% 0.4% 0.9% 0.9% 1.0% 0.9% 1.0% 1.2% 0.9% 0.9% 0.9% 0.9% Public Administration 2.1% 3.2% 2.8% 3.2% 3.0% 2.8% 2.9% 2.7% 2.4% 2.4% 2.3% 2.1% Real Estate & Business 9.2% 6.4% 5.7% 5.8% 5.9% 6.3% 5.7% 6.1% 5.9% 5.6% 5.3% 4.8% Other services 5.6% 5.0% 4.9% 5.2% 4.7% 5.0% 5.6% 6.1% 6.3% 6.3% 6.1% 6.5%

Taxes on Products less Subsidies 2.9% 4.7% 4.7% 5.3% 5.3% 5.1% 6.5% 6.2% 6.2% 6.5% 6.2% 6.5% Less: Subsidies 0.0% 0.1% 0.1% 0.2% 0.1% 0.1% 0.1% 0.2% 0.2% 0.3% 0.3% 0.3% Less: Finance Service Charge 0.2% 0.2% 0.9% 0.7% 1.2% 1.1% 0.9% 1.1% 0.9% 0.9% 0.9% 0.9%

GROSS DOMESTIC PRODUCT (GDP) 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Source: National Institute of Statistics, National Account - June 2005 p/ - preliminary estimates bhlco mm> Q)

adwbbmhl awwodaco m-m w r m w d r Table A3.10: Share of GDP by Industrial Origin and By Economic Sector in Constant 2000 Prices (In percent)

~~ In percent 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004,

AGRICULTURE, FISHERIES 8 FORESTRY 45.6% 45.9% 44.6% 42.8% 42.8% 42.8% 39.4% 35.9% 35.0% 32.3% 33.9% 30.9%

Crops 18.8% 17.7% 18.4% 18.0% 16.8% 17.4% 17.8% 16.5% 15.7% 14.2% 16.7% 15.0% Livestock & Poultry 8.9% 7.9% 7.9% 7.6% 7.2% 7.3% 6.4% 5.4% 5.7% 5.3% 5.2% 5.0% Fisheries 13.6% 13.0% 13.2% 13.1% 13.1% 12.8% 11.1% 10.8% 10.8% 10.3% 9.8% 8.8% Forestry & Logging 4.3% 7.3% 5.0% 4.0% 5.6% 5.4% 4.1% 3.3% 2.8% 2.5% 2.2% 2.0%

INDUSTRY 13.0% 13.6% 15.1% 15.0% 16.6% 16.8% 18.1% 21.8% 23.0% 25.7% 26.8% 28.9%

Mining 0.2% 0.2% 0.3% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.3% 0.3% 0.3% Manufacturing 8.0% 8.0% 8.9% 9.5% 11.4% 12.5% 13.3% 16.0% 17.6% 19.1% 20.0% 21.8% Food, Beverages & Tobacco 4.2% 3.7% 4.1% 3.9% 3.8% 3.9% 3.6% 3.2% 3.1% 3.0% 2.9% 2.6% Textile,Wearing Apparel & Footwear 0.7% 0.9% 1.3% 2.0% 3.6% 4.7% 5.9% 9.2% 11.2% 12.9% 14.1% 16.3% Wood, Paper & Publishing 0.8% 1.3% 1.1% 1.2% 1.6% 1.3% 1.1% 0.9% 0.6% 0.6% 0.5% 0.4% Rubber Manufacturing 0.3% 0.2% 0.2% 0.3% 0.3% 0.5% 0.5% 0.5% 0.5% 0.4% 0.4% 0.3% Other Manufacturing 2.0% 1.9% 2.0% 2.1% 2.1% 2.1% 2.2% 2.2% 2.1% 2.2% 2.2% 2.1% Electricity, Gas & Water 0.3% 0.3% 0.4% 0.4% 0.4% 0.4% 0.4% 0.4% 0.4% 0.5% 0.5% 0.5% Construction 4.4% 5.0% 5.6% 4.9% 4.5% 3.6% 4.1% 5.2% 4.8% 5.8% 6.1% 6.4%

