CAMBODIA POVERTY ASSESSMENT Toward A More Inclusive and Resilient Cambodia © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. RIGHTS AND PERMISSIONS The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; email: pubrights@worldbank.org. Cover art by iStock/Therd oval/Sakorn Sukkasemsakorn/CHARTGRAPHIC/appleuzr Cover and interior design by Hanna Chang PHOTO CREDITS Page 8 - Song Sopheakdey / World Bank Page 10 - Chhor Sokunthea / World Bank Page 17 - Chhor Sokunthea / World Bank Page 23 - Chhor Sokunthea / World Bank Page 25 - Bou Saroeun / World Bank Page 28 - Chhor Sokunthea / World Bank Page 38 - Chhor Sokunthea / World Bank Page 40 - Chhor Sokunthea / World Bank Page 57 - Bou Saroeun / World Bank Page 58 - Dominic Chavez / Global Financing Facility Page 85 - Chhor Sokunthea / World Bank Page 86 - Chhor Sokunthea / World Bank Page 106 - Chhor Sokunthea / World Bank Page 125 - Chhor Sokunthea / World Bank CAMBODIA POVERTY ASSESSMENT Toward A More Inclusive and Resilient Cambodia Wendy Karamba and Kimsun Tong with Isabelle Salcher Poverty and Equity Global Practice East Asia and the Pacific Region November 2022 Contents OVERVIEW______________________________________________________________________________________________________ 7 EXECUTIVE SUMMARY________________________________________________________________________________________ 11 Cambodia Achieved a Decade of Growth, Poverty Reduction, and Shared Prosperity____________________________ 11 Macroeconomic Stability and Openness to Investment and Trade Facilitated Cambodia’s Economic Performance___13 Trade and Investment-led Growth Were Key Drivers for Structural Transformation and Poverty Reduction_________ 15 COVID-19 Exposed Cambodia’s Pre-existing Economic Vulnerabilities__________________________________________ 18 Government Social Assistance Scaled Up to Poor and Vulnerable Households During COVID-19_________________ 20 An Uneven Recovery, Challenges, and Opportunities Lie Ahead________________________________________________ 21 Policy Directions for a More Inclusive and Resilient Recovery___________________________________________________ 22 INTRODUCTION_______________________________________________________________________________________________ 26 CHAPTER 1 POVERTY AND INEQUALITY TRENDS___________________________________________________________ 29 1.1 Poverty Reduction Progress_____________________________________________________________________________ 29 1.2 Growth and Inequality Patterns__________________________________________________________________________ 34 1.3 Key Poverty Reduction Contributors_____________________________________________________________________ 37 1.4 Conclusion_____________________________________________________________________________________________ 39 CHAPTER 2 MULTIPLE POVERTY DIMENSION TRENDS_____________________________________________________ 41 2.1 Living Conditions and Asset Ownership__________________________________________________________________ 41 2.2 Human Capital Trends__________________________________________________________________________________ 49 2.3 Conclusion_____________________________________________________________________________________________ 57 CHAPTER 3 POVERTY AND INEQUALITY PROFILE___________________________________________________________ 59 3.1 Poverty, Inequality, and Vulnerability______________________________________________________________________ 59 3.2 Demographic and Labor Market Characteristics of Poverty________________________________________________ 68 3.3 Education, Housing Conditions, and Access to Services__________________________________________________ 78 3.4 Conclusion ____________________________________________________________________________________________ 84 2 CAMBODIA POVERTY ASSESSMENT CHAPTER 4 FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA_____________________ 87 4.1 Taxes and Social Spending in Cambodia ________________________________________________________________ 89 4.2 Overall Tax and Spending Implications on Poverty and Inequality__________________________________________ 92 4.3 Fiscal Impoverishment and Fiscal Gains to the Poor______________________________________________________ 96 4.4 Contribution of Individual Interventions to Poverty and Inequality Reduction________________________________ 96 4.5 Heterogeneity of Effects Across Deciles__________________________________________________________________ 99 4.6 Limitations and Interpretations__________________________________________________________________________103 4.7 Conclusion____________________________________________________________________________________________103 CHAPTER 5 COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS______________________________107 5.1 COVID-19 Timeline and Developments in Cambodia_____________________________________________________108 5.2 Employment and Income Shocks_______________________________________________________________________109 5.3 COVID-19 Poverty and Inequality Effects and Mitigation Policies__________________________________________117 5.4 Conclusion____________________________________________________________________________________________123 REFERENCES________________________________________________________________________________________________126 APPENDICES ________________________________________________________________________________________________131 Appendix A: Additional Figures and Tables___________________________________________________________________131 Appendix B: Poverty Measurement Methodology_____________________________________________________________137 Appendix C: Estimating Poverty Trends using Survey-to-Survey Imputations___________________________________150 EXECUTIVE SUMMARY 3 Currency Equivalents Exchange Rate Currency Unit = Cambodian Riels (KHR) US$ 1 = KHR 4061.15 (2019 period average) US$ 1 = KHR 4092.78 (2020 period average) Abbreviations and Acronyms ASEAN Association of Southeast Asian Nations BCG Bacillus Calmette-Guerin CARD Council for Agriculture and Rural Development CCT Conditional cash transfers CDF Cumulative distribution functions CDHS Cambodia Demographic and Health Survey CEQ Commitment to Equity CEA Cambodia Economic Association CGCC Credit Guarantee Corporation of Cambodia COVID-19 Coronavirus disease 2019 CPI Consumer Price Index CPS Centre for Policy Studies CSES Cambodia Socio-Economic Survey CSO Central Statistical Organization CSSE Center for Systems Science and Engineering DPT Diphtheria, pertussis and tetanus DVD Digital Video Disc EAP East Asia and Pacific region EU European Union FAO Food and Agriculture Organization of the United Nations FDI foreign direct investment FGP Fiscal Gains to the Poor FI Fiscal Impoverishment FIES Food Insecurity Experience Scale GDP Gross Domestic Product GE Generalized Entropy GFT Garment, Footwear and Travel GNI Gross National Income HCI Human Capital Index 4 CAMBODIA POVERTY ASSESSMENT HFPS High-Frequency Phone Survey HIC High-income country IDPoor Identification of Poor Households Program Kcal Kilocalories KHR Cambodian riel LIC Low-income country LMIC Lower middle-income country LPG Liquefied Petroleum Gas LSMS+ Living Standards Measurement Study Plus MAFF Ministry of Agriculture, Forestry and Fishery MEF Ministry of Economy and Finance MoEYS Ministry of Education, Youth and Sport MI Multiple Imputation MoH Ministry of Health MoP Ministry of Planning MoSVY Ministry of Social Affairs, Veterans and Youth Rehabilitation MRD Ministry of Rural Development NGO Non-governmental organization NAS National Accounts Statistics NIS National Institute of Statistics NSDP National Strategic Development Plan NWGPM National Working Group on Poverty Measurement PP Percentage Points PPP Purchasing Power Parity PSM Propensity Score Matching RDB Rural Development Bank RGC Royal Government of Cambodia SARS-CoV-2 Severe, acute respiratory syndrome coronavirus 2 SDGs Sustainable Development Goals SESC Socioeconomic Survey of Cambodia SME Small and medium-sized enterprises SNEC Supreme National Economic Council UNDP United Nations Development Programme UMIC Upper Middle-Income Country UTC Universal Cash Transfers VAT Value Added Tax VIF Variance inflation factors WB World Bank WHO World Health Organization WTO World Trade Organization EXECUTIVE SUMMARY 5 Acknowledgements The report was written by Wendy Karamba and Kimsun The report benefitted from detailed peer review from Tong with substantial contributions from Isabelle Salcher. Nadia Belhaj Hassine Belghith, Rose Mungai, Gabriela We are grateful for the invaluable contributions of Michal Inchauste, and Claire Honore Hollweg, and many Myck, Kajetan Trzcinski, Monica Oczkowska on fiscal colleagues in the World Bank who provided insights, policy effects on poverty and inequality. The team is grateful suggestions, and improvements to the report process. to Sodeth Ly for discussions on macroeconomic issues; The team wishes to thank the National Working Group Abla Safir, Robert Palacios, and Kenichi Nishikawa Chavez for Poverty Measurement (NWGPM) under the leadership for discussions and suggestions on social protection; of the Ministry of Planning for close collaboration with Ziauddin Hyder and Simeth Beng for suggestions on health the World Bank in reviewing the poverty measurement and education; and to Maryam Salim, Kim Alan Edwards, methodology in 2017 and active engagement in developing Phaloeuk Kong, Sokbunthoeun So, Viet Anh Nguyen for a new poverty measurement methodology using the additional suggestions. Excellent administrative support 2019/20 Cambodia Socio-Economic Survey. In particular, was provided by Mildred Gonsalvez and Socheat Ath. the team wishes to thank H.E. Tuon Thavrak (Secretary of Editorial support was provided by Aldo Morri. Saroeun State of the Ministry of Planning and Chairperson of the Bou helped with the press release, web display, and NWGPM), H.E. Theng Pagnathun (Director General of the dissemination events. Ministry of Planning and Permanent Deputy Chairperson The report was prepared under the guidance of Hassan of the NWGPM), H.E. Hang Lina (Director General of the Zaman (Regional Director for East Asia and the Pacific); National Institute of Statistics and Deputy Chairperson of Mariam J. Sherman (Country Director for Cambodia, the NWGPM); and all members of the NWGPM. The team Myanmar and LAO PDR); Maryam Salim (Country Manager is also grateful to participants in consultation meetings for Cambodia); Inguna Dobraja (former Country Manager with the Ministry of Planning and Ministry of Economy for Cambodia); Kim Alan Edwards (Program Leader for and Finance for feedback on various analytical outputs EAP Equitable Growth, Finance and Institutions Practice included in the report. Group), Rinku Murgai (Manager, Poverty and Equity Global Practice) and Matthew Wai-Poi (Lead Economist, Poverty and Equity Global Practice). EAP Regional Vice President Manuela V. Ferro EAP Regional Director Hassan Zaman Country Director for Myanmar, Cambodia, Lao PDR Mariam J. Sherman Country Manager for Cambodia Maryam Salim Practice Manager Rinku Murgai Task Team Leaders Wendy Karamba and Kimsun Tong OVERVIEW In the decade to 2019, Cambodia’s rapid economic as industry and services expanded to create more and growth, and structural change toward higher better jobs in which the poor were able to participate. productivity and better paying sectors, grew labor Cambodia’s inclusive growth correspondingly boosted earnings and reduced poverty. Cambodia’s economy income and standard of living for broad population has been among the fastest-growing in the world, segments. Poverty rates almost halved between 2009 propelling the country to achieve lower middle-income and 2019 with almost 2 million Cambodians escaping status in 2015. Gross domestic product (GDP) grew by an poverty (Figure 0.1). Figure 0.2 describes factors that average of 7 percent annually over the decade, and GDP enabled Cambodia’s strong growth and poverty reduction per capita grew by an average of 5.4 percent annually. The performance. structure of Cambodia’s economy changed considerably Figure 0.1 Rapid and sustained growth and dramatic poverty reduction 8 40 7 Poverty headcount 6 (percent of population) Percent of population 30 5 4 20 Percent 3 GDP growth (%) 2 GDP per capita growth (%) 10 1 0 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2009 2014 2019/20 -1 -2 Source: Cambodian authorities. Source: Ministry of Planning (2021) Despite remarkable performance, Cambodia’s The significant COVID-19 effects on Cambodia’s economy and households showed signs of economy and households had roots in challenges vulnerability, which materialized following the predating the pandemic. Although Cambodia’s growth coronavirus pandemic (COVID-19). Cambodia’s had been remarkable, it increasingly concentrated in a few economy contracted 3.1 percent in 2020 due to the industries—export-oriented manufacturing, tourism, and pandemic, marking the sharpest decline in Cambodia’s construction—alone accounting for 70 percent of growth recent history and among the most pronounced in East and 39 percent of total paid employment in 2019. While Asia. Many households experienced declines in household poverty reduction was impressive, many households income due to employment loss and/or reduced hours were in a precarious position before the pandemic; they and wages. were economically vulnerable, had low savings, and few safety nets. OVERVIEW 7 Figure 0.2 How Cambodia achieved inclusive growth and poverty reduction between 2009 and 2019 ENABLERS OF GROWTH ■ Macroeconomic stability (low and predictable inflation, stable exchange rate, effective management of public finances) ■ Openness to investment and trade DRIVERS OF GROWTH AND PROSPERITY ■ Trade and investment-led growth ■ Economic growth that outpaced population growth ENABLERS OF POVERTY REDUCTION AND SHARED PROSPERITY ■ Structural change (and job growth in higher-productivity sectors) ■ Growth in labor earnings (non-agricultural) ■ Improved delivery of key social services to underserved regions and populations DEVELOPMENT PROGRESS ACHIEVED ■ Poverty reduction ■ Shared prosperity ■ Rising standards of living ■ Inclusive human development 8 CAMBODIA POVERTY ASSESSMENT Cambodia has an opportunity to shape policy to resiliently and sustainably, and (ii) more can be done to support a more inclusive and resilient recovery. protect poor and vulnerable households, build household Reflecting on Cambodia’s progress before and during resilience, and enhance human development. Figure 0.3 the pandemic, key lessons can be drawn; (i) Cambodia’s outlines policy actions needed to achieve these goals. prosperous and inclusive growth can be developed more Figure 0.3 Policies to support a more inclusive and resilient recovery POLICIES FOR POVERTY REDUCTION AND SHARED PROSPERITY ■ Leverage cash transfers to protect poor and vulnerable households, support their recovery, and build their resilience ■ Strengthen Cambodia’s post-pandemic social protection system ■ Invest in people through health and education POLICIES FOR INCLUSIVE, RESILIENT, AND SUSTAINABLE ECONOMIC GROWTH ■ Improve capabilities of Cambodia’s firms and workers (to deepen structural transformation and sustain long-term growth) ■ Diversify trade ■ Harness domestic investments POLICIES FOR STRENGTHENING PRO-POOR EFFECTS OF FISCAL POLICY AND FINANCING POVERTY REDUCTION STRATEGIES SUSTAINABLY ■ Integrate an equity lens in the development and implementation of the upcoming Revenue Mobilization Strategy ■ Broaden the tax base to allow fiscal policy to be more countercyclical in the future ■ Reexamine spending priorities, reallocate spending, and maximize spending efficiency OVERVIEW 9 10 CAMBODIA POVERTY ASSESSMENT EXECUTIVE SUMMARY Cambodia Achieved a Decade of Growth, Poverty Reduction, and Shared Prosperity This Poverty Assessment evaluates Cambodia’s COVID-19 socio-economic effects on Cambodian poverty reduction progress between 2009 and 2019 Households in 2020. and contributing factors. Based on our understanding Cambodia’s rapid economic growth from 2009 to of contributing factors, the assessment asks what the 2019, combined with structural change, grew labor impact of the coronavirus disease 2019 (COVID-19) earnings and reduced poverty. Cambodia’s economy, has been, and what will be needed to support inclusive among the fastest-growing in the world, achieved lower recovery. The Royal Government of Cambodia (RGC) middle-income status in 2015. Macroeconomic stability— recently updated the national poverty lines for Cambodia. low and predictable inflation, stable exchange rate, and Prompted by  Cambodia’s transition  to  lower middle- sustainable fiscal policy—created favorable conditions for income status in 2015, the RGC revisited the poverty investment and growth. Trade and investment liberalization measurement methodology in 2017; the review confirmed since the 1990s has fostered private investment and high that the way Cambodians live and spend today has foreign direct investment (FDI), creating jobs in Cambodia’s changed considerably as the country became richer, and non-farm sector and markets for its exports. Cambodia’s that the national poverty lines needed revising to better gross national product (GDP) grew about 7 percent annually reflect economic realities. This Assessment uses the new over the decade. Cambodia’s population also grew at a poverty lines to evaluate Cambodia between 2009 and pace far slower than the economy, helping GDP per capita 2019, coupled with other data sources. grow about 5.4 percent annually over the same period. This Poverty Assessment covers 5 chapters. Chapter The structure of Cambodia’s economy changed rapidly 1 examines the progress Cambodia made in reducing as workers moved from agriculture to higher-productivity, poverty and boosting shared prosperity between 2009 better-paying manufacturing and service jobs, boosting and 2019. Chapter 2 examines the evolution of non- worker’s earnings. Growth boosted income and standard monetary poverty between 2009 and 2019. Chapter 3 of living broadly, and poverty rates almost halved between examines the profile of poverty and inequality in 2019/20. 2009 and 2019/201 from 33.8 percent to 17.8 percent (see Chapter 4 examines the 2019 fiscal system and its effects Chapter 1). Almost 2 million Cambodians escaped poverty, on poverty and inequality in 2019/20. Chapter 5 examines many of whom lived in rural areas. 1 The Royal Government of Cambodia (RGC) redefined national poverty lines using data from the Cambodia Socio-Economic Survey (CSES) 2019/20. This Poverty Assessment uses the new poverty lines to evaluate Cambodia’s poverty reduction progress between 2009 and 2019. EXECUTIVE SUMMARY 11 Figure ES.1 Robust growth in per Figure ES.2 Poverty headcount declined capita GDP rapidly 8 40 38.0 35 33.8 Percent of population 30.2 6 30 26.3 25 22.8 Percent 4 20 17.8 15.1 GDP growth (%) 15 12.3 9.6 2 GDP per capita growth (%) 10 5 0 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Cambodia Urban Rural 2009 2014 2019/20 -2 Source: Cambodian authorities. Source: Ministry of Planning (2021); CSES 2009, 2014, 2019/20. Note: Urban=Phnom Penh and other urban areas. Poverty estimates are based on the revised poverty methodology. Living standards improved, helping Cambodia to vaccination improved, and in 2015 Cambodia achieved narrow urban-rural gaps; but low-income and rural most of health-related Millennium Development Goals. households still lag in access to basic services and Cambodia increased primary education enrollment to earning opportunities. Housing conditions improved levels comparable with high-income economies and with more durable dwellings, and the proportion of increased secondary school enrollment. In 2019/20, households with access to electricity more than tripled almost 9 in 10 primary school-aged boys and girls were from 26 to 86 percent, with rural households seeing a enrolled, making young Cambodians more educated seven-fold increase. Access to improved water almost and literate than their parents, and the country achieved doubled from 44 to 74 percent, and access to improved gender parity in primary schooling and literacy. Increased sanitation more than doubled from 35 to 83 percent. ownership of communication devices, transport assets, Maternal and child health improved substantially as and household appliances are further evidence of rising comprehensive health services scaled up along with living standards. pro-poor social assistance. Childhood nutrition and Figure ES.3 Access to electricity tripled 100 95 97 86 87 88 79 80 73 Percent of households 62 58 60 48 43 40 31 26 20 12 12 0 Cambodia Urban Rural Non-poor Poor Area Poverty 2009 2014 2019/20 Source: CSES 2009, 2014, 2019/20. Note: Urban=Phnom Penh and other urban areas. 12 CAMBODIA POVERTY ASSESSMENT Figure ES.4 Maternal mortality declined Figure ES.5 Child mortality declined 700 100 Deaths per 100,000 live births Deaths per 1,000 live births 600 605 83 80 500 472 66 400 60 54 338 300 288 45 40 246 239 35 200 206 28 170 154 124 20 16 100 95 69 12 0 0 2005 2010 2014 2021/22 2005 2010 2014 2021/22 Infant mortality Under-5 mortality Source: CDHS 2005, 2010, 2014, and 2021/22. Source: CDHS 2005, 2010, 2014, and 2021/22. Macroeconomic Stability and Openness to Investment and Trade Facilitated Cambodia’s Economic Performance Cambodia’s remarkable economic growth and low inflation, real wages and financial assets of the poor poverty reduction took place in an environment were less likely to be eroded. characterized by macroeconomic stability and Openness to trade and investment—supported by 3 prudent fiscal management. Macroeconomic stability decades of reforms—and preferential trade access in Cambodia—reflected in low and predictable inflation, also helped drive growth. Trade and investment stable exchange rate, and sustainable fiscal policy— liberalization since the 1990s has helped Cambodia created favorable conditions for investment and growth. foster private investment by attracting high inflows of Inflation kept at single digits and the nominal exchange rate foreign direct investment (FDI) and creating markets for was broadly stable at about KHR 4,000 per U.S. dollars Cambodia’s exports. Capital accumulation—aided largely for several years. The authorities had pegged the Khmer by FDI—has been the primary driver for economic growth riel (KHR) to the United States dollar (US$), providing a in Cambodia, accounting for nearly three-quarters of real nominal anchor for stable prices. Following expansion GDP growth since 2011. Employment growth, human- in the public-sector deficit in 2009, the authorities also capital accumulation, and productivity expansion have pursued fiscal consolidation and prudent management of played a secondary role. Between 2009 and 2019, public finances, contributing to overall macroeconomic Cambodia continued to benefit from membership in stability.2 These factors reduced the uncertainty that may the Association of Southeast Asian Nations (ASEAN) cause potential investors to avoid new and long-term community and the World Trade Organization (WTO), as projects in Cambodia. In addition to a high growth rate, well as the preferential access to the European Union macroeconomic stability helped the poor; by maintaining market under the “Everything but Arms” agreement. 2 World Bank Group (2019). EXECUTIVE SUMMARY 13 Figure ES.6 Growth was driven by capital accumulation, GDP growth decomposition, share in total, 1995-2017 100% 80% Contribution to growth (percent) TFP 60% Employment 40% Avg hours Labor quality 20% Capital 0% -20% -40% 1995-2017 1995-1999 1999-2004 2004-2007 2007-2011 2011-2017 Source: World Bank (2021). Cambodia’s growing economy sustained high population (WAP) share grew slightly, helping to boost employment rates and income growth; but economic output and income per capita. Cambodia’s Cambodia’s slower population growth further helped young population offered a ready supply of labor to be boost per capita incomes. Cambodia’s population employed in more productive industry and services sector grew at a pace far slower than the economy, substantially with higher returns to labor. Cambodia’s labor market increasing the levels of per capita GDP and per capita maintained high and stable labor force participation and income over the decade. Cambodia’s age structure was employment rates between 2009 and 2019/20. also favorable for development. Cambodia’s working-age Figure ES.7 Working-age population grew slightly and labor force participation remained high 100 80 60 Percent 40 WAP (% population) LFP (% of WAP) 20 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2019/20 Source: CSES 2009, 2014, 2019/20. Note: LFP = labor force participation, WAP = working-age population. 14 CAMBODIA POVERTY ASSESSMENT Trade and Investment-led Growth Were Key Drivers for Structural Transformation and Poverty Reduction Structural transformation was crucial to Cambodia’s transformation was achieved mainly through movement poverty reduction success. Growth in per capita income of workers from low to higher-productivity sectors, and to supported poverty reduction in Cambodia; but “structural some extent the migration of workers from low to higher- transformation”3 helped to distribute growth and income productivity areas, and workers staying within existing sectors to broader segments of the population. In Cambodia, but benefitting from within-sector productivity growth. Figure ES.8 Real GDP growth and sectoral contributions (2009–2019) Growth was driven by garment manufacturing, …aided by high FDI inflows into those sectors tourism, and recently construction and real estate 8 1500 6 4 1000 KHR billions Percent 2 0 500 -2 -4 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2013 2014 2015 2016 2017 2018 2019 Agriculture Industry Agriculture, mining and quarrying Manufacturing Services Net taxes on production Construction and real estate GDP Growth Financial activites Source: Cambodian authorities. Source: World Bank (2021). CEM. …resulting in declines in agriculture’s share in … and openness to trade value-added. 80 100 70 Imports 80 Percent of GDP Percent of GDP 60 40 40 39 50 60 40 Exports 21 26 30 40 34 20 20 34 29 10 21 0 0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2009 2014 2019 Agriculture Industry Services Source: World Development Indicators (2021). Source: Cambodian authorities. Investment and trade spurred Cambodia’s structural garment, construction, and tourism sectors—creating transformation away from agriculture towards higher better jobs and helping move Cambodia from agriculture value-added activities. FDI and exports supported dependence toward light manufacturing and services. non-agricultural job creation, driving structural change, Cambodia saw consequent declines in agriculture’s and poverty reduction. FDI flowed mainly into Cambodia’s share in employment and output, and increases in the more productive, non-agricultural sectors—especially the combined share of industry and services. As workers 3 Structural transformation occurs as people and resources shift from lower to higher productivity activities. It raises household incomes and living standards, lifting people out of poverty (World Bank, 2022). EXECUTIVE SUMMARY 15 exited from agriculture, services progressively became 4 in 10 jobs were non-farm wage jobs, up from 2 in 10 the largest employer, while industry became the largest a decade earlier. Workers were enticed to industry and producer. Cambodia’s key export sectors contributed services by the wage differences. For example, in 2009, more than one-third of all paid employment. In 2019, non-agricultural wages were about 4 times higher than garment and footwear manufacturing accounted for agricultural wages, while labor productivity in non- about 1 million jobs (21 percent of paid employment), agriculture was 3 times that of agriculture. Workers who while tourism accounted for 620,000 jobs (13.9 percent of remained in agriculture consequently benefited from paid employment).4 Travel and tourism sectors combined productivity growth within the sector. Value added per accounted for 2 million jobs and one-quarter of GDP. worker increased most in agriculture, where declining numbers of workers—coupled with crop yield growth Cambodian workers moved from agriculture to and smallholder land-use expansion—resulted in a near higher-productivity and better-paying manufacturing doubling of labor productivity in over a decade. Despite and services sectors. In 2019/20, 6 in 10 jobs were the gains, agricultural productivity remained at low levels. non-farm jobs, up from 4 in 10 a decade earlier. About Figure ES.9 Workers moved out of agriculture and those who remained enjoyed some benefits from rising within-sector productivity Workers moved to the non-farm sector Workers moved to non-farm wage employment Percent of total employment Percent of total employment 100 100 27 30 23 17 21 80 38 80 16 20 38 60 24 60 43 26 40 40 58 48 45 36 20 35 20 28 0 0 9 9 7 2009 2014 2019/20 2009 2014 2019/20 Agriculture Industry Services Agriculture wage Agricultural self-employed Non-agricultural wage Non-agricultural self-employed Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Workers moved to more productive sectors Workers who stayed in agriculture saw large productivity increases 20 10 Employment growth 2009-17 (Million KHR at 2000 prices) Output per worker per year (average annual change) 8 10 6 4 0 2 -10 0 0 5 10 15 20 2009 2014 2019/20 Value added per worker (average million riels in 2000 prices) Agriculture Industry Services Source: World Bank (2021). Source: Authors’ calculation based on National Accounts and CSES. 4 World Bank (2021). CEM. 16 CAMBODIA POVERTY ASSESSMENT Workers not only moved across sectors, but spatially Urban-to-urban migration doubled over the decade. from rural areas to towns and cities. Between 2008 The proportion of households receiving remittances and 2019, about one-third of Cambodians migrated out originating from within Cambodia more than doubled from of rural areas, with the pace of migration increasing after 17 percent in 2009 to 43 percent in 2019/20, and the 2014 (Figure ES.10, Figure ES.11), driven by economic volume of remittances grew six-fold largely due to flows opportunities in cities and transport connectivity. to rural areas. Figure ES.10 Rural outmigration and Figure ES.11 Urbanization is rising urban-to-urban migration increased 60 100 Percent of population Percent of population 50 80 40 60.6 60 80.5 77.5 30 40 20 10 20 37.2 18.5 20.7 0 0 2008 2019 2009 2014 2019/20 Rural-to-rural Rural-to-urban Population Urban-to-rural Urban-to-urban Urban Rural Source: Cambodia Population Census 2008, 2019. Source: CSES 2009, 2014, 2019/20. Growth in non-farm labor earnings, especially wages, was most responsible for poverty reduction.5 Manufacturing (mainly garments) and services attracted low-skilled workers from agriculture, offering higher and faster-growing wages. Real wages increased over the decade, with those at the bottom of the wage distribution benefitting. Average real non-agricultural wages quadrupled between 2009 and 2019/20, and average real non-agricultural self-employed income doubled. Most garment and footwear jobs paid at least the minimum wage and offered a range monetary and non-monetary benefits. The minimum wage, applicable only to garment and footwear industries, was hiked up 8 times between 2009 and 2019, increasing from US$50 in 2009 to US$182 in 2019. Cambodia’s labor-intensive manufacturing and export jobs played a crucial role in poverty reduction, with the poverty rate falling from 33.8 percent in 2009 to 17.8 percent in 2019/20. While non-farm job and wage growth contributed to the decrease in poverty rates, nearly 9 in 10 in the labor force are still employed in the informal sector, which typically means lower-quality jobs. 5 We use the decomposition method of Azevedo et al. (2013) to examine the contribution of demographics, labor income, public transfers, and remittances to poverty reduction and identify which factors contribute most to observed poverty reduction. Although the decompositions do not identify causal effects, they are useful to focus attention on income components that are quantitatively more important for changes at the bottom of the distribution. EXECUTIVE SUMMARY 17 Figure ES.12 Growth in non-farm wages Figure ES.13 Growth in remittances led to substantial poverty reduction supported some poverty reduction 20,000 2,000 KHR (2019/20 prices) KHR (2019/20 prices) 15,000 1,500 10,000 1,000 5,000 500 0 0 Agricultural Agricultural Non-agricultural Non-agricultural Remittances Public transfers Other nonlabor wage self-emplyed wage self-employed income 2009 2014 2019/20 2009 2014 2019/20 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Remittances growth contributed to poverty transfers have even less effect on poverty reduction since reduction but marginally; public transfers to a far they provide low coverage and are small. However, there less extent. Despite substantial growth in remittances, has been a steady growth in public transfers over the they account for a small share of household income, and decade – reflecting a commitment of the RGC to help thus contribute marginally to poverty reduction. Public the poor. COVID-19 Exposed Cambodia’s Pre-existing Economic Vulnerabilities COVID-19 exposed Cambodia’s pre-existing construction, tourism, and merchandise exports, which economic vulnerabilities. Although Cambodia’s growth alone accounted for 70 percent of growth and 39 percent has been remarkable, it was insufficiently diversified of total paid employment in 2019. Cambodia’s tourism across products, markets, and financing sources. Five and hospitality sector—estimated to account for about products—garments, footwear, rice, cassava, and 2 million jobs and one-quarter of GDP pre-pandemic— tourism—accounted for 80 percent of total exports; 2 collapsed due to travel restrictions. International arrivals fell markets—the European Union (EU) and the United States 72 percent in the first 8 months of 2020 compared to the (U.S.)—accounted for 69 percent of merchandise exports; prior year. Garment, footwear, and travel goods exports fell and FDI and official development assistance accounted 11.5 percent by August 2020 due to weakened demand. for 72 percent of gross fixed capital formation in 2018.6 The construction and real estate sectors weakened. The economy also faced other negative shocks, including In 2020, Cambodia’s economy contracted 3.1 floods in late 2020 that reduced agricultural production, percent; the sharpest decline in Cambodia’s recent and the partial withdrawal of preferential trade access on history, and among the most pronounced in the East Asia some Cambodian exports (including garments, footwear and Pacific (EAP) region. In 2020, Cambodia’s growth and travel goods). The authorities responded with fell an estimated 10 percentage points (pp) from its measures to support households and firms during COVID pre-pandemic average growth rate. COVID-19 severely (see Chapter 5). Despite extensive government support, negatively affected Cambodia’s growth-driving sectors: the economy and households suffered enormously. 6 World Bank (2021). CEM. 18 CAMBODIA POVERTY ASSESSMENT Figure ES.14 COVID-19 had a negative effect on Cambodia’s economy The pandemic induced economic slowdown…. …among the most pronounced in the EAP region. COVID and Post- COVID period Difference in the average real growth between 2010-2019 and 2020 10 a sia di lia R ia d m PD bo an go ys ne na a m ala ail in on do o et Ch Ca Th La M M Vi In 0 5 Percent -2 -4 0 Percent -6 -8 -5 2017 2018 2019 2020 2021e 2022p 2023p 2024p -10 Agriculture Industry -12 Services Net taxes on production GDP Growth -14 Source: Cambodian authorities and World Bank staff projections. Source: Cambodian authorities and Macro-Poverty Outlook. Note: e = estimate; p = projection. Many Cambodian households were in a precarious time. COVID-19 weakened labor markets considerably position before COVID-19. Although poverty reduction and household incomes declined due to employment had been impressive, many Cambodians remained loss and/or reduced hours and wages (see Chapter economically vulnerable, placing them at risk of falling into 5). As the crisis prolonged, household finances were poverty (see Chapter 1). About 15 percent of Cambodians stressed, pushing more households to assume debt to were near-poor with incomes marginally above the cover living expenses, potentially leaving scarring effects poverty line (see Chapter 3). Household savings were low for future consumption.7 Hitting the poorest people before the pandemic, limiting household means to cope hardest, the adverse COVID-19 effects eroded the shared with shocks, let alone endure months of hardships from prosperity and contributed to rising inequality as low- a pandemic (see Chapter 5). Few social safety nets were income households suffered most from job losses and available to Cambodians to mitigate negative welfare declines in welfare. For the bottom 20 percent, per capita effects (Figure ES.13, see Chapter 4). consumption dropped 3 times more than for the top 20 percent, and low-income households were more likely to COVID-19 partially reversed Cambodia’s poverty reduce food consumption and eat less nutritious meals reduction progress. In 2020, poverty rose for the first (see Chapter 5). 7 Potentially scarring mechanisms include increasing household debt, selling assets, and pulling children out of school. EXECUTIVE SUMMARY 19 Government Social Assistance Scaled Up to Poor and Vulnerable Households During COVID-19 The Government leveraged the existing “IDPoor” Cash transfers provided valuable income support registry to deliver relief cash transfers to registered to poor households during the pandemic and IDPoor households. Launched in June 2020, the curbed the increase in poverty and inequality. The program has disbursed US$714 million in cash transfers program mitigated 40 percent of the increase in poverty. as of July 2022. The cash transfer program has been the We estimate about 460,000 entered poverty in 2020, largest component of the government’s support package. raising the poverty rate 2.8 pp from the official rate in Spending on cash transfers rose from less than 0.1 2019/20.8 In the absence of cash transfers, around percent of GDP in 2019 to 0.7 in 2020 and 1.4 in 2021. 750,000 Cambodians could have entered poverty in The program has reached about 690,000 households 2020, raising the poverty rate 4.7 pp. Increased inequality and 2.7 million individuals, or about 17 percent of the was also mitigated, with the Gini index projected to have population, up from 2 percent of the population pre- increased by 0.2 pp in 2020 rather than by 0.4 pp had COVID-19. cash transfers not been provided. Cash beneficiaries also valued the income support, using the cash to buy food and essential items (see Chapter 5). Figure ES.15 Cash transfers mitigated poverty increase The number of poverty entrants was curbed… …and increased poverty mitigated. 800,000 748,406 5.0 4.7 Percentage points change 4.0 Number of the poor 600,000 458,226 3.0 2.8 400,000 2.0 200,000 1.0 0 0.0 No government With government No government With government intervention intervention (CT) intervention intervention (CT) Source: Authors’ calculations based on CSES 2019/20. Source: Authors’ calculations based on CSES 2019/20. While assistance rapidly scaled up, cash transfer of households saw income declines than those receiving coverage and adequacy left room for improvement. social assistance under the current IDPoor program Per capita consumption of the poorest quintile declined by which covers households at the bottom of the distribution 25 percent even with transfers. Larger cash transfers and (see Chapter 1, Figure 1.7). Expanded coverage to newly better inclusion of the near poor is needed as Cambodia’s poor and near poor is needed to protect households from economic recovery remains uneven and households face falling deeper into poverty. increased risk of poverty despite economic recovery. The crisis exposed holes in the current social protection system for those in the middle of the distribution; twice the share 8 Simulations were implemented using the ADePT macro-simulation module. 20 CAMBODIA POVERTY ASSESSMENT An Uneven Recovery, Challenges, and Opportunities Lie Ahead Cambodia’s policy makers have several risks, challenges and opportunities to consider ahead: Cambodia’s households face increased risk of ranks lowest among comparator countries in foundational poverty despite economic recovery. COVID-19 has human capital, as measured by the World Bank Human affected low-skilled workers, informal workers, and Capital Index (HCI) (see Chapter 2). Cambodia’s HCI score low-income households more than proportionately and of 0.4910 demonstrates shortcomings in early childhood Cambodia’s uneven economic recovery further risks nutrition and in the education system—reflected by high slowing the pace of household recovery. Growth in 2022 stunting rates, late school entry, high grade repetition, is projected at 4.8 percent and will mostly depend on 9 and high dropout rates (see Chapter 2 and Chapter 3). manufacturing and other services given the slower Cambodia’s moderate productivity performance and recovery of travel and tourism. Inflation, led by rising fuel insufficiently diversified economy could reduce its and food prices triggered by the Russia-Ukraine conflict, long-term growth and poverty reduction potential. could slow the pace of poverty reduction as it weighs Productivity is crucial for an economy to grow, and on household budgets. Cambodian households also therefore an important determinant of living standards. face increased climate vulnerability as climate change But overall labor productivity in Cambodia remains increasingly threatens rural and non-agricultural livelihoods low for its level of economic development. Productivity through disruptions to agricultural production, water, and growth within manufacturing and services sector has transport. Investment in climate change adaptation is been suboptimal, in part due to low human capital. The thus increasingly important to protect employment and World Economic Forum Report confirms Cambodia has mitigate poverty. a low share of highly skilled workforce due to weak skills Cambodia has several development challenges to development (see Chapter 2). Cambodia’s inability to address, including spatial disparities in poverty and grow its product basket is linked to low productivity.11 low human development. Despite notable progress Sustainability of Cambodia’s garment sector is also under on health, Cambodia’s overall progress in human threat. In recent years, garment sector competitiveness development has been slow. Education is Cambodia’s has weakened due to rapid increase in legislated minimum biggest obstacle; over the past decade, Cambodia raised wages coupled with stagnant productivity. Moreover, net primary education enrollment to levels comparable partial suspension of the country’s preferential access to to high-income economies; but net enrollment in lower- the EU (European Union) market under the “Everything secondary education (47 percent in 2019/20) remain well but Arms” agreement further narrows the export market. below other ASEAN countries. Learning outcomes and Cambodia’s ability to diversify trade and improve human overall educational attainment also remain low. COVID-19 capital will be crucial to sustained long-term growth. further threatens educational outcomes and nutrition in poor and newly poor households (Chapter 5). Cambodia 9 World Bank (2022). MPO. 10 Indicates that a child born today is expected to be 49 percent as productive in adulthood as he or she would have been with complete education, good health, and a well-nourished childhood (see Chapter 2) (WBHCI). 