74910 2012 Vietnam Poverty Assessment Well Begun, Not Yet Done: Vietnam’s Remarkable Progress on Poverty Reduction and the Emerging Challenges World Bank in Vietnam Hanoi, 2012 Acronyms AC Agricultural Census ADB Asian Development Bank ASEAN Association of Southeast Asian Nations CAF Center for Analysis and Forecasting CBN Cost of Basic Needs CPI Consumer Price Index CPRGS Comprehensive Poverty Reduction and Growth Strategy CPS Country Partnership Strategy CSA Country Social Analysis DFID Department for International Development (UK) DOLISA District-level MOLISA staff DPT1 Diptheria, Pertussis, and Tetanus, first immunization EA Enumeration Area EAP East Asia and Pacific (WB) ELL Elbers, Lanjouw, and Lanjouw FDI Foreign Direct Investment FGT Foster-Greer-Thorbecke FGT0 Poverty headcount FGT1 Poverty gap FGT2 Squared poverty gap GAPAP Governance and Poverty Policy Analysis and Advice GDI Gender Development Index GDP Gross Domestic Product GSO General Statistics Office HCMC Ho Chi Minh City HCR Headcount Rate HDI Human Development Index HOI Human Opportunity Index ILSSA Institute of Labour, Science, and Social Affairs IMF International Monetary Fund L Large M Medium MCP Monetary Child Poverty (rate) MDCP Multi-dimensional Child Poverty (rate) MDG Millenium Development Goal MICS Multi-Indicator Cluster Survey MOC Ministry of Construction MOET Ministry of Education and Training MOH Ministry of Health MOLISA Ministry of Labor, Invalids, and Social Affairs MPI Ministry of Planning and Investment MPI Multi-dimensional Poverty Index NGO Non-Governmental Organization NHDR National Human Development Report (UNDP) NSS National Sample Survey NTP-PR National Targeted Program for Poverty Reduction NTP-SPR National Targeted Program for Sustainable Poverty Reduction PA Poverty Assessment PAPI Public Administration Performance Index PM Prime Minister POVCALNET PovcalNet, the WB’s online poverty analysis tool PPA Participatory Poverty Assessment PPP Purchasing Power Parity PREM Poverty Reduction and Economic Management PRSP Poverty Reduction Strategy Paper RAFC Rural Agriculture and Fishery Census RCS Ravallion, Chen, and Sangraula RIM Rural Impact Monitoring S Small SCOLI Spatial Cost of Living Index SEDP Socio-Economic Development Plan SEDS Socio-Economic Development Strategy SOE State-owned enterprise SPB Social Policy Bank TFESSD Trust Fund for Environmentally and Socially Sustainable Development UNDP United Nations Development Program UNFPA United Nations Population Fund UNICEF United Nations Children’s Fund USAID United States Agency for International Development VASS Vietnam Academy of Social Sciences VBA Vietnam Bank for Agriculture VDR Vietnam Development Report VHLSS Vietnam Household Living Standards Survey VLSS Vietnam Living Standards Survey VND Vietnam Dong VPHC Vietnam Population and Housing Census WB World Bank WDI World Development Indicators WHO World Health Organization WTO World Trade Organization XL Extra large XS Extra small Acknowledgements This report was prepared in partnership by the World Bank and the Center for Analysis and Forecasting, Vietnam Academy of Social Sciences (VASS), with substantial inputs and comments provided by national researchers and experts as well as international partners, including the United Kingdom (DFID), the United Nations (UNDP, UNICEF, UNFPA, UN Resident Coordinators Office), the European Commission, Ireland (IrishAid), and Oxfam GB. Work on new poverty monitoring systems was carried out jointly with the Social and Environmental Statistics Department of the General Statistics Office (GSO), Government of Vietnam, and the Center for Analysis and Forecasting, VASS. Preparation of the report was led by a core team consisting of Valerie Kozel (Task Team Leader) and Nguyen Thang (Director, CAF), Reena Badiani (World Bank), Bob Baulch (RMIT University), Loren Brandt (University of Toronto), Nguyen Viet Cuong (Consultant, NEU), Vu Hoang Dat (CAF), Nguyen Tam Giang (World Bank), John Gibson (Waikato University), John Giles (World Bank), Ian Hinsdale (World Bank), Pham Hung (Consultant, IRC), Peter Lanjouw (World Bank), Marleen Marra (World Bank), Vu Van Ngoc (CAF), Nguyen Thi Phuong (CAF), Paul Schuler (Consultant), Hoang Xuan Thanh (Consultant, Ageless), Le Dang Trung (University of Copenhagen), Phung Duc Tung (IRC), Linh Hoang Vu (World Bank), and Andrew Wells-Dang (Consultant, Oxfam GB). The team from the General Statistics Office included Nguyen Phong (ex-Director, Social and Environmental Statistics Department), Do Anh Kiem (Director, Social and Environmental Statistics Department), Lo Thi Duc, and Nguyen The Quan. Additional inputs were provided by Paul Van Ufford and the team at UNICEF/Hanoi (on child poverty) and Ingrid Fitzgerald (UN Resident Coordinators Office, Vietnam) and Michaela Prokop (UNDP/Hanoi) on the Human Development Index and multi-dimensional poverty indicators. The report benefited from extensive review and inputs at the concept phase, and the team appreciates the many suggestions received at the World Bank concept review meeting and three early consultations workshops (in Hanoi and HCMC) organized by VASS in 2011. The report benefited as well from comments received at two seminars sponsored by the World Bank office in Hanoi in March and June, 2012, and a technical workshop organized by VASS in June, 2012 to discuss the background papers and an early draft of the report. The team is grateful for comments received at the World Bank decision review in June, 2012, including from peer reviewers: Dominque van de Walle; Michael Woolcock; and Salman Zaidi (all from the World Bank); and Dr. Nguyen Thi Lan Huong (Director, ILSSA). More generally, the team would like to acknowledge comments received throughout report preparation from members of the Vietnam country team as well as staff in East Asia PREM including Mette Bertelsen, Christian Bodewig, Quang Hong Doan, Kari Hurt, Steve Jaffee, Andrew Mason, Nguyen Thi Thu Lan, Trang Van Nguyen, Son Thanh Vo, and Myla Williams, A second and final round of consultation workshops was organized by VASS and the World Bank in HCMC and Hanoi in August, 2012 on the revised draft of the report. The team is grateful for comments and suggestions provided by participants at both workshops, including written comments provided in advance of the HCMC workshop by Dr. Jonathan Pincus (Fullbright Program, HCMC); Dr. Huynh Thi Ngoc Tuyet (former researcher from Southern Institute of Sustainable Development); Dr. Nguyen Hoang Bao (HCMC University of Economics); and Dr. Le Thanh Sang (Southern Institute of Sustainable Development). Written comments were received in advance of the Hanoi workshop from Dr. Le Dang Doanh (former Economic Advisor); Dr. Nguyen Hai Huu (MOLISA); Mr. Do Anh Kiem (GSO); Bert Martens (Oxfam/HK); and Dr Trinh Cong Khanh (CEMA). We are also grateful for comments and suggestions provided at the consultation workshops by Nguyen Tien Phong (UNDP); Pham Quang Ngoc (ADB); Madame Pham Chi Lan (former Vice President of VCCI); and Dr. Dang Kim Son (IPSARD). The team would like to thank the GSO for providing excellent logistical assistance as well as timely access to the 2010 VHLSS and other sources of data. This report is one of many products emerging from the long and fruitful collaboration between the World Bank, VASS, and the GSO on poverty measurement, monitoring, and policy analysis. Guidance for the overall work was provided by Victoria Kwakwa, World Bank Country Director in Vietnam; Sudhir Shetty, Poverty Reduction and Economic Policy Sector Director, and Deepak Mishra, Lead Economist, Vietnam Country Program. Their advice and ongoing support is gratefully acknowledged. The advice of many others, both from inside the World Bank as well as outside, who provided valuable inputs and suggestions throughout the process of preparing the background papers and final report is acknowledged and appreciated. The World Bank in Vietnam’s communications team provided excellent just in time support for dissemination and launch of the final report, with particular thanks to Nguyen Hong Ngan, Vu Lan Huong, and Tran Kim Chi. Tuyet Thi Phung, Lynn Yeargin, Mildred Gonsalvez (all World Bank), and Vu Van Ngoc (CAF) provided excellent administrative support over the course of the project, including the production of the final report. Tuyet Thi Phung and Vu Van Ngoc were responsible for organizing numerous consultation and dissemination events, often working late into the night. Many thanks for your efforts. The team would like to thank DFID for substantial financial support provided under the GAPAP trust fund, including Huong Tran Thi Thien and Renwick Irvine, DFID staff in Hanoi, for their ongoing support in preparing the report. We are also grateful to TFESSD donors for supporting new work on perceptions of inequality. Contents Executive Summary CHAPTER 1 Vietnam’s Growth and Poverty Reduction Record: Remarkable Success, but Big Remaining Challenges 9 A. Introduction 10 B. Vietnam’s economy has grown rapidly and has undergone profound structural transformation 10 C. Progress in reducing poverty has been remarkable by any standard 13 D. Despite this remarkable progress, the task of poverty reduction is not finished 20 E. Overview of the report: Vietnam’s old and new poverty reduction challenges 31 CHAPTER 2 Updating Vietnam’s Poverty Monitoring System 36 A. Introduction 37 B. Rethinking Poverty and Poverty Measurement in Vietnam 37 C. Updating Methods for Measuring Poverty 39 D. Constructing a new GSO-WB Poverty Line 47 E. New Poverty Estimates for 2010: GSO-WB and Official Poverty Methodologies 52 Are the New GSO-WB Poverty Lines too High? Are They Consistent with F.  Citizens’ Subjective Views? 54 CHAPTER 3 Poverty Profile: Establishing the Facts about Poverty and the Poor in Vietnam 63 A. Introduction 64 B. The Poor in Vietnam still Predominately Live in Rural Areas and are Increasingly Concentrated in Upland Regions 66 C. Many of the Poor are Farmers Whose Livelihoods are Primarily Linked to Agriculture 67 D. Ethnic Identity Matters even more for Poverty Today 68 E. Poverty is Still Linked to Low Education Attainment 73 F. Housing and Local Infrastructure have Improved Substantially since the Late 1990s 79 Urban Poverty is Low According to GSO-WB Estimates, and Concentrated in Smaller G.  Cities and Towns 80 H. Poverty has Become Less Correlated with Demographic Factors, although Aging is Emerging as an Issue and Child Poverty Remains a Concern 82 I. Poor Households are Still Vulnerable to Weather Shocks 87 Limited Coverage is Provided by Existing Poverty Reduction and Social J.  Protection Programs 87 CHAPTER 4 Spatial Dimensions of Poverty: 1999 and 2009 Poverty Maps 93 A. Introduction 94 B. 2009 Poverty Maps 95 C. Inequality and Wealth Maps 103 D. The Evolution of Spatial Poverty, 1999 to 2009 106 E. In what other Ways can Mapping Methods Inform Policy Design and Evaluation? 