SERVICES 38.4% 35.4% 36.0% 37.3% 36.3% 36.3% 37.0% 37.1% 36.5% 36.2% 33.9% 34.4%

Trade 14.5% 12.6% 12.5% 12.4% 12.4% 11.9% 11.1% 10.7% 10.3% 9.9% 9.4% 9.3% Hotel & Restaurants 2.4% 2.6% 3.4% 3.3% 3.3% 3.1% 3.4% 3.7% 4.3% 4.8% 4.1% 4.7% Transport & Communications 6.3% 6.4% 6.4% 6.8% 6.1% 5.9% 6.7% 6.6% 6.9% 7.0% 6.7% 6.6% Finance 0.3% 0.4% 0.9% 0.9% 1.0% 0.9% 1.0% 1.2% 0.9% 0.9% 0.9% 0.9% Public Administration 2.0% 3.1% 2.7% 3.2% 3.2% 3.3% 2.9% 2.7% 2.4% 2.3% 2.0% 1.8% Real Estate & Business 8.4% 5.9% 5.8% 5.9% 6.0% 6.4% 6.1% 6.1% 5.7% 5.3% 4.9% 4.7% Other services 4.4% 4.4% 4.3% 4.7% 4.4% 4.9% 5.7% 6.1% 6.0% 6.0% 5.8% 6.4%

Taxes on Products less Subsidies 3.3% 5.4% 5.2% 5.6% 5.6% 5.2% 6.4% 6.2% 6.3% 6.6% 6.2% 6.7% Less: Subsidies 0.0% 0.1% 0.1% 0.2% 0.1% 0.1% 0.1% 0.2% 0.2% 0.3% 0.3% 0.3% Less: Imputed Bank Charges 0.2% 0.2% 0.8% 0.7% 1.2% 1.1% 0.9% 1.1% 0.9% 0.8% 0.9% 0.9%

GROSS DOMESTIC PRODUCT (GDP) 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Source: National Institute of Statistics, National Account - June 2005 p/ - preliminary estimates Table A3.11: Real GDP Growth Rate by Industrial Origin and By Economic Sector in Constant 2000 Prices (Percentage change)

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004,

AGRICULTURE, FISHERIES & FORESTRY 9.9% 3.5% 1.2% 5.5% 5.1% 3.7% -1.1% 2.7% -2.8% 12.2% -2.0%

Crops 2.7% 11.2% 3.2% -1.4% 8.1% 15.5% 0.6% 0.4% -4.7% 25.4% -3.4% Livestock & Poultry -2.9% 6.3% 1.3% -0.2% 6.3% -1.2% -9.1% 11.1% -1.2% 5.1% 4.3% Fisheries 4.1% 8.2% 4.6% 5.6% 2.4% -2.3% 5.0% 5.9% 0.6% 1.7% -3.3% Forestry & Logging 87.5% -26.5% -15.5% 46.3% 0.5% -13.7% -13.2% -9.5% -8.6% -4.5% 0.2%

INDUSTRY 14.2% 18.9% 4.4% 16.8% 6.2% 21.2% 31.2% 11.2% 17.3% 11.9% 16.1%

Mining 29.2% 19.6% -13.5% -2.1% -13.7% 32.3% 25.9% 0.3% 23.8% 7.7% 9.2% Manufacturing 9.0% 17.5% 12.5% 27.5% 15.2% 19.7% 30.3% 15.8% 14.5% 12.1% 17.4% Food, Beverages & Tobacco -4.9% 20.0% 0.3% 2.6% 6.1% 4.7% -3.9% 4.1% -1.2% 5.5% -2.1% Textile,Wearing Apparel & Footwear 25.0% 65.0% 57.3% 94.9% 37.1% 40.9% 68.2% 28.4% 21.2% 16.9% 24.9% Wood, Paper & Publishing 84.3% -9.7% 8.8% 44.4% -13.2% -4.6% -9.6% -29.8% 0.6% -14.0% -4.8% Rubber Manufacturing -6.5% 14.9% 21.1% 4.0% 76.2% 14.8% 10.0% 0.7% -0.7% -9.8% -1 1.4% Other Manufacturing 4.5% 11.2% 9.4% 4.7% 7.1% 16.1% 8.6% 2.6% 9.8% 4.7% 4.8% Electricity, Gas & Water 8.6% 24.8% 20.2% 11.8% 9.0% 7.1% 6.8% 5.2% 16.5% 16.0% 4.6% Construction 23.4% 20.8% -8.5% -2.5% -15.7% 27.4% 36.8% -1.8% 27.1% 11.1% 13.2%