11 World Bank (2021). CEM. EXECUTIVE SUMMARY 21 Cambodian growth can benefit from deeper “demographic dividend” with a pool of potential workers structural change and a key demographic transition, expected to expand for the next 35 years.12 Cambodia provided investments are made in people and can harness economic growth stemming from changes firms today. Structural transformation is well underway in the population’s age structure provided it continues to in Cambodia but further productivity gains remain to create good jobs and invest in youth health and education. be achieved. Cambodia is also in the early phases of a Policy Directions for a More Inclusive and Resilient Recovery We recommend several policy recommendations to foster a more inclusive and resilient Cambodian economy and society: (I) Social assistance remains an immediate priority. It will be imperative to deliver health and (III)  Pandemic support to families is crucial, but support education services to the poor. Human capital after the pandemic will also be needed to support is often the only asset the poor have to generate household recovery, rebuild livelihoods, and build income and improve their quality of life. Cambodia’s resilience. Going forward, Cambodia’s policymakers policymakers can increase spending on Cambodia’s can assess if cash transfers will remain a temporary nutrition, health, education, and water and sanitation measure to protect citizens from falling into poverty or (WASH) services. Policy makers can also improve if they can promote future poverty reduction as part of the quality of services, focusing on rural areas to a broader strategy. If part of a broader strategy, cash further narrow spatial and socioeconomic disparities. transfers can be used to build long-term household Spending on health and education was responsible for resilience by (i) boosting individual and family most fiscal redistribution in 2019. Policymakers can financial well-being, and (ii) encouraging savings, (iii) also incentivize Cambodian uptake of social services. expanding financial inclusion and literacy (expand For instance, expanding maternal and child cash access to regulated financial products and services transfer program can encourage pregnant women and broaden use through financial awareness and and mothers to seek health services, or expanding the education) and (iv) encouraging uptake of health and cash transfer program can encourage rural and low- education services. income households to stay in school. Adaptive (II)  social protection systems go a Policy makers can strengthen the pro-poor (IV)  step further by helping ensure that critical effects of fiscal policy while financing poverty investments in human capital are not reduction strategies sustainably. Balancing undermined by future crises or shocks.13 This short-term emergency assistance with medium- will require preparing and enhancing social protection term growth and debt sustainability will be vital in systems ahead of large-scale shocks to build the maintaining macroeconomic stability. Cambodia’s resilience of poor and vulnerable households before, ability to increase investment in social assistance, during, and after shocks.14 health, and education will hinge on its ability to 12 World Bank (2019a). Rutkowski (2018). Using adaptive social protection to cope with crisis and build resilience. (https://blogs.worldbank.org/voices/using- 13  adaptive-social-protection-cope-crisis-and-build-resilience). 14 Bowen et. al. (2020). Adaptive Social Protection: Building Resilient to Shocks. 22 CAMBODIA POVERTY ASSESSMENT generate new sources of revenue; but even without of fixed capital formation. The actions will enhance increased spending, Cambodia can reprioritize economic growth along with resilience, while spending and improve spending efficiency through improving the quality and inclusiveness of jobs to better targeted poverty alleviation efforts. support income growth and poverty reduction. The recent Cambodia Country Economic Memorandum Action (V)  is needed to support Cambodia’s (CEM) lays out recommendations for diversifying COVID-19 economic recovery and enhance Cambodia’s growth model, which we reflect in this economic resilience by diversifying economic report in the Cambodia Policy Matrix.15 activity. Cambodia’s policymakers can forge a new growth path by increasing firm and worker The Cambodia Policy Matrix outlines some of these policy productivity, diversifying exports, and harnessing actions in more detail: domestic investment to sustain rapid expansion 15 World Bank (2021). CEM. EXECUTIVE SUMMARY 23 Figure ES.16 Cambodia Policy Matrix OBJECTIVES ACTIONS Protect the poor and vulnerable, support their recovery, and build their resilience Leverage cash transfers for ■ Continue income support to families during the pandemic and as it winds down. immediate and ■ Increase the transfer amount a family receives. future poverty ■ Expand social assistance coverage to newly poor and near-poor. reduction ■ Use cash transfers to build household financial resilience, livelihoods, and human capital. Build adaptive social protection systems for future crises ■ Increase spending on social protection to a steady-state above 0.1% of GDP. ■ Institute strong and adequate post-pandemic social assistance, benchmarking amount Strengthen of assistance to the poverty line. Cambodia’s ■ Introduce shock-responsive elements to social assistance programs, pre-establishing post-pandemic social protection programs, financing and payment mechanisms, IT systems, and institutional system arrangements. ■ Build on the IDPoor program to develop a social registry, expanding coverage and improving the scoring system used for targeting. ■ Develop adaptive social protection systems for future crises. Improve health and education service delivery Improve human ■ Increase spending on health and education in rural areas. capital through health and ■ Improve quality of health and education service delivery. education ■ Deploy resources more efficiently and effectively based on assessments of where development lags and investments are needed most. Support inclusive, resilient, and sustainable economic growth Diversify ■ Improve capabilities of Cambodia’s firms and workers to sustain long-term growth. Cambodia’s growth model16 ■ Diversify exports. ■ Harness domestic investment to finance the next phase of growth and job creation. Strengthen the Strengthen the redistributive effect of fiscal policy and finance poverty reduction redistributive strategies sustainably effect of fiscal ■ Integrate an equity lens in the development and implementation of the upcoming policy and finance Revenue Mobilization Strategy. poverty reduction strategies ■ Broaden tax base to diversify sources of revenue. sustainably ■ Reexamine spending priorities, reallocate spending, and maximize spending efficiency. 16 See World Bank Cambodia Country Economic Memorandum (CEM) 2021a for detailed policy actions. 24 CAMBODIA POVERTY ASSESSMENT CHAPTER 1. POVERTY AND INEQUALITY TRENDS 25 INTRODUCTION Cambodia’s rapid economic development prompted The new methodology revises the consumption the Royal Government of Cambodia (RGC) to expenditure aggregate to incorporate use-value of durable revisit its poverty measurement methodology and goods and imputed house rent, and better capture redefine national poverty lines. Led by the Ministry education expenses. Poverty line estimates also apply a of Planning (MoP), the National Working Group for cost-of-basic needs and a common basket approach for Poverty Measurement (NWGPM) undertook this task the whole country. in consultation and agreement with all working group Recently, the RGC released new poverty lines and members comprising numerous ministries and institutions. new poverty estimates for 2019/20. Until this, poverty In 2017, the NWGPM, in collaboration with the World estimates used the 2009 poverty lines (updated based Bank, reviewed the poverty estimation methodology, on the annual inflation rates).17 The first official national revealing that rapid economic development in Cambodia poverty lines for Cambodia, defined in 1997 using data over the decade had significantly changed Cambodian from the 1993/94 CSES, served as the basis to estimate consumption patterns, including the poor. Moreover, and track poverty in Cambodia until 2008. Cambodia was no longer a low-income country, having graduated to lower middle-income status in 2015. This The new national poverty line for Cambodia is raised concern that the 2009 poverty lines might be out KHR10,951 per person per day. Considering cost-of- of date and no longer reflected economic realities. The living differences between locations, this translates to KHR review also highlighted the need to revise the poverty 10,951 in Phnom Penh, KHR 9,571 in other urban areas, measurement methodology to align more with international and KHR 8,908 in rural areas. Under the new poverty best practices. It was vital to revise Cambodia’s poverty line, about 17.8 percent of the population was “poor” in lines to use as a new benchmark to monitor the National 2019/20. Poverty rates vary by area of residence, with 4.2 Strategic Development Plan 2019-2023 and Cambodia percent considered poor in Phnom Penh, 12.6 percent in Sustainable Development Goals 2030. other urban areas, and 22.8 percent in rural areas. To revise the poverty lines and to improve poverty Poverty trends are robust to the recent changes in measurement, the RGC decided to: the survey instrument and poverty measurement methodology. Changes in the survey instrument Improve the consumption module of the Cambodia ( i)  and poverty measurement methodology make direct Socio-Economic Survey (CSES) 2019/20 to better comparisons of the consumption aggregate and poverty collect consumption highly disaggregated data and rates between surveys impossible. To compare poverty record quantities. trends, a survey-to-survey imputation approach is used, ( ii)  Revise poverty lines and the consumption aggregate in which consumption is imputed in previous household to make them more consistent with international best surveys to estimate poverty headcount based on the practices. new poverty line. Imputed poverty estimates under the new poverty lines provide a consistent picture of poverty decline as with previous lines. 17 2009 poverty lines were defined in 2013 using CSES 2009. 26 CAMBODIA POVERTY ASSESSMENT Based on the new poverty estimates, this Poverty The Assessment covers 5 chapters. Chapter 1 Assessment examines Cambodia’s poverty examines progress in poverty reduction and shared reduction progress between 2009 and 2019. The prosperity between 2009 and 2019. Chapter 2 examines report largely draws on information from the nationally the evolution of non-monetary poverty between 2009 representative CSES. The Assessment also evaluates and 2019. Chapter 3 examines the profile of poverty the distributional implications of the Cambodian fiscal and inequality in 2019/20. Chapter 4 examines the 2019 system in 2019, relying on both the CSES and the fiscal system and its effects on poverty and inequality in Ministry of Economy and Finance’s administrative data on 2019/20. Chapter 5 examines COVID-19 socio-economic government taxes and spending. effects on Cambodian Households in 2020. The Assessment also examines the COVID-19 socioeconomic effects households between 2020 and 2021. The analysis draws information from the Cambodia High-Frequency Phone Surveys (HFPS) of households and macroeconomic projections to examine COVID-19 effects on poverty, inequality, and other socioeconomic outcomes since 2020. INTRODUCTION 27 28 CAMBODIA POVERTY ASSESSMENT CHAPTER 1 POVERTY AND INEQUALITY TRENDS Cambodia achieved significant poverty reduction in of inequality. Last, the chapter examines key drivers of the last decade, with the pace of poverty reduction poverty reduction in Cambodia. increasing after 2014. Today, COVID-19 threatens those Between 2009 and 2019, the proportion of Cambodians gains. This report first examine trends and patterns in living living below the national basic needs poverty line fell standard improvements in the decade up to 2019; and from 33.8 percent in 2009 to 17.8 percent in 2019/20. based on an understanding of key factors that contributed About 2 million Cambodians in rural areas escaped to this progress, looks at COVID-19 effects and what poverty, the leading factor for the decline in poverty Cambodia will need to support inclusive recovery. headcount. Between 2009 and 2019, Cambodia’s rapid This chapter documents poverty and inequality economic growth averaged 7 percent annually, helping to trends and patterns in Cambodia between 2009 and improve incomes and living standards and reduce poverty. 2019. The chapter first traces the evolution of poverty and While growth was inclusive and average incomes of all inequality as well as depth and severity of poverty. The Cambodians grew, inequality increased slightly in recent chapter then traces the evolution of economic growth and years as consumption for poorer Cambodians grew slower identifies its contribution to poverty reduction. The chapter compared to richer Cambodians. Cambodia’s success in then examines the patterns of consumption growth to reducing poverty can largely be attributed to growth in non- understand the extent to which growth was inclusive farm labor earnings. of large segments of the population and the evolution 1.1 Poverty Reduction Progress Poverty in Cambodia declined rapidly between 2009 and 2019/20 than during the preceding 5 years. The pace and 2019. The proportion of Cambodians living below the of poverty reduction was most rapid in Cambodia’s rural national basic needs poverty line—set at Cambodian riel areas and Coastal region. The proportion of extremely (KHR) 10,951 per person, per day based on the 2019/20 poor Cambodians who cannot afford to buy food to meet Cambodia Socio-Economic Survey (CSES)—declined minimum nutrition requirements of 2,200 kilocalories (Kcal) from 33.8 percent in 2009 to 17.8 percent in 2019/20. per person, per day also declined from 2 to 0.6 percent. The pace of poverty reduction was faster between 2014 CHAPTER 1. POVERTY AND INEQUALITY TRENDS 29 The depth and severity of poverty in Cambodia trillion (US$547 million or 2 percent of GDP) of perfectly decreased. From 2009 to 2019/20, the depth of poverty targeted transfers to escape poverty.18 The amount (or poverty gap) declined from 8.2 to 3.5 percent. The averages KHR1.88 trillion (US$445 billion) in rural areas poverty gap index captures the average shortfall in and KHR0.4 trillion (US$99 million) in urban areas. The consumption of the population from the poverty line, severity of poverty, which gives an indication of inequality expressed as a percentage of the poverty line. It thus among those living in poverty, also fell between 2009 gives an indication of resources needed to be transferred and 2019/20 from 2.8 to 1.1 percent. This indicates a to poor persons to eliminate poverty. This implies that in a reduction in inequality in consumption among poor given year, the poor in Cambodia would require KHR2.23 households. Figure 1.1 Poverty Trends, 2009–2019 Figure 1.2 Food Poverty Trends, 2009–2019 40 38.0 3 33.8 2.1 2.1 35 30.2 2 30 Percent of population Percent of population 26.3 25 22.8 2 1.2 1.2 20 17.8 15.1 1 15 0.8 12.3 9.6 0.6 0.6 10 0.4 0.4 1 5 0 0 Cambodia Urban Rural Cambodia Urban Rural 2009 2014 2019/20 2009 2014 2019/20 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Note: Urban=Phnom Penh and other urban areas. Note: Urban=Phnom Penh and other urban areas. Poverty estimates are based on the revised poverty methodology. Poverty estimates are based on the revised poverty methodology. Figure 1.3 Poverty Depth and Severity, 2009–2019 10 9.3 8.2 8 6.8 5.9 6 4.7 4 3.5 3.3 3.2 2.8 2.5 2.3 1.9 2 1.6 1.5 1.1 1.1 0.8 0.5 0 Cambodia Urban Rural Cambodia Urban Rural Depth of poverty Severity of poverty 2009 2014 2019/20 Source: CSES 2009, 2014, 2019/20. Note: Urban=Phnom Penh and other urban areas. Poverty estimates are based on the revised poverty methodology. 18 The most common approach to estimate the minimum cost of eliminating poverty is to measure the amount of consumption (or income) shortfall from the poverty line of all households across the entire population in a country - the non-poor is considered as having zero shortfall. The average of the 2019 and 2020 exchange rate of KHR 4077 per US$ is used for the conversion. 30 CAMBODIA POVERTY ASSESSMENT Box 1.1 Poverty Measurement in Cambodia The RGC recently redefined the poverty lines for Cambodia using the 2019/20 CSES and cost-of-basic need and common basket approaches. The national poverty line is KHR 10,951 per person, per day; considering cost-of-living differences between locations, this translates to KHR 10,951 in Phnom Penh, KHR 9,571 in other urban areas, and KHR 8,908 in rural areas. Cambodia’s poverty reduction is “robust” in statistical terms; that is, changes in the survey and poverty line measurement methodologies do not change the direction of the poverty trend. Assessments of poverty over time can suffer comparability problems stemming from changes in survey design and methodological improvements, such as those made in the 2019/20 CSES (see Appendix B for details). The challenges were addressed using survey-to-survey “imputation” techniques. Poverty estimates backcast to previous survey rounds thus account for adjustments in design changes. The results support the finding that poverty has declined over the last decade in Cambodia (see Appendix C for details). Although the results support observed poverty reduction, methodological improvements reflect higher living standards compared to the previous methodology. Note, currently poverty rates based on international poverty lines are not available which is why international comparisons of poverty rates and of growth elasticities of poverty are not made in the report. Poverty declined faster in rural than in urban areas. areas fell 5.5 pp from 15.1 to 9.6 percent. Despite poverty Poverty in rural areas fell 15.2 pp from 38 to 22.8 percent reduction progress, 1 in 5 people live below the poverty between 2009 and 2019/20. In contrast, poverty in urban line, with poverty in rural areas almost twice the urban rate. Figure 1.4 Poverty Reduction, 2009–2019 Area Region Plateau and Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Mountains 0 -1.5 Percentage point change -2.8 -2.7 -2.6 -5 -7.5 -7.4 -7.6 -7.4 -7.5 -7.8 -8.0 -8.5 -8.4 -8.8 -10 -8.9 -12.0 -15 2009-2014 2014-2019/20 Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. CHAPTER 1. POVERTY AND INEQUALITY TRENDS 31 Between 2009 and 2019/20, 1.9 million Cambodians During the 2009–2019, the number of poor declined 5 escaped poverty. The number of poor fell nationally percent annually while the population grew at about 1.3 from 4.7 million in 2009 to 4.0 million in 2014 and 2.8 percent annually.19 From 2009 to 2019, the total number million in 2019/20 (Figure 1.5). The rapid reduction of of “food poor” Cambodians declined from about 293,000 the number of poor against a slowly rising population to 101,000. resulted in the substantial reduction in the poverty rate. Figure 1.5 Total Population and Number of Poor People, 2009–2019 A. Population size (million) B. Poor population (million) 20 6 5 4.7 16.0 15 14.0 15.2 4.0 4 4.3 11.8 3.6 10 11.2 3 2.8 9.7 2 2.2 5.9 5 2.6 3.1 1 0.4 0.4 0.6 0 0 2007 2009 2011 2013 2015 2017 2019 2021 2007 2009 2011 2013 2015 2017 2019 2021 Cambodia Urban Rural Cambodia Urban Rural Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. As millions of Cambodians in rural areas rose out of at 9 percent. Migration is a key driver of population growth poverty, the number of urban poor increased with in urban areas as about one-third (34 percent) of migration rural-urban migration. The number of poor in rural in Cambodia is from the rural to urban areas.20 Fertility and areas decreased by 2.1 million from 4.3 million in 2009 to mortality—the other contributors of population growth— 2.2 million in 2019/20. The rural poverty rate decreased increased only marginally over the decade. According significantly because rural poverty fell faster relative to rural to the 2019 General Population Census, fertility in urban population; while the number of rural poor fell annually by areas rose slightly from 1.9 to 2.2 children per woman about 6 percent between 2009 and 2019/20, the rural between 2008 and 2019,21 while life expectancy rose from population fell annually by about 2 percent. Meanwhile, the 66 to 69.8 years.22 number of poor people rose slightly in urban areas from Although poor Cambodians overwhelmingly reside in about 400,000 people in 2009 to about 600,000 people in rural areas, they are increasingly living in cities. The 2019/20. The urban poverty rate decreased nonetheless growing number of poor in cities and declining number in because the urban population grew faster relative to urban rural areas shifted the geographic distribution of poverty. poverty; while the number of urban poor increased annually Still, nearly 80 percent of poor Cambodians still lived in by about 4 percent between 2009 and 2019/20, people rural areas in 2019/20 compared to 90 percent in 2009. living in urban areas grew annually at about twice the rate 19 Based on CSES 2009 to 2019/20 estimates. 20 Ministry of Planning, 2020. General Population Census of the Kingdom of Cambodia 2019. 21 Ibid. Note: Rural total fertility rate is 2.8 children per woman in 2019. 22 World Bank World Development Indicators, 2021. 32 CAMBODIA POVERTY ASSESSMENT Figure 1.6 Distribution of the Population and Distribution of the Poor, 2009–2019 100 80 63 60 82 78 80 92 89 40 20 37 18 22 20 8 11 0 2009 2014 2019/20 2009 2014 2019/20 Population Poor Urban Rural Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. Many Cambodians cluster around the poverty line. in consumption led many people to move out of poverty While this means that many could escape poverty with (Figure 1.7). At the same time, many Cambodians are pro-poor policy interventions, it also means that many “near poor” and have just enough to live above the Cambodians are vulnerable to economic shocks. Rapid poverty line, underscoring their risk of falling into poverty poverty reduction indicates many Cambodians live in the event of adverse shocks, such as the COVID-19 around the poverty line, which enabled many to escape pandemic (see Chapter 5, Figure 5.18, Figure 5.19).23 poverty in the last decade. The cumulative density of However, concentration of people around the poverty line average consumption per capita is steep at the bottom has diminished over the decade with improvements in end of the distribution, which implies small improvements living standards. Figure 1.7 Cumulative Distribution of Per Capita Consumption, 2009–2019 Source: CSES 2009, 2014, 2019/20. Note: Near poor are defined as those living between the poverty line and 1.25 of the poverty line. 23 Near poor are individuals whose consumption (or income) is marginally above the poverty line. The term “marginally above” is defined differently. In Lao PDR, near poor persons are defined as those living from 100 to 150 percent of poverty thresholds, while in the United States they are defined as those living from 100 to 125 percent of poverty thresholds (Saczewska-Piotrowska, 2016). As the first analysis of near poor in Cambodia, we propose defining the near poor as those living from 100 to 125 percent of poverty line. CHAPTER 1. POVERTY AND INEQUALITY TRENDS 33 1.2 Growth and Inequality Patterns Cambodia sustained average real GDP per capita tourism, and more recently construction. Consequently, growth of 5.4 percent annually during 2009–2019, growth has been inclusive, with consumption per capita raising per capita income from US$700 in 2009 for the poorest 40 percent of the population growing by to US$1,530 in 2019 (current US$), and achieving an average 2.4 percent annually and poverty headcount lower-middle income status in 2015. Contributing rate falling 33.8 percent to 17.8 percent between 2009 to Cambodia’s success are labor-intensive and export- and 2019/20.24 oriented sectors including garment manufacturing, Figure 1.8 Poverty Headcount and GDP Figure 1.9 Average Daily Per Capita Per Capita, 2009–2019 Consumption (KHR), 2009–2019 1,500 40 40,000 33.8 KHR/day/person 30 30,000 1,000 26.3 Percent US$ 20 20,000 17.8 500 10 10,000 0 0 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 GDP per capita (constant 2010 USD) Poverty headcount Cambodia Urban Rural Bottom 40 Top 60 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. GDP per capita from World Development Indicators. Note: Urban = Phnom Penh and other urban areas Robust economic growth has sustained strong between 2009 and 2014 and -1.25 between 2014 and poverty reduction in Cambodia. Economic growth 2019.26 This implies that a 10 percent increase in annual was strong in the pre-COVID decade from 2009 to 2019 per capita income was associated with a 9.3 percent and helped drive the significant decline in poverty over decrease in annual poverty rate during 2009–2014 and a the same period. Poverty fell by 5.9 percent annually over 12.5 percent decrease during 2014–2019. Among other this time (Table 1.1), and at a faster rate in the second factors, the growth elasticity of poverty depends on the half of the period (6.9 percent) than in the first half (4.9 initial level of income inequality;27 countries with a more percent). Partly this reflected slightly higher growth in equal distribution of income, in other words, tend to the second half but each point of growth also resulted have greater poverty reduction for a given increase in per in greater poverty reduction in the second half as well. capita income. The growth elasticity of poverty was estimated at -0.93 25 24 Cambodia’s consumption per capita grew by 2.9 percent annually while GDP per capita grew about 5.4 percent annually, which suggests a gap between survey means and national accounts system estimates of private consumption. 25 The percentage reduction in poverty associated with a percentage change in GDP per capita. 26 2019 is selected as the end period for economic growth calculations to avoid the impact of COVID-19 In 2020. The household survey underpinning the poverty estimates was collected between July 2019 and June 2020, so the poverty estimate overlaps the start of COVID-19, but this sub- sample represents only a quarter of the full survey and occurs only at the start of the pandemic when the economic effects were likely lightest. 27 Ravallion (2005). 34 CAMBODIA POVERTY ASSESSMENT Table 1.1 Growth Elasticity of Poverty Annual GDP growth Annual growth in Annual poverty Growth elasticity of GDP per capita reduction poverty 2009–2014 7.0 5.3 4.9 -0.93 2014–2019 7.1 5.5 6.9 -1.25 2009–2019 7.0 5.4 5.9 -1.10 Source: National Accounts. CSES 2009, 2014, 2019/20. Despite accelerating poverty reduction after 2014, the “growth effect” dominated the “redistribution effect”, rising inequality prevented the positive contribution the increase in inequality (redistribution) from 2014–2019 of growth to poverty reduction being even higher. partially offset the positive effect of growth in household Poverty headcount decreased more from 2014–2019/20 consumption on poverty reduction (growth effect). While than 2009–2014, in line with higher household consumption growth in household consumption reduced poverty by growth (Figure 1.10, Figure 1.11). During 2009–2014, real 11.4 pp, increased inequality offset the reduction by per capita GDP grew at an average 5.3 percent annual rate 2.9 pp, resulting in total reduction in poverty of 8.5 pp. but poverty declined 7.5 pp. The 2014–2019 period was That is, if the growth in household consumption between marked by sightly higher real per capita GDP growth of 2014–19 had been shared equally, poverty would have 5.5 percent and a poverty decline of 8.5 pp. While overall fallen by nearly 3 percentage points more. Figure 1.10 Growth and Redistribution Figure 1.11 Growth and Redistribution Effects on Poverty Reduction, 2009–2014 Effects on Poverty Reduction, 2014–2019 4 4 2.9 0.5 0 0 -4 -4 -8 -8 -7.5 -8.0 -8.5 -12 -12 -11.4 -16 -16 Change in poverty headcount Growth Redistribution Change in poverty headcount Growth Redistribution Source: CSES 2009, 2014, 2019/20. Note: The contribution of growth and redistribution to poverty reduction is based on the decomposition method of Datt and Ravallion (1992). While growth in household consumption was consumption growth in favor of the wealthier top 60 inclusive, there are signs of faster but less “pro- percent in rural areas. The negative “shared prosperity poor” growth in recent years. In absolute terms, the premium”—income growth among the bottom 40 poor have benefitted from overall growth in the economy, percent relative to average growth—eroded in recent experiencing average consumption growth above 2 years, especially in the rural areas. While Cambodia’s percent per year. Consumption growth accelerated in the poverty reduction over the last decade has been strong second half of the decade for the population. However, by any standards, if inequality continues to increase in the consumption growth among the nation’s bottom 40 decade to come, a strong possibility given the unequal percent has not kept pace with consumption growth impacts of COVID-19 (see Chapter 5), then it will become of the entire population. This is due to disproportionate harder to sustain poverty reduction in the future. CHAPTER 1. POVERTY AND INEQUALITY TRENDS 35 Consumption growth became less pro-poor in rural during 2014–2019. In urban areas—particularly those areas and slightly more pro-poor in urban areas. outside Phnom Penh—consumption growth was slightly Over the last decade, household consumption grew pro-poor from 2014 to 2019, curbing rising inequality in twice as fast in rural areas than in urban areas, hence urban areas. In contrast, consumption was less pro-poor the rapid reduction in rural poverty. While all Cambodians in rural areas, with the poor benefitting less from growth benefitted from economic growth, consumption across than the rich, contributing to rising rural inequality. the distribution and population groups grew more unevenly Figure 1.12 Consumption Growth of the Figure 1.13 Gini Index, 2009–2019 Bottom 40 and Top 60 of the Population, 2009–2019 5 40 33.8 32.2 32.7 32.9 Annualized growth rate (%) 4 29.5 29.9 30 28.0 25.9 26.2 3 20 2 10 1 0 0 2009– 2014– 2009– 2014– 2009– 2014– Cambodia Urban Rural 2014 2019/20 2014 2019/20 2014 2019/20 Cambodia Urban Rural 2009 2014 2019/20 Bottom 40 Top 60 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. The Gini index rose between 2014 and 2019 after previous years, especially in rural areas, where the Gini having remained stable from 2009 to 2014. From index increased from 26 percent in 2009 to 28 percent in 2009 to 2014, inequality remained stable because 2019/20. Although inequality rose in rural areas, inequality consumption grew more or less evenly for all Cambodians. remained much higher in urban areas with the Gini index Inequality increased in 2019 and was higher than in at 33.8 percent. 36 CAMBODIA POVERTY ASSESSMENT 1.3 Key Poverty Reduction Contributors Between 2009 and 2019, Cambodia maintained high averaged 84 percent. Cambodia has among the highest labor force participation and employment rates. labor force participation rate in the East Asia and Pacific Labor force participation rates averaged 85 percent and region. employment in relation to the working-age population Figure 1.14 High and Stable Labor Force Figure 1.15 Low Unemployment Rates Participation Rates 100 1.2 Unemployment (% labor force) (% population ages 15-64) Labor force participantion 80 1.0 0.8 60 0.6 40 0.4 20 0.2 0 0.0 2009 2010 2011 2012 2013 2014 2015 2016 2017 2009 2010 2011 2012 2013 2014 2015 2016 2017 2019/20 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Cambodia underwent a significant jobs created in industry and services, notably export-oriented transformation and created many productive and manufacturing, tourism, and construction. Workers rapidly remunerative non-farm jobs. Cambodia’s investment moved from agriculture toward non-agricultural wage- and export-led growth supported non-agricultural paying employment. In 2019, 4 in 10 jobs were non-farm job creation, driving structural change, and poverty wage jobs as opposed to 2 in 10 a decade earlier. reduction. More productive and better paying jobs were Figure 1.16 Employment by Sector, Figure 1.17 Employment by Type, 2009–2019 2009–2019 Percent of total employment Percent of total employment 100 100 27 30 23 17 21 80 38 80 16 20 38 60 24 60 43 26 40 40 58 48 45 36 20 35 20 28 0 0 9 9 7 2009 2014 2019/20 2009 2014 2019/20 Agriculture Industry Services Agriculture wage Agricultural self-employed Non-agricultural wage Non-agricultural self-employed Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. CHAPTER 1. POVERTY AND INEQUALITY TRENDS 37 industries paid at least the minimum wage and offered a range of additional monetary and non-monetary benefits. Growth in non-farm labor earnings, especially in Growth in remittances also contributed to poverty wages, contributed most to poverty reduction, as reduction, but to a far less extent than the contribution of opposed to growth in employment.28 Notably, average labor income. Despite substantial growth in remittances, real non-agricultural wages quadrupled and average real they account for a small share of household income, and non-agricultural self-employed income doubled. Job hence their modest contribution to poverty reduction. The growth concentrated in manufacturing (mainly garments) contribution of public transfers is even less given that they and tourism, which offered significantly better remuneration are small, provide low coverage, and “leak” to non-poor that in agriculture. Most jobs in the garment and footwear households (see Figure 4.11, Figure 4.12). Figure 1.18 Growth in Labor Income Figure 1.19 Growth in Non-labor Income 20,000 2,000 KHR (2019/20 prices) KHR (2019/20 prices) 15,000 1,500 10,000 1,000 5,000 500 0 0 Agricultural Agricultural Non-agricultural Non-agricultural Remittances Public transfers Other nonlabor wage self-emplyed wage self-employed income 2009 2014 2019/20 2009 2014 2019/20 Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. 28 W  e use the decomposition method of Azevedo et al. (2013) to examine the contribution of demographics, labor income, public transfers, and remittances to poverty reduction and identify which factors contribute most to observed poverty reduction. Although the decompositions do not identify causal effects, they are useful to focus attention on income components that are quantitatively more important in describing changes at the bottom of the distribution. 38 CAMBODIA POVERTY ASSESSMENT 1.4 Conclusion Sustained and inclusive economic growth has driven Cambodia still has further progress to make in progress in Cambodia, boosting income and living reducing poverty and vulnerability, especially standards broadly. Cambodia witnessed strong growth, among rural Cambodians. While notable progress created better employment and income opportunities for was made for the rural population, still, most of the poor the poor, and reduced income poverty. Trade openness and vulnerable reside in rural areas. Ensuring growth is and investment flows into industry and services sectors inclusive and resilient will support continued poverty enabled Cambodia to reallocate economic activities into reduction progress. Diversifying the economy and more productive sectors, creating of jobs that offer higher improving firm and worker productivity will be important to wages than in agriculture. Cambodian workers were achieve deeper structural transformation, income growth, able to move from agriculture to the more productive and faster poverty reduction. There may be a greater role and better-paying jobs in manufaturing and services. for tax and benefit policies to strengthen the poverty- Labor earnings growth, especially non-farm wages, reducing effect of growth. Progressive social spending played a central role in driving poverty reduction in can help enhance poverty reduction and equity. Cambodia. Employment growth played a supporting role since Cambodia sustained high and stable employment throughout the decade. Human capital endowments for adult Cambodians grew only marginally as shown in Chapter 2 suggesting that returns to existing endowments in human capital rose. CHAPTER 1. POVERTY AND INEQUALITY TRENDS 39 40 CAMBODIA POVERTY ASSESSMENT CHAPTER 2 MULTIPLE POVERTY DIMENSION TRENDS This chapter outlines the progress that Cambodia of households with access to basic sanitation more than has made in many aspects of well-being and the doubled. Asset ownership increased, while ownership of deprivations that Cambodians continue to face. agricultural implements remained relatively unchanged, Poverty encompasses many aspects of human well-being and livestock ownership declined. Cambodia has made beyond those captured by standard monetary measures. progress in building human capital through expanded This chapter examines the evolution of non-monetary access to education and health services. dimensions of well-being such as housing conditions, Challenges remain in ensuring equitable access to asset ownership, and human capital. better living conditions, education, and health for Cambodians have experienced fewer non-monetary Cambodia’s poor and rural residents. Despite progress deprivations over the last decade. Since 2009, housing expanding access to education, transition from primary quality improved considerably as dwellings increasingly to secondary school remains incomplete and dropout consisted of more durable materials. Access to electricity rates during secondary school continue to be high. Adult tripled, and electricity is now the primary source of lighting literacy rates and educational attainment are gradually across the country. Households have increasingly shifted rising, but human capital indicators remain moderately low from cooking with wood, charcoal, and other forms of by international comparisons. Expanded access to health biomass to cooking with liquefied petroleum gas (LPG), care has translated into improved health outcomes for both substitution that reduces environmental damage, time children and adults. However, chronic child undernutrition spent collecting fuel, and illnesses from biomass exhaust still prevails. Non-monetary dimensions of poverty have exposure. The proportion of households with access to particularly improved in rural areas, but some spatial and safe drinking water almost doubled and the proportion socioeconomic differences persist. 2.1 Living Conditions and Asset Ownership Improvements in housing conditions are evidence outside elements. Improvements were faster for wall and of rising living standards, even for rural and poor flooring, although from a low baseline. Nationally, the households. Since 2009, housing quality improved proportion of households with improved roof, wall, and considerably, with dwellings increasingly consisting of floor materials increased respectively by 15, 29, and 20 more durable materials improving protection against percentage points (pp) between 2009 and 2019. While CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 41 Figure 2.1 Housing Conditions, 2009–2019, Percentage A. Improved roof material B. Improved wall material C. Improved floor material 100 95 99 97 99 99 94 98 100 100 84 81 80 80 80 62 65 63 62 60 60 60 57 52 46 40 40 34 40 37 30 22 24 21 20 20 17 20 17 13 10 8 0 0 0 Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 D. Dwelling ownership 100 80 60 94 94 92 97 96 97 94 94 92 95 94 95 40 83 86 85 20 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Owned Not owned but no rent is paid Rented Other E. Number of bedrooms 6 5 4.8 4.8 5 3.9 4.1 4 3.6 3.7 3.5 3.5 3.2 3.2 3 3.2 3 2.8 3 1.9 1.9 2 2 1.4 1.5 1.7 1.3 1.3 1.5 1.5 1.5 1.7 1.2 1.2 1.3 1 0 2019/20 2019/20 2019/20 2019/20 2019/20 2009 2014 2009 2014 2009 2014 2009 2014 2009 2014 Cambodia Urban Rural Non-Poor Poor Number of rooms Person/rooms Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. 