111 CHAPTER 5 Reducing Poverty among Ethnic Minorities 121 A. Introduction 122 B. Ethnic Minority Poverty Reduction Varies across Regions, among and within Ethnic Groups 123 C. Disparities in Access to Education, Infrastructure, and Public Services Accompany and Reinforce Ethnic Minorities’ Poverty Reduction Outcomes 127 D. The Experiences of Ethnic Households that have already Escaped Poverty Offer Lessons and an Innovative Orientation for Future Policies and Programs 131 E. Ethnic Minority Poverty Reduction begins with an Agricultural Transformation from Semi subsistence to Commercial Production 132 F. Successful Ethnic Farmers are Beginning to Diversify into Non-agricultural Employment, Particularly in Areas with Access to Major Cities or International Markets 134 G. Most Ethnic Minorities Continue to Live in their Communities of Origin 136 H. Ethnic Minority Poverty Reduction Strategies Follow a Series of Steps from Agricultural Specialization to Diversification and Accumulation of Financial, Social, and Cultural Capital 137 I. Prevailing Narratives of Ethnic Minority Livelihoods, Cultures, and Gender Relations are Shifting along with Diversified Development, although some Stereotypes Persist 140 CHAPTER 6 Is Inequality Rising in Vietnam? Perceptions and Empirics 145 A. Introduction 146 B. A Step Back: Why are we Concerned about Inequality? 147 C. Is Inequality of Outcomes Rising in Vietnam? 149 D. Why has Income Inequality Increased in Vietnam? 152 E. Inequalities in Opportunities that Perpetuate Income Differences across Generations 164 F. Inequalities in Connections, Voice, and Influence 170 Annexes Annex 1.1 New qualitative research carried out for the 2012 Vietnam Poverty Assessment 32 Annex 2.1  Differences between “Temporally Comparableâ€? and Comprehensive Welfare Aggregates 56 Annex 2.2: Spatial Cost-of-living Estimates for 2010 VHLSS 58 Annex 2.3 Subjective Poverty in Vietnam 59 Annex 3.1 Overview of Vietnam’s Eight Economic Regions 98 Annex 4.1 The Spatial Distribution of Poverty and the Gains from Spatial Targeting 113 174 Annex 6.1 Why doâ€? Perceptions of Inequalityâ€? Diverge from Empirical Measures of Inequality? Figures Figure 1.1 Growth and Poverty Reduction in Vietnam, 1993-2008 10 Figure 1.2 Progress at Reducing Poverty using GSO-WB and MOLISA Monitoring Systems 14 National Poverty Lines Rise with Average Per Capita Consumption: Figure 1.3  Developing andTransition Countries (2005 PPP) 22 Kinh and Ethnic Minorities: Average Annual Rates of Real Growth in Per Capita Figure 1.4  Expenditures, 1998–2010 26 Figure 1.5 Ethnic Minority Poverty Rates and Changing Composition of the Poor, 1993–2010 27 Figure 1.6 Growth in Income Per Capita by Income Group, 2004-10 28 Figure 1.7  Ratio of Ethnic Minority to Kinh Majoirty Enrolment Rates in Public Schools by Level of Education, 1998 and 2010 29 Figure 1.8  Out-of-pocket Spending per Student, by Education and Expenditure Quintile, 2004 and 2010 30 Figure 2.1 Composition of Per Capita Expenditures, 2010 VHLSS 45 Figure 2.2  Composition of Per Capita Expenditures by Per Capita Expenditure Quintile, 2010 VHLSS 45 Figure 2.3 Nutrition Norms Used to Anchor Poverty Lines in Different Countries 49 Figure 2.4 Measuring Subjective Poverty 54 Figure 2.5 Perceived Sufficiency of Consumption by Urban and Rural, 2010 55 Figure 3.1 Level and Composition of Poverty by Region, 1998 67 Figure 3.2 Level and Composition of Poverty by Region, 2010 67 Figure 3.3 Household Income by Expanded Quintile, 2010 68 Figure 3.4 Composition of Income by Expanded Quintile, 2010 68 Figure 3.5 Composition of Poor and Better-off Households in 2010, by Ethnicity 69 Figure 3.6 Distribution of Welfare for Kinh and Ethnic Minorities, 2010 71 Figure 3.7 Level and Composition of Poverty by Region, for Kinh/Hoa 71 Figure 3.8 Level and Composition of Poverty by Region, for Ethnic Minorities 71 Figure 3.9 Composition of Income for Extreme Poor, Poor, and Top Quintile in 2010: Comparing Kinh/Hoa and Ethnic Minority Households 73 Figure 3.10 Schooling Achievement by Age Cohort, 1998 and 2010 73 Figure 3.11 Education Achievements by Expanded Quintiles (persons age 21 and older) 75 Figure 3.12 Population Pyramids for Vietnam: 1999 and 2009 82 Figure 3.13 Monetary and Multidimensional Child Poverty in Vietnam, 2006-10 85 Figure 3.14 Multidimensional Child Poverty in Vietnam by Selected Sociodemographic Variables, 2006-2010 86 Figure 3.15 Child Poverty Rate by Domain, 2010 86 Figure 3.16 Distribution of Population on the Official Poverty List by Expanded Per-Capita Expenditure Quintile, 2010 88 Figure 4.1 Relationship between the Poverty Rate and Gini Index 100 Figure 4.2 Poverty Rate and Proportion of Urban Population 100 Figure 4.3 Poverty Rate and Proportion of Ethnic Minorities 102 Figure 4.4 Poverty Rates, 1999 and 2009 109 Figure 4.5 Progress at Reducing Poverty, 1999-2009 by Poverty Rate in 1999 109 Figure 4.6 Change in Poverty, 1999-2009, Compared to the Initial Gini Index, 1999 109 Figure 4.7 District Poverty: MOLISA compared to Poverty Map Estimates 112 Changes in Welfare Levels ( per-capita consumption) for different Ethnic Figure 5.1  Groups in Vietnam,1998-2010 123 Figure 5.2 Real Per-capita Expenditures for Five Ethnic Categories, 2006-10 125 Changes in Net School Enrolment Rates for Kinh and Ethnic Minorities Figure 5.3  in Rural Areas, 1998-2010 127 Figure 5.4 Net School Enrolment of Selected Ethnic Minority Groups, 2009 128 Figure 5.5 Stunting among Children under Age 5 in Rural Areas, 1998-2010 129 Figure 5.6 Sources of Income for Majority and Minority Households in Rural Areas, 2010 135 Figure 5.7 Sources of Income by Quintile for Minority Households in Rural Areas, 2010 136 Figure 5.8 Paths to Successful Ethnic Minority Development 137 Figure 6.1 Ratio of Mean Per-capita Income by Percentile, 2004-2010 150 Figure 6.2 Mean Per-capita Rural Income per Year by Rural Income Decile, 2004-10 151 Figure 6.3 Theil Decomposition of the Level and Changes in Income Inequality, 2004-10 151 Figure 6.4 Growth by Income Socurce, 2004-2010, Ethnic Minorities 154 Figure 6.5 Growth by Income Source, 2004-2010, Ethnic Majority 154 Figure 6.6 Mean Annual Per-capita Rural Income per Year by Region, 2004-2010 155 Figure 6.7 Sector of Employment for Working-age Individuals in 1998, 2004 and 2010 157 Figure 6.8 Type of Occupation for Working-age Individuals in 1998, 2004 and 2010 157 Figure 6.9 Composition of Income in Urban Areas, 2010 159 Figure 6.10 Composition of Income in Rural Areas, 2010 159 Figure 6.11 Relative Concentration Coefficients of Different Sources of Income, 2010 160 Figure 6.12 Contribution of different Income Sources to the Gini, 2010 161 Figure 6.13 P  er-capita Income per Year by Occupation of the Household Head in Rural and Urban Areas, 2004 and 2010 161 Figure 6.14 Workers Aged 25-30 by Education Level and Job Type 162 Figure 6.15 Hourly Wage and Labor Income Returns to Schooling 163 Figure 6.16 Per-capita Income per Year by Education of most Educated Working-age Household Member, Urban and Rural Households, 2004 and 2010 164 Figure 6.17 R  atio of Enrolments in Primary, Lower Secondary, and Upper Secondary School by Various Groups, 1998 and 2010 165 Figure 6.18 Average Rank in Math Test, by Wealth Quantile, at Ages 5, 8, and 15 Years 167 Figure 6.19 Average Rank in Math Test, by Initial Test Score and Wealth 167 Figure 6.20 Relative Importance of Circumstances for Health Opportunities 170 Figure 6A.1 District-level Expenditure Inequality, 1999 and 2009 175 Figure 6A.2 District-level Expenditure Inequality, 1999 and 2009 Absolute Gini Coefficients 175 Tables Table 1.1 Two Decades of Progress in Reducing the Number of Poor People 16 Table 1.2 Progress at Reducing Incidence, Depth and Severity of Poverty in Vietnam 17 Table 1.3 Improvements in Non-income Dimensions of Poverty, 1993-2010 18 Table 1.4 Contribution of HDI Components to HDI Growth, 1992-2008 19 Table 1.5 Vulnerability to Poverty Remains High in Vietnam 24 Table 2.1 Comprehensive Consumption Aggregates for the VHLSS, 2004, 2006, 2008, 2010 44 Table 2.2  Temporally Comparable Consumption Aggregates for VHLSS, 2004, 2006, 2008, 2010 44 Table 2.3 Spatial Cost-of-Living Index (SCOLI) for each Region and Sector 47 Table 2.4 Composition of the Reference Food Basket, 1993 and 2010 VHLSS 50 Table 2.5  Poverty Estimates for 2010: Comparing the GSO-WB Methodology and Official Methodology 52 Table A2.1 Reference Food Basket for Different Population Groups 57 Table A2.2 Subjective Welfare Regression and Variables at Country Means 58 Table 3.1 2010 Poverty Headcount and Composition, by Region and Sector 66 Table 3.2  Poverty Headcount and Composition in 2010, by Sector of Employment of Household Head 67 Table 3.3 Ethnic Minority Poverty: Headcount and Composition in 2010, Region and Sector 69 Table 3.4 Kinh Majority Poverty: Headcount and Composition in 2010, by Region and Sector 70 Table 3.5 Poverty Headcount, Gap, and Severity in 2010, Kinh and Ethnic Minorities 70 Table 3.6 Poverty Headcount and Composition in 2010, by Education of Household Head 74 Table 3.7 Distribution of Completed Education in 2010, by Ethnicity and Expanded Quintiles 75 Table 3.8  School Enrolment Rates (net) for Boys and Girls in 2010, by Expanded Quintiles and Region 76 Table 3.9  Net School Enrolment Rates for Kinh/Hoa and Ethnic Minority Boys and Girls in 2010, by Expanded Quintile 77 Table 3.10 Average Landholdings for Rural Households in 2010, by Consumption Quintile 78 Table 3.11 Percentage of Rural Households without Allocated or Swidden Land 78 Table 3.12  Percent of Rural Households without Allocated or Sweden Land in 2010, by Region and Quintile 78 Table 3.13 Household Ownership Rates of Durables in 1998 and 2010 (Percent) 79 Table 3.14 Percentage of Households with Access to Housing and Neighborhood Amenities in 2010, by Quintile 80 Table 3.15 Poverty by City Size 81 Table 3.16 Percent of Households with Specific Characteristics, by City Size 81 Table 3.17 Demographic Characteristics and Scale Economies for the Poor 84 Table 3.18 Percent of Households Experiencing Natural Disasters, 2003-08 87 Percentage of Households Officially Classified as Poor, by Expanded Table 3.19  Quintile, 2010 88 Table 3.20 Coverage of Social Protection and Poverty Reduction Policies by Expanded Quintiles 89 Table 3. 21 C overage of Social Protection and Poverty Reduction Policies by Urban/Rural and Ethnicity 90 Table 4.1 Poverty Rate, Depth and Severity: Estimates from the 2010 VHLSS and the Small Area Estimation Approach 95 Table 4.2 Per-Capita Expenditure and Poverty Rate by Province and Region 96 Table 4.3 Inequality and Wealth Measures for Provinces in 2009 103 Table 4. 4 Rural Employment and Percent of the Working Population in Sector 110 Impact on FGT2 of Targeting at Different Levels of Geographic Disaggregation Table A4.1  Optimal Targeting Scheme 117 Impact on FGT0 of Targeting at Different Levels of Geographic Disaggregation Table A4.2  Optional Targeting Scheme 117 Table 5.1 Poverty and Median Expenditures of Major Ethnic Groups in Rural Areas, 2009 124 Table 5.2 Access to Public Utilities by Ethnicity in Rural Areas, 2004-10 130 Boxes Box 1.1 How does Vietnam Monitor Progress at Reducing Poverty? 13 Box 2.1 Do India’s New Official Poverty Lines Measure Up? What are Lessons for Vietnam? 38 Box 2.2 How is Poverty Measured? 40 Box 2.3 How to value Housing Services in the VHLSS 43 Box 3.1 Defining Characteristics of Poor Households at the end of the 1990s 65 Box 4.1 Overview of Program 30A 112 Box 5.1 Six “ Pillars of Disadvantageâ€? 112 Box 5. 2 An Ede Coffee “Hotspotâ€? 133 Box 5.3 Pineapples along the Border 135 Box 5.4 Equity in the Khmer Heartland 140 Box 5.5 Emerging Policy Recommendations: Ethnic Minority Poverty 142 Box 6. 