SERVICES 0.6% 8.3% 9.2% 2.9% 5.0% 14.6% 8.9% 3.8% 4.4% 0.2% 9.2%

Trade -5.4% 5.6% 5.1% 5.5% 0.8% 5.1% 4.5% 1.5% 0.6% 2.2% 6.2% Hotel & Restaurants 19.5% 37.7% 3.0% 6.0% -3.1% 24.3% 19.0% 22.6% 18.8% -10.3% 23.6% Transport & Communications 10.8% 6.7% 11.3% -5.9% 2.4% 28.6% 6.1% 9.6% 7.5% 2.3% 6.2% Finance 22.9% 155.8% 10.6% 10.3% -6.7% 28.3% 36.4% -21.0% 5.0% 0.9% 6.6% Public Administration 71.3% -6.4% 23.9% 4.2% 10.8% -1.1% -0.8% -4.8% -0.4% -4.3% -6.7% Real Estate & Business -23.1% 4.1% 7.4% 7.5% 10.8% 8.4% 7.3% -0.4% -2.3% -0.5% 3.3% Other services 7.5% 4.3% 15.7% -1.4% 16.7% 31.2% 16.7% 3.0% 5.2% 4.8% 18.5%

Taxes on Products less Subsidies 80.8% 1.8% 14.3% 5.3% -2.4% 39.2% 4.5% 7.3% 10.7% 1.0% 15.7% Less: Subsidies 104.3% 127.7% 91.6% -61.5% 92.4% -24.3% 171.3% 1.3% 55.7% 16.3% 0.7% Less: Imputed Bank Charges 22.7% 288.4% -15.3% 89.9% -8.0% -8.7% 36.5% -17.6% 3.7% 11.2% 9.2%

GROSS DOMESTIC PRODUCT (GDP) 9.2% 6.5% 5.3% 5.7% 5.0% 12.6% 8.4% 5.5% 5.2% 7.0% 7.7% Source: National Institute of Statistics, National Account - June 2005 p/ - preliminary estimates Table A3.12: GDP Deflator, 1994-2004 Annual Percentage Changes

Annual percentage changes 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004,

AGRICULTURE, FISHERIES 81 FORESTRY -4.4% 20.7% 0.1% 4.4% 10.4% 1.8% -6.9% -3.0% 5.1% -0.6% 8.2%

Crops 9.4% 36.6% -5.7% 7.4% 9.0% -9.2% -9.0% -2.6% 0.7% 1.2% 6.2% Livestock & Poultry -6.3% 7.9% 11.2% 2.3% 13.5% 23.0% -1.3% -2.8% 4.9% -3.8% 2.2% Fisheries -19.2% 0.2% 8.6% -1.0% 16.8% 9.6% -8.4% -6.2% 12.4% -0.6% 15.6% Forestry & Logging -2.4% 30.4% -5.3% 8.2% -1.9% 0.7% -1.3% 6.5% 3.0% -0.6% 4.1%

INDUSTRY -0.9% 3.7% 9.8% 3.2% 10.9% 1.7% -2.8% 2.2% -2.4% 1.8% 2.0%

Mining -5.8% 1.6% 16.7% 6.8% 9.1% 3.9% 0.8% 6.4% -4.8% 2.7% 2.8% Manufacturing -1.7% 4.3% 7.0% -0.3% 9.1% -0.6% -2.0% 0.3% -1.4% 1.4% 0.9% Food, Beverages & Tobacco -0.2% 7.8% 3.1% 2.9% 10.9% 3.6% -2.9% -2.7% 1.4% 1.4% 7.5% Textile,Wearing Apparel & Footwear -10.2% -6.2% 2.1% -1.7% 12.3% -3.9% -2.4% 0.9% -3.3% -0.4% -0.6% Wood, Paper & Publishing -22.6% 1.9% 20.2% -18.9% 18.3% -7.2% -3.2% 10.7% 7.9% 9.0% 13.2% Rubber Manufacturing 1.3% 22.9% 11.2% 11.O% -46.1% 4.7% 12.8% -1 1.5% 20.3% 66.4% 21.8% Other Manufacturing 5.8% 3.1% 6.5% 5.5% 6.8% 3.3% -1.4% 0.6% -1.4% 1.3% 1.3% Electricity, Gas &Water 17.9% -6.4% 1.6% 0.1% 10.4% -10.4% -4.3% 14.0% 5.3% 4.6% 11.O% Construction 2.2% 4.1% 12.4% 9.3% 14.0% 11.9% -5.3% 8.2% -6.7% 2.7% 5.1%