42 CAMBODIA POVERTY ASSESSMENT almost all Cambodians have adequate roof conditions, 54 Some remaining barriers to electricity access and percent of Cambodians still live in houses with poor-quality consumption include capacity, availability, reliability, walls made of natural, salvaged, or plastic materials with and affordability. Access to high-capacity and reliable insufficient insulation and waterlogging, and 63 percent electricity is skewed in favor of urban and higher-income do not have adequate flooring. households. According to an Energy Access Diagnostic, rural households are less likely to receive high-capacity Housing conditions improved faster in rural areas electricity as well as 24/7 electricity supply, and face more and among the poor, but dwellings are smaller and power disruptions.29 About 1 in 3 rural households do are made of lower-quality materials. Homeownership not receive over 2 kilowatt of electricity (high-capacity), is particularly high among rural and poor households. about 1 in 3 rural grid-connected households receive less Ninety-seven percent of rural households own their than 16 hours of electricity per day, and about 3 in 4 rural dwelling, compared with 85 percent of urban households. grid-connected households face frequent, unpredictable Similarly, 95 percent of poor households own their power outages. Lack of affordable electricity is also dwelling compared to 92 percent of non-poor households. more critical among rural households than among urban Despite high homeownership, dwellings of rural and poor households. households are much smaller and made of lower-quality materials. Rural and poor dwellings have around 0.5 fewer Energy sources for cooking further illustrate rooms than urban and non-poor dweilling, and house 2 electricity access constraints. Only 1 percent of more people per room. While improved roof materials are Cambodian households use electricity for cooking, wth now equally common among urban and rural households, almost 60 percent using firewood or charcoal. Rural and urban households are still roughly twice as likely to have poor households are more likely to use carbon fuels at improved wall materials and nearly 3 times as likely to 75 and 85 percent, respectively. Smoke released by have improved flooring. incomplete combustion of these carbon fuels put rural and poor households at greater risk for respiratory Cambodia has made rapid progress expanding problems. While grid and off-grid electricity systems are electricity access to its citizens, particularly in rural still underused for cooking across the country, households areas. In 2019, 86 percent of Cambodian households increasingly use other fuels such as liquefied petroleum had access to electricity, with the share tripling from 26 gas (LPG). The proportion of Cambodians using gas for percent in 2009. Most progress occurred in rural areas, cooking quadrupled within a decade from 10 percent where access increased nearly 7-fold from 12 percent to 42 percent, with use skewed toward higher-income in 2009 to 79 percent in 2019. Phnom Penh maintained households. However, LPG can also create unsafe levels high levels of access for 99 percent of households, while of indoor air pollution if not used in well-ventilated spaces. other urban areas also saw an increase from 77 percent to 96 percent. At the same time, some forms of off-grid Access to water and sanitation improved over the electricity, such as solar power, are growing in use. About decade for all Cambodians irrespective of poverty 8 percent of Cambodian households use solar for lighting, status, with strong gains in rural areas. The proportion with twice as many rural and poor households using this of households lacking access to safe drinking water was source of energy. The rise in rural electrification has also halved in 1 decade from 56 percent in 2009 to 26 percent led to reduction in use of kerosene and off-grid power in 2019. Most improvements were in rural areas where (batteries) as a source of lighting. the proportion of households using safe water sources increased from 37 percent in 2009 to 67 percent in 2019. 29 Dave et al. 2018. See https://openknowledge.worldbank.org/handle/10986/29512. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 43 Figure 2.2 Access to Electricity, 2009–2019, Percentage A. Access to electricity 100 95 97 86 87 88 79 80 73 62 58 60 48 43 40 31 26 20 12 12 0 Cambodia Urban Rural Non-poor Poor Area Poverty 2009 2014 2019/20 B. Lighting source of energy 1 1 2 1 0 1 4 1 1 1 5 3 4 2 1 4 100 2 0 2 4 2 0 8 7 2 0 0 2 0 1 7 0 2 7 0 9 13 6 16 5 0 12 4 1 26 3 1 80 31 6 37 28 44 9 31 38 60 39 95 97 39 40 39 86 87 88 79 73 46 58 62 40 20 48 43 26 31 12 12 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Electricity Battery Kerosene lamp Solar Other C. Cooking source of energy 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 10 2 5 1 3 16 25 13 19 15 42 45 46 57 70 98 95 98 96 90 83 86 85 74 79 57 53 52 40 29 0 1 1 2 2 1 0 1 1 0 1 1 0 0 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Electricity Firewood/charcoal Gas (LPG) Other Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. 44 CAMBODIA POVERTY ASSESSMENT Most households with access to safe drinking water source dwelling more than doubled between 2009 and 2009, from it from public taps, boreholes, and protected wells. Access 27 percent to 51 percent (see Figure A.7). At the same time, to piped water in the dwelling or on premises is growing, the proportion of households traveling at least 30 minutes although remaining limited in rural areas and among poor round-trip to collect water declined from 27 percent to 4 households, with only 13 percent of households having percent. While there is growing installation of piped water access to piped water in rural areas compared to 57 systems, water on premises largely reflects usage of public percent in urban areas. More than twice as many non-poor taps, boreholes, and protected wells and use of rainwater than poor households have access to piped water. collection and tanker trucks to bring water to the household. However, truck-supplied water is not always safe as there is Distance to main source of water has declined. The no guarantee that it is free from contamination. proportion of households with access to water at their Figure 2.3 Access to Water and Sanitation, 2009–2019, Percentage A. Access to drinking water 100 18 15 26 23 24 80 33 37 49 16 53 46 56 15 29 63 57 63 58 60 44 43 40 54 29 51 30 67 29 30 62 57 34 33 20 33 33 30 25 32 14 21 18 4 9 13 4 8 13 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Piped into dwelling/premises Improved water Unimproved water B. Access to sanitation 100 7 16 13 10 13 21 80 36 34 41 49 54 61 61 72 60 79 90 92 86 40 83 85 77 57 63 64 20 49 43 35 36 24 17 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Basic sanitation Unimproved sanitation None Source: CSES 2009, 2014, 2019/20. Notes: Figure 4.3A: Improved water: Drinking water from public tap/standpipe, tubed/piped well, borehole, protected dug well, and bottled water. Unimproved water: Drinking water from unprotected wells, springs, rainwater collection, surface water, and alterna- tive methods such as carts with small tanks or tanker-trucks. The lower proportion of households having access to piped water in 2019/20 was largely due to the expansion of Phnom Penh city and the change of question. Figure 4.3B: Basic sanitation: Flush (or pour-flush) toilets connected to sewerage, septic tank, pit, or elsewhere, and pit latrine with slab. Unimproved sanitation: Pit latrine without a slab, open pit, latrine overhanging field/water. Urban = Phnom Penh and other urban areas. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 45 Access to basic sanitation also improved, Cambodians increasingly own communication particularly in rural areas. Over the last decade, most devices, transport assets, and some household progress occurred in the form of improved sanitation appliances. In 2019, 93 percent of Cambodian systems, wider use of flush (or pour-flush) toilets, and households owned at least 1 mobile phone, up from 83 improved pit latrines. However, 16 percent of Cambodian percent in 2014. Mobile phone ownership increased most households lack any sanitation facility. This is most acute for poor (14 pp) and rural households (11 pp), but slightly in rural areas where 21 percent have no sanitation facility lag ownership rates for non-poor and urban households. and 2 percent rely on unimproved facilities. These rates Similarly, though still lower for poor and rural households, rise to 34 percent and 2 percent, respectively, for poor ownership of motorcycles increased by about 20 pp for households. In comparison, only 7 percent of households both poor and rural households. Ownership of household in urban areas lack access to basic sanitation. appliances including electric fans, stoves and refrigerators increased for all households. In turn, ownership of more traditional electronics is falling. Figure 2.4 Household Ownership of Assets, 2014–2019, Percentage 100 93 50 83 83 84 80 40 66 68 66 61 60 54 56 30 44 40 33 20 24 20 25 23 16 18 20 10 9 10 9 8 7 5 52 43 6 3 1 21 5 2 3 0 0 -20 -10 -40 -20 er o cle ne eo ) ne on r n r sh ne ne r cle s) n eo te Ca to iro fa di ga ay di ho hi hi ho er isi pu ra cy cy Ra vid c ac ac c/ pl c St lev ge e tri lep llp Bi or m tri e/ llit tri m m D ec Co fri ot Ce Te ec ur Te te lec DV Re g ng M El El ict Sa in (e o/ hi w p de e as a( Se ov er W Vi St m Ca 2014 2019/20 Change (percentage points - right axis) Source: CSES 2014 and 2019/20. Ownership of agricultural implements remained highlighting an increase in agricultural mechanization unchanged between 2014 and 2019. Most Cambodian and shift from animal-drawn equipment. However, households possess small agricultural hand tools such as ownership of large agricultural equipment remains limited harrows, rakes, spades, and hoes. Ownership of other in Cambodia. Consistent with reliance on agricultural agricultural implements is much less common. Ownership livelihoods, ownership of agricultural implements is higher of carts and ploughs— main tools for land preparation— in rural than urban areas. The differences between poor declined, while ownership of hand tractors increased, and non-poor households are less pronounced. 46 CAMBODIA POVERTY ASSESSMENT Table 2.1 Percentage Change in Asset Ownership, 2014–2019     Cambodia Area Poverty       Urban Rural Non-poor Poor Radio -16 -16 -16 -18 -13 Television 2 -11 2 1 -3 Telephone -4 -5 -4 -5 -3 Cellphone 10 2 11 8 14 Video/DVD player -21 -28 -19 -23 -12 Stereo -3 -10 -1 -3 0 Small appliances Camera(picture/video) -1 -3 0 -1 0 Satellite dish 6 3 8 6 4 Computer 2 -4 1 2 0 Sewing machine 0 -2 -1 -1 -1 Electric iron 6 -13 4 5 0 Electric fan 40 8 46 38 43 Refrigerator 13 8 8 14 1 Large appliances Stove (electric/gas) 31 10 30 32 19 Washing machine 6 7 3 7 1 Bicycle -7 -4 -6 -7 -5 Transportation Motorcycle 17 6 19 15 20 assets Car 6 4 3 6 0 Source: CSES 2014 and 2019/20. Figure 2.5 Ownership of Agricultural Implements, 2014–2019, Percentage 100 95 100 90 80 80 60 60 40 40 20 14 11 13 14 10 15 20 2.5 2 0.2 0.1 2 2 1 1 0.4 1 0 0 -20 -20 gh rt w r ill ne r p or lle to m m Ca rro ct hi ou ac /ro pu ce tra ac Ha Pl Tr er Ri er m nd oz at g Ha lld W in Bu sh re Th 2014 2019/20 Change (percentage points - right axis) Source: CSES 2014 and 2019/20. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 47 Livestock ownership declined, except in urban households from 80 to 64 percent. While household areas where poultry ownership increased. Between ownership declined, livestock owners increased their 2009 and 2019, the proportion of households owning number of livestock. Small poultry raised for food such as livestock declined 18 pp nationally and 12 pp in rural chicken, ducks, and quails represents about 90 percent areas. Livestock ownership also fell 16 pp for poor of total animals owned. Figure 2.6 Ownership of Livestock, 2009–2019, Percentage A. Livestock ownership 100 82 80 80 72 70 68 67 63 64 60 56 54 54 52 40 27 21 20 14 0 Cambodia Urban Rural Non-poor Poor Area Poverty 2009 2014 2019/20 B. Number of livestock owned 25 23.4 Average number of livestock owned 20.9 20 19.6 14.5 15.2 15 13.2 13.3 12.8 12.5 10.3 10.3 10.1 10.5 10 5 3.9 2.2 0 2009 2014 2019 2009 2014 2019 2009 2014 2019 2009 2014 2019 2009 2014 2019 Cambodia Urban Rural Non-poor Poor Cattle Pig Chicken Buffalo Horse Sheep Goat Duck Quail Source: CSES 2009, 2014, 2019/20 Note: Urban = Phnom Penh and other urban areas. 48 CAMBODIA POVERTY ASSESSMENT 2.2 Human Capital Trends Cambodia sustained high gross and net primary enrollment rate to about 90 percent by 2019. A stagnant school enrollment rates between 2009 and 2019. gross enrollment rate (or slightly decreasing for rural Gross primary enrollment rates remained above 100. areas) combined with rising net enrollment suggests While gross enrollment remained marginally unchanged, more children are enrolled in classes for their age and net enrollment increased 8 pp nationally, 5 pp in urban decreased grade repetition and student drop out. areas and 8 pp in rural areas bringing primary school age Figure 2.7 Gross and Net Enrollment Rates, 2009 and 2019/20, Percentage A. Gross enrollment rate 120 113 109 106 106 109 108 100 80 69 62 57 54 60 47 43 40 41 40 30 25 18 20 13 0 Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Primary Lower secondary Upper secondary 2009 2019/20 B. Net enrollment rate 100 90 92 89 87 82 81 80 60 52 52 47 45 38 40 40 31 30 27 24 17 20 13 0 Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Primary Lower secondary Upper secondary 2009 2019/20 100 90 92 82 85 86 78 80 61 57 60 47 43 37 40 31 33 30 24 17 20 8 8 0 Cambodia Non-poor Poor Cambodia Non-poor Poor Cambodia Non-poor Poor Primary Lower secondary Upper secondary 2009 2019/20 Source: CSES 2009 and 2019/20. Note: Urban = Phnom Penh and other urban areas. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 49 Children’s secondary school enrollment improved, in enrollment occured in rural areas, narrowing the particularly in rural areas, but transition from urban-rural enrollment gap. Yet, fewer rural children are primary to secondary school remains far from enrolled in secondary education, with a larger gap in complete. Around 4 in 10 children fail to progress from upper secondary school enrollment. Secondary school primary to lower secondary education. Dropout rates are enrollment also continues to differ by poverty status; while high during the transition from primary to lower secondary an increased share of poor children is enrolled in lower and upper secondary school. About 5 in 10 children of secondary school, their enrollment trails that of children lower secondary school age and 3 in 10 children of upper from wealthier households. secondary school age are enrolled. Most improvements Figure 2.8 Literacy and Educational Attainment of Adults 15+, 2009-2019, Percentage A. Literacy of adults 15+ 100 89 89 88 81 83 77 78 80 80 74 77 73 70 69 64 66 60 40 20 0 Cambodia Urban Rural Non-poor Poor Area Poverty 2009 2014 2019/20 B. Educational attainment of adults 15+ 100 4 7 7 6 10 10 9 12 12 12 13 26 80 14 28 19 17 30 26 31 18 29 30 30 33 60 31 31 31 35 32 39 30 36 36 39 40 33 33 36 30 32 31 28 30 20 23 19 24 26 31 28 26 22 19 22 19 19 17 16 10 9 10 14 0 2009 2014 2019 2009 2014 2019 2009 2014 2019 2009 2014 2019 2009 2014 2019 Cambodia Urban Rural Non-poor Poor No schooling Some primary Completed primary Completed lower secondary Completed upper secondary Completed vocational training University degree Source: CSES 2009, 2014, 2019/20. Note: Urban = Phnom Penh and other urban areas. 50 CAMBODIA POVERTY ASSESSMENT Adult literacy rates rose gradually between 2009 from 17 to 25 percent and 8 to 15 percent, respectively. and 2019, particularly for rural adult Cambodians. Still, adult education achievement is biased toward urban Nationally, adult literacy increased from 74 percent in 2009 areas, where 34 percent have a minimum lower secondary to 81 percent in 2019. Literacy in rural areas improved from education. Given educational gains among the non-poor, 70 percent in 2009 to 77 percent in 2019, helping narrow the educational attainment gap between poor and non-poor the urban-rural gap from 19 pp to 11 pp. Adult literacy Cambodians increased, with about 1 in 4 poor Cambodian also improved among poor Cambodians but not fast adults still not receiving any schooling. enough to outpace literacy gains among the non-poor. As Younger generations are more educated and literate such, the 14-pp literacy gap between poor and non-poor than older Cambodians and have attained gender Cambodians did not change between 2009 and 2019. parity in primary schooling and literacy. Expanded Slightly more Cambodian adults now attain higher access to basic education and improved gender parity in levels of education, although 8 in 10 adults attain primary education has contributed to high literacy rates at most primary education. Nationally, between 2009 among the younger generation. Few young Cambodians and 2019, the proportion of adults aged 15 and older have never attended school. In contrast, older generations with no education dropped from 22 to 16 percent, and have lower literacy and higher proportion of older people the proportion with at least lower secondary education with no schooling. Older women are significantly more increased from 14 to 21 percent. Gains were more evident disadvantaged in schooling and literacy skills compared among non-poor adults and adults in rural areas where the to older men. proportion with at least lower secondary education increased Figure 2.9 Age-specific Literacy Rate and Deprivation of Schooling, 2019, Percentage A. Literacy B. No schooling 100 60 Percent of population without schooling 80 50 Percent of population 40 60 30 40 20 20 10 0 0 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65+ Age group Age group Male Female Cambodia Male Female Cambodia Source: CSES 2019/20. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 51 Cambodia’s human capital indicators are moderately “Deployment” subindex (4th) due to its high degree of human low by international comparison. Cambodia has capital utilization, but trails on the “Know-how” subindex a Human Capital Index (HCI) score of 0.49 in the 2020 (121st) due to a very low share of high-skilled employment World Bank Human Capital Project, indicating the score to and skilled employees. Despite active participation in the improve effectiveness of policy implementation amidst the workforce, Cambodian skill sets and opportunities for high- increase of education spending over the last few years. 30 skilled work are limited. Secondary and higher educational Cambodia’s HCI standing would benefit from improvement attainment remains low, as evidenced by the weak in learning outcomes, expected years of schooling and performance on the “Capacity” subindex. Current efforts child health (stunting). Students in Cambodia achieve to educate and develop student skills are constrained by a harmonized test score of 452 on a scale where 652 poor quality primary schools, low secondary and tertiary represents advanced attainment and 300 represents enrollment, and limited skills diversity of graduates, as minimum attainment. This score is marginally higher than evidenced by the “Development” subindex score. the average score for East Asia and the Pacific’s 432. Better performing ASEAN economies have stronger However, a child in Cambodia can expect to complete educational attainment, a solid outlook for building 9.5 years of schooling by her 18th birthday compared to future human capital, and skill-intensive utilization an average of 11.9 years for children in East Asia and the of human capital. Upper-middle income countries, Pacific. Further, 22 percent of Cambodian children under Malaysia and Thailand maintain high skill-intensive age 5 suffer stunting placing Cambodian children at risk utilization of human capital, as evidenced by strong for lifelong cognitive and physical limitations. performance on the “Know-how” subindex. Thailand The 2017 World Economic Forum Human Capital has complemented skills utilization with high deployment Report ranked Cambodia 92 out of 130 countries of the labor force, while Vietnam has built its education included in the HCI.31 Cambodia performs well on the stock while boosting future human capital potential. Figure 2.10 Human Capital Index Human Capital Index 100 80 60 40 Know-how Capacity 20 0 Development Deployment Cambodia Lao PDR Myanmar Vietnam Thailand Source: World Economic Forum Global Human Capital Report, 2017. 30 The World Bank HCI measures the contribution of current health and education outcomes to the productivity of future workers. 31 The World Economic Forum HCI measures a country’s current human capital, current investments, and current outcomes in the labor market. 52 CAMBODIA POVERTY ASSESSMENT Child malnutrition declined in Cambodia over 2 in 2021/22. Stunting reflects cumulative growth deficits decades, but chronic undernutrition remains severe, stemming from inadequate nutrition over a prolonged particularly in rural areas. The proportion of children period, usually associated with poor living conditions, under age 5 with nutritional deficits has been decreasing frequent illness, poor feeding practices, and poor for the last 20 years, but too many of these children are maternal health and nutrition. The problem is most acute still stunted (considered too short for their age). Nationally, in rural areas, where 25 percent of children are stunted, 22 percent of children under age 5 suffered from stunting compared with 17 percent of urban children. Figure 2.11 Child Health Status, 2005–2021/22 A. Nutritional status of children under 5 B. Nutritional status of children under 5 in 2021/22 50 50 43 40 40 40 32 Percentage Percentage 30 28 28 30 25 24 22 22 19 20 20 16 17 16 12 10 10 11 8 10 10 10 10 8 0 0 2005 2010 2014 2021/22 Cambodia Urban Rural Stunted Wasted Underweight Stunted Wasted Underweight D. Vaccinations by age 12 months C. Child mortality among children 12–23 months Percent of children ages 12–23 months 100 90 83 80 80 Deaths per 1,000 live births 79 76 73 70 66 67 60 60 54 50 45 40 40 35 30 28 20 20 16 10 12 7 4 0 0 2 3 2005 2010 2014 2021/22 2005 2010 2014 2021/22 Infant mortality Under-5 mortality All vaccines (basic antigens) No vaccinations Source: CDHS 2005, 2010, 2014, and 2021/22. Note: “All vaccines” includes receiving a vaccination against tuberculosis (BCG), 3 doses each of vaccinations against diphtheria, pertussis, and tetanus (DPT) and polio; and a vaccination against measles by age 12 months. Child survival has improved due to improved and Health Survey (CDHS), under-5 mortality fell from 83 maternal health and increased vaccination coverage deaths to 16 deaths per 1,000 live births between 2005 against 6 vaccine-preventable diseases. Between and 2021/22, while infant mortality fell from 66 deaths 2005 and 2021/22, under age-5 mortality (the probability to 12 deaths per 1,000 live births. These estimates are of dying between birth and age 5) decreased, along with better than average under-5 and infant mortality rates infant mortality (the probability of dying between birth in lower-middle income countries (49 and 37 deaths per and age 1). According to the Cambodia Demographic 1,000 live births, respectively), but slightly worse than the CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 53 averages in upper-middle income countries (13 and 11, female mortality rate. Women tend to live longer than men respectively). Because most children reside in rural areas, worldwide in part due to inherent biological advantages Cambodian national under-5 and infant mortality rates but also due to behavioral differences, as men tend to reflect rural rates, with urban under-5 and infant mortality engage in more unhealthy and risky behaviors.32 at least 20 pp and 10 pp lower. Following massive The maternal mortality ratio declined. In the 2021/22 expansion in vaccination coverage between 2000 and Cambodia Demographic and Health Survey (CDHS) 2005, fewer Cambodian babies are not vaccinated. In estimated 154 maternal deaths per 100,000 live births for 2014, infants receiving all vaccines declined because of the 6-year period preceding the survey. A comparison of reduced vaccination coverage against polio and measles, maternal mortality ratios from the 2005 to 2021/22 shows which has since recovered. a large decline between 2005 and 2010 and marginal Adult mortality decreased. Between 2005 and 2014, decline from 2010 to 2021/22 (as the overlapping adult mortality fell from 3.1 to 2 deaths per 1,000 confidence intervals in Figure 2.12 B indicate).33 Declining population in women ages 15–49, and from 5.2 to 3.5 maternal mortality suggests a decline in obstetric risk deaths per 1,000 population in men ages 15–49 (Figure associated with each live birth, often thanks to improved 2.12 A). The male mortality rate continues to exceed the obstetric care. Figure 2.12 Health Status of Adults (ages 15–49), 2005–2021/22 A. Adult mortality per 1,000 individuals B. Maternal deaths per 100,000 live births 6 5.2 700 5 600 605 4.1 4 3.5 500 472 3.1 400 3 2.5 338 2.0 300 288 2 246 239 200 206 170 154 1 100 124 95 69 0 0 2005 2010 2014 2005 2010 2014 2021/22 Female Men Source: CDHS 2005, 2010, 2014 and 2021/22. Women’s access to maternal health care during expanded during this period. The percentage of mothers pregnancy, childbirth, and postnatal periods receiving antenatal care from a skilled provider increased expanded, especially for delivery. Obstetric care 26 pp from 69 percent to 95 percent, and the percentage by a skilled provider during delivery is crucial for of mothers receiving their first postnatal care from a skilled reduction of maternal and neonatal mortality as trained provider increased by 46 pp from 41 percent to 87 percent. personnels can identify and manage childbirth and The timing of the postnatal checkup also improved, which postnatal complications. Between 2005 and 2014, births is equally critical as most maternal and neonatal deaths attended by trained providers increased significantly, with occur within the first 48 hours after delivery. In 2014, 90 percentage of births delivered by a skilled provider (doctor, percent of mothers received postnatal care in the first 2 nurse, midwife) doubling over the decade 2005 to 2014 days after giving birth, an increase of 26 pp since 2005. from 44 to 89 percent. Antenatal and postnatal care also 32 W  orld Health Organization (2020): “World health statistics 2020: monitoring health for the SDGs, sustainable development goals.” Geneva: World Health Organization. 33 At times, more recent CDHS data are not incorporated since disaggregated data are not yet available. 54 CAMBODIA POVERTY ASSESSMENT Figure 2.13 Access to Maternal Health Care and Child Health Outcomes, 2005–2021/22 A. Access to maternal health care for women ages 15–49 Percent of women who had a live birth in the five years preceeding the survey 100 95 97 99 95 95 98 98 89 88 89 88 90 89 89 79 80 71 74 69 68 70 70 67 67 64 62 62 60 57 44 39 40 38 34 32 31 32 28 26 20 0 Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Antenatal care received from a Births delivered by a skilled provider Postnatal checkup received in skilled provider first 2 days after birth 2000 2005 2010 2014 B. Access to maternal health care C. Nutritional outcomes of children under-5 by wealth quintile, 2014 by wealth quintile, 2021/22 Percent of women who had a live birth in the five years preceeding the survey 100 40 Percent of children under age 5 30 80 20 10 60 0 Lowest Second Middle Fourth Highest Lowest Second Middle Fourth Highest wealth wealth wealth wealth quintile quintile quintile quintile Antenatal care from a skilled provider Birth delivered by a skilled provider Stunted Underweight Wasted Postnatal care in first 2 days after birth Source: CDHS 2005, 2010, 2014 and 2021/22. Note: Antenatal and postnatal care: Skilled provider includes doctor, nurse, and midwife. Delivery: Skilled provid-er includes doctor, nurse, midwife, or auxiliary nurse/midwife. Note: Antenatal and postnatal care: Skilled provid-er includes doctor, nurse, and midwife. Delivery: Skilled provider includes doctor, nurse, midwife, or auxiliary nurse/midwife. More recent CDHS data on access to maternal health are available but are not incorporated here due to changes in definitions that make comparisons of indicators across time inconsistent. See CDHS 2021/22 for more recent 2021/22 estimates. Due to strong expansion efforts in rural areas, socioeconomic status. Obstetric care remains favorably Cambodia narrowed the urban-rural gap in access to biased toward urban women: urban women are more likely maternal health care. The gap between urban and rural than rural women to receive skilled attendance at delivery areas in births delivered by a skilled provider, for instance, (98 vs. 88 percent), postnatal care (98 vs. 89 percent), narrowed from a 31-pp gap in 2005 to a 10-pp in 2014. and antenatal care with a diagnostic component. Similarly, wealthier and more educated women have better access Despite notable gains, Cambodians’ access to to obstetric care than poorer and less educated women. health care and their health outcomes continue This uneven access contributes to poorer maternal and to differ depending on where they live and their CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 55 child health outcomes among low-income and rural be a main barrier for seeking medical care and treatment. households, who also tend to be less educated: Children Rural women are more likely to face cost constraints than under age 5 with less educated mothers and from poorer urban women. Cambodian women also perceive not households have worse nutritional and health outcomes having someone to accompany them to a health facility than those with better educated mothers and from as a challenge to accessing health care. The proportion wealthier households. of women who report having problems with accessing health care decreases with increasing education and Cambodian women continue to face barriers wealth, with 87.8 percent of women in the bottom 20 accessing health care, largely due to cost. According percent of the wealth index reporting difficulty accessing to the 2014 CDHS, 69 percent of women perceive costs to health care, 34.3 pp higher than the top 20 percent.34 Figure 2.14 Perceived Challenges in Accessing Health Care among Women (15–49), Percentage 80 69 64 60 50 45 40 45 40 35 27 24 20 21 14 11 0 Getting permission to go Getting money for Distance to health facility Not wanting to go alone for treatment treatment Urban Rural Cambodia Source: CDHS 2014. 34 N  early 7 in 10 ill or injured Cambodians who need medical treatment visit private medical sector. The utilization of private medical sector increases significantly from 48.2 percent in 2005 to 67.1 percent in 2014 (CDHS, 2005 & 2014). The recent CSES shows similar trends. More than 70 percent of household members who seek medical treatment visit private medical sector. Between 2009 and 2019/20, the share of private medical sector visitors increased from 56 percent to 71 percent. The poor are likely to use public medical providers than the non-poor due to the lower cost of medical treatment compared to private medical providers. 56 CAMBODIA POVERTY ASSESSMENT 2.3 Conclusion Rising incomes and improved service delivery nutrition, or health. Cambodians today are also more supported gains in many aspects of wellbeing, educated and more literate. Considerable improvements especially for low-income and rural households. in rural areas have helped narrow Cambodia’s urban-rural Consequently, we would expect that the proportion of gap in poverty, standards of living, health, and education Cambodians experiencing multiple and simultaneous outcomes. kinds of poverty has fallen over the decade. The proportion Despite notable progress on health, Cambodia’s of Cambodians living below the national poverty line has overall progress in human capital development has markedly declined over the past decade. During the same been slow. Still, many children are stunted. While overall period, the average number of non-monetary deprivations school enrollment has risen across all levels, children are Cambodians suffer also declined. Cambodians today not transitioning from primary to secondary education have better housing conditions, better access to in large numbers, reflecting high dropout rates linked to electricity and improved water and sanitation facilities poor performance and economic pressures. Low-income than a decade prior. They are more connected because of and rural households remain at a greater disadvantage better communications and transport infrastructure and in accessing basic services and earning opportunities, assets. Women are healthier and less likely to die pre- resulting in poorer standards of living. In 2019, poverty maturely during and after childbirth. Children are healthier was twice as prevalent in rural areas than in urban areas and less likely to die from poor standards of living, and about 80 percent of the poor lived in rural areas. CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 57 58 CAMBODIA POVERTY ASSESSMENT CHAPTER 3 POVERTY AND INEQUALITY PROFILE This chapter overviews poverty in Cambodia in (4.2 percent) and other urban areas (12.6 percent), and 2019/20. The chapter presents estimates of poverty and highest in rural areas (22.8 percent). Households with inequality and the distribution of the poor across area of many members, high child dependency ratios, female residence and regions. The chapter also examines the heads, and less-educated heads are more likely to be profile of poor households against non-poor houesholds, poor. Households with agricultural wage as their main further examining the incidence of non-monetary well- source of income are the poorest (39.5 percent), making being such as educational attainment, labor market up 12 percent of the poor nationwide. Households with participation, housing conditions, and access to basic non-agricultural wage as their main source of income are services across poverty status and geographic area. poor (18.8 percent), but they make up 52 percent of the poor nationwide. Poor households have limited access to Under the new national poverty line, about 17.8 improved water, improved sanitation, and electricity. percent of the population (2.8 million people) was poor in 2019/20. Poverty is lowest in Phnom Penh 3.1 Poverty, Inequality, and Vulnerability Nearly 1 in 5 Cambodians live in poverty, most of Plateau and Mountains region have the highest incidence whom reside in rural areas. About 17.8 percent of of poverty at 23.8 percent and 22.5 percent, respectively. the population in Cambodia (2.8 million) lived below This compares to a poverty incidence of 17.8 percent the national poverty line of KHR10,951 in 2019/20 in Plains and of 8.5 percent in the Coastal region. (Figure 3.1). Poverty is highest in rural areas and lowest Phnom Penh has the lowest incidence of poverty at 4.2 in Phnom Penh. The poverty headcount in rural areas percent. Food poverty35 is rare in Cambodia, with only 6 (22.8 percent) is twice the poverty headcount in urban in 1,000 Cambodians are unable to meet minimum food areas (9.6 percent). Rural livelihoods depend largely on consumption needs, and does not vary much across area less productive subsistence activities. Tonle Sap and the of residence (Figure 3.2). 35  n individual lives in food poverty when his/her total consumption and expenditure falls below the food poverty line defined as the cost of A buying food that meets the minimum nutritional requirements of 2,200 kilocalories (Kcal) per person per day. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 59 Figure 3.1 Poverty Rate Figure 3.2 Food Poverty Rate 30 2.0 Percent of population Cambodia Cambodia Percent of population 1.5 20 1.0 10 0.5 9.6 22.8 4.2 17.8 23.8 8.5 22.5 0.4 0.8 0.3 0.4 0.7 0.5 1.4 0 0.0 Urban Rural Phnom Plains Tonle Coastal Plateau Urban Rural Phnom Plains Tonle Coastal Plateau Penh Sap and Penh Sap and Mountains Mountains Area Region Area Region Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Poverty headcount rates are higher in the also have high poverty rates. The poverty rates also vary predominantly rural northwestern region of Tonle Sap widely within districts. For example, the Battambang and northeastern region of Plateau and Mountains province of Tonle Sap has a poverty rate of 23.7 percent, (Figure 3.3 maps). The poverty rate is the highest in the with district-level rates ranging from 11.7 percent (Krong northeast Ratanak Kiri province (29.6 percent). Mondul Battambang) to 28.8 percent (Rukhak Kiri). Commune- Kiri (28.1 percent), Stung Treng (27.3 percent), and Preah level poverty rates show even larger variations, from 4.9 Vihear (25.3 percent) provinces in Plateau and Mountains percent (Svay Pao) to 48.6 percent (Kaoh Chiveang). Poverty Headcount at the Figure 3.3  Figure 3.4 Poverty Density Commune Level Source: CSES 2019/20 and General Population Census of Source: CSES 2019/20 and General Population Census of Cambodia 2019. Cambodia 2019. Note: Preliminary results. Note: Preliminary results. Each dot represents 1000 poor individuals. Although the poverty rate is higher in northeast (Figure 3.4). Dense populations in the southeast, and Cambodian provinces, the number of poor people Phnom Penh in particular, give rise to high poverty density predominantly concentrate in the more densely (number of poor per geographic location) despite relatively populated southeast and northwest provinces low poverty headcounts. In contrast, the northwest 60 CAMBODIA POVERTY ASSESSMENT provinces are relatively poor causing high poverty density. gap index measures the extent to which individuals fall Provinces in the country’s northeast are less populated, below the poverty line, as a percentage of the poverty which explains why some of the poorest provinces have line. The poverty gap index can indicate the minimum relatively low poverty density. As a result, these locations budget required to eliminate poverty for a given year, require different poverty-reduction targeting approaches, as a percentage of the poverty line, assuming transfers with poverty alleviation efforts focusing on areas with high are perfectly targeted (Figure 3.5). About KHR139,899 poverty rates and density. (US$34) per person per year in transfers would be needed to eliminate poverty, amounting to KHR2.23 trillion The depth and severity of poverty is also highest (US$547 million or 2 percent of GDP). 36 The severity of in the predominantly Tonle Sap and Plateau poverty (or squared poverty gap) reflects both how far the and Mountains regions with the highest poverty poor are from the poverty line and consumption inequality incidence. In Tonle Sap and Plateau and Mountains, among the poor. In 2019/20, the poverty severity index in the poverty gap index is about 5 percent compared to 2019/20 was 1.1 percent, lower than in 2014 and 2009 a national poverty gap index of 3.5 percent. The poverty (Figure 3.6). Figure 3.5 Poverty Gap Figure 3.6 Poverty Severity 2.5 6.0 Poverty severity index Cambodia 2.0 Cambodia Percent of poverty line 5.0 4.0 1.5 3.0 1.0 2.0 0.5 1.0 1.6 4.7 0.8 3.3 4.7 1.6 5.2 0.5 1.5 0.2 1.0 1.5 0.5 1.8 0.0 0.0 Urban Rural Phnom Plains Tonle Coastal Plateau Urban Rural Phnom Plains Tonle Coastal Plateau Penh Sap and Penh Sap and Mountains Mountains Area Region Area Region Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Inequality in Cambodia is moderate compared and other urban areas is significantly higher than rural with neighboring countries, but varies across area areas—and thus poverty is lower—urban consumption is of residence and regions. The Gini index was 32.2 highly unequal. Across regions, the Gini index ranges from percent for Cambodia in 2019/20 (Figure 3.7). Inequality 28.5 percent in Plains to 34.5 percent in Phnom Penh is highest in Phnom Penh and lowest in rural areas, with (Figure A.3). Inequality in Cambodia is lower compared to the Gini index estimated at 34.5 percent in Phnom Penh, neighboring countries, with the Gini indices estimates for 30.8 percent in other urban areas, and at 28 percent in Lao PDR, Thailand, and Vietnam exceeding Cambodia’s rural areas. While average consumption in Phnom Penh index by between 3.5 to 6.6 pp (Figure 3.8). 36 Average exchange rate for 2019/20 period is 4077. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 61 Figure 3.7 Inequality Figure 3.8 Inequality by Countries 40 40 Gini index (0 -100) 30 30 Gini index (0 -100) 20 20 10 10 32.2 34.5 30.8 28.0 32.2 38.8 30.0 37.0 35.7 0 0 Cambodia Phnom Penh Other urban Rural Cambodia Lao PDR Myanmar Thailand Vietnam (2019/20) (2018/19) (2017) (2018) (2018) Source: CSES 2019/20. Source: CSES 2019/20. WB (2019a), and CSO, UNDP and WB (2020). Inequality based on household survey data may be The bottom 40 percent of Cambodian households underestimated. Cambodia’s average annual real GDP spend 2.5 times less than those in the top 60 percent. per capita growth rate was 5.4 percent over 2009 to In 2019/20, national average daily per capita consumption 2019, but per capita consumption based on the CSES measured at Phnom Penh prices was KHR21,053. grew only 2.9 percent, 2.5 percentage points (pp) lower Households in the bottom 40 percent of the consumption than GDP per capita growth. This may be because distribution spent, on average, KHR11,155 per capita household surveys in all countries often fail to capture per day compared to KHR27,815 households in the top top income households, thus potentially underestimating 60 percent spent (Figure 3.9). Consumption differences inequality. Note that gaps between private consumption between the bottom 40 and top 60 percent are larger among in national accounts and household surveys exist in all urban (2.7 times) compared to rural households (2.2 times). countries; some differences are to be expected due to definitional differences, but part of the gap is likely due Rural per capita consumption is lower than in urban to the surveys missing some top incomes.37 This is not a areas across the consumption distribution. This flaw in the household survey which was not designed to indicates that regardless of where a poverty line is set, capture the very top of the income distribution. Instead, there is more poverty in rural areas than in urban areas. some researchers have used income tax microdata to Per capita consumption in rural areas is lower across the better quantify the number of richer households and more entire consumption distribution, as evidenced by the 2 accurately estimate inequality.38 distributions not intersecting. In other words, poverty incidence rates are higher in rural than in urban areas, no matter where the poverty line lies (Figure 3.10). 37 Pryddz, Jolliffe, Serajuddin (2021) policy research working paper “Mind the Gap” (available on the PRWP website). 38 S  ee Lustig (2019), Ravallion (2022), Piketty (2003), Piketty and Saez (2003), Atkinson and Piketty (2007, 2010) and Blanchet, Flores, and Morgan (2022). 