1 Emerging Policy Recommendations: Inequality 173 Maps Map 3.1 Spatial Distribution of Poor Minorities 72 Map 3.2 Spatial Distribution of Poor Kinh 72 Map 4.1 Predicted Poverty Rates of Provinces and Districts, 2009 98 Map 4.2 Distribution of Poverty ( Number of Poor People), 2009 99 Map 4.3 Urban and Rural Poverty Rates in 2009 101 Map 4.4 Poverty Rates of Kinh/Hoa and Ethnic Minority Population in 2009 102 Map 4.5 Expenditure Gini Indices, 2009 105 Map 4.6 Ratio of the 90th Expenditure Percentile to the 10th Expenditure Percentile 105 Map 4.7 Proportion of People in the Richest Expenditure Quintile 106 Map 4.8 Provincial Poverty Rates 107 Map 4.9 District Poverty Rates 107 Map 4.10 Distribution of Poverty (number of poor people) in 1999 and 2009 108 Map 5.1 Regional Patters of Poverty and Wealth for Ethnic Minorities 126 2012 Vietnam Poverty Assessment: Executive Summary EXECUTIVE SUMMARY FramingtheChallenges:WellBegunbutNotYetDone Well Begun but Not Yet Done Vietnam’s record on economic growth and poverty reduction over the last two decades has been Vietnam’s record on economic growth and poverty reduction over the last two decades has been remarkable. Using a “basic needsâ€? poverty line initially agreed in the early 1990s, the poverty headcount remarkable. Using a “basic needsâ€? poverty line initially agreed in the early 1990s, the poverty fell from 58 percent in the early 1990s to 14.5 percent by 2008, and by these standards was estimated to headcount fell from 58 percent in the early 1990s to 14.5 percent by 2008, and by these standards be well below was 10 percent estimated to be by well belowSimilar 2010. progress 10 percent in the by 2010. face of Similar steadily progress inrising incomes the face is evident of steadily rising when assessed by “internationalâ€? incomes is evident when standards assessedof $1.25 by and $2.00 international person/day standards (2005 of $1.25 andPPP). Progress has $2.00 person/day also been (2005 PPP). Progress has also been substantial in other dimensions of well-being, ranging substantial in other dimensions of well-being, ranging from high primary and secondary enrolments to from high primary and improvements secondary in health enrolments status to improvements and reduced morbidity and in mortality. health status and reduced Vietnam morbidity has achieved and andin some mortality. Vietnam has achieved and in some cases surpassed many of the Millennium Development cases surpassed many of the Millennium Development Goals (MDGs). Goals (MDGs). Economic Economic Growth Growth andPoverty and PovertyReduction in Vietnam: Reduction in Vietnam: Two Decades of Two Decades ofProgress Progress 100 18,000 1996Ͳ2000 SEDP 2001Ͳ2005 SEDP 2006Ͳ2010 SEDP 90 16,000 80 Per capita GDP (Thousand Jan. 2010 VND) 14,000 70 12,000 Poverty headcount (%) 60 10,000 50 8,000 40 6,000 30 4,000 20 10 2,000 0 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 $1.25/day 2005 PPP HCR $2.00/day 2005 PPP HCR GSOͲWB poverty line HCR Per capita GDP Market-based reforms to promote high and sustained economic growth were critical to Vietnam’s success, buttressed by egalitarian policies in the provision of basic services, access to land, and Market-based reforms to promote high and sustained economic growth were critical to Vietnam’s success, investments in infrastructure to ensure widespread access to opportunities for the poor. But Vietnam buttressed by egalitarian policies in the provision of basic services, access to land, and investments in is still in the midst of economic and social transformation, from a low-income subsistence economy infrastructure in the 1990sto ensure widespread to a diversified access to opportunities middle-income economy; from for a the poor. But traditional Vietnam agrarian is still culture in the midst with family of economic and social and community of support tofrom transformation systems a low-income one that subsistence is rapidly urbanizing and economy in the becoming more 1990s to a socially diversified. The rate of economic growth has slowed, in part due to external factors, diversified middle-income economy; from a traditional agrarian culture with family and community but also as result systemsof persistent of support high to one thatand inflation macro instability is rapidly urbanizing beginning in late 2007. and becoming moreVietnam started socially to articulate diversified. Thearate of series of structural reforms in 2011 that aim to get the economy back on a path of rapid, sustainable, economic growth has slowed, in part due to external factors, but also as result of persistent high inflation and poverty-reducing growth. and macro instability beginning in late 2007. Vietnam launched a series of structural reforms in 2011 that get the remarkable aim toDespite progress, economy back the task on a path of poverty of rapid, reduction sustainable, is not and complete, and in poverty-reducing some respects it growth. has become more difficult. Vietnam’s “basic needsâ€? poverty line, agreed in the early 1990s, is very Despite lowremarkable progress, by international andof the task standards, poverty the methods reduction is not complete, used to monitor and poverty since in early the some respects are it has 1990s outdated. become The standards more difficult. that applied Vietnam’s to low-income “basic needsâ€? povertyVietnam in the in line, agreed 1990s are no the early longeris 1990s, relevant very low by to modern day, rising middle-income Vietnam. Although tens of millions of Vietnamese households have risen out of poverty, many have incomes very near the poverty line and remain vulnerable to falling back into poverty as a result of idiosyncratic shocks, (such as job loss, accidents, or death or illness of a household member), or related economy-wide shocks, (such as the effects of climate change on rainfall and temperatures, human and animal influenza pandemics, and impacts of the 1 2008–09 global financial crisis). Citizen aspirations are rising, and Vietnam’s development policies must reflect both new economic realities and rising aspirations for shared prosperity and greater economic security. Vietnam’s success has created new challenges. The remaining poor are harder to reach; they face difficult challenges—of isolation, limited assets, low levels of education, poor health status—and poverty reduction has become less responsive to economic growth. Ethnic minority poverty is a persistent challenge. Although Vietnam’s 53 ethnic minority groups make up less than 15 percent of the population, they accounted for 47 percent of the poor in 2010, compared to 29 percent in 1998. Using an updated poverty line that reflects living conditions in 2010, 66.3 percent of minorities are poor compared to only 12.9 percent of the Kinh majority population. Rapid structural transformation and Vietnam’s ongoing transition to a market economy have given rise to new patterns of development that bring additional challenges for poverty reduction. Inequality in incomes and opportunities are rising, underpinned by continuing disparities in human development between urban and rural areas and widening disparities within rural areas and across different socioeconomic groups. Poorer areas are not well connected to markets. While there is good coverage of local infrastructure and basic services in most regions of the country, reliability (for example, of electricity) and quality of services is uneven. Urbanization is accelerating and a growing number of workers from rural areas are migrating to the cities to work in private industry and services. Many of these jobs are informal and lack the benefits historically provided by the public sector and state-owned enterprises. There is a growing demand for young, skilled workers; many older workers do not, however, have the training or skills to compete for jobs in the expanding modern economy. Context of the Report Vietnam has a strong track record of monitoring poverty and undertaking poverty assessments.1 The first report, Vietnam Poverty Assessment and Strategy was prepared in the mid-1990s when Vietnam was still a very low-income country (with GDP per capita of less than US$200) but transitioning rapidly toward a market economy building on economic reforms initiated in 1986. The first poverty profile was developed based on the 1993 Vietnam Living Standards Survey (VLSS). A second report, the Vietnam Country Economic Memorandum Attacking Poverty, was launched in 2000. It was prepared by a joint government-donor-nongovernmental organization (NGO) Poverty Working Group, and drew extensively on the 1993 and 1998 VLSS surveys, combined with four in-depth participatory poverty assessments. The report documented the rapid progress achieved over the course of the 1990s in creating economic opportunities for poor men and women, promoting measures to ensure that growth and access to services is fair and equitable, and reducing the vulnerability of the poor. Vietnam’s third Poverty Assessment, the Vietnam Development Report: Poverty, was produced in 2003, again based on a new round of Vietnam Household Living Standards Surveys (VHLSS) data (2002) and nine new Participatory Poverty Assessments led by local NGOs and think tanks. The report again documented sound progress and highlighted several continuing concerns, such as about spending and investment priorities and the role of voice and participation, and contributed substantially to the design of Vietnam’s Comprehensive Poverty Reduction and Growth Strategy (CPRGS). Vietnam’s fourth Poverty Assessment, Poverty Reduction in Vietnam: Achievements and Challenges (2010), was led by the Vietnam Academy of Social Sciences. While similar to earlier reports in highlighting Vietnam’s impressive achievements over the last two decades, the 2012 report 1 Vietnam’s poverty monitoring system is based on a series of nationally representative household Living Standards Surveys carried out by the General Statistics Office (GSO) in 1993, 1998, 2002, 2004, 2006, 2008, and 2010. Successive rounds of the Vietnam Household Living Standards Surveys (VHLSS), combined with data from special purpose surveys, and supplemented by participatory poverty assessments (in 1998, 2003, and 2008) have been used to document progress. 