SERVICES -4.7% 3.4% 5.2% 4.4% 9.2% 1.4% 1.3% 1.7% 2.9% 2.5% 5.7%

Trade -1.8% 2.8% 8.4% 2.1% 9.7% 3.2% -2.0% -0.9% 2.0% 2.4% 9.6% Hotel & Restaurants 5.8% -13.2% 4.0% 9.8% 14.8% 4.1% 4.4% 7.6% 2.5% 0.1% 0.6% Transport & Communications -4.5% 4.7% 6.3% 14.5% 10.3% -0.8% 8.1% -1.7% 0.5% 4.9% 10.8% Finance -6.2% 9.2% 2.3% 2.5% 10.0% 0.6% -0.8% 0.3% 3.3% 1.4% 5.4% Public Administration -6.2% 9.2% 2.3% -1.1% -1.4% 18.0% -2.3% 0.2% 9.1% 8.5% 10.7% Real Estate & Business -5.7% 2.8% 3.4% 4.2% 10.2% -3.9% 4.2% 3.1% 5.1% 3.2% -0.5% Other services -12.4% 11.4% -0.1% 0.1% 7.5% -3.0% -1.8% 5.4% 2.4% -0.2% 2.4%

Taxes on Products less Subsidies -4.9% 16.1% 7.0% 5.5% 13.4% 4.6% -3.8% -1.4% 2.1% 1.4% 3.9% Less: Subsidies -5.2% 17.0% -33.8% 9.1% 16.4% 11.2% -3.5% 0.3% 3.3% 1.4% 5.4% Less: Imputed Bank Charges -6.2% 9.2% 2.3% 2.5% 10.0% 0.6% -0.8% 0.3% 3.3% 1.4% 5.4%

GROSS DOMESTIC PRODUCT (GDP) -4.4% 11.7% 3.4% 4.3% 10.2% 1.7% -3.1% 0.0% 2.3% 1.1% 5.3% Source: National Institute of Statistics, National Account - June 2005 p/ - preliminary estimates Table A3.13: Cambodia: Monetary Survey, 1997-2004

December 1997 1998 1999 2000 2001 2002 2003 2004

(In billions of riels: end of period)

Foreign assets (net) , 1,172 1,728 2,019 2,588 3,081 3,737 4,027 4,797 Foreign assets ,597 2,203 2,508 3,047 3,587 4,279 4,741 5,482 Foreign liabilities 425 475 489 459 507 542 714 685 National Bank (net) 808 1,425 1,649 2,102 2,429 3,220 3,494 4,114 Assets ,033 1,675 1,924 2,389 2,740 3,598 3,906 4,506 Liabilities 225 251 275 286 31 1 379 412 392 Deposit money banks (net) 363 304 370 486 651 517 533 682 Assets 564 528 585 659 847 681 835 975 Liabilities 200 224 214 173 196 164 302 293

Domestic assets (net) -109 -498 -576 -758 -877 -849 -684 -467 Credit to the economy (net) 697 839 876 904 868 942 1,223 1,608 Government (net) 54 179 103 3 -75 -1 19 -128 -209 National Bank 58 182 107 4 -75 -78 -37 -118 Deposit Money Banks -4 -4 -4 0 0 -4 1 -91 -90 Public enterprises 6 6 10 3 7 2 14 0 Private sector 637 655 763 898 936 1,059 1,337 1,817 Other items (net) -805 -1,337 -1,453 -1,662 -1,744 -1,791 -1,907 -2,075 Capital and reserves -875 -1,387 -1,451 -1,541 -1,959 -1,943 -2,089 -2,192 Other 70 50 -2 -121 314 248 29 1 218

Broad money 1,063 1,230 1,443 1,831 2,204 2,888 3,329 4,329 Narrow money 385 543 532 540 610 813 937 1,153 Currency in circulation 356 509 490 495 578 766 908 1,115 Demand deposits 29 34 42 45 32 47 29 38 Quasi-money 678 687 91 1 1,291 1,594 2,075 2,391 3,176 Time deposits 13 20 32 46 56 74 82 97 Foreign currency deposits 665 667 879 1,245 1,539 2,001 2,310 3,079