62 CAMBODIA POVERTY ASSESSMENT Average Consumption Per Capita Per Day (‘000 KHR) Figure 3.9  50 40 Thousands (KHR) 30 20 10 0 Cambodia Urban Rural Phnom Plains Tonle Sap Coastal Plateau and Penh Mountains Area Region Bottom 40 Top 60 Cambodia Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Figure 3.10 Cumulative Distribution of Per Capita Consumption by Urban and Rural Areas Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. About 15 percent of Cambodians are near-poor and A negative shock—such as a price hike, unemployment, at risk of falling into poverty in the event of shock. illness, or natural disaster—that reduces per capita daily The 15 percent of Cambodians considered near-poor consumption by KHR2,723 (US$0.70) will nearly double are non-poor but are vulnerable to falling into poverty the poverty rate. Chapter 5 illustrates the negative effects as their daily per capita consumption lies between the of COVID-19 on poverty. poverty line and 1.25 times the poverty line (Figure 3.11). CHAPTER 3. POVERTY AND INEQUALITY PROFILE 63 Share of Near-poor Figure 3.11  25 Cambodia Percent of population 20 15 10 5 10.2 18.0 6.7 16.8 18.0 11.0 14.8 0 Urban Rural Phnom Plains Tonle Sap Coastal Plateau and Penh Mountains Area Region Source: CSES 2019/20. Note: Near-poor = those whose daily per capita consumption lies between the poverty line and 1.25 times the poverty line. Urban = Phnom Penh and other urban areas. Rural households are more vulnerable to falling Near-poor and poor households differ in some into poverty than urban households. Rural residents characteristics, but often not substantially. Per capita are more likely to cluster just above the poverty line than consumption of near-poor households exceeds that of urban residents. About 18 percent of rural residents have poor households by only 4 percent (Table 3.1). In terms daily per capita consumption between the poverty line of demographic characteristics, near-poor households and 1.25 times the poverty line, compared to 10 percent are smaller in size than poor households because they of urban residents. This suggests an adverse shock to have fewer children. Relative to poor households, near- incomes, and thus consumption, will push many more poor households are also more likely to be headed by rural Cambodians into poverty than urban residents. someone who completed primary school, own durable assets, and have access to basic infrastructure such as improved water, improved sanitation, and electricity. 64 CAMBODIA POVERTY ASSESSMENT Table 3.1 Characteristics of Poor, Near-poor and Non-poor Households 2019/20 t-statistics Non-poor Poor Near-poor secure (1) (2) (3) (1) vs (2) (2) vs (3) Household size 5.6 5.0 4.0 8.97*** 19.18*** Children <15 years 2.0 1.7 1.1 7.92*** 20.38*** Adult 15-64 years 3.2 3.0 2.7 4.17*** 6.40*** Elder > 64 years 0.3 0.3 0.3 0.32 2.47** Dependency ratio 1.0 0.8 0.6 3.83*** 14.07*** Age of household head (years) 48.8 49.0 48.7 0.12 -0.24 Female household head (percent) 24.1 22.4 20.8 0.91 1.56 Highest level of household head educational attainment (percent) No education 32.4 26.1 13.7 5.20*** 10.68*** Some primary 45.1 45.1 35.2 -1.09 7.02*** Completed primary 19.0 24.2 31.5 -3.44*** -5.23*** Completed lower secondary 3.0 3.4 11.5 -1.2 -8.54*** Completed upper secondary 0.5 0.9 2.5 -1.86* -3.94*** Vocational training 0.0 0.2 0.8 -1.49 -2.68*** University 0.0 0.1 4.9 -1.05 -7.67*** Household income per capita per month (‘000 KHR) Agricultural self-employed 59.8 75.3 117.0 -1.48 -1.51 Non-agricultural self-employed 33.3 52.6 260.6 -4.13*** -8.53*** Agricultural wage 43.2 39.0 18.8 0.84 5.88*** Non-agricultural wage 204.6 221.7 326.5 -2.58*** -7.18*** Remittances 15.6 21.1 32.5 -3.20*** -3.80*** Other income 3.7 4.8 14.9 -1.076 -4.97*** Total income 360.1 414.5 770.4 -4.68*** -9.55*** Basic infrastructure (percent) Access to improved water 69.7 72.1 83.1 -1.71* -10.64*** Access to improved sanitation 61.6 73.8 85.2 -7.27*** -11.51*** Access to electricity 73.3 81.2 88.9 -5.40*** -9.00*** Durable index -1.55 -1.02 0.48 -12.60*** -21.72*** Source: Authors’ calculation based on CSES 2019/20. Note: Significance level: * p<0.10, ** p<0.05, *** p<0.01; Near-poor is defined as those whose daily per capita consumption lies between the poverty line and 1.25 times the poverty line while non-poor secure as those whose daily per capita consumption is higher than 1.25 times the poverty line. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 65 Near-poor and non-poor households, however, with 51 percent for non-poor households (Figure 3.12). differ in their profiles. Near-poor households have more Near-poor households are less likely to have access to child dependents, educational attainment of household basic infrastructure, whith more than 72 percent of near- heads, and access to public services. The proportion of poor households having access to improved water and household heads who at least completed primary school improved sanitation, 10 pp below access of non-poor is 29 percent for near-poor households, compared households. Figure 3.12 Educational attainment of poor, Figure 3.13 Income share of poor, near near poor and non-poor household heads poor and non-poor households 100 3 3 100 19 24 12 80 80 42 32 57 53 60 60 Percent Percent 45 45 2 40 40 35 12 9 34 20 9 13 32 20 26 14 17 18 15 0 0 Poor Near-poor Non-poor Poor Near-poor Non-poor No education Some primary Agricultural self-employed Non-agricultural self-employed Completed primary Completed Lower secondary Agricultural wage Non-agricultural wage Completed Upper secondary Vocational training Remittances Other income University Source: CSES 2019/20. Note: Near-poor = those whose daily per capita consumption lies between the poverty line and 1.25 times the poverty line. Near-poor households also have not only lower non-poor households generate a larger share of income income per capita but also less diversified incomes from non-agricultural self-employment than near-poor than non-poor households. With average earnings households (34 percent vs. 13 percent). This implies that at KHR 414,500 (US$102) per person per month, near- non-agricultural self-employment plays an important role in poor households earn about half of non-poor households. raising total household income and reducing the likelihood Near-poor households derive more than 50 percent of of falling into poverty. Holding other factors constant, their income from non-agricultural wage work, whereas non-agricultural self-employment significantly reduces the non-poor households derive only 42 percent of their likelihood of being poor or near-poor (Table 3.2). income from such sources (Figure 3.13). Conversely, 66 CAMBODIA POVERTY ASSESSMENT Table 3.2 Correlates of Poverty Status (Multinomial logit) Marginal effects Non-poor Poor Near-poor secure Household size Children <15 years 0.057*** 0.040*** -0.097*** Adult 15-64 years 0.035*** 0.018*** -0.054*** Elderly > 64 years 0.040*** 0.032*** -0.072*** Dependency ratio 0.031** 0.005 -0.035** Age of household head 0.000 0.000 0.000 Female household head 0.039*** 0.015 -0.054*** Household head educational level (base: no education) Some primary -0.021** -0.004 0.025* Completed primary -0.065*** -0.027* 0.091*** Completed lower secondary -0.099*** -0.091*** 0.190*** Completed upper secondary -0.113* -0.064 0.177*** Main income sources (base: agricultural self-employed) Non-agricultural self-employed -0.082*** -0.047*** 0.128*** Agricultural wage 0.070*** 0.039** -0.109*** Non-agricultural wage 0.015 0.008 -0.024* Remittances -0.020 0.033 -0.013 Other income 0.000 0.016 -0.015 Basic infrastructure Access to improved water -0.025** -0.039*** 0.064*** Access to improved sanitation -0.088*** -0.027** 0.115*** Access to electricity -0.031*** -0.009 0.040*** Region (base: Phnom Penh) Plains 0.099*** 0.053** -0.151*** Tonle Sap 0.121*** 0.062*** -0.183*** Coastal 0.007 0.002 -0.009 Plateau and Mountains 0.086*** 0.015 -0.102*** Sample size (N=10,075) 1,489 1,334 7,243 Pseudo R2 0.1854 Log likelihood -6355.1795 Source: CSES 2019/20. Surprisingly, female-headed households are this as nearly 60 percent of female-headed households less likely to be near-poor than male-headed received remittances compared to 45 percent of counterparts. The proportion of near-poor households male-headed households. In addition, female-headed headed by women was slightly higher than by men households received on average KHR 49,000 (US$12) in 2009; however, this was not the case in 2019/20. per person, per month in remittances, more than double Differences in remittance receipt could partially explain that of male-headed households. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 67 3.2 Demographic and Labor Market Characteristics of Poverty Women head only 1 in 5 Cambodian households. headed households (30 percent) relative to the Coastal Female household heads are slightly more common in region (16 percent). Women head a slightly larger urban areas (24 percent) than in rural areas (20 percent) proportion of poor (24 percent) than non-poor (21 percent) (Figure 3.14). Phnom Penh tends to have more female- households. Figure 3.14 Female-headed Households Figure 3.15 Poverty Rate by Gender of Household Head 40 60 Percent of households Cambodia 30 Female Male Percent of population 40 20 10 20 0 Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor 0 Cambodia Urban Rural Married separated Widowed Single Divorced/ Poverty Area Region status Area Marital status Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Women are more likely to head poor households. Higher risk of poverty for female-headed households The poverty rate for female-headed households is 21.5 is associated with lower labor force participation percent, 4.5 pp higher than for male-headed households and educational attainment. Female household heads (Figure 3.15). The difference in poverty rate by gender are more likely to be economically inactive compared to of household head is only statistically significant in rural male household heads (21 percent vs. 6 percent). When areas, where the female-headed household poverty rate female household heads participate in the labor market, is 29.1 percent compared to 21.5 percent for male- they are more likely to be unemployed.39 Secondly, nearly headed households. It is worth noting that feminization of two-thirds of female household heads had only some poverty—defined as a widening gap in poverty between primary schooling, suggesting that most are confined to female-headed and male-headed households—was low-wage work. Households headed by a widowed female observed in rural areas and among widowed heads (see have a higher poverty headcount rate than households Box 3.1 for details). However, when examining the gap in headed by a widowed male. poverty between men and women, we only observe this feminization in urban areas. 39 A person is considered underemployed if he/she is not working but seeking employment. 68 CAMBODIA POVERTY ASSESSMENT Box 3.1 “Feminization” of Poverty in Cambodia In the last decade, there is a growing tendency for female-headed Cambodian households to live in poverty compared to male-headed households, especially in rural areas. The increase in gender gap in poverty is known as the “feminization” of poverty. The incidence of poverty among female-headed households was higher relative to male-headed households throughout the decade (Figure A.4). Second, the gender gap in poverty widened over time (Box Figure 3.1). Decomposing the gender gap in poverty by area of residence and marital status, feminization of poverty was only observed in rural areas and among widowed household heads (Figure A.6). The evidence suggests a need to target poverty alleviation efforts to female-headed households, especially those in rural areas or those that have been widowed. Feminization of poverty, when defined as the widening gap in poverty between women and men, is not apparent (Box Figure 3.2). The poverty incidence of women and men was almost identical over the decade (Figure A.5). This is because most of Cambodian women do not live-in female-headed households. Box Figure 3.1 Difference in poverty Box Figure 3.2 Difference in poverty rate between female and male rate between women and men household heads 8 2009 2014 2019/20 1.0 2009 2014 2019/20 Percentage points 6 0.5 Percentage points 4 0.0 2 -0.5 0 -1.0 Cambodia Urban Rural Cambodia Urban Rural Area Area Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Heads of poor households tend to have little to no primary schooling decreases the incidence of poverty to schooling. As in other East Asian countries, living in 21 percent. Only 12 percent of households whose head poverty is strongly associated with limited educational completed primary school live below the povery line. Having attainment of the household head, with primary school a household head who completed secondary school completion constituting a key threshold in Cambodia. reduces the risk of falling into poverty to below 10 percent. Around 32 percent of households headed by someone Overall, 77 percent of the poor live in households headed with no schooling are poor (Figure 3.16). Obtaining some by someone who had not completed primary school. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 69 Figure 3.16 Poverty Rate by Educational Attainment of the Household Head 40 Percent of population 30 Cambodia 20 10 0 g y y y y g e ar ar ar ar re lin in in im im nd nd g o de tra ho pr pr co co sc y l e ed se se na it m rs No et tio er r So ive pe pl w ca m up Un lo vo Co ed ed ed et et et pl pl pl m m m Co Co Co Source: CSES 2019/20. Household size and child dependency ratios population is relatively young, with nearly one-third being correlate with poverty. The average Cambodian children between ages 0–14. On average, Cambodian household has 4.4 members, but poor households households have 1.3 children ages 0–14, with 2.0 average 5.6 persons, 1.4 more than non-poor children in poor households and 1.2 children in non-poor households. Poverty increases with household size, and households. Consistent with poverty incidence increasing households with 7 or more members have the highest with household size, households with 3 or more children poverty incidence (36.2 percent), exceeding the national have the highest poverty headcount (36.6 percent), 18.8 average by 18.4 pp (Figure 3.17). The Cambodian pp higher than the national average (Figure 3.18). Figure 3.17 Poverty Rate by Figure 3.18 Poverty Rate by Number Household Size of Children 50 50 Cambodia Percent of population Cambodia Percent of population 40 40 30 30 20 20 10 10 0 1-2 3-4 5-6 7+ 0 Number of household size None 1 2 3+ Number of children Source: CSES 2019/20. Source: CSES 2019/20. 70 CAMBODIA POVERTY ASSESSMENT Children are more likely to live in poverty than youth compared with Phnom Penh and the Coastal region. Girls and adults. Over one-fifth (22.5 percent) of children and boys face a similar incidence, depth, and severity (ages 0–14) live in poverty, compared to 18.5 percent of of poverty. The incidence of youth poverty is 4 pp lower youth (ages 15–24) and 15.1 percent of adults (ages 25+). than the incidence of child poverty, with similar patterns Children constitute 29 percent of the total population but of regional variation (Figure 3.19). Youth represent 17 account for 37 percent of the poor. Child poverty in rural percent of the total population and 17 percent of them live areas is twice the rate in urban areas (27.2 percent vs. in poverty. Of all age groups, the poverty headcount rate 13.4 percent). Children are more likely to live in poverty is the lowest among adults (15.1 percent), even slightly in the Tonle Sap and the Plateau and Mountains regions lower than the national headcount rate of 17.8 percent. Figure 3.19 Poverty Rate by Age Group 25 Cambodia Percent of population 20 15 10 5 0 0-14 15-24 25+ Age group Source: CSES 2019/20. Figure 3.20 Child Poverty Figure 3.21 Youth Poverty Percent of children (aged 0 -14) Percent of youth (aged 15 -24) 40 40 Cambodia Cambodia 30 30 20 20 10 10 0 0 Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Area Region Area Region Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Poverty rates are higher among households headed non-agricultural self-employment (8.7 percent) (Figure by someone engaged in agricultural activities. The 3.22). Although households engaging in agricultural self- poverty headcount ratio is highest among households employment are not the poorest, they constitute 43.3 headed by agricultural wage workers (37.9 percent), percent of the poor while making up only 35.5 percent followed by those engaged in agricultural self-employment of the population. As a result, more than half of poor activities (21.8 percent). Poverty headcount is lowest Cambodians live in households headed by engaged in among households headed by someone involved in agricultural work (self and wage employment). CHAPTER 3. POVERTY AND INEQUALITY PROFILE 71 Figure 3.22 Poverty Rate by Household Figure 3.23 Poverty Rate by Household Head Employment Status Main Income Source 50 50 Percent of population 40 Percent of population Cambodia Cambodia 40 30 30 20 20 10 10 0 0 Agricultural self-employed Non-agricultural self-employed Agricultural wage Non-agricultural wage Remittances Other Agricultural self-employed Non-agricultural self-employed Agricultural wage Non-agricultural wage Economically Source: CSES 2019/20. inactive Source: CSES 2019/20. A household’s poverty status tends to be determined Labor force participation is high for both Cambodian by its source of livelihood. In developing countries, men and women. Nearly 9 in 10 Cambodians aged including in Cambodia, the household head tends to be 15–64 participated in the labor market during the 7 the household’s most important income earner, making days preceding the 2019/20 CSES (Figure 3.24). This his/her main occupation the key livelihood source. is partly due to the high female labor force participation However, the main occupation of the household head of 84 percent,compared to a regional average of 61 is no longer the household’s main income source for percent and global average of 50 percent (World Bank, the poor. 40 The poverty headcount rate estimated by 2019b). Despite high female participation, a 7-pp gender households’ main income sources is, to a great extent, gap remains. Labor force participation also varies by in line with a household head’s main occupation (Figure geographic area, with a participation rate of 89 percent 3.23). However, the share of the poor among households in rural areas, 4 pp higher than urban areas. The Plateau dependent on agricultural activities—with self-employed and Mountains region registers the highest participation agricultural or agricultural wage employment—amounts rate of 91 percent, compared to 81 percent in Phnom to 33 percent, or 19 pp lower than those who live in Penh. Moreover, youth (ages 15–24) are 20 pp less households with non-agricultural occupations as their likely to participate in the labor force than adults (ages main source of income. 25–64). In contrast, poor Cambodians are more likely to be economically active than the non-poor, suggesting that labor force participation is not sufficient by itself to stay out of poverty. That is, members of poor households cannot afford to stay idle or not work. 40 Main income is defined as an income source accounting for more than half of the household’s total income. 72 CAMBODIA POVERTY ASSESSMENT Figure 3.24 Labor Force Participation Rates 100 11 9 11 7 Percent of population aged 15-64 13 15 16 13 27 80 60 87 89 91 87 89 93 40 85 84 73 20 0 Cambodia Urban Rural Male Female Non-poor Poor 15-24 25-64 Area Gender Poverty status Age group Labor force Out of labor force Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. What distinguishes the poor from the non-poor is of residence, gender, poverty status, or age group. While not lack of jobs but lack of well-paid jobs. Among women and youth are more likely to be economically the economically active, employment rates are universally inactive than men and adults, they are more likely to high. Of the 87 percent of Cambodians aged 15–64 be unemployed. Poor individuals are equally likely to participating in the labor force, 98 percent were employed participate in the labor force, and when they participate in the 7 days preceding the CSES interview (Figure 3.25). they are equally likely to be unemployed with only 1 pp This constitutes one of the highest employment rates difference to non-poor individuals. This suggests that the among countries with similar levels of development. In poor are in lower-quality jobs with insufficient pay to lift turn, only 2 percent of active, working-age Cambodians them out of poverty. are unemployed. Employment rates differ slightly by area Figure 3.24 Percentage of Employed Among Economically Active Persons 100 Percent of active population aged 15 - 64 2 3 2 1 4 2 3 3 2 80 60 98 97 98 99 96 98 97 97 98 40 20 0 Cambodia Urban Rural Male Female Non-poor Poor 15-24 25-64 Area Gender Poverty status Age group Employed Unemployed Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 73 A significant share of the labor force is not educated, completed at least primary school than older age groups. especially women. About 42 percent of the labor force Female youth tend to be more educated than male youth has not completed primary school, of which 58 percent (Figure 3.26, Figure 3.27), reflecting outcomes of higher are female and 42 percent are male workers. Youth (ages net primary, lower-secondary, and upper-secondary 15–24) entering the labor force are more educated than school enrollment rate of girls compared to boys. older cohorts, with a larger proportion of youth having Figure 3.26 Men’s educational attainment Figure 3.27 Women’s educational in the labor force by age group attainment in the labor force by age group 100 4 100 1 13 13 9 5 3 4 20 8 16 16 7 5 12 12 Percent of age groups 80 11 21 Percent of age groups 23 22 80 12 14 24 28 60 33 33 60 46 35 37 39 45 45 40 40 41 41 30 35 20 23 20 26 20 38 20 16 19 26 7 12 11 9 0 4 3 0 15-24 25-34 35-44 45-54 55-64 15-24 25-34 35-44 45-54 55-64 No schooling Some primary No schooling Some primary Completed primary Completed lower secondary Completed primary Completed lower secondary Completed upper secondary Completed upper secondary Source: CSES 2019/20. Source: CSES 2019/20. Nearly half of Cambodian workers are in paid workers are more likely to be female. Older and female employment. In 2019/20, 47.1 percent of workers were workers are therefore more vulnerable to negative shocks in paid employment, 37.8 percent are own-account than younger and male workers respectively. Cambodian workers, 14.6 percent are unpaid family workers, and the workers are more vulnerable to economic shocks and lack remaining 0.4 percent are employers. Younger workers of entrepreneurship than other countries in the region. are more likely to be in paid employment than older The vulnerable employment rate in Cambodia is 52 workers. Sixty-two percent of youth start working as an percent, higher than the regional average of 41 percent. employee, 24 percent as an unpaid family worker, and In contrast, the share of employers in total employment 14 percent as own-account workers. Paid employment is less than 1 percent, lower than the regional average of is higher for men than for women and increases with 2.7 percent. age (Figure 3.28, Figure 3.29). In contrast, unpaid family 74 CAMBODIA POVERTY ASSESSMENT Figure 3.28 Employment type for men Figure 3.29 Employment type for women by age group by age group 100 100 Percent of employment type 16 Percent of employment type 24 17 24 25 80 29 80 45 13 56 14 32 60 67 60 47 56 40 40 62 63 63 61 51 53 20 41 20 35 29 20 12 0 0 15-24 25-34 35-44 45-54 55-64 15-24 25-34 35-44 45-54 55-64 Employee Employer Employee Employer Own account worker Unpaid family worker Own account worker Unpaid family worker Source: CSES 2019/20. Source: CSES 2019/20. Cambodia’s poor are mostly low-earning employees. Importantly, poverty headcounts are much higher for About 51 percent of the poor are in paid employment, employees in the private sector than in the higher-paying compared to 46 percent of the non-poor (Figure 3.30). public sector (23.6 percent vs. 3.5 percent).41 In urban areas, the proportion of the poor engaged in Nearly 30 percent of wage workers earned less paid employment reaches 67 percent, 8 pp higher than than the minimum wage of US$182 (KHR 739,129) the non-poor. On average, the gross monthly salary was per month in 2019. Most wage workers who earned about KHR 1.1 million (US$270) per month, while poor less than minimum wage were poor workers (45 percent). employees earned only KHR 0.8 million (US$196) per Wage workers with no schooling were more likely to earn month. For example, the average monthly income of a less than minimum wage (56 percent), while those who family of 5 with 2 income earners amounts to KHR 1.6 completed higher levels of education were less likely to million (US$392), equivalent to 99 percent of the poverty earn minimum wage. The proportion of wage workers line. As a result of low-paying jobs, poverty rates are higher who earned less than minimum wage and lived below the among employees than among employers or own-account national poverty line was 32.3 percent, almost 2 times workers. Approximately 21.3 percent of employees are higher than those who earned at least a minimum wage. poor, compared to only 12.7 percent of employers, and Cambodians can find employment, but the payment may 17.5 percent of own-account workers (Figure 3.31). not be sufficient to meet basic needs. 41 E  mployees in private sector earned on average of KHR 1 million (US$245) per month, while those in public sector earned KHR 1.4 million (US$343) per month. Public sector workers account for only 5 percent of total labor force or 13 percent among non-agricultural wage workers. Note, the currency conversion used is KHR 4077 per US$, which is the average of the 2019 and 2020 exchange rate. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 75 Figure 3.30 Employment type by poverty Figure 3.31 Poverty rate by employment status 100 30 7 Percent of employed people 16 14 9 18 19 Cambodia Percent of population 80 26 32 33 39 35 20 60 44 40 67 10 59 51 46 47 20 37 0 0 Poor Non-poor Poor Non-poor Poor Non-poor Employee Employer Own Unpaid Inactive Private Public Cambodia Urban Rural account family worker worker Employee Employer Employment status Employee Own account worker Unpaid family worker Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Households receiving remittances are more likely to the Coastal region, remittances receipt is associated with live in poverty than those that do not. The poverty rate of higher poverty rates, as households with remittances tend remittance-receiving households is 19.3 percent compared to be poorer than those without remittances. Overall, the with 16.4 percent for households without remittances poverty gap for households receiving remittances is 3.9 (Figure 3.32). While poverty rates are higher among rural percent, only 0.7 pp higher than that for households not households, receiving remittances does not seem to affect receiving remittances (Figure 3.33). a household’s poverty status in urban or in rural areas. In Figure 3.32 Poverty rate Figure 3.33 Poverty gap by remittance-receipt by remittance-receipt 30 Percent of poverty line 8 Percent of population 20 6 4 10 2 0 0 Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Area Region Area Region No remittances Received remittances No remittances Received remittances Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. 76 CAMBODIA POVERTY ASSESSMENT Female-headed households tend to receive Remittances play an important role in supporting remittances more than male-headed households. female-headed households. Remittances help keep More than 60 percent of female-headed households the poverty rates of female-headed and elderly-headed received remittances over the past 12 months prior to households on par with male-headed counterparts. In the CSES interview compared to 46 percent of male the absence of remittances, poverty rates of female- headed households (Figure 3.34). In addition, female- headed and elderly-headed households could be higher. headed households received remittances on average of Estimations based on Propensity Score Matching KHR 1,235 (US$0.3) per person, per day representing 6 (PSM) suggest that in the absense of remittances, percent of per capita consumption, more than twice the remittance-receiving households with heads aged 55 amount male-headed households received (Figure 3.35). and above would have had a poverty rate that is 3.5 pp higher. Similarly, female-head households would have Households headed by elderly people are more experienced a 6.1 pp higher poverty rate (Table 3.3).42 likely to receive remittances. More than 80 percent Remittances also reduced the poverty gap and poverty of household heads aged 75 and above received severity. These results are consistent with previous remittances, compared to nearly 30 percent of household empirical studies finding that remittances can contribute heads aged 15–24 (Figure 3.34). Remittances per capita to poverty reduction.43 Slightly different from the literature, per day for households with older heads amounted to this finding highlighted that the remittances need to be KHR 1,646 (US$0.4), equivalent to 8 percent of daily large enough to have a significant poverty reducing effect; per capita consumption, 3 times higher than households that is, at least KHR 1,200 (US$0.3) per person, which headed by a younger person (Figure 3.35). represents 6 percent of daily per capita consumption. Figure 3.34 Share of household with Figure 3.35 Remittances per capita per remittances by household head’s gender day by household head’s gender and age and age group group 100 2,000 Percent of households 80 1,500 60 KHR 1,000 40 20 500 0 0 Male 75+ 15-24 25-34 35-44 45-54 55-64 65-74 Male 15-24 25-34 35-44 45-54 55-64 65-74 75+ Female Female Gender Age group Gender Age group Source: CSES 2019/20. Source: CSES 2019/20. 42  ooking at the full sample shows that remittance-receiving households tend to have 1.2 pp higher poverty rate in the absence of remittances, L but this is statistically insignificant. However, households with remittances would experience higher poverty gap and poverty severity—that is, worse living condition—if they did not receive remittances. On average, remittance per capita is relatively too small amounting to KHR 713 (US$0.2) per day, about 3.4 percent of per capita consumption. About 40 percent of households received remittances less than KHR 400,000 (US$98) during the 12 months prior to the 2019/20 CSES, only 1 percent of per capita consumption. 43 Chan (2009); Hing and Lun (2011); Tong (2011); MoP (2012); Vathana and Luca (2016). CHAPTER 3. POVERTY AND INEQUALITY PROFILE 77 Table 3.3 The Impacts of Remittances on Poverty (Propensity Score Matching) One-to-one 5-nearest matching matching Kernel matching All sample Poverty rate -1.16 -1.22 -0.34 Poverty gap -0.97** -0.54** -0.18 Poverty severity -0.43** -0.23** -0.04 Household head ages 55+ Poverty rate -5.45** -3.52** -1.41 Poverty gap -1.88** -1.44*** -0.71* Poverty severity -0.70** -0.55*** -0.26 Female headed household Poverty rate -6.40* -6.10*** -4.08** Poverty gap -2.23** -1.36** -0.94** Poverty severity -0.95** -0.46* -0.29 Source: Authors’ calculations based on CSES 2019/20. 3.3 Education, Housing Conditions, and Access to Services More than 80 percent of Cambodians ages 15 and (Figure 3.36). The highest and lowest literacy rates can above can read and write, but poor and elderly be found in Phnom Penh and Plateau and Mountains, Cambodians are more likely to be illiterate. Younger respectively (93 vs. 76 percent). While there is no gender generations have become more literate, with literacy rate literacy gap among younger cohorts, overall literacy rates highest among those aged 15–24 (above 90 percent), are still higher for men than women (87 percent vs. 77 and gender gaps have closed. More adults are literate percent). Poor households are less literate than non-poor in urban (89 percent) than in rural areas (77 percent) households (84 percent vs. 70 percent) (Figure 3.37). Figure 3.36 Literacy Rate by Group (15+) Figure 3.37 Literacy Rate by Group and Gender (15+) Percent of population aged 15+ Percent of population aged 15+ 100 100 80 80 60 60 40 40 20 20 Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor 15-24 25+ Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor 15-24 25+ Poverty Age Poverty Age Area Region status group Area Region status group Male Female Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. 78 CAMBODIA POVERTY ASSESSMENT The net enrollment rate in primary education is 90 in lower secondary among poor children (32 percent) is percent, with significant differences between the almost half that of non-poor children (52 percent) (Figure poor and non-poor. 44 Socio-economic status, rather 3.39). Net enrollment in lower secondary education is than geographical disparities or availability of schools, highest in urban areas (52 percent) and lowest in rural appears to be driving differences in primary school net areas (46 percent). Girls are more likely to be enrolled in enrollment. The net enrollment rate for poor households lower secondary school than boys, with about 52 percent is 86 percent, 6 pp lower than for non-poor households of girls aged 12–14 enrolled, compared to 43 percent of (Figure 3.38). Children in poor households do not attend boys aged 12–14. The gender gap in schooling is highest primary school mainly because they are not performing in the Coastal region, where 38 percent of boys aged 12– well in school (43 percent), do not want to go to school 14 enrolled compared with 50 percent for girls. IIona and (28 percent), or are too poor (20 percent) (Figure 3.40). Sunjum (2020) suggest that economic factors and school under-performance are strongly linked with boys leaving The net enrollment rate in lower secondary education school early at lower secondary level. One-third of all is 48 percent and varies widely by poverty status, children did not enroll in lower-secondary school because gender, and location. Only 48 percent of children aged they “must contribute to household income” (Figure 3.41). 12–14 are enrolled in lower-secondary school. National Children from poor households are more likely not to Student Assessments suggest that student learning attend because they need to support household income. outcomes remain low and did not improve over the period Girls face the added constraint of having to help with 2013-2017, with students in grade 6 and 8 not being able household chores. to answer half of questions correctly.45 Net enrollment Figure 3.38 Net Primary Enrollment Rate Figure 3.39 Net Lower-Secondary by Area Enrollment Rate 100 100 Percent of children aged Percent of children aged 12-14 enrolled in school 6-11 enrolled in school 80 80 60 60 40 40 20 20 0 0 Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Poverty Area Region status Area Region status Male Female Male Female Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. 44 S  ome discrepancies between survey and administrative data exist: Net enrollment rates in primary and secondary school estimated from the 2019/20 CSES are lower than estimates from administrative data published by the Ministry of Education, Youth and Sports. In 2018, net enrollment rates were 97.8 percent and 59 percent in primary and secondary school, respectively (World Bank, 2020a). 45 MoEYS (2017) and MoEYS (2018) CHAPTER 3. POVERTY AND INEQUALITY PROFILE 79 Figure 3.40 Reasons Children of Primary School Age Do Not Attend School 100 Percent of children aged 6 -11 not enrolled 13 14 13 11 16 10 17 0 0 1 7 13 10 80 13 20 10 21 14 0 5 10 60 34 42 37 43 30 24 40 17 20 30 33 28 31 28 32 28 0 Cambodia Urban Rural Boys Girls Non-poor Poor Area Gender Poverty status Don't want to Did not do well in school Must help with household chores Too poor Too young Other Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Figure 3.41 Reasons Children of Lower Secondary School Age Do Not Attend School 100 Percent of children aged 12 -14 not enrolled 12 10 7 7 12 15 19 13 3 80 4 7 4 8 12 33 60 32 31 37 37 29 32 40 24 22 16 24 24 20 20 20 21 24 27 23 20 20 13 0 Cambodia Urban Rural Boys Girls Non-poor Poor Area Gender Poverty status Don't want to Did not do well in school Must contribute to household income Must help with household chores Too poor Other Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Most Cambodian households have dwellings poor households. About 37 percent of poor households made of improved roofing materials, while less have dwellings constructed of permanent and semi- than half have improved walls or floors. Ninety- permanent materials, which is 10 pp lower than non- seven percent of poor households have access to poor households (Figure 3.43). Poor households are even improved roofing materials, compared with 99 percent less likely to have dwellings constructed with improved of non-poor households (Figure 3.42). However, poor floor materials, withh only 12 percent of poor households households are less likely to have walls and floors made having improved floor construction materials compared to of improved construction materials compared to non- 41 percent of non-poor households (Figure 3.44). 80 CAMBODIA POVERTY ASSESSMENT Figure 3.42 Improved Roof 100 Percent of households 80 60 40 20 0 Cambodia Urban Rural Phnom Plains Tonle Sap Coastal Plateau Non-poor Poor Penh and Mountains Area Region Poverty status Source: CSES 2019/20. Note: Improved roof: tiles, fibrous cement, galvanized iron, mixed but predominantly made of galvanized, and concrete. Urban = Phnom Penh and other urban areas. Figure 3.43 Improved Wall Figure 3.44 Improved Floor 100 100 Percent of households Percent of households 80 80 60 60 40 40 20 20 0 0 Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Poverty Area Region status Area Region status Source: CSES 2019/20. Source: CSES 2019/20. Note: Improved wall: concrete or brick or stone, galvanized iron Note: Improved floor: cement, parquet polished wood, pol- and fibrous cement. ished stone marble, vinyl and ceramic tiles. Urban = Phnom Penh and other urban areas. Urban = Phnom Penh and other urban areas. Most Cambodian households have a roof made iron, or concrete, as compared with 67 percent of non- of cement, iron, or concrete (68 percent); a wall poor households. The share of poor households with made of wood, log, or plywood (48 percent); and a walls made of wood, log, or plywood is 54 percent, 7 floor made of wooden planks or bamboo strips (58 pp higher than the share of non-poor households (Figure percent). Roofs made of cement, iron, or concrete are 3.46). Most poor households (83 percent) have a floor more common in Phnom Penh than in the Plains region made of wooden planks or bamboo strips, compared (87 percent vs. 56 percent) (Figure 3.45). Around 76 with over half of non-poor households (57 percent) who percent of poor households have a roof made of cement, use materials of higher quality (Figure 3.47). CHAPTER 3. POVERTY AND INEQUALITY PROFILE 81 Figure 3.45 Type of Roof Figure 3.46 Type of Wall 100 100 21 34 37 Percent of households Percent of households 80 80 46 45 40 50 47 56 68 63 71 70 67 65 60 76 78 76 60 87 81 40 40 75 58 54 48 51 48 46 47 20 35 43 20 32 30 27 28 32 23 22 22 13 17 0 0 Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Poverty Area Region status Area Region status Thatch/leaves/grass Tile Cement/iron/concrete Other Bamboo/thatch Wood Concrete/brick/stone Other Source: CSES 2019/20. Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Note: Urban = Phnom Penh and other urban areas. Figure 3.47 Type of Floor 100 3 8 14 12 11 7 22 10 22 25 13 Percent of households 80 10 12 44 12 13 19 70 60 15 83 40 75 71 68 68 61 57 53 20 37 11 16 0 Cambodia Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Area Region status Earth/clay Wood/bamboo Cement Ceramic Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. Access to basic infrastructure services such as households with improved water sources is highest in water, sanitation, and electricity improves quality of Phnom Penh (95 percent) and lowest in the Plateau and life. While water and sanitation are especially important Mountains region (63 percent) (Figure 3.48). determinants of health and nutritious status, while About 8 in 10 Cambodian households also have electricity influences children’s study and economic access to improved sanitation, depending on development. economic and geographical disparities. Improved Around 8 in 10 Cambodian households have access sanitation is extremely important for health as poor to improved water, but the poor have less access, sanitation is often related to transmission of diseases such and geographical disparities remain. Safe and reliable as cholera or diarrhea, which in turn can reduce socio- water is essential for health and well-being. About 70 economic development and educational opportunities. percent of poor households having access compared Nationwide, 80 percent of households have access to to 81 percent of non-poor households. The share of improved sanitation (Figure 3.49). Access to improved 82 CAMBODIA POVERTY ASSESSMENT sanitation is lower among poor than non-poor households Electricity plays a significant role in Cambodian daily life, (62 percent vs. 83 percent). Since most poor live in rural especially for lighting. Unstable and unreliable electricity areas, the share of households with improved sanitation is can limit children’s ability to study and adults’ ability to lower in rural than urban areas (74 percent vs. 91 percent). work. Overall, 86 percent of Cambodian households have While access to improved sanitation in Phnom Penh is access to electricity (Figure 3.50). Electricity access is nearly universal (98 percent), access in other regions nearly universal in Phnom Penh (99 percent) and lowest is much lower. About 65 percent of households in the in Plateau and Mountains (66 percent). Considerable Plateau and Mountains region have improved sanitation. differences in electrification between poor and non- poor households remain. While 88 percent of non-poor Nearly 9 in 10 Cambodian households are connected households have electricity access, only 73 percent of to the electric grid, but access remains unevenly poor households have it. distributed between poor and non-poor households. Figure 3.48 Access to Improved Water Figure 3.49 Access to Improved Sanitation 100 100 Cambodia Cambodia Percent of households 80 Percent of households 80 60 60 40 40 20 20 Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Poverty Area Region status Area Region status Source: CSES 2019/20. Source: CSES 2019/20. Note: Improved water: drinking water from public tap/standpipe, Note: Improved sanitation: flush (or pour-flush) toilets connected tubed/piped well, borehole, protected dug well, and bottled water. to sewerage, septic tank, pit and pit latrine with slab. Urban = Phnom Penh and other urban areas. Urban = Phnom Penh and other urban areas. Figure 3.50 Access to Electricity 100 Cambodia Percent of households 80 60 40 20 Urban Rural Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Non-poor Poor Poverty Area Region status Source: CSES 2019/20. Note: Urban = Phnom Penh and other urban areas. CHAPTER 3. POVERTY AND INEQUALITY PROFILE 83 3.4 Conclusion These findings can be used to inform the design of Households with 7 or more members have a poverty poverty programs, including improving allocations of incidence fourfold higher than households with 3 to 4 limited resources, and improving targeting of poor areas members. Similarly, households with 3 or 4 children have and people. a poverty rate threefold higher than those with only 1 child. The incidence of monetary poverty in Cambodia is Household headed by someone working in 17.8 percent in 2019/20. Poverty is lowest in Phnom agriculture is strongly associated with poverty. Penh and highest in rural areas. The poverty headcount Households are more likely to be poor when headed rate was estimated at 9.6 percent in Phnom Penh and by someone engaged in agricultural wage employment 22.8 percent in rural areas. Poverty is also highest in the (poverty rate of 37.9 percent) than engaged in non- predominantly rural Tonle Sap and in the Plateau and agricultural self-employment (poverty rate of 8.9 percent). Mountains region. Nationally, the poverty gap index is 3.5 Although pursuing non-agricultural livelihoods decreases percent and the poverty severtity index is 1.1 percent. the odds of living in poverty, more than half of the poor The depth and severity of poverty is highest in rural areas live in households deriving their income primarily from and the regions of Tonle Sap and Plateau and Mountains. non-agricultural wage employment. This is because more households engage in non-agricultural income-earning Inequality is moderate in Cambodia but varies activities than in agriculture in Cambodia. across geographical areas. Inequality, as measured by the Gini index, is highest in Phnom Penh, and lowest Labor force participation and employment rates in rural areas. The Gini index estimate is 34.5 percent in Cambodia are very high, but the educational in Phnom Penh, 30.8 percent in other urban areas, and attainment of Cambodia’s workforce remains low 28 percent in rural areas. Across regions, the Gini index and gendered. Of the 87 percent of Cambodians aged ranges from 28.5 percent in Plains to 34.5 percent in 15–64 participating in the labor force, 98 percent were Phnom Penh. employed in the 7 days preceding the CSES household survey interview. About 42 percent of the labor force Poverty in Cambodia remains overwhelmingly rural. has not completed primary school, of which 58 percent Of the 2.8 million poor people, 2.27 million (80 percent) are female workers and 42 percent are male. More than live in rural areas, and only 92,665 (3 percent) live in the half of the poor work in low-wage employment. Poor capital of Phnom Penh, while 0.47 million (17 percent) live employees earned only KHR 0.8 million (US$196) per in other urban areas. The highest incidence of poverty month, nearly 30 percent less than non-poor workers. is found in the predominantly rural Tonle Sap and in the Thus, the poor can find employment, but the payment Plateau and Mountains region. However, the largest may not be sufficient to meet basic needs. number of poor people concentrate in the northwest and southeast provinces, where population density is highest. Educational attainment is lower in rural areas and the lowest among poor households. Rural enrollment Poor households tend to be those with less- rates in lower and upper-secondary school lag urban educated heads, female heads, many household enrollment rates by 6 and 15 pp, respectively. The members, and many young members. The poverty enrollment rate gap between the poorest and richest rate is almost 3 times higher among households headed households widened to 31 pp for lower-secondary school by someone with no schooling than in households and reached 45 pp for upper-secondary school. headed by someone who has completed primary school. 