2 raised concerns—for the first time—about a slowdown in progress. Beginning in 2008, Vietnam struggled with rising macro instability and bouts of high inflation, and the impacts of exogenous shocks (for example, the financial crisis in 2008–09). The report notes that both economic opportunities and risks have been heightened since Vietnam joined the World Trade Organization in late 2007, and macro and financial risks need to be carefully managed. Moreover, the report raised concerns about rising income inequality, that poverty reduction may be becoming less responsive to economic growth, and again raised questions about voice and participation. This new report—Well Begun, Not Yet Done: Vietnam’s Remarkable Progress on Poverty Reduction and the Emerging Challenges—takes a fresh look at the lives of poor men, women, and children and explores the constraints and opportunities they face today in rising out of poverty. It builds on a rich body of poverty analysis and an excellent base of knowledge from previous reports and aims to do three things. First, it proposes revisions to Vietnam’s poverty monitoring system—via better data, updated welfare aggregates, and new poverty lines—to bring these more in line with economic and social conditions in present-day Vietnam. Second, it revisits the stylized facts about deprivation and poverty in Vietnam, and develops an updated profile of poverty using data from the 2010 VHLSS and new qualitative field studies. Third, it analyzes some of the key challenges for poverty reduction in the next decade, including changing regional patterns of poverty and wealth, high and persistent poverty among ethnic minorities, and rising inequality in outcomes and opportunities. Updating Systems for Poverty Monitoring Vietnam has a robust system for monitoring poverty, based on a long-running system of nationally representative, comparable living standards surveys (the Vietnam Household Living Standards Surveys, or VHLSS), consistent estimates of household welfare, and a poverty line that has been kept constant in real purchasing power since 1993. Consistency in methodology and comparability over time are two of the great strengths of Vietnam’s poverty monitoring system. However, by 2008, it was clear that key aspects of Vietnam’s poverty monitoring system had become outdated. The methods used to measure household welfare and construct the original poverty line 2 were based on economic conditions and the consumption patterns of poor households in the early 1990s. Conditions have changed and Vietnam today is very different from Vietnam in the 1990s. In particular, the consumption patterns of poor households in 2010 were substantially different from consumption patterns in 1993, the reference period used to calculate the original poverty line. Beginning in 2009, a team of World Bank staff and consultants worked in collaboration with the GSO to revise and update Vietnam’s poverty monitoring system. The design of the 2010 VHLSS (and subsequent rounds) was improved and a new sample framework developed on the basis of the 2009 Population and Housing Census. The welfare aggregate (per capita consumption) was revised to make it a more comprehensive measure of well-being, and new spatial cost-of-living indexes were calculated using a new survey of consumer prices carried out in conjunction with the 2010 VHLSS. A new poverty line was constructed using a cost-of-basic-needs approach similar to that of the original line, but based on new information on consumption patterns from the 2010 VHLSS. The new 2010 poverty line is VND 653,000 per person per month, which is substantially higher than the original GSO-WB poverty line. The increase reflects improvements in the quality of the food reference basket (fewer calories from rice, more consumption of proteins, vegetables, and fats) and a higher allocation for basic nonfood spending, including housing and durables. The poverty rate in 2010 based on the new poverty line applied to the 2010 VHLSS is 20.7 percent, which compares to a poverty rate of 14.2 percent using official urban and rural poverty lines (VND  eferred to as the GSO-WB poverty line, developed jointly by the General Statistics Office and the World Bank in the 2 R mid-1990s. 3 The poverty rate in 2010 based on the new poverty line applied to the 2010 VHLSS is 20.7 percent, which compares to a poverty rate of 14.2 percent using official urban and rural poverty lines (VND 500,000 per person per month and VND 400,000 per person per month, respectively). The GSO-WB poverty rate is substantially higher in rural areas, in part due to differences between official poverty lines and the new GSO-WB poverty 500,000 perline, but also person due to per month differences and VND 400,000 inper the overall person permethodological approach. month, respectively). The GSO-WB For example, the GSO-WB poverty rate is calculated on the basis of the VHLSS but official poverty rates are calculated poverty rate is substantially higher in rural areas, in part due to differences between official poverty lines and the new GSO-WB poverty line, but also due to differences in the overall methodological at the commune level using a short-form questionnaire and local consultations, then aggregated up to approach. For example, the GSO-WB poverty rate is calculated on the basis of the VHLSS but province and national official povertylevels. rates are calculated at the commune level using a short-form questionnaire and local consultations, then aggregated up to province and national levels. New Poverty Estimates Poverty for 2010 Estimates by Region for 2010 and by Region Urban/Rural and Areas Urban/Rural Areas GSOͲWBPovertyRate OfficialPovertyRate Contributionto Contributionto Incidence(%) total(%) Incidence(%) total(%) AllVietnam(national) 20.7 100 14.2 100 Urban 6.0 9 6.9 14 Rural 27.0 91 17.4 86 RedRiverDelta(Hanoi) 11.4 12 8.4 13 EastNorthernMountains 37.7 21 24.2 20 WestNorthernMountains 60.1 9 39.4 9 NorthCentralCoast 28.4 16 24.0 20 SouthCentralCoast 18.1 7 16.9 10 CentralHighlands 32.8 10 22.2 9 Southeast(HCMC) 8.6 7 3.4 4 MekongDelta 18.7 17 12.6 17 Neither set of poverty lines is inherently better than the other; they are designed to serve different and of poverty Neither setequally lines is The valid objectives. inherently strength better than the of the GSO-WB other; they approach lies inare designed consistent to serve over measurement different and equally valid objectives. time The strength and its independence of the GSO-WB from budgetary or political approach liesItin considerations. consistent serves measurement monitoring over time an important function. In contrast, Vietnam’s official poverty lines are primarily and its independence from budgetary or political considerations. It serves an important intended to help set targets andmonitoring determine resource allocations for the government’s poverty reduction programs and policies. In this function. In contrast, sense, they areVietnam’s official administrative poverty lines lines, constrained are primarily by resource intended availability. The to help official poverty setare lines targets and updated each time Vietnam develops a new five-year Socio-Economic Development Plan. Revisiting the Facts about Poverty and the Poor The new GSO-WB poverty line was used to construct an updated profile of poverty based on the 2010 VHLSS, complemented by new information collected through participatory poverty assessments (PPAs) and qualitative field studies. The poverty rate—defined as the proportion of the population living below the poverty line--is a widely understood and frequently reported measure of poverty. But it ignores the fact that all poor people are not the same; some have incomes or consumption levels very close to the poverty line, while others live in much poorer conditions, well below the standards captured in the poverty line. The poverty profile discriminates between the total poor (individuals living below the GSO-WB poverty line) and the extreme poor (individuals whose per-capita expenditures are less than two-thirds the poverty line). In 2010, 20.7 percent of the population lived in poverty and 8 percent lived in extreme poverty.  he official lines were last updated in 2010 as part of the preparation for the 2011-2015 Socio-Economic Development 3 T Plan (SEDP) 4 The updated poverty profile shows that many of the factors that characterized the poor in the 1990s still characterize the poor today: low education achievement and limited job skills, heavy dependence on subsistence agriculture, physical and social isolation, specific disadvantages linked to ethnic identity, and exposure to natural disasters and risks. Over the past decade, rising levels of education and diversification into off-farm activities have been powerful forces for poverty reduction in Vietnam. The remaining poor still predominately reside in rural areas and their livelihoods depend on agriculture and related activities. But some of the stylized facts have changed. Ethnic minority poverty was only emerging as a concern in the late 1990s, but it has become a much greater concern today as the gap continues to widen between minority populations and the Kinh majority is not narrowing. The report documents wide diversity across Vietnam’s 53 ethnic minority groups, and there are encouraging signs of progress for some groups and in some regions. But the concentration of minorities among the poor is rising; in 1993, poverty was widespread and minorities comprised only 20 percent of all poor households. By 1998, the share of minorities among the poor had increased to 29 percent, and by 2010 minorities accounted for 47 percent of the total poor and a resounding 68 percent of the extreme poor. 66.3 percent of minorities were poor in 2010 compared to only 12.9 percent of the Kinh majority. Many ethnic minorities continue to live in more isolated and less productive upland regions of Vietnam, and three-quarters of their total income comes from agriculture and allied activities. In contrast, poor Kinh have more diversified labor and earnings portfolios and live in coastal and delta regions. Our analysis suggests that agriculture will continue to be an important source of income for many of the poor, including but not limited to the ethnic minority poor. Compared to many other countries, agriculture land is equitably distributed in Vietnam. However the relationship between landlessness and poverty has increased, particularly for households living in the Mekong Delta, despite the rapid expansion in opportunities for off-farm employment and concomitant income diversification over the last decade. Vietnamese today are far better educated than they were a decade ago. Primary school completion rates were high already by the end of the 1990s. Since then, there has been a rapid increase in enrolments at lower and upper secondary levels, leading to an increase in the number of students who attend colleges and universities. Lack of education continues to be an important determinate of poverty: 46 percent of poor households and 58 percent of extreme poor households were headed by persons who had not completed primary school. Gaps in enrolments between children from poor and better-off households persist. Most primary-school-aged children—rich and poor, minority and majority—are enrolled in school. But enrolments among (poor) minorities drop off at the lower secondary level, and children from lower-income households are much less likely to be enrolled in upper secondary schools than children from better-off households, perpetuating the intergenerational transmission of poverty in Vietnam. Differential enrolments also contribute to rising inequality. According to the 2010 VHLSS, 40 percent of persons 21 years and older in the richest quintile had completed a university degree; in contrast, less than 2 percent in the poorest quintile were university graduates. In fact, more than a quarter of those in the poorest quintile had not even completed primary school by 2010. The impacts of demographic factors on poverty have changed since the late 1990s. Child poverty continues to be a concern, although less so than in the 1990s, when poor rural households frequently had many children and struggled to feed and educate them. As a result of family planning policies initiated in the early 1990s, most households now have only one or two children, and the adult children from the erstwhile large families in the 1990s often help to support their parents and siblings. Aging is a new demographic risk; Vietnam’s population is aging and new analysis suggests that the elderly, particularly those who live alone, may be increasingly at risk of poverty in the future. Although targeting is good, existing poverty and social protection programs provide only partial coverage and limited benefits to poor and at-risk individuals. In 2010, only half of the extreme poor were eligible for benefits under the Ministry of Labor, War Invalids, and Social Affairs (MOLISA) 5 Emerging Challenges: Changing Spatial Patterns of Poverty and Rising Inequality New poverty maps4 (see below) were developed based on the 2009 Housing and Population Census and the 2010 VHLSS. The maps indicate that poverty is becoming more concentrated in upland regions of Vietnam, including the North East and North West Mountains and parts of the Central Highlands. In contrast, household wealth is heavily concentrated in the Red River Delta (near Hanoi) and Southeast (near Ho Chi Minh City) and in urban centers along the coast. Although poverty rates are low in urban areas, urban residents struggle to cope with the rising cost of living (including increases in electricity and water tariffs and rising fuel prices), and many work in the informal sector without social protection or employment benefits. Urban poverty is most prevalent in Vietnam’s small cities and towns, which lag behind Vietnam’s larger cities in terms of basic infrastructure and public services.  