(Changes in oercent of broad money stock) Memorandumitems : Foreign assets (net) 32.2 52.4 23.6 39.5 26.9 29.8 10.0 23.1 Foreign assets 40.4 57.1 24.8 37.4 29.5 31.4 16.0 22.3 Foreign liabilities 85.0 4.7 1.2 -2.1 2.6 1.6 5.9 -0.9

Domestic assets (net) -15.6 -36.6 -6.3 -12.6 -6.5 1.3 5.7 6.5 Domestic credit (net) 14.2 13.4 3.0 2.0 -2.0 3.4 9.7 11.6 Credit to government (net) -8.1 11.7 -6.1 -6.9 -4.3 -2.0 -0.3 -2.4 Other 22.3 1.7 9.2 8.9 2.3 5.4 10.0 14.0 Other items (net) -29.7 -50.1 -9.4 -14.5 0.1 -6.7 -6.1 0.1

Broad money 16.6 15.7 17.3 26.9 20.4 31.1 15.3 30.0 Money 6.1 14.9 -0.9 0.5 3.8 9.2 4.3 6.5 Quasi-money 10.5 0.8 18.2 26.4 16.6 21.8 11.0 23.6

Foreign currency deposits (in millions of US. dollars) 193 176 233 318 395 508 580 764 Nominal GDP (billions of riel) 10,130 11,719 13,408 14,089 14,860 15,994 17,311 19,630 Velocity I/ 10.8 10.6 9.9 7.9 7.2 6.1 5.7 5.0

Exchange rates (riels/U.S. dollar; end of period) 3,452 3,780 3,775 3,910 3,900 3,935 3,980 4,031 Source: National Bank of Cambodia (NBCJ. I/ Ratio of nominal GDP to average stock of broad money. Table A3.14: Cambodia: Central Government Operations, 1998-2004 11

Actual Est. Actual Est.