84 CAMBODIA POVERTY ASSESSMENT Low levels of human development among the poor provinces. Improving service delivery of water, sanitation, reflect unequal access to services. Access to utilities and electricity in rural areas will contribute to closing is far lower among the poor than the non-poor, and urban-rural differences in living standards. Closing urban- poor households in rural areas are the most deprived. rural gaps in secondary education and improving gender For example, 69 percent of rural poor households have parity will not only close disparities but also help unleash access to electricity, 69 percent to improved water, and 58 the potential of the Cambodian workforce. percent to improved sanitation. Urban poor households It is important to provide social assistance not only lag non-poor households but have better access to basic to the poorest but also to the vulnerable near-poor, services than poor rural households. since many households clustered above the poverty line Poverty alleviation efforts must target rural areas have limited resources to absorb even a small shock (see were most of the poor and vulnerable reside. Poverty Chapter 5 on COVID-19 effects and social assistance). alleviation needs to target the northwest and southeast CHAPTER 3. POVERTY AND INEQUALITY PROFILE 85 86 CAMBODIA POVERTY ASSESSMENT CHAPTER 4 FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA Fiscal policy is central for macroeconomic stability poverty? Within the limits of fiscal prudence, what could and growth, as well as for poverty and inequality be done to make taxes and transfers more “pro-poor”? reduction. In the decade to 2019, Cambodia sustained The 2019 fiscal system in Cambodia, and many macroeconomic stability, including prudent fiscal of its elements, reduced inequality but also left management, and strong growth. Strong and inclusive poor households out-of-pocket in the short term. growth contributed to substantial poverty reduction and According to Fiscal Incidence Analysis using the CEQ shared prosperity. Fiscal policy also plays a direct role in methodology, the modeled fiscal system in Cambodia poverty and inequality reduction through distributional decreases the Gini index by 0.95 percentage points (pp). effects. How resources are collected and spent influences Spending on education shows the largest benefit on whether the fiscal toolkit achieves poverty-reducing and inequality, complemented by direct taxes, which the CEQ equalizing goals, but a comprehensive analysis of the approach estimates to be largely progressive. Despite distributional implications of fiscal policy in Cambodia the decrease in inequality, the degree of redistribution was missing. is small by international comparison. At the same time, This chapter assesses the distributional effects the analysis estimates that poorer households pay more of pre-COVID-19 fiscal policy and its individual in indirect taxes than they receive in cash benefits in the elements in Cambodia.46 This chapter applies an short-term;the size of targeted transfers to the poor are internationally recognized methodology developed by the not large enough to offset poverty-increasing effects of Commitment to Equity (CEQ) Institute to assess how taxes indirect taxes. The fiscal system increases the short- and government spending affect poverty and inequality term measure of the poverty headcount by an estimated (see Box 4.2). Such evidence can inform policymakers 2.1 pp. This increase excludes the longer-term of health and other stakeholder assessment of existing fiscal and education spending which children today will gain instruments and in designing reforms. The chapter seeks in the future; but it does highlight that in cash terms to answer: How much are social spending, subsidy, and many poorer households are out-of-pocket today due to tax policies contributing to redistribution and poverty the fiscal system. It will therefore be crucial to increase reduction goals? How are specific taxes and government progressive social spending, including targeted cash spending contributing to equalization and poverty transfers, to improve the net position of poor households. reductions? How are fiscal policy reforms that change Other components of spending, such as on infrastructure, the size a particular tax or benefit affecting inequality and can make the overall system progressive by improving inclusivity of economic growth. 46 This chapter heavily draws from the Working Paper of Karamba et. al (forthcoming) on the Distributional Impact of Fiscal Policy in Cambodia. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 87 Box 4.1 RGC Interventions in Response to COVID-19 In response to COVID-19, the RGC introduced emergency interventions to mitigate the adverse COVID-19 impacts on the economy and households and to ensure economic and social stability in 2020. The interventions include social assistance measures, measures to stabilize the livelihoods of businesses and workers, measures to finance businesses, and fiscal-related measures to increase financing and restore and promote growth in the post-pandemic context. In addition, the RGC is implementing a COVID-19 masterplan to address the health response to the crisis that has been directly supported through World Bank financing, including procurement of medical supplies and equipment for treatment of infected cases and testing. Announced measures include: ■ introduction of COVID-19 cash transfers to poor and vulnerable households; ■ tax relief for the tourism and garment, footwear and travel (GFT) goods manufacturing sectors; ■ retraining and upskilling programs for laid-off workers in tourism and GFT goods sectors; ■ unemployment benefits for suspended workers in GFT goods sector of US$70 per month (US$40 paid by the government and US$30 paid by the factory); ■ unemployment benefits of 20 percent of minimum wage for suspended workers employed in the tourism sector (paid by the government); ■ exemption of property registration tax for purchases below US$70,000; ■ additional capital injection for the Rural Development Bank (RDB) to support agroprocessing firms; ■ establishment of a new SME Bank designed to support SMEs through co-financing and risk sharing with commercial banks; ■ establishment of a new Credit Guarantee Corporation of Cambodia (CGCC) that has launched the first guarantee scheme; ■ establishment of a new cash for work program targeting small infrastructure improvement projects; ■ measures to improve the ease of doing business; ■ actions to improve trade facilitation, including post audit clearance; and ■ measures to inject liquidity into the financial sector through the temporary lowering of capital and reserve requirements as well as regulatory forbearance. The total fiscal cost of RGC’s COVID-19 response amounted to US$823 million in 2020 and US$689 million in 2021, out of which US$882 million is expected to be spent on social protection programs (US$500 million for cash transfers, US$260 million for cash for work, US$122 for wage subsidies and training). SME support (RDB, SME Bank, and CGCC) amounts to another US$600 million over 2020-2021. 88 CAMBODIA POVERTY ASSESSMENT Note the fiscal assessment predates COVID-19 and July 2022 and has been the largest component of the the results may be different due to the increase government’s support package. Spending on cash in cash transfers in response to COVID-19. During transfers rose from less than 0.1 percent of GDP in 2019 the pandemic, Government scaled up social assistance to 0.7 in 2020 and 1.4 in 2021. The program has reached to poor and vulnerable households (see Chapter 5 about 690,000 households and 2.7 million individuals, or discussion on COVID-19 effects and social assistance). about 17 percent of the population, up from 2 percent of Launched in June 2020, the cash transfer program the population pre-COVID-19. Box 4.1 summarizes the has disbursed US$714 million in cash transfers as of government’s fiscal response to the crisis. 4.1 Taxes and Social Spending in Cambodia Revenue collection in Cambodia is highly by direct taxes (5 percent of GDP). Value added tax (VAT) centralized with a small number of taxes collected is the largest component of indirect taxes and excise tax at the municipal or provincial level. In total, central is the second largest. The system of direct taxes and government revenue amounted to 23 percent of GDP in contributions covers taxes on income, profits, capital 2019, while provincial government revenue amounted to gains, and property, of which corporate income tax and 2 percent (Table 4.1). Indirect taxes are the main source salary tax account for the largest components. Table 4.1 of government revenues (13 percent of GDP), followed outlines Cambodia’s Government revenue structure. Table 4.1 Cambodia Domestic Government Revenues, 2019 Percent of Included Amount Percent of Government in CEQ Source (billion KHR) GDP Revenue analysis Total Domestic Revenues (A+B) 27,730 25 100 A. Central Government Revenue (I+II+III) 25,576 23 92 I. Tax Revenue 22,053 20 80 Direct taxes of which 5,045 5 18 Profit tax 4,197 4 15 n.a. Payroll tax 848 1 3 Yes Tax on interest, royalties, and dividends -- -- -- No Land and property -- -- -- Yes Indirect taxes of which 14,095 13 51 VAT (domestic and on imports) 7,410 7 27 Yes Excise duties (domestic and on imports) 6,478 6 23 Yes Others 207 0 1 Yes Taxes on international trade 2,913 3 11 n.a. revenue II. Non-tax revenue 3,342 3 12 n.a. III. Capital revenue 181 0 1 n.a. B. Provincial revenue 2,153 2 8 n.a. Source: Ministry of Economy and Finance (MEF). Note: Social Security Contributions revenue is recorded under the revenue of the National Social Security Fund, hence there is no record under Budgetary Central Government Level. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 89 Social sector spending in Cambodia remains low by education. Thailand and Vietnam spent, on average, more international comparison. Cambodia is spending less than 2 percent of GDP on health. Cambodia, together on education and health as a share of GDP than other with Myanmar and Lao PDR, lag other ASEAN countries ASEAN countries. Education spending accounted for respectively. Cambodia lags on the HCI due to weaker 3 percent of GDP in 2019, of which primary education education attainment, skill utilization, and maternal and absorbed more than half of overall spending. Cambodia’s child health and nutrition outcomes.47 Cambodia needs to constitution mandates public schools provide 9 years of catchup and align its education and health outcomes with free basic education. Health spending was only 2 percent best performers within ASEAN and the world. As the CEQ of GDP in 2019, almost unchanged since 2009. Best analysis demonstrates, increases in education and health performers in ASEAN spend twice the amount as a share expenditures to match top performing ASEAN countries of GDP. For example, between 2009 and 2018 Malaysia could help Cambodia redistribute wealth and support and Vietnam spent, on average, 5 percent of GDP on low-income households. Table 4.2 Central Government Spending, 2019 Amount Percent Included in CEQ Type (billion KHR) of GDP Analysis? Central Government Expenditure 27,317 24.8   Social Benefits 2,310 2.1   Social security 898 0.8 Partial Social assistance to citizens 1,219 1.1 Partial Social assistance to social and cultural entities 138 0.1 No Other social benefits 56 0.1 No Health (MoH spending) 1,924 1.7 Yes Education (MoEYS spending) 3,679 3.3 Yes Other Spending 19,404 17.6 No Source: Ministry of Economy and Finance (MEF), Ministry of Health (MoH), Ministry of Education, Youth, and Sports (MoEYS). Social assistance spending, in particular cash primarily for former civil servants as pensions. The transfers, is very low in Cambodia albeit better other half is devoted to social assistance. Cash transfer than other developing countries in East Asia and spending in 2019 was minimal despite full national roll out the Pacific. 48 In Cambodia, social protection spending of conditional cash transfers (CCT) for pregnant women consisting of social assistance and social security of and children below age 2. However, Cambodia’s social about KHR 2,255 billion in 2019 amounted to 2 percent assistance spending has been increased significantly of GDP, among the lowest in ASEAN. About half of during the COVID-19. social protection spending is devoted to social security, 47 See World Bank (2020b) for more details. 48 T  he National Social Protection Policy Framework (2016–2025) has two main pillars: social assistance and social security. The social assistance pillar includes emergency response, human capital development, vocational training, and welfare for vulnerable people. The social security pillar includes pension, health insurance, employment injury insurance, unemployment insurance, and disability insurance. 90 CAMBODIA POVERTY ASSESSMENT Figure 4.1 Total Social Assistance Figure 4.2 Total Cash Transfer Spending Spending (EAP non-HIC) (EAP non-HIC) 6 6 5 5 Percent of GDP Percent of GDP 4 4 3 3 2 2 1 1 0 0 Sa dia ala a a a ilip sia Ti iet es or m te Th hina do d C an ia bo ar Ki a N ao P ti G R a on te Vi olia Ki m Ch ti Th ina b d ia ilip Fiji do es Sa sia a DR G r a C iji w ma a M odi M goli o ne a La mo ne In ilan am n ew D F M nes ys es m na am m M es na V in rib rib In pin m C aila Ph lay ne pu M o P g ui ui p Ne an -L -L et on a y or y M L Ph m Ti a a pu Pa Pa Source: CEQ and World Bank databases and World Bank calculations. Note: Total social assistance spending excluding health fee waivers. Cash transfers include CCT and UCT. HIC = high-income country, CCT = conditional cash transfers, UCT = universal cash transfers. In Cambodia, the CEQ methodology covers 19.3 income combined with an imputation of employment percent of total revenues and 23.2 percent of status. Corporate income tax is excluded from the analysis government spending. The CEQ analysis relies on because the allocation of the tax burden to individual the 2019/20 Cambodia Socio-Economic Survey (CSES) households is generally not possible and is further and 2019 administrative data on taxes, transfers, and complicated by the fact that a substantial proportion of expenditures. Indirect taxes and subsidies are calculated businesses in Cambodia operate outside the formal sector. based on general rules applied to consumption taxation Direct transfers are divided into scholarships and other and the CSES information on household expenditures. welfare transfers based on available CSES information. Income taxes, health and social security insurance are “In-kind” transfers are allocated to households based on simulated using the detailed regulations based on the either age eligibility (education) or declared usage of these information on employment earnings and self-employment services (health). Box 4.2 Commitment to Equity (CEQ) Methodology This chapter applies an internationally recognized methodology developed by the Commitment to Equity (CEQ) Institute49 to assess how taxes and government spending affect poverty and inequality. The CEQ approach combines detailed household survey data with tax and benefit regulations to estimate what each household paid in taxes and received in benefits. The CEQ approach then describes the distributional consequences of the fiscal system sequentially through 4 income measures. The baseline concept is Market Income, the income that an individual (or a household) earns before any taxes are subtracted and prior to inclusion of any transfers. Three “post-fiscal” income measures—Disposable Income, Consumable Income, and Final Income—stem from sequentially subtracting and adding different forms of fiscal interventions to pre-fiscal income. Comparisons of the 4 income aggregates shed light on how fiscal interventions affect income distribution sequentially from the pre-fiscal income to final income. We can then compare poverty and inequality estimates derived from pre-fiscal and post-fiscal income. Sequentially quantifying the poverty and inequality consequences of direct taxes and transfers, indirect taxes and subsidies, and social spending on health and education combine to provide a comprehensive incidence analysis. 49 See Lusting and Higgins (2018) for description of the CEQ methodology. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 91 4.2 Overall Tax and Spending Implications on Poverty and Inequality In Cambodia, the overall fiscal system somewhat 32.2 percent at Disposable Income. Indirect taxes and reduces inequality. Bearing in mind the caveat that subsidies have limited effect, and their consideration top income households are often underrepresented leaves the Gini index of Consumable Income at 32.2 in household survey data—a common problem with percent. “In-kind” transfers from health and education, on most surveys, especially in developing countries—the the other hand, have the largest effect on inequality with fiscal system in Cambodia reduces inequality by 1 pp. the Final Income Gini at 31.4 percent. This reflects the Inequality, as measured by the Gini coefficient, falls fact that in-kind transfers represent a significant share of between Market Income and Final Income (Figure 4.3). pre-fiscal income and proportionally benefit lower-income Before any fiscal interventions, the Market Income Gini households more than those at the upper end of the index is 32.4 percent. Once considering direct taxes income distribution. and direct transfers, the Gini index reduces slightly to Figure 4.3 Gini Index Before and After Fiscal Interventions in Cambodia 34 33 Gini coefficient (%) 32.4 32.2 32.2 32 31.4 31 30 Market income Disposable income Consumable income Final income Source: Authors’ calculations based on CSES 2019/20. While Cambodia does reduce inequality through middle-income countries achieve inequality reduction of taxes and transfers, the degree of redistribution up to 9 percentage points (pp) from the pre-fiscal income. is small in international comparison. Figure 4.4 When in-kind transfers are excluded, the redistributive demonstrates that the redistributive effect of fiscal policy effect in Cambodia is even lower. This is because of the (including in-kind transfers) is lower in Cambodia than very low impact of direct transfers on inequality. in some lower-middle income countries. Some lower- 92 CAMBODIA POVERTY ASSESSMENT Figure 4.4 Change Gini After Fiscal Policy (Including In-Kind Transfers) 0 -5 Point Change -10 -15 -20 -25 Poland Croatia Panama Uruguay Chile Romania Namibia Argentina South Africa Costa Rica Brazil Mexico Iran Mexico Mexico Ecuador Belarus Colombia Colombia Venezuela Russia Albania Dominican Republic Paraguay Armenia Peru Peru Indonesia Guatemala Jordan Ukraine Tunisia Kenya El Salvador Bolivia Tanzania Nicaragua Egypt Ghana Comoros Honduras Sri Lanka Cambodia Burkina Faso Uganda Ethiopia Low High income Upper middle income Lower middle income income Source: CEQ and World Bank databases and World Bank calculations (see Rodriguez and Wai-Poi 2020). Note: HIC = high-income country, UMIC = upper middle-income country, LMIC = lower middle-income country, LIC = low-income country. At the same time, poorer households pay more in on consumption are negative and large, thereby increasing tax than they receive in cash transfers; short-term poverty measured today as all households face higher poverty is slightly higher due to indirect taxation. prices on goods from indirect taxes. For poor and near- Overall, the modelled fiscal system increases short-term poor households, the higher price reduces purchasing poverty by 2.1 pp, as shown by the transition from Market power, potentially reducing net expenditures below the Income to Consumable Income (Figure 4.5). The net poverty line. It is important to note that poorer households transfer achieved through direct transfers and direct taxes are only out-of-pocket in the short-term; this analysis does on income are too small to change average household not include the economic benefits of health and education incomes, and thus poverty at Disposable Income. However, services which households will enjoy in the longer-term and the net transfer resulting after subsidies and indirect taxes which will help reduce poverty. Figure 4.5 Poverty Headcount Before and After Fiscal Interventions 25 Poverty headcount (%) 19.8 20 17.7 17.8 15 10 Market income Disposable income Consumable income Source: Authors’ calculations based on CSES 2019/20. Note: Per CEQ technical recommendations, in-kind health and education transfers are not included when estimating the impact on poverty. Cambodia is not alone in having fiscal policies that short-term poverty headcount ratio in the majorirty leave the poor out-of-pocket in the short-term. of developing countries. Most fiscal systems in these Fiscal policy (excluding in-kind transfers) increases the countries rely heavily on indirect taxes. For instance, CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 93 Tanzania generates 60 percent of its revenues from are VAT exempt. However, nearly every person is affected indirect taxes. 50 Cambodia’s fiscal system is similarly by the Cambodian indirect tax system because at least reliant on indirect taxes, with revenues collected through one of the items they consume regularly has an implicit or consumption taxes (value-added tax and excise duties) explicit indirect tax. Net purchasing power for households amounting to twice the revenues collected through along the income distribution therefore decreases after income taxes. Although excise duties cover socially receiving transfers (direct and subsidies) and paying taxes costly goods such as alcohol and tobacco, these taxes (direct and indirect). Although many lower-middle income also cover more necessary goods, such as bottled water countries’ fiscal systems increase short-term poverty, and communication services (telephone and internet). others perform better (Figure 4.6). Most necessities such as meat, seafood, grains, and fruit Figure 4.6 Fiscal Policy’s Impact on Poverty Across Countries 10 Percentage point change 5 0 -5 -10 -15 Poland Panama Chile Uruguay Iran, Islamic Rep. Ecuador Mexico Jordan Paraguay Russian Federation Mexico Mexico Belarus Russian Federation Indonesia South Africa Colombia Peru Argentina Costa Rica Peru Venezuela, RB Dominican Republic Colombia Guatemala Brazil Albania Armenia Tunisia El Salvador Honduras Bolivia Nicaragua Ghana Cambodia Sri Lanka Tanzania Uganda Ethiopia Burkina Faso High Low income Upper middle income Lower middle income income Market to Disposable Market to Consumable Source: CEQ and World Bank databases and World Bank calculations. Note: Pensions as deferred income treatment. Relevant international poverty line used for income classes: US$1.90 for LIC, US$3.20 for LMIC, US$5.50 for UMIC and HIC. Fiscal policy has larger repercussions on rural than policy, meaning there are more likely to fall under the urban poverty in Cambodia. Areas with higher levels poverty line when indirect taxes are paid. Notably, three of pre-fiscal poverty are disproportionately effected. In features of rural life are accounted for the results: (i) home other words, the net benefits of the fiscal system are less production of food and other staples is not subject to concentrated in rural and other urban areas where more indirect taxation; (ii) current preferential indirect tax rates individuals are impoverished. This is due to the impact and exemptions are modelled; and (iii) indirect taxes on of indirect taxes. The modelling indicates that the current food expenditures are modelled in urban areas but not (pre-COVID) fiscal system increases short-term poverty rural ones, reflecting the likely informal nature of food in rural areas by 2.5 points (Figure 4.7), higher than the purchases in rural areas. For these three reasons, the 1.8 points in non-Phnom Penh urban areas, primarily greater rural poverty increase is driven by the underlying driven by indirect taxation. The larger tax result reflects welfare distribution and not modelling characteristics. the greater rate of near-poor in rural areas before fiscal 50 Younger et al. (2016). 94 CAMBODIA POVERTY ASSESSMENT Figure 4.7 Fiscal Policy’s Impact on Poverty by Area of Residence 3.0 2.5 Percentage point change 2.0 1.5 1.0 0.5 0.0 Full system Direct Indirect interventions interventions Cambodia Phnom Penh Other Urban Rural Source: Authors’ calculations based on CSES 2019/20. Note: Pre-fiscal poverty headcount (%): Cambodia=17.7, Phnom Penh=4.1, Other urban = 12.4, Rural = 22.6. Table 4.3 Fiscal Policy’s Impact on Inequality and Poverty by Area of Residence Consumable Market income Disposable Income income Final income Gini coefficient (%) Cambodia 32.4 32.2 32.2 31.4 Phnom Penh 34.8 34.5 34.6 34.0 Other Urban 30.9 30.8 30.9 30.2 Rural 28.0 28.0 28.0 27.5 Poverty headcount ratio (%) Cambodia 17.7 17.8 19.8 n.a. Phnom Penh 4.1 4.2 5.0 n.a. Other Urban 12.4 12.6 14.3 n.a. Rural 22.6 22.8 25.1 n.a. Source: Authors’ calculations based on CSES 2019/20. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 95 4.3 Fiscal Impoverishment and Fiscal Gains to the Poor Comparison of poverty before and after taxes and post-fiscal incomes (Consumable Income) was below transfers only broadly indicates the gains or losses the poverty line. The average reduction in income as a the fiscal system creates. But this can fail to capture proportion of Market Income among Cambodians fiscally an important aspect: that some poor are made poorer (or impoverished was 3.7 percent. non-poor are made poor) by the tax and transfer system. Conversely, some of the poor gained through the Fiscal Impoverishment (FI) traces the number of people fiscal system. Fiscal gains to the poor (FGP) traces who are impoverished after execution of fiscal policy. pre-fiscal poor households to determine how many Almost 20 percent of Cambodians experience fiscal experienced an income increase following fiscal policy impoverishment. This means that: (i) some individuals execution. Fiscal policy made only about 2 percent of whose pre-fiscal incomes were below the poverty line pre-fiscal poor better off; that is, were net beneficiaries were net contributors to the fiscal system such that this of the Cambodian fiscal system. Among pre-fiscal poor reduced their net cash position, and (ii) some individuals beneficiaries, the average increase in market income was whose pre-fiscal incomes were above the poverty line 10.6 percent. were net contributors to the fiscal system such that the Table 4.4 Fiscal Impoverishment and Fiscal Gains to The Poor Market income Change in poverty Fiscally Fiscal impoverished Fiscally better- poverty headcount headcount impoverished (% of (% of consumable off (% of market (%) (percentage points) population) income poor) income poor) 17.7 2.1 19.5 98.4 1.9 Source: Authors’ calculations based CSES 2019/20. 4.4 Contribution of Individual Interventions to Poverty and Inequality Reduction In addition to overall effects of fiscal policy, if a transfer is progressive in relative terms. To analyze policymaking must assess the contribution of each whether a tax or transfer is equalizing, we use the marginal fiscal intervention to poverty and inequality. We can contribution to income inequality measured by the Gini analyze an intervention in terms of its progressivity and coefficient. The marginal contribution measures the marginal contribution to poverty and inequality. To do additional change in inequality due to a tax or transfer. An this, we use the Kakwani index (see Box 4.3) to analyze 51 intervention is equalizing when the marginal contribution whether a tax or transfer is progressive. A Kakwani index is positive. Comparing the marginal contribution and the for a tax will be positive (negative) if a tax is globally Kakwani will reveal whether a tax or transfer is equalizing progressive (regressive), and a Kakwani index is positive (unequalizing) despite being regressive (progressive). 51 Kakwani (1977). 96 CAMBODIA POVERTY ASSESSMENT Box 4.3 Kakwani Index of Progressivity and Marginal Contribution We use the Kakwani index (Kakwani 1977) to analyze whether a tax or transfer is progressive.52 A tax is considered progressive if the burden proportionally increases with income. For each tax, the Kakwani index is defined as the difference between the concentration coefficient of the tax and the Gini coefficient of Market Income. Similarly, a transfer is considered progressive if the entitlement proportionally decreases with income. In the case of transfers, the Kakwani index is the difference between the concentration coefficient and the Gini coefficient of Market Income (see Equation 1). A Kakwani index for taxes will therefore be positive (negative) if a tax is globally progressive (regressive), and a Kakwani index is positive if a transfer is progressive in relative terms. We define the Kakwani index based on the following terms: xi is the income of household i; T(xi) is the tax liability of household i; X = ∑xi is the total pre-tax income; T = ∑T(xi) is the total tax collected; g=— T X is the total tax ratio. The Kakwani Progressivity Index for a tax T is KT = CT – GX, where CT and GX represent, respectively, the concentration coefficient of the tax and the Gini coefficient of pre-tax income. If the calculated KT > 0, then the distribution of the tax burden is more unequal than that of pre-tax income, meaning that low-income households’ share of the tax burden is lower than their share in original income, so the tax is progressive. If KT = 0, the tax is neutral and if KT < 0 the tax is regressive. To analyze whether the tax or transfer is equalizing, we use the marginal contribution to income inequality measured by the Gini coefficient, which measures the additional change in inequality due to a tax or transfer. The marginal contribution is the difference between the Gini without the fiscal intervention and the Gini coefficient of all income components combined. An intervention is equalizing when the marginal contribution is positive. When there is only a single intervention in the fiscal system, the Kakwani index would have been sufficient to determine whether the intervention is unambiguously equalizing. With multiple fiscal interventions, the relationship between inequality outcome and progressivity is complex. Progressivity can also be assessed using Lorenz curves and concentration curves. Comparing the marginal contribution and the Kakwani will reveal whether a tax or transfer is equalizing (or not equalizing) despite being regressive (or progressive). Tax and Transfer Effects on to Gini coefficient reduction. Social security and health Inequality insurance contributions are regressive and unequalizing; the burden of these contributions decreases with income Direct taxes are the most progressive and most and contributes to Gini coefficient increase. Lower-income equalizing type of tax in Cambodia. Taxes on income households face a higher burden of social security due and property are both progressive and equalizing; the to the low maximum threshold (KHR 1,200,000). Indirect burden of taxation increases with income and contributes taxes are neutral and less equalizing. 52 Progressivity can also be assessed using Lorenz curves and concentration curves. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 97 Figure 4.8 Progressivity and Redistributive Effect of Taxes Marginal contribution Kakwani Coefficient 0.2 0.8 Marginal contribution to equality 0.6 (Change in Gini Points) 0.1 0.4 Kakwani 0.2 0.0 0.0 -0.2 -0.1 -0.4 x x ns e x x s x x x ns rib s C op HI x, T ta ta ta ta xe ta ta ta e cis VA SS ax io o ta ry ty n g oT l n i ex ut a ut tt io tio in Re ent er la ct rib M Pu dat ht an irec Sa ra re R nt nt lig ta st di o Pr D co o gi m ific ic l in C bl m d ec Al co Sp Ac s xe ta ct re Di Source: Authors’ calculations based on CSES 2019/20. Note: SSC=Social Security Contributions, HI=Health Insurance. All social transfers are progressive in Cambodia, spending reduces inequality the most. Except for except electricity subsidies and tertiary education tertiary education, all other education spending is both spending. Direct transfers are the most progressive but progressive and inequality-reducing. Primary education contribute little to reducing inequality due to their small has the most equalizing effect. amount and poor targeting (see Figure 4.12). Education Figure 4.9 Progressivity and Redistributive Effect of Transfers and Subsidies Marginal contribution Kakwani Coefficient Marginal contribution to equality 1.0 1.0 (Change in Gini Points) 0.8 0.8 0.6 0.6 Kakwani 0.4 0.4 0.2 0.2 0.0 0.0 -0.2 -0.2 rs b. s y y y y s ar ar ar ar r r fe fe fe su im im nd rti ns ns ns Te y Pr Pr co tra tra tra cit e- Se tri ct n lth Pr ec io ire a at El He ld uc ta Ed To Source: Authors’ calculations based CSES 2019/20. 98 CAMBODIA POVERTY ASSESSMENT Taxes and Transfers Effects on to higher overall increases in poverty than do direct taxes. Poverty VAT has the highest poverty-increasing effect, increasing poverty by about 1.76 pp. On the direct tax side, social The degree of poverty reduction from direct security contributions have the highest poverty-increasing transfers in Cambodia is not sufficient to offset effect, increasing the poverty rate by 0.4 pp. Direct transfers the poverty-increasing effect of taxes. Taxes have reduce poverty by a limited 0.06 pp. the strongest effects on poverty, and indirect taxes lead Figure 4.10 Marginal Contribution of Taxes and Transfers to Poverty Reduction Salary tax SSC HI Property tax MoT tax Rental tax Registration tax Contributions Direct taxes Direct taxes and contributions VAT Specific tax, excise Accommodation tax Public lighting tax All indirect taxes Total direct transfers Electricity sub. -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 Source: Authors’ calculations based CSES 2019/20. 4.5 Heterogeneity of Effects Across Deciles The burden of taxes and benefits varies across transfers do not offset the tax burden. While in absolute Market Income distribution deciles. In absolute terms, terms, Cambodian households in the upper end of the the burden of direct taxes, in particular the salary tax, income distribution bear a greater burden of direct and concentrates at the upper end of the income distribution. indirect taxes, relative to Market Income, the proportional In relative terms, the direct tax burden as a share of Market tax burden does not vary considerably across the Market Income does not vary considerably across the distribution Income distribution. except for the poorest and richest deciles. Cash transfers Strong progressive, in-kind education and health are not sufficient to offset the burden of direct taxes transfers boost the final position of low-income across all deciles of the distribution except the poorest. Cambodian households. The bottom 40 percent are net As such, the net cash position for most households is beneficiaries of the fiscal system, while richer households negative before considering indirect taxes and subsidies. are net contributors. The reduction in the Gini index from After indirect tax and subsidies, the net cash position 32.2 to 31.4 percent when looking at Final Income is due is negative for each Market Income decile. Cash to in-kind transfers (Figure 4.3). CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 99 Distribution of Taxes and Benefits Across Deciles of the Market Figure 4.11  Income Distribution Market Income to Disposable Income A. Direct Tax Burden/Benefit by Decile B. Direct Tax Burden/Benefit by Decile (KHR/month) (% of market income) 5,000 1.5 Proportion of income (%) 0 1.0 -5,000 0.5 KHR/month -10,000 0.0 -0.5 -15,000 -1.0 -20,000 -1.5 -25,000 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Market income deciles Market income deciles SSC - employer HI - employer Salary tax SSC - employer HI - employer Salary tax Withholding tax Rental tax Registration tax Withholding tax Rental tax Registration tax MoT tax Property tax School stipends MoT tax Property tax School stipends Welfare transfers Total Welfare transfers Total Market income to Final income A. Overall Tax Burden/Benefit by Decile B. Direct Transfers and Subsidies by Decile (KHR/month) (% of market income) 40,000 12.0 Axis Proportion of income (%) 10.0 20,000 8.0 0 6.0 KHR/month -20,000 4.0 2.0 -40,000 0.0 -60,000 -2.0 -80,000 -4.0 -6.0 -100,000 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Market income deciles Market income deciles Education transfers Health transfers Education transfers Health transfers Total transfers - cash Electricity subsidy Total transfers - cash Electricity subsidy Total direct taxes Total indirect taxes Total direct taxes Total indirect taxes Total Total Source: Authors’ calculations based on CSES 2019/20. SSC=Social Security Contributions, HI=Health Insurance. Welfare transfers have limited poverty-reducing benefit from improved targeting, as Figure 4.11 shows; and redistributive effects, however, because they non-poor households possess the IDPoor equity card to provide low coverage, are small, and are imperfectly access CCT benefits if they have small children, although targeted. In 2019, about 10 percent of Cambodian at a lower incidence than poor households. The CCT households possessed a IDPoor “equity” card and only 2 for pregnant women and children provided a total of percent received the CCT available for pregnant women US$190 during pregnancy and a child’s first 2 years, or and children under age 2. 53 The welfare transfers could approximately US$63 each year. 53 CSES normally underestimates the number of households with an IDPoor equity card. In June 2020, the Ministry of Planning announced that households with equity card amounted to 562,534, which was equivalent to 16 percent of Cambodian households–6 percentage points higher than the estimates generated from the 2019/20 CSES. 100 CAMBODIA POVERTY ASSESSMENT Figure 4.12 Proportion of Households with IDPoor Equity Card and CCT Recipients 40 30 HEF CCT 20 10 0 1 2 3 4 5 6 7 8 9 10 Market income decile Source: Authors’ calculations based on CSES 2019/20. Because in-kind tertiary education benefits increase primary and secondary public education slightly more with income, they are regressive and not equalizing. than non-poor children. However, due to direct costs and Enrollment in tertiary public education is not favorable related foregone incomes of remaining in school, these for poorer segments of the population; about 60 percent households do not benefit much from tertiary education. of enrollment in tertiary education is among the top 20 In-kind health benefits are progressive but not percent income households in comparison to 4 percent equalizing. While poorer households utilize public health among the bottom 40 percent. Poor children clearly more than richer households, they capture smaller shares benefit from having basic education and appear to access of in-kind public health benefits. Figure 4.13 Number of Users of Public Education and Health A. Percent of Publicly Enrolled Students B. Percent of Public Health Facility Users by Decile and Level of Education by Decile 70 25 60 20 50 40 15 30 10 20 5 10 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Income decile Income decile Preschool Primary Secondary Tertiary Source: Authors’ calculations based on CSES 2019/20. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 101 Figure 4.14 Total In-kind Expenditures by Decile 50 40 Billion KHR 30 20 10 0 1 2 3 4 5 6 7 8 9 10 Income decile Education Health Source: Authors’ calculations based on CSES 2019/20. Most electricity subsidies accrue to non-poor and even where electricity is present, they often cannot households. Non-poor households are more likely to afford connection or monthly fees. Even if they can afford be connected to the electricity grid and thus consume connection fees, their electricity consumption tends to be more electricity. The poor are more likely to live in low. Overall, the implicit subsidy embedded in the tariff neighborhoods without access to an electricity network, structure is regressive. Figure 4.15 Proportion of Households with Access to Electricity 100% 80% 60% 40% 20% 0% 1 2 3 4 5 6 7 8 9 10 Income decile 0-100 101-200 201+ Not connected to grid/not consuming Source: Authors’ calculations based on CSES 2019/20. 102 CAMBODIA POVERTY ASSESSMENT 4.6 Limitations and Interpretations The CEQ analysis provides useful insights about the included. This is not a flaw in the survey data, which are effects of Cambodia’s tax and benefits system, but not designed to accurately measure the very top of the we must acknowledge some limitations of the CEQ income distribution, but it does have implications for fiscal methodology and data, and their implications on incidence analysis. Another limitation is the challenge of interpreting the CEQ results. First, the CEQ looks at assigning direct taxes to individuals when the data does only part of the fiscal system and may not fully capture not explicitly distinguish those formal employment from some taxes paid by households or benefits received from those in informal employment. As such, we needed to public goods. For instance, public infrastructure could impute formal employment based on proxy characteristics be progressive through its impacts on employment and from the data to model the incidence of direct taxes, as economic activity. Second, while the data are nationally these are not paid by individuals working in the informal representative, they are not designed to fully represent economy. high-income households, a phenomenon standard As suggested by the macro-validation of the CEQ with household surveys in all countries. While missing exercise for Cambodia, discrepancies exist between top incomes does not affect our ability to measure official statistics and modelled aggregate values of poverty, it has important consequences for inequality several elements of the tax and benefit system.54 measures. If many richer households are not captured These discrepancies could be lessened with more in the survey, “true” inequality will be higher. This might detailed micro-level household data and more detailed attenuate distributional implications and redistributive administrative information on how particular elements of effects of various instruments since the higher burden the system operate. of taxes on non-represented rich households is not 4.7 Conclusion The 2019 fiscal system in Cambodia, per the CEQ system than they receive from it in cash. It is therefore model, reduces inequality but leaves the poor crucial to increase progressive social spending to offset out-of-pocket in the short-term. Note, these results the revenue collection repercussions on the poor. predate the increase in cash transfers in response to Cambodia is not alone among developing countries COVID-19. Of all fiscal interventions, in-kind education in seeing a short-term poverty increase due to and health transfers have the largest effect on inequality, the modelled elements of fiscal policy. The CEQ but the degree of redistribution is small by international methodology has been consistently applied in over comparison. At the same time, poverty increases in the 60 countries, allowing comparison of Cambodia’s short-term following execution of fiscal policy because performance with peer lower middle-income countries, direct transfers to the poor are not large enough to several of whom experience net poverty increases due mitigate the effect of indirect taxes. Without reform, to fiscal policies. This is not surprising due to their heavy poor households will continue paying more into the fiscal reliance on indirect taxation. Nonetheless, some low 54  everal elements of tax system are simulated relatively well when compared to administrative statistics. For example, the exercise captures S about 96 percent of employment injury insurance, 99.2 percent of property tax, 99.2 percent of education expenditure, and 93.1 percent of healthcare costs. Simulations of other fiscal elements are less precise. This applies to instruments such as the salary tax (23.1 percent of the administrative aggregate) or means of transportation tax (35.3 percent). The limited precision in modelling social security contributions, health insurance, and salary taxes stems from the high degree of progressivity of tax contributions and lack of highest earners in the data. CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 103 and lower middle-income countries apply fiscal policies their current material position and their medium and long- more effectively in reducing poverty. The findings for term prospects. Cambodia suggest some fiscal system reforms are To reduce poverty, fiscal policy must look to spend needed to improve its distributional effect, drawing on the more but also to spend better. This can be achieved international lessons for progressive fiscal policy in World through significant and well-targeted direct transfers. Bank (2022). Improving identification of poor households through the Sustaining high revenue growth while addressing IDPoor program will improve social spending targeting. equity will require new approaches. Cambodia’s Subject to available funding, increasing the amount Revenue Mobilization Strategy outlines strategies to of cash transfers to the poor will help offset loss in improve revenue collection to meet spending priorities. purchasing power from indirect taxation. For example, Evidence generated from the CEQ can help bring an equity electricity subsidies do not benefit poorer households lens to fiscal policy discussions and help inform design very much and cost 0.1 to 0.15 percent of GDP; if spent of the next phase of the Revenue Mobilization Strategy, instead on targeted cash transfers this would double alongside the National Strategic Development Plan. the pre-COVID budget. Further, increasing the amount of cash transfers to the poor and vulnerable will not Cambodia may need to spend more on health and only mitigate negative COVID-19 repercussions (see education to improve the inequality-reducing effects Chapter 5), maintaining some of the increased spending of fiscal policy. Countries that see significant reductions will bring Cambodia’s social assistance budget closer to in inequality usually do so through health and education regional and international levels while offsetting much of spending. Higher education spending should be targeted the short-term increase in poverty from fiscal policy. As towards pre-primary and primary education, as these a rough calculation, the poverty reducing effects of the strongly benefit poorer households. However, more and government’s COVID-19 response (discussed in Chapter improved targeting of spending on secondary and tertiary 5), if sustained after COVID, would largely eliminate education is likely to bring significant medium and long- the short-term impact of fiscal policy in Cambodia on term benefits to poorer households and the economy. poverty.55 Even if social assistance spending was not Transfers could be designed to support education beyond reduced to COVID-19 levels but sustained at half the primary school; tertiary education scholarships would current level, the imapct of fiscal policy on short-term significantly benefit those who decide against it because would put Cambodia closer to the middle of the upper of material constraints. Better access to publicly funded middle income country range (Figure 4.16). healthcare among the poorest households would improve 55  s Chapter 5 documents, Cambodia increased its social assistance spending in response to the COVID-19 crisis from 0.1 percent to 0.7 A percent of GDP. Chapter 5 estimates that this has mitigated the poverty-increasing impact of the pandemic by 1.9 percentage points, which is applied to the 2019 CEQ results of this chapter to roughly update Cambodia’s fiscal policy impact on short-term poverty. This is a rough approximation only. 104 CAMBODIA POVERTY ASSESSMENT Figure 4.16 Impact of fiscal policy on short-term poverty in LMICs 6 Pre-COVID 4 impact Half COVID response* COVID Percentage Point Change 2 response* 0 -2 -4 nd * Ho ia* e i L ia ka m m Ni dia* a Co va os Tu a Sa ia bo r as Es ia Le ini m o a a t a as am o in gu di di Ivo an ny Ca oth liv s b Ca tna C lvad at do ur an or d ni ra In m bo Co Ke Bo h ra bo w m s Uk Za ol G e ca Vi ry M Sr El CHAPTER 4. FISCAL POLICY EFFECTS ON POVERTY AND INEQUALITY IN CAMBODIA 105 106 CAMBODIA POVERTY ASSESSMENT CHAPTER 5 COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS The coronavirus disease 2019 (COVID-19) is a examines pandemic effects on employment, income, serious health threat that has had far-reaching and food insecurity using evidence from 5 rounds of economic consequences and reversed several the Cambodia High-Frequency Phone Survey (HFPS) years of global poverty reduction. Because close conducted in Cambodia between May 2020 and March and prolonged contact with an infected person increases 2021 (see Box 5.1), complemented with other sources. the risk of exposure to the virus, the early global and The chapter also examines coping strategies households community response was to limit transmission through adopted during the pandemic and the role public safety travel restrictions, school and business closures, and nets played in mitigating the increase in poverty and social distancing, among others. In the absence of a inequality. vaccine or treatment for COVID-19, these measures COVID-19 is expected to increase poverty in relieved pressure on health care systems. However, the Cambodia for the first time in a decade. An estimated economic and social consequences of these measures 460,000 individuals fell into poverty in 2020, raising reached even economies that had few COVID-19 cases. the poverty rate by 2.8 percentage points. Household Worldwide, the pandemic caused a large recession and welfare and food security declined for many Cambodian the global economy contracted by 3.1 percent in 2020,56 households due to employment and income shocks, increasing global poverty for the first time in over 20 especially at the onset of the pandemic. Most households years.57 New virus variants have emerged, increasing reduced food and non-food expenditures to cope with uncertainty about how quickly the pandemic can be income losses or income uncertainties. The introduction overcome. of a new government social assistance cash-transfer This chapter examines the effects of the COVID-19 program to support the income of poor and vulnerable pandemic and policy response on the socioeconomic households helped mitigate negative pandemic effects, outcomes of households in Cambodia. The chapter curbing the increase in poverty that would have occurred.  56 I MF World Economic Outlook October 2021 (https://www.imf.org/en/Publications/WEO/Issues/2021/10/12/world-economic-outlook- october-2021). 57 World Bank Poverty and Shared Prosperity 2020 (https://www.worldbank.org/en/publication/poverty-and-shared-prosperity). CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 107 5.1 COVID-19 Timeline and Developments in Cambodia Cambodia mobilized its response to contain the 5.4). However, after the first community transmission spread of COVID-19 within 2 months of its onset. cluster was detected in Phnom Penh in November, some Cambodia recorded its first case of COVID-19 on January restrictions were reinstated. On November 29, 2020, 27, 2020. By March 2020, Cambodia had established Cambodia reported its first known local transmission a national response committee and an inter-ministerial cluster in Phnom Penh based on a patient with no response committee to combat COVID-19, led by the history of international travel, but the outbreak remained Prime Minister and the Ministry of Health. Cambodia contained. restricted international arrivals, closed education COVID-19 cases and deaths have since surged as institutions, garment factories, and entertainment venues, more infectious variants emerged. On February 20, and imposed restrictions on public gatherings and 2021, Cambodia detected its largest COVID-19 outbreak cancelled major public holidays. in Phnom Penh. Until then, Cambodia had reported little Cambodia successfully prevented a major disease more than 500 infections and no COVID-related deaths.59 outbreak in the first year of the pandemic. For more By early May 2021, Cambodia had almost 20,000 than a year after its first case in January 2020, Cambodia confirmed cases and more than 100 deaths (Figure 5.1). registered a limited number of COVID-19 cases and no Cases further rose to more than 30,000 cases in June, deaths. Most of the cases recorded throughout 2020 were more than 50,000 cases in July, and almost 80,000 in through direct contact with infected persons or imported August. By mid-April 2022, Cambodia had confirmed cases. 58 Given the limited spread of infection, most over 136,000 cases and over 3,000 deaths. restrictions were lifted by September 2020 (see Figure Figure 5.1 Cumulative Number of Confirmed Figure 5.2 Percentage of Population COVID-19 Cases and Deaths in Cambodia Vaccinated against COVID-19 in Cambodia 160,000 100 At least 1 dose Fully vaccinated Cumulative 140,000 cases 119588 80 120,000 100,000 60 80,000 60,000 40 40,000 Cumulative 20 20,000 deaths 2872 0 0 7 7 7 20 -27 7 7 7 7 7 21 -27 7 7 -0 7 27 20 -03 0 20 -04 0 20 -05 0 20 -06 0 20 -07 0 0 20 09 0 20 -10 0 20 -11 0 20 -12 0 0 20 -0 0 20 -03 0 -0 0 10 20 -2 20 -2 20 -2 20 -2 21 -2 21 -2 21 -2 21 -2 21 -2 22 -2 22 -2 21 2-1 21 -1 21 -1 21 -1 21 -1 21 -1 21 -1 21 -1 21 -1 21 -1 22 -1 22 -1 22 2-1 22 -1 3- 4- 20 -01 20 -03 20 -05 20 -07 20 -09 20 -11 20 -01 20 -03 20 -05 20 -07 20 -09 20 -11 20 -01 20 -08 20 -01 20 -0 - 0 21 2 20 20 Source: John Hopkins University CSSE COVID-19 Data. Source: John Hopkins University CSSE COVID-19 Data. (ourworldindata.org/coronavirus; accessed February 24, 2022). (ourworldindata.org/coronavirus; accessed February 24, Note: Dates are in Month/Day/Year format. 2022). Note: Dates are in Month/Day/Year format. 58 Reuters. January 2020. Cambodia confirms first case of coronavirus – health minister. (https://www.reuters.com/article/us-china-health- cambodia/cambodia-confirms-first-case-of-coronavirus-health-minister-idUKKBN1ZQ1C1) 59 Cambodia: Coronavirus Country Pandemic Profile (https://ourworldindata.org/coronavirus/country/cambodia) - Last accessed May 11, 2021. 108 CAMBODIA POVERTY ASSESSMENT Following the development of COVID-19 vaccines, 4, 2021 and began its vaccination program on February Cambodia raced to vaccinate its citizens and made 10. Since then, vaccination progress has been impressive. impressive strides. Launched in February 2021, the As of mid-April 2022, about 88 percent of Cambodians vaccination program helped counter the spread of the had received at least one COVID-19 vaccine dose and virus and severity of illness. Cambodia granted emergency about 83 percent were fully vaccinated against COVID-19 use authorization of the Sinopharm vaccine on February (Figure 5.2). 5.2 Employment and Income Shocks COVID-19 severely negatively affected Cambodia’s tourism sector, estimated to account for about 2 million economy and key sectors driving economic growth: jobs and one-quarter of GDP pre-pandemic, virtually construction, tourism, and merchandise export. collapsed. Cambodia’s key export sectors contracted These sectors had contributed to more than 70 percent due to weakened demand for garment, footwear, and of growth and 39 percent of total paid employment in travel goods exports. Orders from the United States 2019. 60 COVID-19 hit Cambodia’s economy in 2 waves and the European Union decreased considerably. The with the collapse in external demand in 2020 and construction and real estate sectors weakened after community spread of the virus in 2021. In 2020, the having contributed over one-third of real growth in 2019. economy contracted 3.1 percent, the sharpest decline in The agricultural sector grew merely 0.4 percent. Although Cambodia’s recent history (Figure 5.3). The service sector the economy is forecast to recover, growth in 2022 and was hit hardest due to travel restrictions. The travel and 2023 is expected to remain below pre-pandemic levels. Figure 5.3 Real GDP Growth and Sectoral Contributions to Growth 10 8 7.5 7.0 7.1 6.3 6 5.2 4.8 4 3.0 2 0 -2 -3.1 -4 2017 2018 2019 2020 2021e 2022p 2023p 2024p Agriculture Industry Services Net taxes on production GDP Growth Sources: Cambodia authorities and World Bank staff projections. Note: e=estimate; p = projection. 60 World Bank. 2020c. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 109 The pandemic caused widespread job losses and Mobility declined by more than from pre-pandemic suspensions. According to World Bank estimates in May levels when the first restrictions were imposed 2020, the pandemic put at least 1.76 million Cambodian in March 2020. This adversely affected the services jobs at risk. In October 2020, CNV International reported 61 sector profoundly, in addition to weakened consumer that between 35,000 and 40,000 garment and footwear demand from shocks to incomes and employment. Visits workers out of a total of 800,000 had been laid off since to public transit stations declined significantly (Figure pandemic onset. 62 Job losses were also magnified by A.10), decreasing visits to retail and recreational centers, an increased number of migrant workers returned to workplaces, and parks as Cambodians minimized social Cambodia. In August 2020, the International Labour interactions and increasing stayed home. Although Organization (ILO) estimated 100,000 Cambodian migrant mobility periodically improved, COVID-19 outbreaks and workers returned home since the start COVID-19. Upon 63 ensuing stringency measures decreased mobility (Figure return, many faced unemployment. The United Nations 5.4). Overall mobility, measured anonymously by Apple Development Programme (UNDP) predicted that the and Google using mobile devices, negatively correlated COVID-19 crisis could increase the unemployment rate to with government restrictions, measured by the Oxford 4.8 percent in 2020. 64 stringency index. A COVID-19 outbreak in February 2021 triggered the strictest restrictions and disrupted Workers in the service and hospitality, construction, mobility. High-risk cities introduced 8 p.m to 5 a.m. night and garment sectors were most affected. The travel curfews, which later became the focus of the country’s and tourism sector, estimated to have provided about 2 first lockdowns66 and restrictions non-essential business million pre-pandemic jobs, was brought to a standstill. activity and travel (except for medical emergencies). International arrivals fell 64.6 percent in the first 6 months Even food markets and supermarkets closed in certain of 2020. Mass factory closures due to COVID-19 affected neighborhoods in Phnom Penh with very high numbers thousands of livelihoods. Garment, textile, and footwear of COVID-19 cases.67 Among recent signs of economic factories, employing around 1 million Cambodians (largely stabilization, mobility has since been gradually recovering women from poor rural areas), closed starting in March as global and domestic stringency measures relax. 2020.65 Dragged down by the decline in foreign direct investment in Cambodian construction, Cambodia’s construction boom slowed, resulting in job suspensions and losses and reduced hours for construction workers. 61 World Bank (2021a), Cambodia Economic Update: Cambodia in the time of COVID-19, Phnom Penh, Cambodia. 62  NV International is recognized as a public benefit organization in the Netherlands, with a mission to contribute to decent work C opportunities in developing countries, including Cambodia. 63 ILO (2020), COVID-19: Impacts on Cambodian Migrant Workers. Phnom Penh, Cambodia. 64 UNDP (2020), COVID-19 economic and social impact assessment in Cambodia: CGE >AP simulation exercises, Phnom Penh, Cambodia. 65 World Bank (2020c). 66  hmer Times. May 5, 2021. (https://www.khmertimeskh.com/50851438/curfew-hours-in-phnom-penh-and-takhmao-city-to-be-revised- K to-8pm-and-3am/). 67  Reuters. April 24, 2021. (https://www.reuters.com/world/asia-pacific/cambodia-closes-markets-curb-covid-19-thousands-plead- food-2021-04-24/). 110 CAMBODIA POVERTY ASSESSMENT Figure 5.4 COVID-19 Stringency Index and Mobility Trends in Cambodia 100% 80% 60% 40% 20% Baseline -20% -40% -60% -80% -100% 20 0-0 3 20 -03 3 20 -0 3 20 -05 3 20 -0 3 20 -07 3 20 -08 3 20 -0 3 20 -10 3 20 -1 3 20 -12 3 20 -01 3 20 1-0 13 20 -03 13 20 -0 3 20 -05 3 20 -0 3 20 -07 3 20 -08 3 20 -0 3 20 -10 3 20 -1 3 20 -12 3 20 -01 3 -0 3 13 2 -1 20 2-1 20 -1 20 4-1 20 -1 20 6-1 20 -1 20 -1 20 9-1 20 -1 20 1-1 21 -1 21 -1 21 4-1 21 -1 21 6-1 21 -1 21 -1 21 9-1 21 -1 21 1-1 22 -1 22 -1 2 - 21 2- 2- 20 -01 20 20 Stringency index (0–100) Driving Walking Source: Hale, et al. (2021) COVID-19 Stringency Index (ourworldindata.org/coronavirus). Apple COVID-19 Mobility Trends 2020/21 (covid19.apple.com/mobility). Note: Dates are in Month/Day/Year format. Mobility is defined as requests for directions in Apple Maps. Mobility is relative to January 2020 baseline and expressed in percentage. Stringency index is a composite measure based on 9 indicators, rescaled to a value from 0 to 100 (100 = strictest). Indicators include school closures; workplace closures; cancellation of public events; restrictions on public gatherings; closures of public transport; stay-at-home requirements; public information campaigns; restrictions on internal movements; and international travel controls. Box 5.1 Cambodia High-Frequency Phone Survey (HFPS) The COVID-19 pandemic created an urgent need for timely data to help monitor and mitigate effects on Cambodians. The World Bank, in consultation with Cambodia’s National Institute of Statistics, designed the Cambodia High-Frequency Phone Survey (HFPS) to monitor the evolving socioeconomic effects of the COVID-19 pandemic on households and inform policy responses and interventions. The nationally representative survey covers employment and income, public safety nets, coping strategies, children’s engagement in learning activities, and food insecurity. The survey tracked Cambodian households over 10 months between May 2020 and March 2021. Selected respondents—typically the household head—completed interviews every 8 weeks. The Cambodia HFPS consists of 2 separate samples: (i) Living Standards Measurement Study Plus (LSMS+), and (ii) “IDPoor” households (IDPoor is Cambodia’s national poverty identification program and official targeting mechanism for programs that support the poor). The HFPS-LSMS+ sample was drawn from the nationally representative LSMS+ household survey implemented in October–December 2019. The HFPS-IDPoor sample was drawn from a list of beneficiaries of the conditional cash transfer program for pregnant women and children under age 2 with an IDPoor equity card. Cambodia High-frequency Phone Surveys (HFPS) workers and industry workers stopped working in the also highlight the negative initial effects of the early pandemic; in May 2020, 43 percent of those who pandemic on service and industry sector jobs. had stopped working were last employed in the services Between May 2020 and March 2021, about 11 percent and 30 percent were last employed in industry (Figure of Cambodian households had an adult who had stopped 5.5). Most of these work stoppages were in wholesale working at some point (Figure A.11). Mostly service sector and retail trade, construction, and manufacturing. Given CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 111 services employed nearly twice the adults industry survey round had engaged in agricultural work prior to before the pandemic, the number of service jobs lost the pandemic, suggesting that seasonality in employment was substantial. Since August 2020, at least half of all also played a role in work stoppages. respondents who had stopped working since the last Figure 5.5 Sectoral Distribution of Work Figure 5.6 Distribution of Work Stoppages Stoppages by Educational Attainment 100 100 13 27 24 24 80 35 31 35 % of work stoppages 80 % of work stoppages 43 43 42 38 9 17 23 60 60 30 38 34 40 30 40 29 28 67 49 56 53 20 20 27 35 31 28 31 30 0 0 May '20 Aug '20 Oct '20 Dec '20 Mar '21 May '20 Aug '20 Oct '20 Dec '20 Mar '21 Agriculture Industry Service Complete secondary/higher Complete primary Incomplete primary Source: Cambodia HFPS 2020/21. Source: Cambodia HFPS 2020/21. Poor and less-educated workers suffered most from because of seasonality in employment or because job losses, including women who disproportionately Cambodians needed to care for a sick relative. Flooding work in the garment, retail, and hospitality sectors in October affected 14 provinces including Phnom Penh (see Figure 5.6, Figure A.12). According to the HFPS, and disrupted farming activities. main earners in IDPoor households were more likely to Employment remained below pre-pandemic levels have left manufacturing and construction jobs. despite restrictions being lifted and mobility slowly Most work stoppages in the early months of the picking up in 2020. According to the HFPS, 82 percent pandemic were due to COVID-19-related business of respondents had been working prior to the COVID-19 closures or temporary layoffs. Although these work outbreak. In May 2020, 71 percent of households had an stoppages diminished in August 2020 as businesses adult working (Figure 5.7). In contrast, employment for the began to open, they rose again in December 2020 and main earners in poor households increased between May March 2021, possibly because of the COVID-19 outbreak and December 2020 (Figure 5.7). This possibly reflects in November 2020 and February 2021. Over time, an the poor taking on more income-generating work to cope increasing share of work stoppages also occurred with income shocks (see Figure 5.15). 112 CAMBODIA POVERTY ASSESSMENT Figure 5.7 Respondents (LSMS+) or Household’s Main Earner (IDPoor) Working in Last 7 Days 100 17 14 16 29 30 28 31 27 80 35 60 % 40 83 86 84 71 70 72 69 73 65 20 0 May '20 Aug '20 Oct '20 Dec '20 Mar '21 Aug '20 Oct '20 Dec '20 Mar '21 LSMS+ IDPoor Currently working Not currently working Source: Cambodia HFPS 2020/21. Cambodians who remained employed often suffered Non-farm household business revenues also fell reduced hours, pay cuts, and unpaid wages. due to weakened consumer demand. At the beginning Although most businesses were open by September of the pandemic, 81 percent of household businesses 2020, continued low sales caused firms to lay off workers reported having earned “less” or “no revenues” relative and reduce hours or wages. 68 Workers in the garment, to the month prior (Figure 5.8). In March 2021, even after footwear, and tourism industries worked on average the initial shock diminished, 50 percent of household 20 days in July 2020, 6 fewer days than in July 2019. businesses reported having earned less or no revenues. Reduction in working hours decreased earnings, with Meanwhile, the share of households reporting their the average monthly salary for garment, footwear, and business revenues “stayed the same” tripled between tourism workers declining from US$220 (KHR 893,453) in May 2020 and March 2021. July 2019 to US$190 (KHR 777,629) in July 2020. 69 Figure 5.8 Changes in Sales Revenues in Non-farm Household Businesses Relative to Last Month 5 5 5 5 4 8 5 10 7 100 15 80 31 29 40 37 46 44 42 46 60 % 73 40 61 62 52 58 46 48 48 40 20 8 3 3 4 2 4 5 3 0 May '20 Aug '20 Oct '20 Dec '20 Mar '21 Aug '20 Oct '20 Dec '20 Mar '21 LSMS+ IDPoor No revenue Less The same Higher Source: Karamba, Salcher and Tong (2021b). 68 Cambodia Business Pulse Survey 2020. Round 1 (June 2020) and 2 (September 2020). 69 Ngo et al. (2021). The currency conversion used is KHR 4061.2 per US$ for 2019 and KHR 4092.8 per US$ for 2020. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 113 Shocks to employment, working hours, and earnings more on wage employment, were more likely to report led to significant declines in household income. reduced wage income (Figure 5.10). About 40 percent Over 40 percent of households reported lower income in of farm households and almost 60 percent of non- March 2021 compared to pre-pandemic January 2020. farm business households reported reduced income. Most Cambodian livelihoods depend on labor income, Households experiencing income losses from wages, especially from wage employment, farming, and family farming, and businesses reported average 45 percent businesses (Figure 1.17). In March 2021, about 30 percent income decline relative to January 2020. Remittances of wage employment households reported reduced wage and assistance from friends and family also decreased, income (Figure 5.9). Low-income households, who rely but affected few households. Figure 5.9 Incidence of Changes in Figure 5.10 Incidence of Changes in Household Income Relative to January Household Income Relative to January 2020 (LSMS+) 2020 (IDPoor) Total HH income 41 40 18 Total HH income 23 45 32 Family farm 40 49 11 Family farm 41 53 6 Non-farm family biz 57 40 3 Non-farm family biz 59 31 10 Wage empl. 30 59 11 Wage empl. 47 49 4 Assistance Govt/ NGO 4 11 85 Assistance Govt/ NGO 33 95 Assistance family /… 42 58 Assistance family /… 10 35 55 Remittances 49 34 17 Remittances 67 33 Pension 3 85 13 Pension 83 17 Income from property 100 Income from property 100 0 20 40 60 80 100 0 20 40 60 80 100 % % Reduced Stayed the same Increased Reduced Stayed the same Increased Source: Cambodia HFPS Round 5 (March 2021). Source: Cambodia HFPS Round 5 (March 2021). Household income losses remained widespread Moderate food insecurity was elevated at in 2020, but losses slowed as economic activity the beginning of the pandemic because most showed signs of recovery. Restrictions lifted and households worried about not having enough mobility slowly picked up throughout 2020 (Figure 5.4). to eat due to lack of money. Despite the COVID-19 Employment showed signs of recovery when the percent pandemic, markets remained functional and access to of employed respondents in December 2020 returned food staples remained robust, but some households to May 2020 levels (Figure 5.7). In March 2021, fewer experienced food insecurity due to lack of money.70 Poor households (45 percent) reported declines in income households most often shared experiences of worrying, compared to May 2020 (83 percent) (Figure 5.11). rationing foods, and changing diets because they did not Households also reported their household labor income have enough money or resources. In August 2020, the decreased by 37 percent since the last round, somewhat prevalence of moderate or severe food insecurity was lower than reported reduction of 40 percent in August higher among IDPoor households (67 percent) compared 2020 (Figure 5.12). to all Cambodian households (48 percent) (Figure 5.13). 70 n Cambodia, food insecurity typically takes the form of low-quality diets and reduced food quantities. More severe forms of food insecurity, I such as not eating for entire days, are rare, thus the prevalence of severe food insecurity is very low. 114 CAMBODIA POVERTY ASSESSMENT Figure 5.11 Incidence of Changes in Total Figure 5.12 Reduction in Household Labor Household Income since the Last Survey Income since the Last Survey Round Round 100 1 1 5 LSMS+ IDPoor 8 7 9 8 12 9 9 16 11 0 80 29 42 32 43 43 47 46 45 -10 60 % 88 -20 40 83 % 63 57 51 48 46 52 -30 20 45 44 -40 0 -37 May Aug Oct Dec Mar Jun Aug Oct Dec Mar -41 -40 -40 -43 -44 -40 '20 '20 '20 '20 '21 '20 '20 '20 '20 '21 -50 -45 LSMS+ IDPoor Reduced Stayed the same Increased Aug '20 Oct '20 Dec '20 Mar '21 Source: Karamba, Salcher and Tong (2021b). Source: Karamba, Salcher and Tong (2021b). Percentage of Population Experiencing Food Insecurity Figure 5.13  100 80 67 60 48 % 39 35 40 34 17 17 20 20 5 4 3 1 1 1 1 1 0 Aug '20 Oct '20 Dec '20 Mar '21 Aug '20 Oct '20 Dec '20 Mar '20 LSMS+ IDPoor Moderate-or-severe food insecurity Severe food insecurity Source: Karamba, Salcher and Tong (2021b). Note: Food insecurity prevalence rates are based on the application of the Food Insecurity Experience Scale (FIES), an experience- based metric of food insecurity severity that relies on people’s direct responses to 8 questions about their access to food during the 30 days preceding the survey. Cambodian households reduced consumption and (NGOs), and relying on insurance. Potentially scarring adopted low-cost and potentially scarring strategies coping mechanisms—selling assets, taking children out to cope with the initial COVID-19 shock. Reducing of school, taking loans from formal/informal financial food and non-food consumption was the most common institutions, crediting purchases, and delaying payment coping mechanism households adopted. In May 2020, 65 obligations—were the third most common strategy percent of households reduced their food consumption adopted by 58 percent of households. and 61 percent reduced non-food consumption (Figure Poor households were more likely to reduce 5.14). Low-cost coping mechanisms were the second consumption and adopt low-cost and potentially most common strategy adopted by 64 percent of scarring coping mechanisms. This suggests they had households. These included engaging in additional a limited financial cushion on the eve of the COVID-19 income-generation, borrowing, or receiving aid from crisis, let alone to endure months of a pandemic. For family/friends, relying on savings, relying on assistance instance, in May 2020, 89 percent of IDPoor households from the government or non-governmental organizations CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 115 reduced food consumption in response to COVID-19 to 65 percent of Cambodian households. Further, 79 compared to 65 percent of all Cambodian households percent of IDpoor households resorted to potentially (Figure 5.15, Figure 5.14). Another 83 percent of IDPoor scarring coping strategies—crediting purchases, delaying households resorted to low-cost coping strategies— payment obligations, and taking out loans—compared largely borrowing from friends and family—compared with 58 percent of Cambodian households. Figure 5.14 Actions Adopted by Cambodian Figure 5.15 Actions Adopted by IDPoor Households in Response to COVID-19 crisis Households in Response to COVID-19 crisis Reduce food consumption 65 Reduce food consumption 89 68 86 Reduce non-food consumption 61 Reduce non-food consumption 84 59 72 Credit purchases 36 Credit purchases 65 43 59 Borrow from friends & family 27 Borrow from friends & family 45 29 51 Assistance from government 5 Assistance from government 45 27 96 Delay payment obligations 20 31 26 Delay payment obligations 35 Additional income generation 17 22 25 Additional income generation 36 Loan from financial institution 13 16 22 Loan from financial institution 28 17 23 Sale of assets 15 Sale of assets 18 9 May '20 23 May '20 Assistance from NGO 13 Assistance from NGO 32 Dec '20 Dec '20 14 7 Children took on HH chores 7 Children took on HH chores 8 5 9 Sell harvest in advance 6 Sell harvest in advance 6 18 16 Assistance from friends & family 6 Assistance from friends & family 4 2 6 Take advanced payment from employer 4 Take advanced payment from… 8 20 10 Rely on savings 3 Rely on savings 2 Covered by insurance policy Covered by insurance policy 0 20 40 60 80 100 0 20 40 60 80 100 % % Source: Karamba, Salcher and Tong (2021b). Source: Karamba, Salcher and Tong (2021b). IDPoor sample LSMS+ sample. of households with valid equity card. Households increasingly used potentially scarring additional income-generating activities. Poor household coping strategies as the pandemic progressed, enrolled in the IDPoor program had access to much but poor households relied mostly on low-cost needed income support, mitigating the extent to which strategies instead, thanks to increased government they resorted to potentially-scaring coping strategies. The income support. In December 2020, signs emerged share of households receiving government assistance that households were crediting purchases, delaying significantly increased between May and December 2020 payments obligations, or taking out loans in response to (Figure 5.15). In addition, more households received the COVID-19 crisis, actions that could have potentially assistance from NGOs, although government assistance scarring effects. Households also increasingly engaged in continued to play a larger role. 116 CAMBODIA POVERTY ASSESSMENT 5.3 COVID-19 Poverty and Inequality Effects and Mitigation Policies In June 2020, the RGC launched a new cash transfer The proportion of IDPoor households with a valid “equity” program to support poor and vulnerable households card (henceforth eligible IDPoor households) receiving during the pandemic.71 Launched on June 24, 2020, government social assistance increased considerably the COVID-19 relief cash transfer has disbursed monthly from 50 percent in June 2020 to 91 percent in August 2020 cash payments to households identified as poor through (Figure 5.16). Nearly all social assistance was delivered in the pre-existing “IDPoor” national poverty identification the form of cash. Demand for the program was high with program. As of July 2022, the program had disbursed almost all registered beneficiary households, regardless US$714 million in cash transfers, reaching almost 690,000 of region of residence, collecting their cash payments. In households (2.7 million individuals) or about 17 percent of March 2021, only 5 percent of IDpoor households had the population.72 This represented a dramatic increase in not received cash transfers (Figure 5.17). Some eligible social assistance compared to pre-COVID-19 levels. households remained uncovered because they did not register for the program. The COVID-19 cash transfer program has had high reach and high take-up among IDPoor households. Figure 5.16 Eligible IDPoor Households Figure 5.17 Eligible IDPoor Households Receiving Social Assistance from the that Received the Relief Cash Transfers Government 100 95 96 100 93 95 91 92 90 92 80 80 60 60 50 % % 40 40 20 20 0 0 Jun '20 Aug '20 Oct '20 Dec '20 Mar '21 Aug '20 Oct '20 Dec '20 Mar '21 Source: Karamba, Salcher and Tong (2021b). Source: Karamba, Salcher and Tong (2021b). In the absence of government intervention, the without fiscal intervention, thereby increasing the poverty pandemic could have increased poverty by 4.7 rate 4.7 percentage points (Figure 5.18, Figure 5.19). pp in 2020, which would have reversed 3 years of Households reliant on non-agricultural wage income, Cambodia’s poverty reduction progress. About particularly from the garment and construction sectors, 750,000 individuals could have entered poverty in 2020 were hit hardest (Figure 5.20). The number of new poor 71 C  ambodia implemented several key policy measures in response to COVID-19 including the: (1) COVID-19 relief cash transfers for poor and vulnerable households, (2) unemployment benefits for laid-off workers in the garment/footwear/travel goods manufacturing and tourism sectors, (3) tax relief for firms in the garment/footwear/travel goods manufacturing and tourism sectors, (4) tax exemption for property purchases below US$70,000, and (5) establishment of a new bank to provide financing to small and medium enterprises. 72 M  oSVY, Notification of the implementation result of cash transfer programs to poor and vulnerable household during the fight against COVID-19, Phnom Penh, Cambodia. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 117 was the highest among manufacturing wage employees, (Figure 5.21). Geographically, the poor and new poor followed by garment employees, construction workers, concentrate in 2 of the most populated areas, the Plains wholesale and retail, and hotel and restaurant workers and the Tonle Sap regions. Figure 5.18 Simulated Number of Poor Figure 5.19 Simulated Poverty Rate in 2020 in 2020 4.0 3.7 25 23.1 2.9 20 18.4 3.0 2.8 17.8 Percent of population Million people 15 2.0 10 1.0 5 0.0 0 CSES 2019/20 No shock Shock CSES 2019/20 No shock Shock Source: Authors’ calculations based on CSES 2019/20. Note: Following Bourguignon and Ferreira (2005) and Bourguignon (2008), poverty is simulated based on actual GDP, sectoral growth rates, and estimated employment figures for 2020. Simulations were implemented using the ADePT macro-simulation module. GDP is classified into 3 broad economic sectors: agriculture, industry, and services. Figure 5.20 Simulated Number of Poor, by Main Income Source 1500 1410 Number of the poor ('000) Already poor New poor 1000 597 640 500 357 232 171 88 101 28 43 18 4 0 Agriculture Non-agriculture Agriculture Non-agriculture Remittances Other (self-employed) (self-employed) (wage) (wage) Figure 5.21 Simulated Number of Poor, by Wage Income Source 500 422 Number of the poor ('000) 400 381 339 Already poor New poor 300 190 200 147 115 100 39 36 47 25 0 Manufacture Construction Whole & retail sale Hotel & restaurant Garment Source: Authors’ calculations based on CSES 2019/20. Note: Simulations based on actual GDP, sectoral growth rates, and estimated employment figures for 2020. Simulations were implemented using the ADePT macro-simulation module. GDP is classified into 3 economic sectors: agriculture, industry, and services. 