Poverty Rates (percent poor) in 1999 and 2009 Poverty Rates (percent poor) in 1999 and 2009 1999 2009 The report looks at inequality through two lenses-- a qualitative study of “perceptions of inequalityâ€? and empirical analysis based on various rounds of the VHLSS. The chapter documents widespread concerns across the population about rising inequality The perceptions study draws on a number of rich focus group discussions that describe which inequalities are viewed as unacceptable in the eyes of Vietnamese people, and also captures less easily measured inequalities, such as inequalities in connections, voice, and 2009 poverty maps were estimated using the 2009 Population and Housing Census and the 2010 VHLSS; 1999 poverty 4  influence. The quantitative rates are obtained from analysis examines Minot, Baulch, the and Epprecht factors chapter (2003). See 4 forthe driving rise a more in inequality, detailed description. including geographic variations in growth processes, growth in the non-agricultural sector, and disparities in education and ethnic identity. The rise in income inequality is in part a reflection of growth processes that have altered the relative returns to assets, such as education and productive capital in the economy. 6 Growth has interacted with existing inequalities in opportunities—inequalities in education, access to good jobs, patterns of social exclusion, geographic disparities—to increase income inequality and welfare The report looks at inequality through two lenses-- a qualitative field study in sites throughout Vietnam of “perceptions of inequalityâ€? and empirical analysis based on various rounds of the VHLSS. The chapter documents widespread concerns across the population about rising inequality The perceptions study draws on a number of rich focus group discussions that describe which inequalities are viewed as unacceptable in the eyes of Vietnamese people, and also captures less easily measured inequalities, such as inequalities in connections, voice, and influence. The quantitative analysis examines the factors driving the rise in inequality, including geographic variations in growth processes, growth in the non-agricultural sector, and disparities in education and ethnic identity. The rise in income inequality is in part a reflection of growth processes that have altered the relative returns to assets, such as education and productive capital in the economy. Growth has interacted with existing inequalities in opportunities—inequalities in education, access to good jobs, patterns of social exclusion, geographic disparities—to increase income inequality and welfare gaps between rich and poor households. The persistent and rising gap between the welfare of ethnic minorities and Kinh majorities also contributes to rising inequality: inequality is particularly high in regions with a high concentration of ethnic minorities. The report identifies many new avenues for future research and policy analysis. For example, more work is needed to better understand old and new sources of vulnerability, including urbanization and changing patterns of employment, and new research is needed on aging and health shocks. In addition, a more in-depth analysis of Vietnam’s targeted poverty reduction policies and programs is needed, with particular focus on policies designed to reduce poverty among ethnic minorities. Although Vietnam has successfully eradicated extreme poverty and hunger in all but a few isolated regions, there are widespread concerns about rising inequality in opportunities and outcomes. New work is needed to better understand various sources of inequality and, more importantly, to understand what is the role of public policy in addressing rising inequality. Emerging Policy and Program Implications The report is primarily focused on poverty and inequality diagnostics, and as such aims to support a better informed debate on policy and program responses with key stakeholders in Vietnam (government ministries, the National Assembly, local researchers and research institutes, INGOs and NGOs, and international partners). The findings of the report suggest four emerging areas of policy focus. â—?â—? First, it is essential for Vietnam to reduce volatility and macro instability, and undertake complementary structural reforms–-restructuring state owned enterprises, reforming the financial sector, raising the effectiveness of public investments and moving to a more transparent and open development process—necessary to put Vietnam back on the path of high and sustained rates of economic growth. But the quality of growth matters as much of the rate of growth. â—?â—? Second, measures are needed to make Vietnam’s economic growth more inclusive, for example by supporting productivity and growth in the rural sector through improving connectivity, strengthening skills, improving the investment climate, expanding access to basic services, also better targeting agriculture support measures (e.g. credit, agriculture extension, and market information) to the needs of poor and ethnic minority farmers. Support for labor intensive industries and SMEs in both formal and informal sectors will also contribute to inclusive growth, including better access to credit and training, expanded vocational training for youth in poor and ethnic minority areas, and incentives for local enterprise development to provide more diversified employment options in local communities. The occupational and geographic mobility of labor should be enhanced: migration of rural workers into Vietnam’s rapidly growing cities and towns as well as to foreign countries such as South Korea, Japan and Malaysia has been a powerful force for inclusive growth and poverty reduction in the past. It is important to reduce inequality of opportunities, including improving the quality of education and promoting skills development, particularly in rural areas. Improving governance through greater 7 transparency and accountability will help to increase local participation and prevent a rise in the inequalities that undermine inclusive growth. â—?â—? Third, policies to promote high and inclusive growth must be complemented by effective social insurance and social assistance policies. Vietnam should protect social spending and social assistance in the process of economic restructuring. To put in place automatic stabilizers to effectively protect the poor in challenging times of economic instability, social benefits and official poverty lines should be inflation-indexed, also adjusted to capture differences in the spatial cost of living, including between rural and urban areas, and to properly take into account the basket of goods and services specific to the poor. Better measures are needed to protect poor and vulnerable households from the rising cost of basic services, particularly rising electricity costs in the context of the planned energy subsidy phase-out . Migrant workers have been hard hit by the rising cost of living in urban areas; they should have equal access to basic services, portable benefits (including health insurance), and better access to social protection programs. â—?â—? Finally, continuing improvements are needed in Vietnam’s poverty monitoring system so that it provides a reliable source of information for policy making in a rapidly changing economy. To this effect, objective resource-independent poverty lines should be used in parallel with resource-linked targeting lines, and the source and appropriate application of the two types of poverty lines should be communicated clearly to policy makers, practitioners, and the public. Furthermore, the construction of future poverty profiles and poverty estimates should be done in a transparent way to facilitate monitoring of progress by independent experts and the public at large. 8 chapter 1 Vietnam’s Growth and Poverty Reduction Record: Remarkable Success, but Big Remaining Challenges Vietnam has made remarkable progress at reducing poverty and promoting prosperity over the last two decades. But the task of poverty reduction is not yet finished: shared growth, ethnic minority poverty, increasing vulnerability, and rising inequality are the major poverty challenges going forward. 9 1.1 Vietnam has experienced high and sustained rates of economic growth over the last two decades, driven by a series of market-oriented reforms launched in the late 1980s. Initial progress was led by reforms in the rural economy, which led to a highly egalitarian distribution of agricultural land to rural households A. and diversification in on-farm activities, reforms that provided the right incentives for Introduction increases in farm production and export orientation. In recent years, job creation in the private sector has become a 1.1 Vietnam has experienced high and sustained rates of economic growth over the last two driving force behind Vietnam’s high economic growth, complemented by increased integration decades, driven by a series of market-oriented reforms launched in the late 1980s. Initial progress of agriculture in the was led marketin by reforms economy, and further the rural economy, whichopening of theegalitarian led to a highly Vietnamese economy distribution to global trade and of agricultural land to rural households and diversification in on-farm activities, investment. Vietnam’s accession to the World Trade Organization (WTO) in early reforms that provided the right 2007 created incentives for increases in farm production and export orientation. In recent years, job creation in the opportunities for private a new sector hasround becomeofa reforms, potentially driving force leading behind Vietnam’s highto economic substantial changes growth, in the policy and complemented business environment, with major by increased integration implications of agriculture for economic in the market economy, and furtherand growth poverty opening of the reduction. Vietnamese But these economy to global trade and investment. Vietnam’s accession to the opportunities are accompanied by new challenges and risks; growth has slowed in recent World Trade Organization (WTO)years, and in early 2007 created opportunities for a new round of reforms, potentially leading to substantial Vietnam has struggled changes in the with policyperiods of macro and business instability environment, withand bouts major of high for inflation. implications economic growth and poverty reduction. But these opportunities are accompanied by new challenges and risks; growth has 1.2 Vietnam’s historical slowed in recent years,growth patterns and Vietnam withremarkably have been has struggled pro-poor; periods of macro growth instability in of per and bouts capita gross high inflation. (GDP) averaged 6.1 percent a year between 1993 and 2008, and poverty fell by an domestic product 2.9 percentage average of1.2 points a Vietnam’s historical patterns1.1). year (figure growth have been remarkably pro-poor; growth in per capita gross domestic product (GDP) averaged 6.1 percent a year between 1993 and 2008, and poverty fell by an average of 2.9 percentage points a year (figure 1.1). Figure 1. 