(In bllllons of rlels) (In percent of GDP) otal revenue 943 1,316 1,438 1,571 1,762 1,775 2,124 8.1 10.0 10.4 10.7 11.2 10.3 11.9 ax revenue 679 948 1,055 1,138 1,245 1,219 1,561 5.9 7.2 7.8 7.8 7.9 7.3 8.2 Direct taxes 56 83 136 140 132 150 158 0.5 0.6 1.0 1.0 0.8 0.9 0.8 Indirect taxes 248 432 500 580 655 635 845 2.1 3.3 3.6 4.0 4.2 3.7 4.3 Of which: Excise taxes (incl. on imporls) 76 92 113 155 210 198 304 0.7 0.7 0.8 1.1 1.3 1.1 1.6 Domestic 16 16 19 20 29 33 49 0.1 0.1 0.1 0.1 0.2 0.2 0.2 Import 60 76 94 135 181 164 255 0.5 0.6 0.7 0.9 1.2 0.9 1.3 VAT (incl. on imports) 2i ... 31 5 371 412 445 437 54 1 ... 2.4 2.7 2.8 2.8 2.5 2.8 Domestic ... 43 73 85 118 161 186 ... 0.3 0.5 0.6 0.8 0.9 0.9 Import 286 313 327 347 316 404 2.2 2.3 2.2 2.2 1.8 2.1 Refund (-) 14 14 -20 -40 -50 0.1 0.1 ... -0.1 -0.2 ... Trade taxes 376 433 391 376 424 395 516 3.2 3.3 2.8 2.6 2.7 2.3 2.6 Tax revenue from provinces 29 41 35 39 42 0.2 0.3 0.2 0.2 0.2 ontax revenue 230 355 353 424 501 525 544 2.0 2.7 2.6 2.9 3.2 3.0 2.8 OF which : Timber royalties 23 36 41 29 15 7 2 0.2 0.3 0.3 0.2 0.1 0.0 0.0 Enterprises and immobile leases 27 20 27 7 8 28 26 0.2 0.2 0.2 0.0 0.0 0.2 0.1 Civil aviation 18 19 25 40 34 22 27 0.2 0.1 0.2 0.3 0.2 0.1 0.1 Tourism income 1 1 6 14 19 20 29 0.0 0.0 0.0 0.1 0.1 0.1 0.1 Operating surplus of SOEs 13 0 13 7 2 5 1 0.1 0.0 0.1 0.0 0.0 0.0 0.0 Royalties (mining, etc.) 4 2 0 3 3 1 1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Royalties (casino) 0 12 21 32 30 37 0.0 0.1 0.1 0.2 0.2 0.2 Post and telecommunications (PTT) 87 109 92 122 123 120 94 0.8 0.8 0.7 0.8 0.8 0.7 0.5 Quota Auction 87 22 29 39 32 42 0.7 0.2 0.2 0.2 0.2 0.2 Garment licenses 21 43 36 27 67 81 0.2 0.3 0.2 0.2 0.4 0.4 Others 42 27 29 115 199 194 204 0.4 0.2 0.2 0.8 1.3 1.1 1.0 apital revenue 33 14 29 9 16 31 19 0.3 0.1 0.2 0.1 0.1 0.2 0.1 otal expenditure 31 1,571 1,825 2.118 2,367 2,820 2,914 3,206 13.5 13.9 15.3 16.3 18.0 17.4 18.0 urrent expenditure 94 1 1,097 1,222 1,381 1,579 1,842 1,902 8.1 8.4 8.9 9.5 10.1 11.0 10.7 Wages 451 525 512 488 587 614 640 3.9 4.0 3.7 3.4 3.7 3.5 3.3 Civil administration 154 194 211 214 304 327 343 1.3 1.5 1.5 1.5 1.9 1.9 1.7 Defense and security 297 331 301 274 282 287 296 2.6 2.5 2.2 1.9 1.8 1.7 1.5 Nonwage 490 572 677 893 969 1,143 1,152 4.2 4.4 4.9 6.1 6.2 6.6 5.9 Operating expenditures 372 410 497 628 690 736 717 3.2 3.1 3.6 4.3 4.4 4.2 3.7 Civil administration 220 276 365 510 579 626 606 1.9 2.1 2.8 3.5 3.7 3.6 3.1 Defense and security 152 134 132 118 111 110 111 1.3 1.0 1.0 0.8 0.7 0.6 0.6 Economic transfers 14 12 31 32 51 60 64 0.1 0.1 0.2 0.2 0.3 0.3 0.3 Social transfers 71 104 104 114 137 165 141 0.6 0.8 0.8 0.8 0.9 1.0 0.7 Civil administration 63 95 90 101 123 151 126 0.5 0.7 0.7 0.7 0.8 0.9 0.6 Defense and security 8 9 14 13 13 14 16 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Interest 15 22 21 22 28 34 49 0.1 0.2 0.2 0.1 0.2 0.2 0.2 Other nonwage 18 21 11 35 72 87 135 0.2 0.2 0.1 0.2 0.5 0.5 0.7 Reserve funds 0 0 13 63 0 35 40 0.0 0.0 0.1 0.4 0.0 0.2 0.2 Election 0 2 ... 14 7 27 6 0.0 0.0 ... 0.1 0.0 0.2 0.0 Cash adjustment -115 -15 -172 22 -0.1 -1.0 0.1 .apital expenditure 630 728 896 1,101 1,256 1,244 1,282 5.4 5.5 6.5 6.7 8.0 7.4 7.3 Locally financed 120 224 303 283 338 334 296 1.o 1.7 2.2 1.9 2.2 1.9 1.5 Externally financed 510 504 593 818 917 910 986 4.4 3.8 4.3 5.6 5.9 5.3 5.0 :urrent balance -32 205 186 160 167 -98 203 -0.3 1.6 1.3 1.1 1.1 -0.6 1.2 lverall balance -628 -509 -680 -796 -1,057 -1,139 -1,082 -5.4 -3.9 -4.9 -5.5 -6.7 -7.0 -6.1 lverall balance (Incl. grants) -286 -167 -297 -297 -606 -727 -720 -2.5 -1.3 -2.1 -2.8 -3.9 -4.5 -3.8 lnancing 628 509 680 796 1,057 1,139 1,082 5.4 3.9 4.9 5.5 6.7 6.6 6.1 oreign financing (net) 505 51 5 708 889 1,060 902 925 4.4 3.9 5.1 6.1 6.8 5.2 4.7 Of which : Project aid 504 51 1 594 839 690 822 905 4.3 3.9 4.3 5.8 5.7 4.7 4.6 Budget support 2 4 113 54 178 139 45 0.0 0.0 0.8 0.4 1.1 0.8 0.2 Amortization 0 0 0 -4 -8 -59 -25 0.0 0.0 0.0 0.0 -0.1 -0.3 -0.1 lomestic financing 117 -45 -11 11 -160 100 -110 1.o -0.3 -0.1 0.1 -1.0 0.6 -0.6 Central bank financing 120 -70 -113 -64 -105 3 24 1.o -0.5 -0.8 -0.4 -0.7 0.0 0.1 Nonbank financing -4 26 103 74 -55 97 -134 0.0 0.2 0.7 0.5 -0.4 0.6 -0.7 lutstanding operations 41 6 38 -20 32 115 8 49 0.1 0.3 -0.1 0.2 0.7 0.0 0.3 fernorandurn items: Defense and security outlays 457 477 447 405 407 41 1 423 3.9 3.6 3.2 2.8 2.6 2.4 2.4 Revenue from forestry 25 49 41 29 15 7 2 0.2 0.4 0.3 0.2 0.1 0.0 0.0 Total revenue (excl. garment quotas) 943 1,229 1,395 1,535 1,735 1,709 2,043 8.1 9.4 10.1 10.6 11.1 9.9 10.4 Health, Education, Rural, Ag. devp. (commitment) 149 232 299 382 512 529 5731 1.3 1.8 2.2 2.6 3.3 3.1 3.81 Customs department revenue 527 796 797 838 952 875 1,176 4.5 6.1 5.8 5.8 6.1 5.1 6.0