118 CAMBODIA POVERTY ASSESSMENT Box 5.2 Cambodia’s COVID-19 Relief Cash Transfers Bracing for the economic impact of COVID-19, the Royal Government of Cambodia (RGC) moved quickly to fund and scale-up a cash transfer program for poor households. The program leveraged the country’s existing “IDPoor” identification system and built on the existing infrastructure of the Cash Transfer Program for Poor Pregnant Women and Children. Households identified as poor through the IDPoor system qualify for several social services. Currently households with an” IDPoor equity” card can access several national programs such as the free access to health care under the Health Equity Fund, financial assistance through the Cash Transfer Program for Poor Pregnant Women and Children (0–2 years old), and the Scholarship Program for Children in Primary and Secondary Schools, in addition to the COVID-19 Relief Cash Transfer Program. COVID-19 Relief Cash Transfers depend on level of poverty, area of residence, and household demographics (presence of elderly or children) (Box Table 5.1). Each eligible household receives (i) KHR 80,000-120,000 (US$20–29) per month depending on area of residence, (ii) an additional allocation ranging between KHR 16,000-52,000 (US$4–13) for each household member, and (iii) KHR 16,000–40,000 (US$4–10) for each child aged zero to 5, elderly person aged 60 and above, and a household member with a disability or HIV/AIDS. By March 2021, most beneficiary households had received 9 installments of the COVID-19 relief cash transfers, totaling an average of US$366 per household.73 Box Table 5.1 Monthly Cash Transfers during COVID-19 (KHR) Phnom Penh Other urban Rural IDPoor 1 IDPoor 2 IDPoor 1 IDPoor 2 IDPoor 1 IDPoor 2 Household 120,000 120,000 120,000 120,000 80,000 80,000 Each member 52,000 36,000 40,000 28,000 24,000 16,000 Additional top-ups for vulnerable members Child aged 0–5 40,000 28,000 40,000 28,000 24,000 16,000 Adult aged 60+ 40,000 28,000 40,000 28,000 24,000 16,000 Disabled member 40,000 28,000 40,000 28,000 24,000 16,000 Member with HIV/AID 40,000 28,000 40,000 28,000 24,000 16,000 Source: Ministry of Social Affairs, Veterans and Youth Rehabilitation (2020). The cash transfer program is implemented primarily through the Ministry of Planning (MoP) and Ministry of Social Affairs, Veterans and Youth Rehabilitation (MOSVY). The MoP is responsible for beneficiary identification through IDPoor, including making data on eligible households available to the MoSVY, which manages registration, benefit calculations, and payment processes. The IDPoor system uses a proxy means test community members implement to identify households living in poverty; a commune administrator interviews households using handheld tablets for data entry and the On Demand-IDPoor Mobile Application (OD-IDPoor App)—a data collection tool designed for conducting household interviews, synchronizing data with the national IDPoor database, and issuing Equity Cards to poor households in the scope of the On Demand Identification Process. Any IDPoor equity cardholder can register themselves for the cash transfers with the commune administrator who verifies household identity. The cardholder receives an account exclusively dedicated to the transfer program from an e-payment provider. The program payment mechanism relies on the WING services, a leading mobile payments provider with a network of more than 9,000 payment agents across Cambodia. WING receives automated access to relevant fields of the IDPoor database so that each account can be associated with every owner’s IDPoor number and receive government beneficiary money. By agreement, WING can establish new agents in any location not well serviced.74 73 Karamba, Salcher and Tong (2021b). 74 World Bank (2021). Cambodia COVID-19 Relief Transfers: a review of the payment mechanisms. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 119 Cash transfers to poor and vulnerable households measured at the national poverty line. The poverty rate in mitigated increased poverty and inequality. The cash 2020 thus increased about 2.8 percentage points from transfers are projected to have curbed the number of new the official rate in 2019/20 instead of the 4.7 percentage poverty entrants to about 460,000 instead of 750,000 points it could have increased without cash relief. The that could have entered poverty in the absense of cash transfers also mitigated increased inequality, with the government intervention (Figure 5.22). That is, COVID-19 Gini index estimated to have increased by 0.2 percentage cash transfers prevented almost 300,000 people from points in 2020 rather than the 0.4 percentage points had falling into poverty in 2020. As such, the income support cash transfers not been introduced. likely mitigated 40 percent of the increase in poverty as Figure 5.11 Simulated poverty change with and without cash transfers A. Simulated Change in the Number of Poor B. Simulated Change in Poverty Rate 800,000 748,406 5.0 4.7 Percentage points change 4.0 600,000 Number of the poor 458,226 2.8 3.0 400,000 2.0 200,000 1.0 0 0.0 No government With government No government With government intervention intervention (CT) intervention intervention (CT) Source: Authors’ calculations based on CSES 2019/20, GDP and sectoral growth rates, and estimated employment figures for 2020. Note: Simulations were implemented using the ADePT macro-simulation module. COVID-19 cash transfers did not fully offset not large enough to fully mitigate COVID-19 negative negative pandemic welfare effects on the poor. consumption effects. The COVID-19 relief cash transfer Even with the cash transfers, per capita consumption of program had no effect on households outside the bottom, the lowest quintile declined by 25 percent (Figure 5.23). who were largely ineligible for the program. This suggests that direct cash transfers to the poor are Figure 5.23 Simulated Welfare Loss in 2020 with and without COVID-19 Cash Transfers, by Consumption Quintile Poorest 20% 2 3 4 Richest 20% 0 -10 Percentage change -10 -10 -13 -13 -12 -13 -12 -14 -20 -30 -25 -29 -40 No government intervention With government intervention (CT) Source: Authors’ calculations based on CSES 2019/20, GDP and sectoral growth rates, and estimated employment figures for 2020. Note: Simulations were implemented using the ADePT macro-simulation module. 120 CAMBODIA POVERTY ASSESSMENT Despite not fully mitigating welfare losses, beneficiary that it was “very important” (Figure 5.24). About 42 households perceived the COVID-19 cash transfers percent of beneficiary households felt the program made a to have been important for economic well-being. In “complete difference” to household economic well-being, March 2021, about 9 months into the program, about while another 37 percent felt it made a “strong difference” 38 percent of beneficiary households reported that the (Figure 5.25). Beneficiary households with labor income program was “extremely important” for their household’s losses expressed similar perceptions. economic well-being, while another 40 percent reported Figure 5.24 Perceived Importance of Relief Figure 5.25 Perceived Difference Relief Cash Transfers for Household Economic Cash Transfers Made to Household Well-being Economic Well-being % of beneficiary households % of beneficiary households 11 23 22 19 19 22 44 40 33 27 37 39 45 38 46 47 42 37 Oct–Nov '20 Dec '20 –Jan '21 Mar '21 Oct–Nov '20 Dec '20 –Jan '21 Mar '21 Extremely important Very important Complete difference Strong difference Moderately important Not so important Moderate difference Slight difference Not important at all No difference Source: Cambodia HFPS 2020/21. Source: Cambodia HFPS 2020/21. The COVID-19 cash transfer program also likely or-severe food insecurity declined and remained stable mitigated food insecurity. In May 2020, prior to the (Figure 5.13). Cash transfers were essential for buying food launch of the cash transfer program, nearly 90 percent of and other necessities. Nearly all beneficiary households IDPpoor households reported reducing food consumption used the cash transfers to purchase food, while more in response to the COVID-19 crisis (see Figure 5.15). In than half of beneficiaries used them to purchase other December 2020, after the launch of the cash transfer essential items. These transfers allowed the average program, the proportion of IDPoor households reducing beneficiary household with 5 members to meet about food consumption did not increase further. During the first 20 percent of monthly food and 10 percent of monthly 6 months of the program, the prevalence of moderate- minimum consumption requirements. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 121 Box 5.3 COVID-19 and Policy Response Effects on Children’s Education Education has been severely affected by the pandemic and the government’s efforts to contain it. On March 16, 2020—days after the World Health Organization (WHO) declared COVID-19 a global pandemic—Cambodia’s Ministry of Education, Youth, and Sports (MoEYS) closed all education institutions, disrupting learning in the country’s 13,482 schools—both public and private—affecting 3,210,285 students and 93,225 teachers.75 This would be 1 of 3 closures between 2020 and 2021, lasting 5, 1, and 7 months respectively (see Figure A.13). While health and safety measures reduce health risks, long school closures have created enormous challenges in reintegrating students and recouping lost learning. School closures during COVID-19 disrupted learning, especially for students from poor families who rely more on in-person instruction. Before the pandemic, 91 percent of Cambodian households with school-age children (ages 6–17) participated in education. This proportion fell to 62 percent when schools closed for the first time in May 2020 (Box Figure 5.1). As schools reopened to in-person instruction, participation in learning activities correspondingly increased. Although remote alternatives were launched, unequal access to, and effectiveness of, remote learning likely will increase the gap in learning outcomes between rich and poor. Households with School-age Children Engaged in Education or Box Figure 5.1  Learning Activities (%) 100 80 60 40 20 0 May/June '20 Aug '20 Oct '20 Mar '21 LSMS+ IDPoor Source: Karamba, Salcher and Tong (2021b). According to the World Bank Cambodia Economic Update, Cambodia’s current cohort of students stands to lose 1.5 learning-adjusted years of schooling—15 percent of pre-pandemic expectations.76 Learning at school could suffer from the pandemic’s effect on household income. Income shocks risk increasing student dropouts if families withdraw children to work or are not able to afford school costs. According to MoEYS, approximately 50 to 75 percent of students at all levels—pre-primary, primary, lower secondary, and upper secondary—returned to school after the first reopening in 2020.77 Cambodia’s total expected years of schooling are already among the lowest in mainland Asia (see HCI discussion, Chapter 2). Cambodia has a long way to go, especially compared to the region’s high-income countries, which perform better in total expected and learning-adjusted years of schooling. Addressing schooling participation and quality gaps will be important for increasing Cambodian worker skills and productivity, and aggregate growth, and other development outcomes. 75 MoEYS, Cambodia Education Response Plan to COVID-19 Pandemic, July 2020, p. 6. The number of affected schools, students, and teachers does not include tertiary/higher education and nonformal education institutions. 76 World Bank (2021c). 77 World Bank (2021c). 122 CAMBODIA POVERTY ASSESSMENT 5.4 Conclusion Cambodia’s heavy reliance on export-oriented wide network of payment service providers helped to manufacturing, tourism, and construction made the quickly get money into the hands of those in need.78 economy and employment susceptible to external The transfer amount was credited electronically to (iii)  shocks. This left Cambodia’s economy especially each household, and could be cashed-out at the vulnerable to reduced global consumption or restricted local WING Bank or agent locations. This payment international and domestic travel. Together, the sectors mechanism offered a quick way to deliver benefits in accounted for 70 percent of growth and 39 percent of a largely unbanked nation and allowed beneficiaries total paid employment in 2019. The shock to export to withdraw cash, a preferred method of payment demand, along with pandemic stringency measures, among poor beneficiaries.79 This promoted greater naturally triggered a downturn in manufacturing, tourism, inclusion among poor populations, of whom 4 in 5 and construction employment, as well as employment adults do not have access to a bank or mobile money in closely linked sectors such as food, retail, and account.80 transportation. Employment in other sectors also suffered due to weakened consumer demand from reduced Social assistance transfers provided valuable incomes. Most jobs lost in the crisis were in industries income support to poor households during the that pay low average wages but are important for low- pandemic and mitigated the increased poverty and income households. inequality. COVID-19 cash transfers prevented almost 300,000 people from falling into poverty in 2020. As such, Government social assistance to poor and the income support likely mitigated 40 percent of the vulnerable households scaled-up considerably increase in poverty as measured at the national poverty during the pandemic. The scale-up was unprecedented line. The poverty rate in 2020 thus increased about 2.8 in terms of Government spending, scale of payments, and percentage points from the official rate in 2019/20 instead the share of the population receiving social assistance. of the 4.7 percentage points it could have increased Several features helped quickly deliver emergency cash without cash relief. The Gini index also increased an transfers to poor and vulnerable populations: estimated 0.2 percentage points in 2020 rather than The existing, well-established “IDPoor” registry of ( i)  0.4 percentage points had the cash transfers not been poor and vulnerable households and identification provided. Beneficiaries also felt the income support was process made it possible to rapidly introduce and needed and made a difference to their economic well- scale-up new emergency social assistance. The being, including to buy food. Government leveraged the IDPoor database to While the magnitude of social assistance is of deliver relief cash transfers to about 90 percent of historical proportions and COVID-19 cash transfers registered households within 2 months of launch. are clearly helping, their adequacy is in question. Barriers to beneficiaries enrollment in the program (ii)  First, although assistance was rapidly scaled up, coverage or accessing funds were low. A simple registration and adequacy left room for improvement. Analysis shows process, payment collection process, and access to a that the poorest quintile saw a 25 percent reduction in per capita consumption during COVID-19 even with cash 78 Karamba, Salcher and Tong (2021b). 79 Karamba, Salcher and Tong (2021a). 80 Karamba, Salcher and Tong (2021a) and Global Findex Database 2017. CHAPTER 5. COVID-19 NEGATIVE EFFECTS ON CAMBODIAN HOUSEHOLDS 123 transfers. Second, the crisis is shedding light on gaps Recovery: As the crisis winds down, it will be in the current social protection system for those in the important to avoid withdrawing much-needed middle of the income distribution. The COVID-19 crisis social assistance prematurely. Households need an negatively affected twice the share of households who opportunity to stabilize their livelihoods and recoup receive social assistance, but current social assistance some financial losses incurred during the pandemic only cover households at the bottom of the distribution in and to repay debts. That is, viewing the cash transfer the IDPoor database, failing to provide safety nets for the program not only as an emergency relief but also a vulnerable, near poor. Third, evaluating program success part of a recovery strategy. based on poverty averted can give a false impression; Rebuilding livelihoods: During recovery, the cash a large proportion of Cambodians cluster around the transfer program can be leveraged to support poverty line, and many could avoid falling into poverty with livelihoods, providing opportunities for employment even small assitance (see Chapter 1, Figure 1.7). Despite or to acquire economically productive assets. challenges, the good news is the COVID-19 cash transfer relief program has lasted the duration of the pandemic, Building resilience: The Government could with authorities extending the program throughout 2022 consider integrating cash transfers into the menu while introducing additional assistance. of social protection schemes as part of long-term development to boost incomes and savings of Policymakers can reassess cash transfer goals. the poor. This can be achieved by providing more Cash transfers provide protection for the poor and will and better targeted cash transfers and promoting be important to support recovery, rebuild livelihoods, and financial inclusion. build resilience. Going forward, policymakers need to consider if cash transfers will only fulfill a protection goal to Authorities must weigh these spending decisions reduce current poverty—and thereby remain a temporary as part of broad financial analysis. Of course, the measure—or if they have a role as part of a broader future Government must review its sources of revenue and poverty reduction strategy. Options include: level of fiscal space and weigh the cost of cash transfers against other priorities, and without compromising on Immediate protection for poor and vulnerable education and health investment. This may require households: Increasing the amount of cash transfers evaluating the tax system to explore sources of revenue. It to the poor will offset losses in household welfare further requires identifying an optimal tax mix to raise anti- and help cover essential food, non-food, and income poverty resources while avoiding disproportionate burden needs. Expanding coverage of social assistance to on the poor. Ongoing tax system review and reforms will households not eligible for the IDPoor program but help, while strengthening humanitarian and development vulnerable to falling into poverty may be needed. community partnerships can mobilize resources for education, health, or large-scale cash transfers. 124 CAMBODIA POVERTY ASSESSMENT CHAPTER 2. MULTIPLE POVERTY DIMENSION TRENDS 125 REFERENCES Alfani Federica, Carlo Azzarri, Macro d’Errico and Vasco Molini (2012), Poverty in Mozambique: New Evidence from Recent Household Surveys. Policy Research Working Paper 6127, World Bank, Washington DC. Arndt Channing and Simler Kenneth (2010), Estimating Utility-Consistent Poverty Lines with Applications to Egypt and Mozambique, Economic Development and Cultural Change, University of Chicago Press, 58(3), 449-474. Azevedo Joao Pedro, João Pedro Azevedo, Gabriela Inchauste, Sergio Olivieri, Jaime Saavedra and Hernan Winkler (2013), Is labor income responsible for poverty reduction? A decomposition approach. Policy Research Working Paper 6414, World Bank, Washington DC. Bourguignon François and Francisco Ferreira (2005), Decomposing changes in the distribution of household incomes: Methodological Aspects. In Microeconomics of income and distribution dynamics in East Asia and Latin America, edited by Bourguignon, François., and Ferreira, Francisco H. G., and Lustig, Nora. Oxford University Press and World Bank, Washington DC. Central Statistical Organization (CSO), United Nations Development Programme and World Bank (2020), Myanmar Living Conditions Survey 2017, Report 3: Poverty Report. CSO, UNDP and WB, Nay Pyi Taw. Chan Sophal (2009), Costs and Benefits of Cross-Country Labour Migration in the GMS: Cambodia Country Study. Working Paper 47, Cambodia Development Resource Institute, Phnom Penh. Cunningham Wendy and Claire H. Hollweg with Johanne Buba, Marong Chea, Dalika Lathapipat, Une Lee, Anne Lopez, Sodeth Ly, Lan Nguyen, Miguel Sanchez, and Maheshwor Shrestha (2019), Cambodia’s Future Jobs: Linking to the Economy of Tomorrow (Vol. 2): Main Report (English). World Bank, Washington, D.C. http://documents.worldbank.org/ curated/en/224811567668159930/Main-Report Dave Rutu, Sandra Keller, Bryan Bonsuk Koo, Gina Fleurantin, Elisa Portale and Dada Rysankova (2019), Cambodia Beyond Connections: Energy Access Diagnostic Report Based on the Multi-Tier Framework. World Bank, Washington DC. https://openknowledge.worldbank.org/handle/10986/29512 Datt Gaurav and Martin Ravallion (1992), Growth and redistribution components of changes in poverty measures: A decomposition with applications to Brazil and India in the 1980s. Journal of Development Economics, 38(2), 275-295. Deaton Angus and Zaidi Salman (2002), Guidelines for Constructing Consumption Aggregates for Welfare Analysis. LSMS Working Paper 135. World Bank, Washington DC. Demirguc-Kunt, Asli; Klapper, Leora; Singer, Dorothe; Ansar, Saniya; Hess, Jake. 2018. Global Findex Database 2017: Measuring Financial Inclusion and the Fintech Revolution. Washington, DC: World Bank. © World Bank. https:// openknowledge.worldbank.org/handle/10986/29510 License: CC BY 3.0 IGO. 126 CAMBODIA POVERTY ASSESSMENT Elbers Chris, Jean O. Lanjouw, and Peter Lanjouw (2003), Micro-Level Estimation of Poverty and Inequality. Econometrica, 71(1), 355-64. Enami Ali, Lustig Nora and Aranda Rodrigo (2017), Analytic Foundations: Measuring the Redistributive Impact of Taxes and Transfers. Working Paper 446, Center for Global Development. Food and Agriculture Organization and the World Bank (2018), Food Data Collection in Household Consumption and Expenditure Surveys Guidelines for Low- and Middle-Income Countries. Rome. I04 pp. License: CC BY-NC-SA 3.0 IGO Ferreira Francisco H.G., Jeremie Gignoux, and Meltem Aran (2010), Measuring Inequality of Opportunity with Imperfect Data: The Case of Turkey. Policy Research Working Paper 5204, World Bank, Washington, DC. Foster James, Joel Greer and Erik Thorbecke (1984), A Class of Decomposable Poverty Measures. Econometrica, 52(3), 761-766. https://doi.org/10.2307/1913475. Fuchs Victor (1986), The feminization of poverty? Working Paper No. 1934, National Bureau of Economic Research, Cambridge. Thomas, Noam Angrist, Rafael Goldszmidt, Beatriz Kira, Anna Petherick, Toby Phillips, Samuel Webster, Emily Hale  Cameron-Blake, Laura Hallas, Saptarshi Majumdar and Helen Tatlow (2021), A global panel database of pandemic policies (Oxford COVID-19 Government Response Tracker). Nature Human Behaviour. (https://doi.org/10.1038/s41562- 021-01079-8)  Hing Vutha, Lun Pide and Phann Dalis (2014), The Impacts of Adult Migration on Children’s Well-being: The Case of Cambodia. Cambodia Development Resource Institute, Phnom Penh. IIona Hamilton and Sunjum Jhaj (2020), Why are boys leaving lower secondary school early in Cambodia?, UNICEF Cambodia, Education Section, Research Report. International Labour Organization (2020), COVID-19: Impacts on Cambodian Migrant Workers. International Labour Organization, Phnom Penh.  Kakwani Nanok (1977), Measurement of Tax Progressivity: An International Comparison. Economic Journal, 87(345), 71-80. Karamba Wendy, Clarence Tsimpo Nkengene and Kimsun Tong (2020), The Socioeconomic Impacts of COVID-10 on Households in Cambodia: Results from the High-Frequency Phone Survey of Households Round 1: Main Report (English). Impacts of COVID-19 on Households in Cambodia. World Bank, Washington, D.C. http://documents.worldbank.org/curated/en/520011610516176623/Main-Report Karamba Wendy, Salcher Isabelle, and Tong Kimsun (2021a), The Socioeconomic Impacts of COVID-19 on Households in Cambodia, Report No. 4 (Vol 2): Results from the High-Frequency Phone Survey of Households Round 4 (17 December 2020-12 January 2021). World Bank, Washington, DC. http://documents.worldbank.org/curated/ en/165331617235367415/Key-Findings Karamba Wendy, Salcher Isabelle, and Tong Kimsun (2021), The Socioeconomic Impacts of COVID-19 on Households in Cambodia, Report No. 5: Results from the High-Frequency Phone Survey of Households Round 5 (1-21 March 2021). World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35868 License: CC BY 3.0 IGO. REFERENCES 127 Lambert Peter (2001), The Distribution and Redistribution of Income. 3rd Edition. Manchester University Press. Lusting Nora and Sean Higgins (2018), The CEQ Assessment: Measuring the Impact of Fiscal Policy on Inequality and Poverty. Chapter 1 in Nora Lusting (editor), Commitment to Equity Handbook: Estimating the Impact of Fiscal Policy on Inequality and Poverty. Brookings Institution Press and CEQ Institute, Tulane University. Pearce Diane (1978), The feminization of poverty: Women, work and welfare. Urban and Social Change Review, 11, 28-36 Pimhidzai Obert and Tong Kimsun (2019), Review of the Poverty Measurement: Technical Note. Official Use Only. World Bank, Phnom Penh. Pressman Steven (1988), The feminization of poverty: Causes and remedies. Challenge, 31(2), 57-61. Mathiassen Astrid (2013), Testing Prediction Performance of Poverty Models: Empirical Evidence from Uganda. Review of Income and Wealth, 59(1), 91-112. Medeiros Marcelo and Costa Joana (2008), Is there a feminization of poverty in Latin America? World Development, 36(1), 115-127. Ministry of Education, Youth and Sports (2018), Results of Grade Eight Student Achievement from the National Assessment in 2017. Ministry of Education, Youth and Sports, Phnom Penh. Ministry of Education, Youth and Sports (2017), Results of Grade Six Student Achievement from the National Assessment in 2016. Ministry of Education, Youth and Sports, Phnom Penh. Ministry of Planning (2012), Migration in Cambodia: Report of the Cambodia Rural Urban Migration Project (CRUMP). Ministry of Planning, Phnom Penh. Ministry of Planning (2013), Poverty in Cambodia – A New Approach: Redefining the Poverty Line. Ministry of Planning, Phnom Penh. Ministry of Planning (2020), General Population Census of the Kingdom of Cambodia, Phnom Penh, Cambodia. Ministry of Social Affairs, Veterans and Youth Rehabilitation. Procedures and process of the implementation of cash transfer programs to poor and vulnerable households during the fight against COVID-19. Ministry of Planning, Phnom Penh. Ministry of Planning (2021), Poverty in Cambodia: Setting the New Poverty Line, Ministry of Planning, Phnom Penh. Ngo Sothath, Khon Leakhena, Sour Mengyou, Nuth Sovannarith, and Khun Sokeng (2021), The COVID-19 pandemic and workers in Cambodia: Magnitude of impacts on suspended workers and implications for policy and programme interventions. Centre for Policy Studies, Phnom Penh.  Northrop Emily (1990), The feminization of poverty: the demographic factor and the composition of economic growth. Journal of Economic Issues, 24(1), 145-160. National Institute of Statistics, Directorate General for Health [Cambodia], and ORC Macro (2001), Cambodia Demographic and Health Survey 2000. National Institute of Statistics, Directorate General for Health, and ORC Macro: Phnom Penh, Cambodia, and Calverton, Maryland USA. 128 CAMBODIA POVERTY ASSESSMENT National Institute of Public Health, National Institute of Statistics [Cambodia] and ORC Macro (2006), Cambodia Demographic and Health Survey 2005. National Institute of Public Health, National Institute of Statistics and ORC Macro: Phnom Penh, Cambodia and Calverton, Maryland, USA. National Institute of Statistics, Directorate General for Health, and ICF Macro (2011), Cambodia Demographic and Health Survey 2010. National Institute of Statistics, Directorate General for Health, and ICF Macro: Phnom Penh, Cambodia and Calverton Maryland, USA. National Institute of Statistics, Directorate General for Health, and ICF International (2015), Cambodia Demographic and Health Survey 2014. National Institute of Statistics, Directorate General for Health, and ICF International: Phnom Penh, Cambodia, and Rockville, Maryland, USA. Ravallion Martine (2005), Pro-poor growth: a primer. Policy Research Working Paper 3242, World Bank, Washington, D.C. Ravallion Martine (1998), Poverty Lines in Theory and Practice. Living Standards Measurement Study. Working Paper 133, World Bank, Washington DC. Ravallion Martine and Bidani Benu (1994), How Robust Is a Poverty Profile? The World Bank Economic Review, 8, 75- 102 http://dx.doi.org/10.1093/wber/8.1.75 Ravallion Martine and Monika Huppi (1991), Measuring changes in poverty: A methodological case study of Indonesia during an adjustment period. World Bank Economic Review 5(1): 57–82. Rodriguez Laura and Matthew Wai Poi (2020), Fiscal Policy, Poverty and Inequality in Jordan: The Role of Taxes and Public Spending. World Bank, Washington DC. Roth Vathana and Luca Tiberti (2016), Economic Effects of Migration on the Left-Behind in Cambodia. The Journal of Development Studies, 53(11), 1787-1805. DOI: 10.1080/00220388.2016.1214718. Royal Government of Cambodia (2019), National Strategic Development Plan 2019-2023. Phnom Penh, Cambodia. Saczewska Piotrowska Anna (2016), Near poverty definition, factors, prediction. Econometrics. Advances in Applied Data Analytic 4(54): 82-94 Tong Kimsun (2011), Migration, remittances and poverty reduction: Evidence from Cambodia. Cambodia Development Resource Institute, Cambodia Development Review, 15(4), 7-12. United Nations Development Programme (2020), COVID-19 economic and social impact assessment in Cambodia: CGE & GTAP simulation exercises. United Nations Development Programme, Phnom Penh. World Bank (1999), Cambodia Poverty Assessment. World Bank, Washington, DC. World Bank (2014), Where Have All the Poor Gone? Cambodia Poverty Assessment 2013. World Bank, Washington, DC. World Bank (2016), Cambodia: Poverty Reduction and Equity in 2013. Phnom Penh, Cambodia. World Bank Group. 2019. Improving the Effectiveness of Public Finance: Cambodia Public Expenditure Review. World Bank, Washington, DC. © World Bank. https://openknowledge.worldbank.org/handle/10986/32034 License: CC BY 3.0 IGO. REFERENCES 129 World Bank (2019a), Lao People’s Republic Poverty Assessment 2020: Catching and Falling Behind. World Bank, Washington, DC. License. Creative Commons Attribution CC BY 3.0 IGO. World Bank (2019b), Cambodia’s Future Jobs: Linking to the Economy of Tomorrow. World Bank, Washington, DC. World Bank (2019c), Country Partnership Framework for Kingdom of Cambodia, For the Period FY2019-FY2023. World Bank, East Asia and Pacific Region World Bank (2020a), Cambodia Economic Update: Cambodia in the time of COVID-19. Special Focus: Teacher Accountability and Student Learning Outcomes. Phnom Penh, Cambodia World Bank (2020b), The Human Capital Index 2020 Update: Human Capital in the Time of COVID-19. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/34432 License: CC BY 3.0 IGO. World Bank (2021). World Bank. 2021. Resilient Development - A Strategy to Diversify Cambodia’s Growth Model: Cambodia Country Economic Memorandum. World Bank, Phnom Penh. World Bank. https://openknowledge.worldbank. org/handle/10986/36908 License: CC BY 3.0 IGO. World Bank (2021a), Cambodia Economic Update: Cambodia in the time of COVID-19. World Bank, Phnom Penh.  World Bank (2021b), Cambodia Economic Update: Road to Recovery. World Bank, Phnom Penh. World Bank (2021d), Macro-Poverty Outlook. World Bank, Phnom Penh. World Bank (2022). Changing Course: 2022 Poverty and Shared Prosperity Report. World Bank, Washington, DC. World Economic Forum (2017), The Global Human Capital Report: Preparing people for the future of work. World Economic Forum, Geneva. https://www.weforum.org/reports/the-global-human-capital-report-2017 World Health Organization (2020), World Health Statistics 2020: Monitoring Health for the SDGs, Sustainable Development Goals. World Health Organization, Geneva. Wright Robert (1992), A feminization of poverty in Great Britain? Review of Income and Wealth, 38(1), 17-25. Younger Stephen D., Flora Myamba and Kenneth Mdadila (2016), Fiscal Incidence in Tanzania. CEQ Working Paper 36. CEQ Institute, Tulane University 130 CAMBODIA POVERTY ASSESSMENT APPENDIX A Additional Figures and Tables Figure A.1 Poverty headcount, Figure A.2 Number of Food Poor, 2009–2019 2009–2019 350 293 50 38.6 300 Percent of population 40 270 250 182 Thousands 30 33.8 25.1 200 26.3 150 165 96 20 22.9 17.8 100 10 13.5 77 50 21 13 24 0 0 2009 2014 2019/20 2007 2009 2011 2013 2015 2017 2019 2021 Revised Method (2019/20 CSES) Cambodia Urban Rural Revised Method (2014 CSES) Old Method (2009 CSES) Source: MoP (2021); CSES 2009, 2014, 2019/20. Source: MoP (2021); CSES 2009, 2014, 2019/20. Gini index by region Figure A.3  40 Cambodia 30 Gini index (0 -100) 20 10 0 Phnom Penh Plains Tonle Sap Coastal Plateau and Mountains Source: CSES 2019/20. APPENDIX A 131 Figure A.4 Poverty rate of female and male Figure A.5 Poverty rate of women household head and men 40 40 30 30 Percent of population Percent of population 20 20 10 10 0 0 2009 2014 2019/20 2009 2014 2019/20 Female Male Female Male Source: CSES 2009, 2014, 2019/20. Source: CSES 2009, 2014, 2019/20. Poverty Rate by Marital Status Figure A.6  15 10 Percentage points 5 0 -5 -10 Married Divorced/ Widowed Single Married Divorced/ Widowed Single separated separated Household head Individual 2009 2014 2019/20 Source: CSES 2019/20. Note: Marital status collected from household members aged 13 and above. Time to Water Source Figure A.7  100 5 4 6 3 5 4 7 20 24 25 18 26 27 20 20 32 34 80 26 41 39 50 55 60 42 40 48 43 45 49 40 51 51 74 75 69 51 54 20 38 41 42 27 28 32 35 25 17 15 0 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 2009 2014 2019/20 Cambodia Urban Rural Non-poor Poor Water on premises <30 minutes >=30 minutes Don't know Source: CSES 2004, 2014, 2019/20. Note: Round-trip travel to get to the source, get the water, and return. No mode of travel is specified in the questionnaire. Urban = Phnom Penh and other urban areas. 132 CAMBODIA POVERTY ASSESSMENT Table A.1 Asset Ownership Cambodia Urban Rural Non-Poor Poor 2014 2019 2014 2019 2014 2019 2014 2019 2014 2019 Small appliances Radio 32.5 16.2 30.4 14.4 33.1 17.3 34.1 16.6 26.5 13.9 Television 65.8 67.9 89.1 78.3 59.5 61.6 70.3 71.5 48.5 46 Telephone 5.3 1 6.5 1.4 5 0.8 5.7 1.1 4 0.7 Cellphone 82.9 92.6 93.2 95.1 80.1 91 85.7 93.6 72.3 86.2 Video/DVD player 23.3 2.8 30.5 2.6 21.4 2.9 25.8 3 13.8 1.7 Stereo 4.6 2.1 12.1 2.1 2.6 2 5.4 2.1 1.6 1.8 Camera(picture/video) 1.5 0.9 5.2 1.9 0.5 0.3 1.9 1 0.1 0.2 Satellite dish 2.1 8.1 1.1 4.3 2.4 10.4 2.4 8.6 1.1 5.3 Computer 6.4 8.7 20.8 17.2 2.5 3.5 7.9 10.1 0.7 0.3 Sewing machine 3.7 3.4 7.6 5.4 2.7 2.1 4.3 3.8 1.4 0.6 Electric iron 18.3 23.9 56.3 43.8 8 11.7 21.9 27 4.4 4.8 Electric fan 43.5 83.7 84.4 92 32.5 78.6 48.6 86.3 23.9 67.3 Large appliances Refrigerator 7 19.7 29.6 37.4 0.8 9 8.7 22.6 0.4 1.7 Stove (electric/gas) 24.7 56.1 67.9 78 13 42.8 29.2 60.9 7.5 26.4 Washing machine 2.8 9.2 12.5 19.8 0.2 2.8 3.5 10.6 0.1 0.6 Transportation assets Bicycle 60.9 53.7 50.3 46.6 63.8 58 60.3 52.9 63.2 58.4 Motorcycle 65.5 82.9 82 88.3 61 79.6 70.4 85.5 46.8 67 Car 4.8 10.4 14.7 19.1 2.2 5.1 6.1 12.1 0.1 0.3 Source: CSES 2014, and 2019/20. Note: Urban = Phnom Penh and other urban areas. Table A.2 Ownership of Agricultural Implements Cambodia Urban Rural Non-Poor Poor 2014 2019 2014 2019 2014 2019 2014 2019 2014 2019 Cart 10.6 2.1 1.3 1.1 13.1 2.7 10.3 2.2 11.9 1.6 Tractor 0.4 0.8 0.2 0.4 0.5 1.1 0.5 0.9 0.1 0.2 Bulldozer/roller 0.2 0.1 0.3 0.2 0.2 0.1 0.2 0.1 0.2 0.1 Plough 13.8 2.5 1.6 0.9 17.1 3.4 13.4 2.4 15.4 2.6 Threshing machine 0.5 0.8 0.2 0.6 0.6 0.9 0.6 0.8 0.4 0.6 Harrow 94.9 90.3 92.7 84.8 95.5 93.6 94.6 89.7 95.9 93.8 Hand tractor 10.2 14.5 1.7 4.5 12.5 20.5 10.3 14.1 9.8 16.9 Rice mill 1.7 1.6 0.1 0.4 2.1 2.3 1.8 1.7 1.2 1 Water pump 13.1 14.3 3 7 15.8 18.7 13.8 14.8 10.3 10.9 Source: CSES 2014, and 2019/20. Note: Urban = Phnom Penh and other urban areas. APPENDIX A 133 Maternal Health Care Figure A.8  120 100 95 9799 95 96 9598 98 98 89 88 89 88 90 89 89 87 90 83 86 85 79 81 80 71 70 74 69 68 70 67 67 64 65 66 62 60 60 54 50 48 44 39 41 40 37 22 20 17 0 Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Cambodia Urban Rural Antenatal care Births delivered Births delivered by Postnatal checkup Postnatal care received from in a health facility a skilled provider received in first received from a skilled provider 2 days after birth a skilled provider 2005 2010 2014 Source: CDHS 2000, 2005, 2010, and 2014. Nutritional Status of Women (15–49) Figure A.9  25 21 20 20 19 18 15 14 11 10 10 6 5 0 2000 2005 2010 2014 Undernutrition (BMI <18.5) Overnutrition (BMI ≥ 25.0 Source: CDHS 2000, 2005, 2010, and 2014. 134 CAMBODIA POVERTY ASSESSMENT Figure A.10 Mobility Trends by Destination Relative to February 2020 Baseline A. Transit B. Retail and recreation 80% 80% -44% compared to baseline -20% compared to baseline 40% 40% Baseline Baseline -40% -40% -80% -80% 5 5 5 5 5 5 5 5 5 5 5 5 5 5 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 2 6 0 2 6 0 2 2 6 0 2 6 0 2 -0 -0 -1 -0 -0 -1 -0 -0 -0 -1 -0 -0 -1 -0 20 20 20 21 21 21 22 20 20 20 21 21 21 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 C. Workplaces D. Residential 80% 80% -21% compared to baseline +11% compared to baseline 40% 40% Baseline Baseline -40% -40% -80% -80% 15 15 15 15 15 15 15 15 15 15 15 15 15 15 2- 6- 0- 2- 6- 0- 2- 2- 6- 0- 2- 6- 0- 2- -0 -0 -1 -0 -0 -1 -0 -0 -0 -1 -0 -0 -1 -0 20 20 20 21 21 21 22 20 20 20 21 21 21 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 E. Parks F. Grocery and pharmacy 80% -19% compared to baseline 80% -16% compared to baseline 40% 40% Baseline Baseline -40% -40% -80% -80% 15 15 15 15 15 15 15 15 15 15 15 15 15 15 2- 6- 0- 2- 6- 0- 2- 2- 6- 0- 2- 6- 0- 2- -0 -0 -1 -0 -0 -1 -0 -0 -0 -1 -0 -0 -1 -0 20 20 20 21 21 21 22 20 20 20 21 21 21 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: Google COVID-19 Community Mobility reports (https://www.google.com/covid19/mobility/; accessed on February 22, 2021). APPENDIX A 135 Figure A.11 Respondents (LSMS+) who Figure A.12 Distribution of Work Stopped Working Since the Last Survey Stoppages by Gender Round 15 13 100 21 11 12 % of work stoppages 80 40 45 10 53 59 10 9 60 % 40 79 5 60 55 20 47 41 0 0 May '20 Aug '20 Oct '20 Dec '20 Mar '21 May '20 Aug '20 Oct '20 Dec '20 Mar '21 LSMS+ Top 60 Bottom 40 Source: Cambodia HFPS 2020/21. Source: Cambodia HFPS 2020/21. School Closures and Stringency Index Figure A.13  1st 2nd 3rd school school school closure closure closure 100 Phased Schools school reopen 80 reopening 60 40 20 0 1-21-2020 2-5-2020 2-20-2020 3-6-2020 3-21-2020 4-5-2020 4-20-2020 5-5-2020 5-20-2020 6-4-2020 6-19-2020 7-4-2020 7-19-2020 8-3-2020 8-18-2020 9-2-2020 9-17-2020 10-2-2020 10-17-2020 11-1-2020 11-16-2020 12-1-2020 12-16-2020 12-31-2020 1-15-2021 1-30-2021 2-14-2021 3-1-2021 3-16-2021 3-31-2021 4-15-2021 4-30-2021 5-15-2021 5-30-2021 6-14-2021 6-29-2021 7-14-2021 7-29-2021 8-13-2021 8-28-2021 9-12-2021 9-27-2021 10-12-2021 10-27-2021 11-11-2021 Source: Hale, et al. (2021). (ourworldindata.org; accessed on November 16, 2021). Note: Stringency index is a composite measure based on 9 response indicators including school closures, workplace closures, and travel bans, rescaled to a value from 0 to 100 (100 = strictest). Dates are expressed in Month/Day/Year format. 136 CAMBODIA POVERTY ASSESSMENT APPENDIX B Poverty Measurement Methodology The poverty lines for Cambodia were defined in 1997 potentially underestimating poverty in the country. In using the 1993/94 Cambodia Socio-Economic Survey 2015, Cambodia became a lower middle-income country, (CSES) (formerly known as Socioeconomic Survey of having achieved US$1,070 gross national income (GNI) Cambodia (SESC)). These poverty lines served as the per capita, surpassing the GNI per capita threshold of basis to measure poverty in Cambodia until 2008. Due US$1,025. to Cambodia’s rapid economic development during the This section documents RGC’s new methodological 1993–2008 period, the Royal Government of Cambodia approach in revising its poverty estimation methodology. (RGC) decided to define a new poverty line for the country Given the desire to reflect changing consumption patterns in 2013 using the 2009 CSES. Up to this point, the 2009 of the poor and improve poverty measurement, the RGC poverty lines (adjusted for inflation) have served as the decided to (i) improve the 2019/20 household survey basis to estimate poverty in Cambodia. instrument to better capture consumption by using a highly In 2017, the World Bank collaborated closely with the disaggregated list of items and recording their quantities, National Working Group for Poverty Measurement and (ii) revise its poverty measurement methodology in Cambodia (NWGPM) under the leadership of the to align more closely with international best practices. Ministry of Planning (MoP) to review Cambodia’s poverty The revised methodology included 2 major changes: (i) estimation methodology. 