1 Growth and Poverty Reduction in Vietnam, 1993-2008 Figure 1. 1 Growth and Poverty Reduction in Vietnam, 1993-2008 100 18,000 1996Ͳ2000 SEDP 2001Ͳ2005 SEDP 2006Ͳ2010 SEDP 90 16,000 80 Per capita GDP (Thousand Jan. 2010 VND) 14,000 70 12,000 Poverty headcount (%) 60 10,000 50 8,000 40 6,000 30 4,000 20 10 2,000 0 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 $1.25/day 2005 PPP HCR $2.00/day 2005 PPP HCR GSOͲWB poverty line HCR Per capita GDP Source: WB-GSO poverty headcount calculated using 1993 and 1998 VLSS and 2004–2010 VHLSS. Dollar-a-day rates Source: come fromWB-GSO poverty Povcalnet. Per headcount capita GDP calculated using GSOusing calculated 1993 population andand GDP1998 data. VLSS and 2004–2010 VHLSS. Note: HCR = Headcount Dollar-a-day rates come of Poverty, Ratefrom that is, incidence Povcalnet. Per capitaof GDP poverty. calculated using GSO population and GDP data. Note: HCR = Headcount Rate of Poverty, that is, incidence of poverty. 1.3 Despite remarkable progress, Vietnam’s task of poverty reduction is not complete, and in important respects, it has become more difficult. This chapter takes stock of Vietnam’s past record 1.3 at reducing Despite poverty and remarkable improvingVietnam’s living conditions—acknowledging progress, remarkableis task of poverty reduction progress judged not complete, and in by any standards—and highlights several remaining and new challenges. It argues that the task respects, important of it has become poverty reduction is by no more complete, This meansdifficult. chapter and that more of takes stock it will become Vietnam’s difficult past record at with growing affluence and reducing poverty and improving livingas rising aspirations, conditions—acknowledging Vietnamese society becomes more remarkable progress heterogeneous, judged by any market- oriented reforms continue, and Vietnam becomes more integrated into the global standards—and highlights several remaining and new challenges. It argues that the task of povertyeconomy. reduction is by no means complete, and that it will become more difficult with growing affluence and B. Vietnam’s economy has grown rapidly and has undergone profound structural transformation 1.4 Comprehensive economic reforms were launched 2 in the second half of the 1980s under Doi Moiand have accelerated over the last two decades. As a result of the reform process, the economy has been liberalized both internally and externally. The passage of the revised Land Law in 1993 and 10 the introduction of the Enterprise Law in 2000 were among the most important milestones in terms of domestic reforms. The accession of Vietnam to the WTO is widely recognized as a key milestone in the country’s external liberalization. Vietnam announced an ambitious plan to restructure the economy and shift into a new growth model in 2011, which is a new and important step in the country’s ongoing transition toward a market economy. 1.5 The Land Law of 1993 marked the continuation of a program of agricultural reforms that were initiated in 1988 with the implementation of Resolution 10. Resolution 10 radically changed the incentive system in the rural sector by recognizing, for the first time, that the household was the basic production unit of Vietnam’s agrarian economy and granted it the needed autonomy. With the aim of consolidating these changes, the 1993 Land Law granted households five basic rights: to transfer, exchange, inherit, rent, and mortgage their land. The law also extended the lease term to 20 years for annual cropland and 50 years for perennial cropland. The implementation of this law resulted in an extensive land titling program in Vietnam. In terms of scale and speed of implementation, it was one of the largest rural titling programs in the developing world (Iyer and Do 2008). Resolution 10 and the Land Law of 1993 together played a crucial role in boosting agricultural growth in the 1990s, thus enabling Vietnam to move from a food deficit country in the 1980s to one of the world’s largest rice exporters by the end of the 2000s. 1.6 A series of additional policy reforms outside the agricultural sector helped lay the foundation for rapid development of the private sector, whose role was officially recognized by Vietnam’s 1992 constitution. The most important milestone in the process was the Enterprise Law of January 2000. It represented a radical change in approach compared to the preceding Private Enterprise Law and Company Law, both of which were approved in 1990. Private enterprises were allowed to operate prior to 2000, but were subjected to a series of government approvals and controls. With the introduction of the new Enterprise Law, citizens were allowed to establish and operate private businesses with limited intervention from government officials. The most important innovation introduced by the Enterprise Law was the simplification of registration procedures and the associated elimination of a large number of business licenses, which sharply reduced transaction costs for businesses and helped install greater business confidence. As a result of these reforms, the number of registered enterprises increased by almost 15 times within only 10 years, from 31,000 in 2000 to 460,000 in 2009, according to the Ministry of Planning and Investment. 1.7 External liberalization has been accelerated at all levels—unilateral, bilateral, regional, and multilateral—over the last two decades. Beginning in the late 1980s, tariffs were unilaterally reduced, and numerous quantitative restrictions on trade abolished. Subsequently, Vietnam actively participated in bilateral and regional trade agreements. Membership in the Association of Southeast Asian Nations (ASEAN) in 1995 and its associated Asian Free Trade Area, and the U.S.-Vietnam Bilateral Trade Agreement in 2001, were important steps in the integration process. After 2003, Vietnam accelerated its negotiations for WTO membership and officially acceded to the WTO in January 2007. Becoming a WTO member has had important implications for Vietnam’s development, because of major changes taking place at the border (a reduction in import tariffs and removal of nontariff barriers to trade), beyond the border (greater access to overseas markets and to the WTO’s dispute settlement mechanism), and behind the border (opening of service sectors and distribution systems, changes in legal and regulatory frameworks,and so forth). Implementation of these agreements not only helped promote exports and restructuring in the domestic economy, but became key drivers for reform of key institutional underpinnings of a market economy, including legal and judicial structures. The Common Investment Law of 2005, for example, helped to harmonize treatment and regulation of all types of businesses including domestic firms, foreign firms, and cooperatives. 1.8 Two decades of reform have helped to sustain high growth in the economy and transform Vietnam in the process. Even with the marked slowdown in economic activity in the last few years in part caused by the international financial crisis, itself a reflection of Vietnam’s growing integration with the rest of the world, the Vietnamese economy has grown at an annual rate of more than 8 percent over the last decade. Today, the Vietnamese economy is four times larger than it was in the early 1990s, and the country now falls into the ranks of lower-middle-income countries. In 2010, per capita gross national income was more than US$3,000 (purchasing power parity [PPP]). 11 1.9 This growth has been accompanied by pronounced structural changes at the aggregate level. Twenty years ago, Vietnam was primarily rural, with nearly 80 percent of the population living in the countryside and only 20 percent residing in cities and towns. Moreover, the urban sector was dominated by two major economic and political hubs, Hanoi in the north, and Ho Chi Minh City in the south. In terms of GDP, slightly more than 40 percent of the economy was generated by agriculture, followed by services and then industry. Growth in the agricultural sector (cropping and farm sidelines) has played an important role in Vietnam’s development success. Nonetheless, its share of GDP has fallen to half of what it was in the early 1990s, and in 2010 contributed 20 percent of GDP. Industry, which includes manufacturing, construction, and utilities, has been the most rapidly growing and dynamic sector and currently makes up 38 percent of GDP. Services contribute 42 percent, modestly higher than the level in 1992. 1.10 These changes in the structure of the economy are largely mirrored in the composition of employment in Vietnam. In 1992, three-quarters of the labor force identified agriculture as their primary source of employment, with only 10 and 15 percent, respectively, in industry and services. Rapid productivity growth in the farm sector has contributed to rising incomes in the countryside; equally important, it has enabled the reallocation of a growing share of labor into even higher-value activities in industry and services. Today, the share of the labor force working in agriculture has fallen below 50 percent, while the share in both industry and services has doubled. 1.11 Accompanying this shift in the composition of employment has been a change in its type, most notably a reduction outside of agriculture in the role of self-employment (largely small, family-run businesses) relative to wage employment. The role of the state in wage employment has also fallen. Overall, however, the state actually employs a slightly larger percentage (upwards of 20 percent) of the labor force than it did in the early 1990s, reflecting the growth in wage employment in the state- owned enterprises sector. Urbanization, aided by increasing migration from the countryside, has also increased, but according to Vietnam’s 2009 census, only 30 percent of the population was classified as urban at that time. This puts urbanization in Vietnam at levels observed elsewhere in Southeast Asia about a decade ago.1 1.12 Thanks to external liberalization, Vietnam’s foreign trade has grown at more than twice the rate of GDP growth, and in 2010 the foreign trade ratio (imports plus exports as a percentage of GDP) was an unprecedented 165 percent. By comparison, and at its peak in China in 2006, it was only 70 percent. The composition of exports has slowly shifted. Exports of oil and agricultural products continue to remain important, but labor-intensive light manufacturing goods now represent the fastest-growing component of exports. Imports of capital machinery and intermediate goods dominate on the other side of the ledger. Export growth has been aided by the run-up in foreign direct investment in Vietnam, which rose from only US$0.5 billion in 1992 to around US$11.0 billion by 2010, with much of this occurring after WTO entry. Rapidly rising wages in China make Vietnam very appealing. Currently, foreign-invested firms are the source of half of Vietnam’s nonoil exports. In terms of employment, however, these firms still employ less than 2 percent of the labor force. 