Sources: National Institute of Statistics. Statistical Year Book 2003, MEF-TOFE and IMF Staff reports. 1i Excludes provincial revenue and expenditure data from 1997-2001 2/ VAT refunds are adjusted. 3/ Total expenditure is based on a mixture of cash and accrual data. 41 Includes expenditure committed but not yet allocated to the accounts of the government agencies that execute the budget

(in Millions of Riel)

Million Riels 2000 2001 2002 2003 2004 Outturn Outturn Outturn Outturn Outturn Total Current Expenditures ('000) 1,215 1,385 1,574 1,758 1,745 I. General 283,438 268,958 298,154 410,752 301,739 ...... Royal Palace 19,775 19,006 23,456 23,065 22,375 ...... National Assembly 20,650 31,657 33,160 34,190 42,120 ...... Senate 10,140 13,999 16,600 17,871 17,917 ...... Constitutional Council 2,165 2,781 3,023 3,453 3,361 ...... Council of Minister 75,275 92,175 73,290 84,644 81,948 ...... Civil Service Secretariat 613 699 1,105 1,234 1,223 ...... CDC 2,030 2,408 2,961 2,770 3,162 ...... Interior-Administration 8,574 10,956 31,511 29,046 17,155 ...... Relations Assembly and Inspections ' 939 1,027 1,663 2,333 2,189 ...... Foreign Affairs and Int'l Cooperation 36,199 41,266 52,244 54,082 49,577 ...... Economy and Finance 95,558 19,331 36,762 111,281 35,664 ...... Planning 3,247 3,632 4,900 5,240 5,072 ...... Justice 5,784 6,298 7,195 11,099 10,487 ...... Nat'l Election Committee 2,490 23,722 7,434 26,781 6,199 National Audit Authority n.a n.a 2,850 3,662 3,288 II. Defense and Security 450,724 417,323 406,753 41 1,014 422,820 ...... National Defense 308,560 277,026 264,615 269,515 271,739 Interior - Security 142,165 140,297 142,139 141,498 151,081 111. Social 368,810 449,417 583,034 615,901 672,287 ...... Information 8,943 8,296 8,795 12,084 12,780 ...... Public Health 101,808 129,749 164,391 172,966 192,062 Education, Youth ...... and Sport 165,816 209,246 289,669 300,458 325,908 ...... Culture and Fine Arts 7,092 8,136 12,065 13,839 12,489 ...... Environment 3,859 4,535 7,131 7,739 7,376 ...... Social Affairs, Labor and Vocational Training 26,452 28,156 33,301 33,388 32,592 ...... Public Worship and Religion 1,860 2,312 3,066 3,705 3,497 Woman Affairs and Veteran 52,980 59,040 64,617 71,721 85,583 IV. Economic 112,505 150,706 159,472 170,513 151,230 ...... Civil Aviation Secretariat 3,826 5,580 5,073 6,514 6,025 ...... Industry, Mines and Energy 5,487 5,972 6,983 7,009 6,185 ...... Commerce 6,239 11,775 12,652 16,186 12,175 Agriculture, Forestry ...... and Fishery 23,415 30,468 39,652 39,024 38,552 ...... Rural Development 7,550 12,442 18,395 16,901 16,627 ...... Posts and Telecommunications 28,766 40,960 36,108 34,710 24,908 ...... Public Works and Transport 20,317 21,271 12,944 14,030 12,521 ...... Tourism 6,396 8,687 9,670 15,462 12,886 ...... Urbanization and Construction 4,311 4,786 6,134 6,833 7,046 Water Resources and Meteorology 6,199 8,764 11,861 13,844 14,305 Sources: PRSP-Annual Progress Report, Royal Government of Cambodia (May 18,2004) and TOFE 2004, Ministry of Economy and Finance Table A3.16: Consumer Price Index and Exchange Rates