81 The review pointed out that construction of a new consumption aggregate, including the rapid economic development in Cambodia over the incorporating the use-value of durable goods, imputed decade may have significantly changed the consumption house rent, and educational expenses, and (ii) adoption patterns of the poor. This raised concerns that the of a cost-of-basic-need and common-basket approach poverty basket and poverty line might be outdated and to define the poverty line. Survey Methodology Since 1993, the National Institute of Statistics (NIS) of primary source for Cambodia’s official monetary poverty Cambodia implemented the CSES which is a nationally estimates. The data are also used to monitor Cambodia’s representative household survey that collects household progress in achieving the goals and targets of the National socio-economic information. The main objective of the Strategic Development Plan (NSDP 2019–2023) and of the survey is to generate data needed to assess welfare Sustainable Development Goals (SDGs). The most recent and living conditions among the Cambodian population round is the 2019/20 CSES, which was implemented and to estimate Gross Domestic Product (GDP) and between July 2019 and June 2020. the Consumer Price Index (CPI). The CSES is the 81 N  WGPM members include the Ministry of Planning (MoP), the Ministry of Health (MoH), the Ministry of Economy and Finance (MEF), the Ministry of Education, Youth and Sports (MoEYS), the Ministry of Social Affairs, Veterans and Youth Rehabilitation (MoSVY), the Ministry of Agriculture, Forestry and Fishery (MAFF), the Ministry of Rural Development (MRD), Ministry of Women’s Affairs, the Supreme National Economic Council (SNEC), Council for Agriculture and Rural Development (CARD), Centre for Policy Studies (CPS) and Cambodia Economic Association (CEA). APPENDIX B 137 The 2019/20 CSES consists of a sample size of 10,075 expenditure, food security, education, housing, household households with a total of 44,549 household members. economic activity, household liabilities, household Stratified 2-stage random sampling method was used to income from other sources, construction activity, construct a nationally representative sample. Stratification durable goods, child health, health care expenditure, took place at the provincial level (24 provinces and 1 capital disability, employment, and victimization. The village level city, Phnom Penh) with further stratification according to questionnaire comprises 3 modules to capture community the urban and rural level. In the first stage of sampling, demographic information, community economic activities enumeration areas/villages were randomly selected using and infrastructure, and retail prices. The 7-day diary a systematic sampling selection method with probability records household expenditure and income daily. proportional to size. A total of 1,008 primary sampling units Fieldwork for the 2019/20 CSES was implemented (enumeration areas/villages) were selected. Households over a 12-month period from July 2019 to June 2020. were then randomly selected in the second stage. This differed slightly from the earlier rounds conducted CSES 2019/20 has 3 instruments: (i) household between 2007 and 2017 which were implemented questionnaire, (ii) village questionnaire, and (iii) diary. annually and within the calendar year. The 2019/20 CSES The household questionnaire comprises 17 modules was implemented about 1 ½ years after the 2017 CSES. including household roster, food consumption, non-food The New Poverty Methodology Revised consumption module in New Consumption Aggregate 2019/20 CSES The revised consumption aggregate for 2019/20 consists To improve the quality of data used to estimate the official of food and non-food expenditures, use-value of durable poverty rate, the NIS revised the consumption module in the goods, and imputed house rent. The new consumption 2019/20 CSES. The NIS disaggregated the broad 20 food aggregate now incorporates the use-value of durable and 13 non-food categories captured in the 2014 CSES goods and imputed house rent, and better captures consumption modules into 64 food items and 40 non-food educational expenses incurred. The construction of the categories respectively. The change from broad categories main components of the consumption aggregate are to more disaggregated categories of consumption allows discussed below. for better coverage of the poverty basket, more reasonable FOOD AND BEVERAGE CONSUMPTION estimation of calorie intake by item, and better valuation of items consumed. In addition, both the quantity and value The recall “Food, Beverage and Tobacco” module (Section of food consumed are captured in the recall consumption 1.B) provides consumption and expenditure information module allowing for recall data to be used to define the for the 64 food items consumed in the 7 days preceding poverty line. While these changes in the survey instrument the survey. The 64 food items were covered under the improve the quality of the data produced, they come with following broader categories: some costs of comparability across surveys as will be (i) rice and other rice products discussed in Appendix C. (ii) fish and seafood (iii) meat, oils or fats, (iv) fruits or fruit products 138 CAMBODIA POVERTY ASSESSMENT (v) vegetables health, and other expenditures such as salary/wages (vii)  (vi) non-alcoholic and alcoholic beverages for housekeepers, child-care expenses, housing (vii) food and drinks consumed away from home insurance and maintenance excluding improvement. (viii) other food expenses. Daily per capita expenditure on non-food items is obtained Section 1.B solicited 4 main questions: by converting expenditures of each item into daily expenditures and then dividing by the household size. (i) total quantity consumed of each food (ii) expenditures on food Construction of the non-food component of the imputed value of home-produced food or food (iii)  consumption aggregate excludes 4 non-food categories received as wages paid in-kind, gifts, or free in Section 1.C following the recommendations of Deaton (iv) total value of each food. 82 and Zaidi (2002). These categories include tax on income or property, bank repayment, and “lumpy” expenditures Weekly total household food consumption is obtained that tend to be financed out of savings such as wedding by summing consumption of 64 food items.83 Daily per and funeral costs. capita food consumption is obtained by dividing total food consumption by 7 and then by the household size. The non-food consumption aggregate also includes educational expenses. Unlike the previous methodology, NON-FOOD EXPENDITURES the new one uses educational expenses reported in The recall “Non-Food Expenditures” module (Section the education module rather than those reported in the 1.C) provides expenditures information for 40 non-food non-food expenditure module. Education expenditures categories. Expenditures for non-food categories are are captured in both the non-food expenditure module reported with recall periods of 1 month, 6-months, or 1 (Section 1.C) and the education module (Section 2). year.84 Non-food expenditures are grouped broadly as Section 1.C only collected information on expenses for follows: books, paper, and other stationery and whether these items were procured for education expenses was unclear. (i) clothing and footwear Section 2 collected information on all types of education (ii) personal goods expenses incurred during the past school year including (iii) transportation school fees, tuition (private lesson), textbooks, other school (iv) communication supplies, allowances for children studying away from (v) household equipment home, transport costs, gifts to teachers, and contributions (vi) recreation and entertainment to school building or development funds. Only 23 percent 82 Data on tobacco consumption in the last 7 days, which classified as a nonfood item, was also obtained from Section 01.B. 83  The 64 food categories include: Rice (1st quality), rice (2nd quality), rice noodles/fried noodles, Chinese noodles/Khmer noodles, other cereals/ flour/ other bakery products, bread, mudfish, catfish, other inland fish, shrimp/lobster, crabs, other seafood, preserved or processed fish/ seafood, pork, beef, duck, chicken, other meat products, eggs/egg-based products, milk/yoghurt, oil/fats, banana, mangoes, longan, papaya, tamarind, coconut, nuts/edible seeds, maize/corn crop, other fresh fruits, dried/preserved fruits, morning glory, spring onion/garlic/ leeks leaves, cabbage/leaves, gourd/cucumber/ pumpkin/eggplant, other fresh vegetables, prepared and preserved vegetables, potato/ sweet potato/carrot/radish, mushrooms/dried mushroom, pea/bean/soy bean/bean sprout, sugar cane/palm sugar, sweets, salt, pepper, monosodium glutamate, fish sources/soy sources/chilly sources, other ingrediencies, nutritive tablets, coffee/tea/ chocolate, bottle water, soft drinks/orange juices/fruit juices, ice cream, beer at home, wine at home, other alcohol not at bar or restaurant, cigarette and other tobacco, food at school, drink at school, food at work, drinks at work, food or snack at restaurant/pub or café, drinks at restaurant/pub or café, prepared meals bought outside and eaten at home, and other food expenses 84 N  on-food expenditures with a 1-month recall period include gasoline, diesel, lubricant, gas, telephone service, internet service, lottery, movie, karaoke, newspapers, magazine, books, papers, other stationaries, salary or wage for housekeeper, medicine, medicine products, medical or dental consultation. Non-food expenditures with a 3-month recall include local travel, hotel, guesthouse. Non-food expenditures with a 6-month recall include clothing, shoes, slipper, household utilities (textile), raincoat, umbrella, toothpaste, toothbrush, tooth care, hair soap, cloth soap, lotion, perfume, international travel, postal service, spoon, fork, knife, chopstick, pets, toys, games. Non-food expenditures with a 12-month recall include jewelry, watches, clocks, car travel insurance, expenditure on motorbikes or cars, dwelling insurance and maintenance, health insurance, social contribution e.g., wedding reception/funeral, the cost of organizing wedding, funeral, or other religious events. APPENDIX B 139 of households reported expenses for books, paper and costs to and from school. To avoid double counting, the other stationery in Section 1.C compared to 60 percent methodology excludes transportation costs to and from in Section 2. For these reasons, education expenditures school in Section 1.C. Annual education expenditures from Section 2 are used for the welfare measurement. are converted to daily per capita expenditures by dividing Both Section 1.C and Section 2 capture transportation them by 365 and then dividing them by household size. Share of Households with Educational Expenses Figure B.1  70 60 60 Percent of households 50 40 30 23 20 10 0 Seciotn 01C Section 02 Source: MoP (2021). Note: Section 1.C represents the percentage of households reporting expenses on books, papers, and other stationery. Section 2 represents the percentage of households reporting expenses on school fees, tuition, textbooks, other school supplies, allowances for children studying away from home, transportation costs, gifts for school building. Housing utilities represent the other group of non-food USE-VALUE OF DURABLE GOODS expenditures. The “Housing” module (Section 4) provides Use-value of durable goods is a recent addition to the monthly expenditures on 10 utilities in the consumption consumption aggregate. Section 9 captures ownership aggregate: of 45 durable items. Of these, the NWGPM selected (i)) water charges 32 items to estimate the use-value of durables for the (ii) sewage or wastewater disposal welfare aggregate. Those included are: (1) radio , (2) (iii) garbage television, (3) telephone, (4) cell phone, (5) video, VCD or (iv) electricity DVD player, (6) stereo, (7) camera, (8) satellite dish, (9) (v) gas bicycle, (10) motorcycle, (11) car, (12) jeep or van, (13) (vi) kerosene sewing machine, (14) refrigerator, (15) electric kitchen or (vii) firewood gas stove, (16) washing machine, (17) dishwasher, (18) (viii) charcoal freezer, (19) vacuum cleaner, (20) electric iron, (21) electric (ix) battery charging fan, (22) air conditioner, (23) suitcase, (24) generator, (25) (x) housing maintenance and minor repairs. battery, (26) sofa set, (27) dinning set, (28) bed set, (29) wardrobe or cabinet, (30) computer, (31) cart, (32) water Monthly expenditure of each item is converted to daily per pump. Some of the durable goods were excluded because capita expenses by divided by 30.4, and then divided by they are not commonly owned. Since the age of durable the household size. goods is not available, the use-value of durable goods is computed by using the depreciation rates published by the tax department of the Ministry of Economy and Finance (Table B.3). To estimate daily per capita use-value of durable goods, the yearly use-value is divided by 365, and then divided by household size. 140 CAMBODIA POVERTY ASSESSMENT Table B.3 Depreciation Rate of Durable Goods Category Durable goods Depreciation method Depreciation rate 1 Buildings, infrastructure of Equal depreciation rates 5% per annum buildings and construction etc. 2 Computers, software, electronic Progressive depreciation rates 50% on the remaining value annually information systems etc. 3 Cars, lorries, furniture, office Progressive depreciation rates 25% on the remaining value annually equipment etc. 4 All other assets Progressive depreciation rates 20% on the remaining value annually Source: MoP (2021). HOUSING RENT for other urban and rural dummies (Table B.5). The monthly imputed house rent is divided by 30.4 to convert The consumption aggregate now includes imputed house it to daily expense, and then divided by household size to rents. The new methodology employs a hedonic regression convert it to daily per capita terms. The 2019/20 CSES approach to impute house rents. First, a regression model solicited 2 questions about house rent: (i) “If rented: How is estimated using rent paid by renter-households as the much did you pay for rent of this house last month?”; (ii) dependent variable against housing characteristics such “If owned: How much would you have to pay per month as floor area; number of rooms; construction materials of to rent a similar dwelling (estimated value)?”. It is widely walls, roofs, and floors; access to electricity, access to noted that the latter estimated rental value is not always improved water, access to improved toilet, and dummy credible or usable because rents known to homeowners variables for other urban and rural areas. The estimated may be subsidized, out of date, or not representative of coefficients of each independent variable are then used the general property in their area. This is the reason for to impute rents for all sample observations. Since the employing a hedonic regression. housing rental market in rural areas is very thin (Table B.4), the regression model is run with the full sample, controlling Table B.4 Legal Status of Housing Sample Share (percent) Not own Not own Owned house not rent Rent house Owned house not rent Rent house Phnom Penh 671 40 214 74.6 4.2 21.2 Other urban 2,534 132 251 91.8 4.0 4.2 Rural 6,089 212 29 96.6 3.1 0.3 Cambodia 9,294 384 395 92.3 3.5 4.3 Source: MoP (2021). APPENDIX B 141 Table B.5 Hedonic Regression Explanatory variables Coefficient Floor area (log) 0.672*** Number of rooms 0.147** Wall (bamboo/thatch/leaves/grass) 0.196 Wall (wood/logs/plywood) -0.671 Wall (concrete/brick/stone/fibrous cement) -0.148 Wall (galvanized iron/aluminum) -0.498 Roof (tile) 0.023 Roof (fibrous cement) -0.212 Roof (galvanized iron/aluminum) -0.139 Floor (wooden planks) -0.195 Floor (bamboo strips) -0.329 Floor (cement/brick/stone) 0.012 Floor (ceramic tiles) -0.005 Light (public electricity) -0.462 Water (piped into dwelling) 0.161 Water (piped into compound/yard/plot) 0.264 Water (public tap) 0.452 Toilet (pour flush) 0.330 Toilet (latrine) 0.040 Other urban -0.180* Rural -0.325 Constant 10.533*** R-squared 0.428 Sample size 394 Source: MoP (2021). Note: *, **, *** indicate statistical significance at 10%, 5% and 1%, respectively; actual rent for 1 household was missing and leaving the number of observations to 394. Spatial and Within-survey Section 1.B.85 Each has its pros and cons. The CPI has Temporal Price Adjustments a comprehensive coverage of items but limited spatial coverage as it only covers Phnom Penh and urban/rural The consumption aggregate is adjusted for spatial price areas in 8 provinces. Conversely, the unit values obtained variations and within-survey temporal price variations. Two from Section 1.B are more limited in item coverage but price data sources are used to construct the spatial price more comprehensive in spatial coverage. The NWGPM index to adjust for price differences across geographic decided to construct the food spatial price index from areas. The spatial price indices used are: (1) the Consumer the unit values obtained in Section 1.B and construct the Price Index (available only for Phnom Penh, other urban non-food spatial price index from CPI. The food spatial areas within 3 provinces, and other rural areas within 5 price index constructed from the survey data reflects the provinces), and (2) the unit value of food items from recall prices that the poor faced better than those from CPI 85 There are 24 provinces and 1 capital city in Cambodia. 142 CAMBODIA POVERTY ASSESSMENT given the urban bias of the CPI. Ideally, non-food spatial was not captured in Section 1.C. These deflators were price index should be estimated from the survey data so used to spatially deflate the nominal daily per capita food that it covers more spatial coverage, but it is not practical consumption and non-food expenditure to Phnom Penh in this case because quantity of non-food expenditure prices (Table B.6). Table B.6 The Number of Food and Non-food Items by Data Sources Food Non-food Non-food unit Data Food unit value CPI prices value CPI prices source (Section 01.B) (Section 01.C) Number of Limited number of More comprehensive items More comprehensive items items items (55 items) (103 items) Quantity was not (155 items) Limited spatial coverage – collected; unit val- Limited spatial coverage – Coverage More comprehensive ue is not available Phnom Penh, other urban Phnom Penh, other urban area spatial coverage and other rural areas* and other rural areas* Source: MoP (2021). Note: * Urban prices are based on 3 provinces, while rural prices on 5 provinces. The capital city, Phnom Penh, is the reference geographic across the 3 regions are larger than the non-food price area. Prices are highest in Phnom Penh, followed by other differences. Rural households generally face lower food urban and rural areas (Table B.7). Food price differences prices as food is often provided by local producers. Table B.7 Spatial Price Index in 2019/20   Food Non-food   2019q3 2019q4 2020q1 2020q2 2019q3 2019q4 2020q1 2020q2 Phnom Penh 100 100 100 100 100 100 100 100 Other urban 77 78 79 82 96 95 96 97 Rural 70 71 76 77 90 90 90 92 Source: MoP (2021). Note: Laspeyres Price Index. The spatially-deflated daily per capita food consumption base (Table B.8). The daily per capita consumption at and non-food expenditure aggregate is temporally Phnom Penh prices is estimated by adding spatially and adjusted for price difference across the survey months temporally price adjusted food consumption and non- by using the average CPI of July 2019–June 2020 as the food expenditure. APPENDIX B 143 Table B.8 Phnom Penh Consumer Price Index in 2019/20 Date All Food Non-food 2019 July 177.5 216.0 146.2 2019 August 178.5 218.2 146.3 2019 September 178.4 218.0 146.4 2019 October 178.3 217.4 146.5 2019 November 178.5 217.8 146.6 2019 December 179.9 219.8 147.6 2020 January 180.8 221.1 148.2 2020 February 180.7 221.0 147.9 2020 March 181.5 222.9 147.8 2020 April 180.4 224.7 144.5 2020 May 181.6 226.0 145.5 2020 June 182.4 226.8 146.3 Source: MoP (2021). New Poverty Lines The third change allowed for a utility-consistent approach to calculating basic needs poverty line that would produce The new poverty line in Cambodia is derived using the a consistent poverty profile. Previously, the non-food cost-of-basic-needs approach and applies a common allowance component of the poverty line was estimated consumption basket for the entire country. This approach separately for the 3 geographic regions (Phnom Penh, differs from the earlier approaches used to determine the Other urban and Rural) without accounting for cost- poverty line along 3 main aspects: of-living differences. This effectively produced region- ■ Uses 7-day recall data to construct the food poverty specific non-food baskets (and poverty lines) that did not line instead of data from a diary. represent the same level of utility (i.e., consistency). ■ Non-food allowance is based on the share of non-food Recalled food consumption data is preferable to diary- expenditure to total consumption around the food based data for poverty measurement. CSES collects poverty line86 instead of actual non-food spending of consumption data using 2 different methods: (i) The recall the bottom 20th–30th percentile of the consumption method, whereby households report their consumption distribution estimated separately for the 3 areas on specified items (7-day for food consumption and (Phnom Penh, Other urban and Rural) as was done in 1-month, 3-month, 6-month, 12-month for non-food the previous methodology (MoP, 2013). expenditures) based on their recollection; and (ii) the 7-day diary method whereby households record expenditures ■ Applies a common basket for Cambodia assuming on consumption items as they happen. The 7-day recall a common reference population for the non-food method to collect food consumption data is in line with allowance component of the poverty line. the food data collection for household consumption and expenditure surveys in low and middle-income countries developed by the Inter-Agency and Expert Group on Food Security, Agricultural and Rural Statistics and endorsed 86 Following the approach outlined in Ravallion (1998). 144 CAMBODIA POVERTY ASSESSMENT by the United Nations Statistical Commission in March baskets) suggest that the poverty headcount in Phnom 2018. The 7-day diary adopts an acquisition approach, 87 Penh in 2013 was higher than in rural areas. However, leading to inconsistent consumption and expenditure of the poverty rate is higher in rural areas compared to some food items households purchase infrequently but in urban areas when applying a common-basket approach large stocks or consume from owned production, such as (Figure B.2). Evidence from other countries—such as in rice. In other words, consumption exceeds expenditure Indonesia, Egypt, and Mozambique—also shows that or vice versa, which makes it difficult to establish a food computing “regional” poverty lines separately without poverty line using 7-day diary data. corresponding adjustments for utility consistency (as done for Cambodia) distorts the poverty profile.88 In some Poverty lines drawn from a common reference population— cases, urban households with much higher consumption rather than separate population groups—produce a more can appear poorer than rural households with much lower consistent poverty profile. Estimates based on the official levels of consumption.89 poverty line defined using the 2009 CSES data (multiple Poverty Profile in 2013 Figure B.2  25 Percent of population 20 15 10 5 0 Multiple poverty baskets Spatial price adjustment to common basket Phnom Penh Other urban Rural Source: Ministry of Planning (undated) and World Bank (2016). 87 FAO and the World Bank, 2018. 88 See Ravallion and Bidani (1994) for a case study on Indonesia and Arndt and Simler (2010) for a case study on Egypt and Mozambique. 89 Alfani et al. 2012. APPENDIX B 145 Table B.9 Summarized the evolution of the poverty line measurement approach. Poverty Line Setting in 1993/94, 2009 and 2019/20 1993/94 2009 2019/20 1. Calorie intake 2,100 2,200 2,200 2. Food poverty line The reference food bundle The reference food bundle The reference food bundle was derived from the third was derived from the bot- was derived from the bot- quintile tom 5th-30th percentile tom 5th-30th percentile 3. Non-food allowance Spending on non-food Spending on non-food Spending on non-food around food poverty line of the bottom 20th-30th around food poverty by regions (regression percentile by regions line (share of non-food approach) expense approach) 4. Imputed expenditure Yes No Yes House rent Imputed house rent Actual house rent Imputed house rent Durable goods Imputed durable goods Exclude durable goods Imputed durable goods Allowance for clean water 5.  No Yes No Source: MoP (2021). The poverty line is set by first estimating a food poverty average quantities consumed for every food item among line based on the cost of a food basket that delivers 2,200 the reference population, then converting consumed calories per person, per day, then adding a non-food quantities into calories using the calorie conversion allowance. The estimated food poverty line and the non- factors from the Food and Agriculture Organization of the food allowance uses the procedures outlined here. United Nations (FAO) and the Association of Southeast Asian Nations (ASEAN). Second, budget shares for each Food Poverty Line: The food poverty line is based on food item are calculated to identify the most common a food basket concept anchored in nutrition. The food food items the reference population consumes. Dropping poverty line for Cambodia (KHR5,266 per capita per day) food items with a budget share of less than 0.7 percent of is based on the cost of a food basket providing 2,200 total food expenditure leaves 28 food items, accounting calories per person, per day. The composition of the for 89 percent of food expenditure and 1,740 calories food basket reflects consumption patterns prevailing intake per capita, per day. The quantity of the 28-item in a reference population—the 5th to 30th percentile of food bundle is scaled up to reach 2,200 calories. The the distribution of total consumption per person. First, reference food basket is valued at national median unit the reference food basket is constructed by taking values of the reference population. 146 CAMBODIA POVERTY ASSESSMENT Table B.10 Food Basket of Reference Population (2019/20) No Food items Calorie Price per Food Unit Unit intake calorie line price 1 Rice 1,020 0.8 796 2,100 kg 2 Bread 9 5.0 46 1,000 piece 3 Mudfish 34 12.1 417 12,000 kg 4 Catfish 16 9.8 161 10,000 kg 5 Other inland fish 56 7.3 411 10,000 kg 6 Preserved or processed fish/seafood 19 3.8 71 10,000 kg 7 Pork 81 4.8 390 18,000 kg 8 Beef 5 29.1 145 34,000 kg 9 Chicken 27 9.8 261 15,000 kg 10 Eggs and egg-based products 15 7.7 114 500 piece 11 Milk or yoghurt 24 3.8 89 1,500 can 12 Oils or fats 101 0.6 62 5,000 kg 13 Banana 53 1.3 71 2,000 set 14 Longan (mien) 5 11.3 53 8,000 kg 15 Maize and corn crop 84 0.7 59 1,000 piece 16 Other fresh fruits 10 8.9 87 6,000 kg 17 Trakun (watercress marsh cabbage) 10 8.6 90 3,000 kg 18 Spring onion/garlic/leeks leaves 4 21.9 80 7,000 kg 19 Cabbage/ leaves 5 16.0 78 4,000 kg 20 Gourd/cucumber/pumpkin/eggplant 8 14.5 120 3,000 kg 21 Other fresh vegetables 15 5.5 83 3,000 kg 22 Sugar cane/palm sugar 64 0.8 49 3,000 kg 23 Sweets 240 0.2 45 500 kg 24 Monosodium glutamate - - 10,000 kg 25 Fish/soy/chilly  16 3.1 50 3,100 liter 26 Soft drinks/orange juices/fruit juices 56 6.7 372 6,700 liter 27 Beer 69 6.1 418 6,100 liter 28 Food away from home 154 4.2 649 2,200 5,266 Source: MoP (2021). APPENDIX B 147 Calculating caloric intake for food consumed away Since these are not available in the diary, the NWGPM from home requires use of the diary and making some assumed the following calorie conversion: (i) 100 milliliters assumptions. Since the recall module does not account of a drink provides 45.2 calories, the average derived for the quantity of food and drinks consumed away from 100 milliliters of Coca-Cola, energy drink and fruit from home, the quantity captured in the 7-day diary is juice; and (ii) 1 bowl of food is assumed to be 300 grams, used to estimate calorie intake. The diary collects both with 100 grams assumed to equal 200 calories.90 These value and quantity—measured in milliliters or liters for assumptions enable estimation of average monetary drinks, and grams, kilograms, or bowls for food—of food value of food away from home per capita, per day and its consumed away from home. Ingredients of drinks or corresponding calorie intake (Table B.11). foods consumed are required to estimate calorie intake. Table B.11 Food Away from Home Calorie Intake   Mean monetary value per capita per day (KHR)   Drink Food Phnom Penh 498 2,034 Other urban 266 1,140 Rural 158 866   Mean monetary value per capita per day (KHR) Drink Food Phnom Penh 146 376 Other urban 52 253 Rural 37 206 Source: MoP (2021). To ensure that the COVID-19 pandemic did not affect Non-food Allowance: Some households cannot results, the poverty line setting used only 9-months of data afford enough food while needing other essential non- (July 2020-March 2021) of CSES 2019/20. It is important food items such clothing, shelter, and housing utilities. to take the coronavirus disease 2019 (COVID-19) Since these items take priority over food, they present pandemic into account when defining the poverty line. very basic household needs that setting the poverty line The Cambodia High-Frequency Phone Survey (HFPS) should consider. In practice, determining the poverty implemented in May 2020 showed that more than line considers several non-food expenses such as 80 percent of respondents experienced reductions in clothing, personal goods, transportation, communication, total household income since the COVID-19 outbreak recreation, education, health, housing utilities, house rent (Karamba et. al, 2020). To cope with these income losses, and use-value of durable goods. The non-food allowance more than half of households reduced both food and non- is the average of an upper-bound, non-food allowance and food consumption. a lower-bound, non-food allowance. The upper non-food 90 Coca-Cola: 330ml=139 calories; Energy drink: 330 ml=152 calories; Fruit Juice: 330 ml=156 calories. 148 CAMBODIA POVERTY ASSESSMENT allowance is the non-food expenditures among people Alternatively, cost-of-living adjustments for the 3 whose food expenditures are within 10 percent of the food geographic areas can be applied to the national poverty poverty line, while the lower non-food allowance is the non- lines. In such a case, the resulting poverty lines for food expenditures among people whose total expenditures Cambodia in 2019/20 would be KHR10,951 per person, are within 10 percent of the food poverty line. per day in Phnom Penh; KHR9,571 for other urban; and KHR8,908 for rural areas. Table B.12 summarizes these The national poverty line for Cambodia of KHR10,951 per poverty lines and their underlying food and non-food person per day is obtained by summing the food poverty allowances. line and the non-food allowance. Table B.12 Poverty Lines in 2019/20 (Per Person Per Day) Food poverty line Non-food allowance Overall Poverty line Phnom Penh 5,266 5,685 10,951 Other urban 4,145 5,426 9,571 Rural 3,822 5,086 8,908 Source: MoP (2021). APPENDIX B 149 APPENDIX C Estimating Poverty Trends using Survey-to-Survey Imputations The household survey instrument and poverty in 2 surveys, 1 of which must contain consumption data, measurement methodology used to estimate the official to predict consumption in another survey. The 2009 and national poverty headcount for Cambodia has recently 2014 CSES are comparable to the 2019 CSES in terms been revised to align better with international best of sampling design and sample size. The 3 surveys use a practices. While necessary, the changes in the 2019/20 stratified, two-stage random sampling method to sample Cambodia Socio-Economic Survey (CSES) instrument, more than 10,000 Cambodian households. The only consumption aggregate, and poverty line methodology differences between the surveys are the sampling frame make it difficult to compare data between surveys and to and consumption module; the sampling frame for the compare poverty rates across time. 2009 and 2014 CSES are drawn from the 2008 Census, while the sampling frame for the 2019/20 CSES is drawn To overcome the incomparability issues, a survey-to-survey from the 2019 Census. imputation approach is used to predict consumption in previous CSES household surveys to enable the estimation The imputation approach consists of two steps: (i) of poverty trends. Survey-to-survey imputation involves estimating the relationship between daily per capita computing “missing data” into 2 or more data sources consumption and explanatory variables using the (usually sample surveys) based on a model developed 2019/20 CSES, and (ii) using the estimated coefficients from another data source collected from the same target to simulate daily per capita consumption in the 2009 and population. In the application here, the 2019/20 CSES 2014 CSES. The explanatory variables comparable in the is used to construct a model with variables (predictors) 3 data sets include: suited to predict consumption in the CSES’ collected in ■ Household demographics (household size, age other years, and hence to estimate poverty. In a similar composition, gender of household head); case for Lao PDR, the World Bank (2019) imputed consumption from the most recent household expenditure ■ Education (household head and other household survey (2018/19) into the previous survey (2012/13). The members); early foundations of this method are derived from poverty ■ Employment (labor force status and sector of mapping introduced by Elbers, Lanjouw, and Lanjouw employment); (2003), who estimated consumption data into the census from a consumption model estimated from a household ■ Housing characteristics (floor area, number of rooms; survey. Most recently, this technique has also been used roof, wall and floor construction materials; access to to impute consumption between household survey and electricity, drinking water and sanitation, and source demographic and health surveys. of cooking fuel); Survey-to-survey imputation requires that at least 1 ■ Land ownership (has agricultural land less than 0.5 ha); comparable survey contains household consumption data. Survey-to survey imputation relies on common variables 150 CAMBODIA POVERTY ASSESSMENT ■ Asset ownership (television, generator, video, cell Looking at poverty incidence rates separately by urban phone, bicycle, motorcycle, car, sewing machine, and rural areas indicates that the models work well. Urban electric kitchen, electric fan, computer, tractor, radio, poverty headcount is in the range of 14–15 percent in 2009 water pump, threshing machine, rice miller, washing and 12–13 percent in 2014. Rural poverty headcount is machine, refrigerator, air condition); roughly 38 percent in 2009 and between 28–30 percent in 2014 (Table C.3). The predicted poverty rates at 1.25 ■ Livestock (cattle, pig, chicken) and some interaction times the poverty line are in the narrow range of 51.7–52.2 terms between agricultural land holding and the percent in 2009 and 42.2–43.7 percent in 2014 (Table C.4). number of employed household members by sector). Among the proposed 4 models, the restrict (using reduced Model specification is as follows: explanatory variables) performs best and produces poverty ln( y i ) = βX i + u i (1) headcount rates closest to the 2014 estimates. Estimated poverty rates closely match the 2014 estimates at the Where y is daily per capita consumption of household i , national and geographical levels, except for Phnom Penh β is the vector of coefficients to be estimated, X is the where the difference is about 4 pp (Table C.3). Despite vector of explanatory variables, u is the error term. The this difference in Phnom Penh, the predicted poverty estimate of β is obtained from two different models: (i) full rates produce a consistent poverty profile where poverty sample, and (ii) urban and rural separately. The latter is incidence is highest in rural areas and lowest in Phnom expected to capture the difference in consumption pattern Penh. Predicted or actual poverty rates based on the between geographical areas. To address multicollinearity common basket method are nearly double those found concerns, which happens when 1 or more explanatory using the multiple basket approach (24.9 or 25.1 percent variables are highly correlated and generates imprecise vs 13.5 percent) (Figure C.1). However, the predicted coefficient estimates, variance inflation factors (VIF) poverty headcount is lower than actual poverty estimates are calculated for all variables. VIF for j th explanatory in 2009 across the 3 main geographical areas (Figure C.2). variable is the reciprocal of 1 minus R2-value obtained This evidence suggests that survey-to-survey imputation by regressing the j th variable on the remaining variables. performs well for estimating poverty headcount in the That is, the greater variation in a variable another variable short term. The predicted daily per capita consumption explains, the higher the VIF. Given a general rule of thumb, and poverty headcount from the restrict model is used a VIF exceeding “10” reflects very high multicollinearity for analyzing the poverty and inequality trends from 2009- requiring correction. This process identified 7 variables for 2019/20 in Chapter 1. removal from the original model for Cambodia. Three checks are conducted to confirm the robustness The consumption aggregate estimate for 2009 and 2014 of the imputations. First, the poverty headcount rates for is imputed using Stata’s Multiple Imputation (MI) package. each geographical area are produced to test whether the Table C shows the consumption models with all explanatory model generates plausible predictions. Second, poverty variables (full model) and restricted explanatory variables headcount rates based on a higher poverty line (1.25 (restricted model) used for survey-to-survey imputation. times poverty line) are estimated to ensure that they are The imputed point estimates show poverty headcount reasonable. Lastly, the predicted poverty headcount rates rates in the range of 30 to 33.4 percent for 2009 and 24.9 for 2009 and 2014 are compared against the poverty to 26.3 percent for 2014 (Table C.2). Over the past decade, headcount rates estimated using the cost-of-basic-needs poverty incidence has decreased by around 16 percentage approach and common basket to define the poverty line points or about 1.6 percentage points per year. based on CSES 2014.91 91 Pimhidzai and Tong (2019). APPENDIX C 151 Table C.1 Consumption Models used for Survey-to-survey Imputation (Ordinary Least Square) Dependent variable: Log (consumption per capita per day) Full Model Restrict Model Household size =3 -0.156*** -0.163*** Household size =4 -0.231*** -0.235*** Household size =5 -0.295*** -0.296*** Household size =6 -0.365*** -0.365*** Household size =7 -0.435*** -0.432*** Household size =8+ -0.433*** -0.425*** Number of children ages 0-14 -0.059*** -0.064*** Age of household head (years) 0.008*** -0.001*** Age of household head (squared) -0.000*** Household head is female -0.055*** -0.052*** Household head is single female -0.053 -0.053 Household head occupation (elementary) -0.036** -0.037** Average years of schooling (15 years +) 0.019*** 0.019*** Number of out-of-school children ages 6-14 years -0.042 -0.039 Household has no children ages 6-14 years -0.053*** -0.058*** Number of wage workers =1 -0.034*** -0.032*** Number of wage workers =2 -0.052*** -0.050*** Number of wage workers =3 -0.103*** -0.098*** Household with agricultural land < 0.5 ha 0.191*** 0.049* Number of wage workers =1 & Agricultural land < 0.5 ha 0.018 0.004 Number of wage workers =2 & Agricultural land < 0.5 ha 0.045 0.030 Number of wage workers =3 & Agricultural land < 0.5 ha 0.058* 0.041 Number of non-agricultural workers =1 0.030 0.024 Number of non-agricultural workers =2 -0.002 -0.011 Number of non-agricultural workers =3 0.003 Number of non-agricultural workers =1 & Agricultural land < 0.5 ha -0.193** -0.059 Number of non-agricultural workers =2 & Agricultural land < 0.5 ha -0.144** -0.005 Number of non-agricultural workers =3 & Agricultural land < 0.5 ha -0.154** Number of agricultural workers =1 -0.030** -0.028* Number of agricultural workers =2 -0.021 -0.018 Number of agricultural workers =3 -0.041* -0.034* Number of agricultural workers =1 & Agricultural land < 0.5 ha -0.065** -0.060** Number of agricultural workers =2 & Agricultural land < 0.5 ha -0.069** -0.055* Number of agricultural workers =3 & Agricultural land < 0.5 ha -0.101** -0.073* Household with no cultivated land -0.029** -0.029** Own house 0.019 0.021 Floor area per capita 0.008*** 0.008*** Number of rooms per capita 0.235*** 0.238*** Roof: thatch/mostly thatch/plastic 0.091 -0.021 Roof: title/cement/iron 0.110 Wall: bamboo/thatch/grass 0.090 0.077*** Wall: concrete/brick/stone 0.018 0.008 Wall: wood/logs/plywood 0.011 Continue on the next page 152 CAMBODIA POVERTY ASSESSMENT Continued from previous page Dependent variable: Log (consumption per capita per day) Full Model Restrict Model Floor: earth/clay -0.083*** -0.082*** Floor: wooden planks -0.097*** -0.095*** Floor: bamboo strips -0.115*** -0.114*** Floor: cement/polished wood -0.042*** -0.041*** Pipe water 0.039*** 0.039*** Flush toilet 0.063*** 0.064*** Light source: kerosene lamp/candle 0.057 0.064 Light source: battery 0.063*** 0.062** Cooking source: firewood -0.095** -0.061*** Cooking source: charcoal/LPG -0.035 Electricity expense (log) -0.001 -0.001 Has television 0.061*** 0.063*** Has generator 0.112*** 0.116*** Has video 0.085*** 0.086*** Has cell phone 0.074*** 0.076*** Has bicycle -0.035** -0.033** Total number of bicycles 0.027*** 0.027*** Has motorcycle 0.120*** 0.122*** Has car 0.393*** 0.394*** Has sewing machine 0.064*** 0.066*** Has electric kitchen 0.074*** 0.073*** Has electric fan 0.034** 0.033** Has computer 0.002 0.005 Has tactor 0.051 0.051 Has radio -0.031 0.034*** Total number of radios 0.066* Has water pump 0.029** 0.030** Has threshing machine 0.032 0.034 Has rice mill 0.044 0.048* Has cattle -0.018 -0.017 Total number of cattle 0.004* 0.004* Has pig 0.054** 0.053** Total number of pigs 0.004 0.004* Has chicken 0.035*** 0.036*** Total number of chickens 0.000 0.000 Number of disabled household member -0.022** -0.025*** Has washing machine 0.098*** 0.100*** Has refrigerator 0.105*** 0.106*** Has air conditioning 0.069*** 0.068*** Total number of computers 0.086*** 0.085*** Provincial dummy Yes Yes Constant 9.233*** 9.540*** Number of observations 10075 10075 Adjusted R-squared 0.672 0.671 Source: MoP (2021). Note: Significance level: * p<0.10, ** p<0.05, *** p<0.01; Restrict model excludes 7 variables suggested by VIF APPENDIX C 153 Table C.2 Predicted Poverty Headcount in 2009 and 2014 (%) 2009 2014 2019/20 Model 1 33.4 24.9 17.8 Model 2 34.0 25.9 17.8 Model 3 33.8 25.5 17.8 Model 4 33.8 26.3 17.8 Source: MoP (2021). Note: Model 1: Full model. Model 2: Full model, separate model for urban and rural. Model 3: Restrict model Model 4: Restrict model, separate model for urban and rural. Table C.3 Predicted Poverty Headcount by Geographical Area in 2009 and 2014 (%) 2009 2014 2019/20 Model 1 13.9 12.4 9.6 Model 2 14.8 12.0 9.6 Urban Model 3 14.5 12.5 9.6 Model 4 15.1 12.3 9.6 Model 1 37.9 28.4 22.8 Model 2 38.3 29.7 22.8 Rural Model 3 38.2 29.0 22.8 Model 4 38.0 30.2 22.8 Source: MoP (2021). Note: Model 1: Full model. Model 2: Full model, separate model for urban and rural. Model 3: Restrict model. Model 4: Restrict model, separate model for urban and rural. Table C.4 Predicted Poverty Headcount at 1.25 times poverty line in 2009 and 2014 (%) 2009 2014 2019/20 Model 1 51.7 42.2 32.9 Model 2 52.2 43.2 32.9 Model 3 52.2 42.8 32.9 Model 4 52.1 43.7 32.9 Source: MoP (2021). Note: Model 1: Full model. Model 2: Full model, separate model for urban and rural. Model 3: Restrict model. Model 4: Restrict model, separate model for urban and rural. 154 CAMBODIA POVERTY ASSESSMENT Figure C.1 Poverty Headcount Rates Based on Different Methodologies 50 38.6 40 Percent of population 30 33.8 25.1 20 26.3 22.9 17.8 10 13.5 0 2009 2014 2019/20 Revised Method (2019/20 CSES) Revised Method (2014 CSES) Old Method (2009 CSES) Source: CSES 2009, 2014 and 2019/20 to MoP (2021). Figure C.2 Poverty Headcount Rates in 2009 50 43.5 40 38.1 38.6 33.8 Percent of population 30 25.0 24.6 22.3 22.9 19.3 20 12.8 10.2 8.3 10 0 Phnom Penh Other Urban Rural Cambodia Old method (CSES 2009) Revised Method (2014 CSES) Survey-to-survey Source: CSES 2009, and Pimhidzai and Tong (2019). Figure C.3 Poverty Headcount Rates in 2014 50 40 Percent of population 30.3 30 28.3 25.1 26.3 20 18.7 17.6 18.5 12.9 12.5 13.5 10.9 10 6.8 0 Phnom Penh Other Urban Rural Cambodia Old method (CSES 2009) Revised Method (2014 CSES) Survey-to-survey Source: CSES 2009, and Pimhidzai and Tong (2019). APPENDIX C 155 Cambodia Country Office Exchange Square Building No. 19-20, Street 106 Sangkat Wat Phnom, Khan Daun Penh Phnom Penh, Cambodia www.worldbank.org/cambodia