1.13 In addition to productivity growth, rising rates of investment in the domestic economy have been an important source of growth. This works through two channels—on the demand side, as an important source of growth in expenditure, and on the supply side, through investment’s role in expanding the country’s productive capacities and introducing new technology and know-how into the economy. Between 1992 and 2010, gross capital formation rose from only 17.6 percent of GDP to 38.9 percent, comparable to levels observed in the Republic of Korea; Japan; and Taiwan, China at their peaks. In 2010, the World Bank put domestic savings at 33.2 percent of gross national income. With the government sector typically running fiscal deficits and state-run firms net borrowers, the huge increase in savings is coming from a more than doubling in the savings rates of households and private enterprise. 1.14 Finally, reform and rising incomes have had a profound impact on household demographic behavior and population growth. In the early 1990s, average fertility rates of 3.4 births per woman  1 These numbers may underestimate the reduction of the share of employment in agriculture because of the growth in the countryside of secondary jobs in industry and services. In absolute terms, labor supply to agriculture is likely smaller today than it was when the reforms began. 12 translated into population rates of growth of nearly 2 percent per year. By 2010, fertility had fallen to 1.8, below replacement levels, and population growth to only 1 percent. Over the same period, average household size declined by nearly one person, from 5 to 4. With the sharp drop in fertility, the percentage of the population of working age has increased, pushing labor force participation rates upward from 50 to 60 percent of the entire population. Vietnam’s falling dependency ratio, that is, the ratio of those not working to those in the labor force, has had a direct impact on per capita incomes, and indirectly affected incomes through rising savings rates and investment and the “demographic dividend.â€? C. Progress in reducing poverty has been remarkable by any standard 1.15 Vietnam’s dramatic decline in poverty is evident across a number of different approaches used to monitor progress, whether assessed in terms of national poverty lines or using internationally comparable lines, or using household surveys or bottom-up community-based methods (box 1.1). The absolute number of poor people living in Vietnam has dropped sharply, and reductions in the poverty headcount have been accompanied by notable reductions in the depth and severity of poverty. However, progress has been uneven across regions and ethnic groups and has started to slow. Box 1.1 How does Vietnam Monitor Progress at Reducing Poverty? Vietnam has used two very different approaches to measure poverty and monitor progress. Both were initiated in the early 1990s and both have evolved over time. The first approach was developed and led by the Ministry of Labor, Invalids, and Social Affairs (MOLISA), identified in the early 1990s as the primary government agency responsible for poverty reduction programs and policies. MOLISA is tasked with proposing official urban and rural poverty lines at the beginning of each five-year Socio-Economic Development Plan (SEDP) and setting the beginning period poverty rate. Using the official lines and the beginning period poverty rate, MOLISA is responsible for assessing changes in poverty and updating its list of poor households on an annual basis, using a “bottom-upâ€? mix of local surveys and village-level consultations to count the number of poor at local (commune) levels, which are then aggregated up to calculate provincial and national poverty rates. Progress is assessed against poverty reduction targets set in the SEDP. The MOLISA lines were initially based on rice equivalents but since 2005 have been calculated (with technical support from General Statistics Office[GSO]) using a Cost-of-Basic-Needs (CBN) methodology similar to the second approach (see below) led by GSO. The official lines are not adjusted for inflation, but are revised in real terms only every five years. MOLISA’s primary objective using this approach is to determine budget allocations and define eligibility for a number of targeted poverty reduction programs (for example, the National Targeted Program for Poverty Reduction, and Program 30a). The second approach is led by the GSO and measures poverty and monitors progress on the basis of nationally representative household surveys. GSO uses two different methods to measure poverty—one based on official poverty lines (adjusted for inflation) applied to per capita incomes, and one using an approach developed by a joint GSO and World Bank team in the late 1990s and first presented in the 2000 Poverty Assessment. The GSO-WB poverty line is constructed using a standard CBN methodology, based on a reference food basket for poor households anchored in caloric norms (through 2008, 2,100 kilocalories per person per day) plus an additional allocation for essential nonfood needs based on consumption patterns of the poor. Unlike Vietnam’s official poverty lines, the GSO-WB lines have been kept roughly constant in real purchasing power since the late 1990s, and applied to per capita consumer expenditures measured in successive rounds of the Vietnam Living Standards Survey (VHLSS) to estimate changes in poverty over time at the national, urban/rural, and regional level. The GSO-WB lines have been used widely in Vietnam and in international discussions to monitor changes in poverty since 1993. We use these poverty rates in figure 1.1. 13 The share of the population living below Vietnam’s national poverty lines has declined dramatically 1.16 Figure 1.2 shows historical poverty trends based on General Statistics Office/World Bank (GSO-WB) estimates and official poverty lines and methods. The continuing use of the two separate systems for measuring and monitoring poverty, producing widely different poverty estimates, has at times complicated the dialogue between the development community and local researchers (who ced by GSOuse thetypically methodology, the GSO estimates) particularly in the and the government 1990s, (which helped has tended keepMOLISA to official to use the poverty high on th estimates). Although the different estimates sometimes caused confusion, the ongoing development nment’s agenda. and insistence on rigorous approaches to measurement has contributed to a better conceptualization of poverty on the part of government and the policy research community in Vietnam. Moreover the higher poverty rates produced by the GSO methodology, particularly in the 1990s, helped to keep poverty high on the government’s agenda. Figure 1.2 Progress at Reducing Poverty using GSO-WB and MOLISA Figure 1.2 Progress Monitoring at Reducing Systems Poverty using GSO-WB and MOLISA Monitoring Systems 70 60 50 Heacount rate of poverty (HCR) (%) GSOͲWB poverty HCR Official MOLISA poverty HCR 40 30 20.7 20 14.2 10 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Sources: WB-GSO poverty headcount calculated using 1993 and 1998 VLSS and 2004–2010 VHLSS. MOLISA estimates based on UNDP 2004; Government of Vietnam 2005; MOLISA 2011; and 2011 Vietnam Statistical Yearbook. Sources: WB-GSO poverty headcount calculated using 1993 and 1998 VLSS and 2004– 1.17 VHLSS. 2010 MOLISA Over time, estimates as the poverty based rate fell on UNDP (narrowing the gap2004; Government between MOLISA andof Vietnam 2005; GSO estimates) and as the poverty estimates produced through the MOLISA 2011; and 2011 Vietnam Statistical Yearbook. Vietnam Household Living Standards Survey (VHLSS) became increasingly recognized as valid and robust, MOLISA’s poverty estimates have become more aligned with those produced by the GSO. As part of the workup to the 2011–2016 Socio-Economic Development Plan (SEDP), the government agreed formally in Prime Minister’s Over time, the poverty as Decision (PM’s) 60/2010 rate 2 fell (narrowing to separate the gap the two important tasksbetween MOLISA of (a) targeting and GSO poor households for estimates) an social assistance, on the one hand; and (b) measuring and monitoring poverty over time on the other. poverty estimates The aim is toproduced through build on the strengths of the Vietnam both systems. Household As part Living of this agreement, the Standards GSO was given Survey (VHLSS formal responsibility for producing national and provincial poverty estimates, based on successive me increasingly recognized as valid and robust, MOLISA’s poverty estimates have become mor rounds of the nationally representative VHLSS. MOLISA would concentrate on the task of identifying d with those whichproduced by the within individual households GSO. As part provinces, of the districts, workup should and communes to the be2011–2016 included on the Socio-Economi opment Plan (SEDP), the government agreed formally in Prime Minister’s (PM’s) Decisio 102 to separate PM the 2  two Decision important 60/2010 tasks “On the Issuance of (a)Criteria, of Principles, targeting poor and Norms for thehouseholds for social Allocation of Development assistance, on th Investment Funding in the State Budget 2011–2015.â€? and; and (b) measuring and monitoring poverty over time on the other. The aim is to build on th ths of both systems. As part of this agreement, 14 the GSO was given formal responsibility fo cing national and provincial poverty estimates, based on successive rounds of the nationall MOLISA poverty list, with a ceiling defined by the provincial poverty rates proposed by the GSO in consultation with MOLISA. The intention over the longer term is to align MOLISA and GSO’s poverty estimates at the national and provincial levels, with the aggregate number of households on the poverty list determined by GSO’s VHLSS-based measures of poverty based on official poverty lines. 