Consumer Price Parallel Official Official Rate Index (CPI) Rate 21 Market Exchange /Parallel (July-December, YO Exchange Rate 41 Rate (%) 2000=100~11 Rate 31 1999 March 140.7 8.0 3,824 3,795 -0.8 June 144.2 4.9 3,823 3,805 -0.5 September 144.0 0.7 3,917 3,900 -0.4 December 142.6 -0.5 3,795 3,775 -0.5 2000 March 140.9 0.2 3,826 3,820 -0.2 June 139.4 -3.3 3,879 3,865 -0.4 Septem ber 144.3 0.2 3,910 3,880 -0.8 December 141.5 -0.8 3,919 3,910 -0.2 2001 (Jul. - Dec. 2000=100) March 100.0 0.0 3,924 3,915 -0.2 June 99.7 -0.3 3,926 3,920 -0.2 September 101.3 1.3 3,960 3,945 -0.4 December 100.7 0.7 3,905 3,900 -0.1 2002 March 102.8 2.9 3,904 3,900 -0.1 June 102.9 3.2 3,924 3,920 -0.1 September 104.5 3.2 3,970 3,950 -0.5 December 104.4 3.7 3,942 3,935 -0.2 2003 March 105.0 2.1 3,953 3,945 -0.2 June 104.6 1.7 3,990 3,990 0.0 September 105.1 0.5 4,002 4,005 0.1 December 104.9 0.5 3,984 3,980 -0.1 2004 March 105.9 0.9 3,995 3,991 -0.1 June 107.7 3.0 4,033 4,024 -0.2 September 112.8 7.4 4,092 4,054 -0.9 December 110.8 5.6 4,038 4,031 -0.2 2005 March 111.5 5.4 4,037 4,031 -0.1 June 115.3 7.0 4,137 4,106 -0.7 September 118.5 5.0 4,203 4,175 -0.7 December 118.1 6.7 4,120 4,116 -0.1

I/End of period 21 Inflation rate is the percentage change in the CPI from the same month of the previous year. 3/ Average end-period market rate (riel/US dollar). 4/ Average end-period official rate (riel/US dollar). Sources: National Bank of Cambodia and National Institute of Statistics Table A3.17: Cambodia - External Assistance to the Budget, 1997-2004 (In billions of riels)

1997 1998 1999 2000 2001 2002 2003 2004

Total official external financing 445.8 505.4 51 5.4 707.6 889.3 1,060.6 901.9 925.1

Budget support 95.9 1.5 4.4 113.2 54.2 178.2 139.1 44.5 Grants 46.2 1.5 4.4 23.8 36.2 91.5 52.2 11.8 Loans 49.7 0.0 0.0 89.4 18.0 86.7 86.9 32.7 World Bank 7.2 0.0 0.0 36.9 18.0 8.9 64.5 8.9 Asian Development Bank 42.5 0.0 0.0 52.5 0.0 77.8 22.4 23.7

Project Aid 350.8 503.9 511 .O 594.4 839.4 890.3 822.1 905.3 Grants 244.6 340.7 337.4 360.0 463.2 360.0 360.0 391.2 Loans 97.0 168.9 166.8 232.5 355.1 501 .I 494.8 536.3 World Bank 66.4 56.3 69.8 97.8 139.3 160.8 178.6 177.6 Asian Development Bank 30.6 112.6 96.6 134.7 191.3 245.1 236.4 243.2 Others 24.5 95.2 79.8 115.5 Pending 9.2 -5.7 6.8 1.9 21 .I 29.2 -32.7 -22.1

Amortization -0.9 0.0 0.0 0.0 -4.3 -8.0 -59.3 -24.8

Sources: Ministry of Economy and Finance and the IMF