1.18 As part of this new arrangement, GSO and MOLISA worked together to develop a common methodology for producing the national and provincial poverty estimates, including the construction of new official urban and rural poverty lines to be used for the period of the 2011–2015 SEDP. The team developed three options for the new official lines, reflecting different requirements and living standards. The higher options included higher allocations for essential nonfood spending, based on consumption patterns of low-income households in the VHLSS. Following intensive discussion, the government chose the lowest of the three options. While the higher option was preferable on strictly methodological grounds, the government operates under a constrained budget and could not extend benefits under the National Target Program for Sustainable Poverty Reduction (NTP-SPR) and other targeted programs to the anticipated large increase in eligible households—the higher-option poverty lines implied national poverty rates of 18 to 20 percent of the population. Given the inevitable tension between resource availability and needs, the MOLISA lines are often referred to as “budgetingâ€? or “planningâ€? lines, and the process of agreeing on official poverty levels at the start of an SEDP, and annual targets for poverty reduction over the course of SEDP implementation, involve a range of technical, financial, and political considerations. As described in chapter 2, other countries face similar challenges. 1.19 In September 2010, Vietnam announced a new official poverty rate of 14.2 percent (figure 1.2). The official poverty line for urban areas was raised from VND260,000 per person per month (US$1.34 person per day, 2005 PPP) to VND500,000 per person per month (US$1.61 per person per day, 2005 PPP). The official line for rural areas was raised from VND200,000 per person per month (US$1.03 per person per day, 2005 PPP) to VND400,000 per person per month (US$1.29 per person per day, 2005 PPP). A second and higher set of official “near-poorâ€? lines was also approved, allowing the government greater leeway in expanding eligibility criteria when deemed desirable, such as for determining eligibility for health insurance subsidies. The near-poor lines are 30 percent higher than the official poverty lines—VND650,000 per person per month (US$2.24 per person per day, 2005 PPP) for households living in urban areas and VND520,000 per person per month (US$1.83 per person per day, 2005 PPP) for rural households—and similar in value (and implied national poverty rate) to the higher of the three poverty line options initially proposed. 1.20 The government set ambitious targets for poverty reduction in the 2011–2015 SEDP; poverty at the national level is targeted to fall by 2 percentage points each year between 2011 and 2015, and by 4 percentage points in the poorest communities, including those with high proportions of ethnic minority households. Achieving these targets will require a substantially higher rate of progress than achieved under the previous SEDP, and may be particularly challenging given the slowdown in economic growth and in the absence of substantially higher spending to support pro-poor policies and spending. Progress is monitored closely down to the commune level, and there are strong incentives for local authorities to meet these targets.3 1.21 New poverty estimates for 2011 were released by GSO in Vietnam’s 2011 Statistical Yearbook based on a new household survey (2011 VHLSS) covering nearly 47,000 households. Poverty in 2011 is estimated to have been reduced to 12.6 percent—a 1.6 percentage point reduction between 2010 and 2011. MOLISA released its own set of 2011 poverty estimates on March 28, 20124. According to these figures, poverty is estimated to have been reduced to 11.8 percent—a 2.4 percentage point reduction between 2010 and 2011. According to MOLISA’s Decision 375, poverty fell most rapidly in Vietnam’s high-poverty regions—the West Northern Mountains (6.4 percentage points), the North Central Coast (5.7 percentage points), the Central Highlands (3.6 percentage points), and the East  etailed work, including field studies carried out as part of the Poverty Assessment, indicate considerable variation in how 3 D resources for poverty reduction are used at the local level. There are incentives to show progress, and in some cases these incentives may cause officials to focus resources on households just below the poverty line (because progress is judged in terms of crossing the poverty line) rather than chronic or extreme poor. 4 MOLISA Decision 375/QÄ?-LÄ?TBXH issued on March 28, 2012. 15 Northern Mountains (3.2 percentage points). Poverty was estimated to fall by only 1.2 percentage points in the Mekong Delta, well below targets set in the SEDP. In response to a new resolution on social protection (Resolution 15) approved by the Central Party Committee in late 2012, MOLISA is developing new average and minimum living standards cut-offs that will provide a more scientific basis for benefit levels linked to future (new) social assistance programs. The methodology used to calculate minimum living standards is similar to that used to calculate the 2010 GSO/WB poverty line. 1.22 For the present, given the differences in 2011 poverty estimates, and pending stronger implementation of agreements reached in PM Decision 60/2010, there is a strong rationale for continuing to use both the MOLISA approach (for targeting) and the GSO approach (for independent monitoring). We return to this issue in Chapter 2. 1.23 As part of the background work for this report, the team worked closely with the GSO to update the GSO-WB poverty line and related methodologies for poverty monitoring, to ensure that Vietnam’s methods for monitoring poverty fully reflect current economic and social conditions. The updated GSO-WB poverty line is VND653,000 per person per month (US$2.24 per person per day, 2005 PPP), which yields a poverty rate of 20.7 percent in 2010 (figure 1.2, blue triangle for 2010). Chapter 2 describes proposed changes to the GSO-WB approach including improvements to the VHLSS, updated welfare aggregates, and construction of a revised 2010 GSO-WB poverty line. Note that poverty estimates using the new 2010 methodology are not strictly comparable to poverty estimates from recent rounds of the VHLSS for reasons presented in Chapter 2 and are explicitly set apart in the tables and figures in the remainder of this chapter. The fraction of the population living below the international standards of US$1.25 and US$2.00 has also declined 1.24 Vietnam’s own poverty line(s) are clearly better for assessing progress and identifying remaining challenges within the country than international poverty lines. However, PPP-adjusted international poverty lines are often used to compare progress across countries. Vietnam’s progress at poverty reduction is equally impressive judged by international standards of US$1.25 and US$2.00 per person per day (2005 PPP). The poverty headcount fell from 63.7 percent using US$1.25 (2005 PPP) in 1993 to 16.7 percent by 2008, and from 85.7 percent using US$2.00 (2005 PPP) in 1993 to 43.3 percent by 2008, the last year for which comparable poverty rates were published by the World Bank (Table 1.2). Thus, poverty fell by an estimated 3 percentage points per year between 1993 and 2008, albeit with faster progress in the 1990s and first half of the 2000s than in recent years. total,nearly In total, nearlyhalf Vietnam’s half population Vietnam’s was lifted population was out of poverty lifted out ofin less thanin poverty two decades less than two decades 1.25 Measured by temporally comparable GSO-WB standards, more than 43 million people were 1.25 Measured by temporally comparable GSO-WB standards, more than 43 million people were lifted out ofout lifted of poverty poverty between between 1993 andand 1993 2008. 2008. A remarkable A remarkable reduction reduction in in the the number number of of poormen, men, women, poor women, and children living in Vietnam is also confirmed using PPP-adjusted international poverty lines. and children living in Vietnam is also confirmed using PPP-adjusted international poverty lines. Table 1.1 Two Decades of Progress in Reducing the Number of Poor People Table 1.1 Two Decades of Progress in Reducing the Number of Poor People Poverty standard Number of poor Change (millions) (millions) (% pts) 1993Ͳ 1993Ͳ 1998Ͳ 1993Ͳ 2008, 1993 1998 2008 1998 2008 2008 Annual Official GSOͲWB poverty line: consumption 39.8 28.2 12.3 Ͳ11.5 Ͳ15.9 Ͳ27.4 Ͳ2.9 $1.25/day (2005 PPP): consumption 43.6 37.5 14.3 Ͳ6.2 Ͳ23.1 Ͳ29.3 Ͳ3.1 $2.00/day (2005 PPP): consumption 58.7 59.0 36.9 0.4 Ͳ22.1 Ͳ21.8 Ͳ2.8 Sources: VASS 2010 for 1993–2008 GSO-WB headcount estimates; POVCALNET for 1993–2008 US$1.25 and US$2.00 Sources: VASS headcount for 1993–2008 2010Population estimates. GSO-WB statistics taken headcount estimates;except from POVCALNET POVCALNET for for 2010, 1993–2008 which US$1.25 come from and World US$2.00 Bank Data on headcount estimates. Population statistics taken from POVCALNET except for 2010, which come from World Bank Data on Vietnam web page, http://data.worldbank.org/country/vietnam. Vietnam web page, http://data.worldbank.org/country/vietnam. The depth and severity of poverty have also fallen sharply 1.26 The poverty headcount is a widely understood 16 and widely reported measure of poverty. However, it ignores the fact that all poor people are not the same; some have incomes or consumption levels very close to the poverty line, while others live in much poorer conditions, well below standards reflected in The depth and severity of poverty have also fallen sharply 1.26 The poverty headcount is a widely understood and widely reported measure of poverty. However, it ignores the fact that all poor people are not the same; some have incomes or consumption levels very close to the poverty line, while others live in much poorer conditions, well below standards reflected in the poverty line. Two additional indicators are used to measure the depth and severity of poverty. The poverty gap (depth) measures the average, across all people, of the gap between the living standards of the poor and the poverty line. The squared poverty gap (severity) is calculated using a similar methodology, but gives greater weight to households whose living standards are further away from the poverty line. 1.27 According to table 1.2, Vietnam has made steady progress in reducing the depth and severity of poverty, whether measured by national or international standards. Living conditions not only have improved for households living near the poverty line, but also for many of Vietnam’s poorest households. Table 1.2 Progress Table at Reducing 1.2 Progress Incidence, at Reducing Depth Incidence, Depth and and Severity Severity of Poverty of Poverty in Vietnam in Vietnam GSOͲWB poverty line $1.25/day 2005 PPP line $2.00/day 2005 PPP line Incidence Depth Severity Incidence Depth Severity Incidence Depth Severity (Squared (Squared (Squared (Headcount (Poverty Poverty (Headcount (Poverty Poverty (Headcount (Poverty Poverty rate, %) Gap, %) Gap, %) rate, %) Gap, %) Gap, %) rate, %) Gap, %) Gap, %) 1993 58.1 18.5 7.9 63.7 23.6 11.0 85.7 43.5 25.7 1998 37.4 9.5 3.6 49.7 15.1 6.0 78.2 34.2 18.0 2002 28.9 7.0 2.4 40.1 11.2 4.1 68.7 28.0 14.1 2004 19.5 4.7 1.7 21.5 5.4 2.0 50.3 17.1 7.8 2006 15.9 3.8 1.4 16.8 4.2 1.5 42.4 13.9 6.2 2008 14.5 3.5 1.2 11.8 2.8 1.0 34.5 10.3 4.3 2010 20.7 5.9 2.4 N.A. N.A. N.A. N.A. N.A. N.A. Sources: VASS, 2010 for 1993–2008 GSO-WB headcount estimates; POVCALNET for 1993–2008 US$1.25 and US$2.00 Sources: VASS, headcount 2010 estimates; for 1993–2008 Statistics GSO-WB for 2010 headcount calculated estimates; by the World POVCALNET Bank using for 1993–2008 the comprehensive US$1.25 consumption and US$2.00 aggregate. headcount estimates; Statistics for 2010 calculated by the World Bank using the comprehensive consumption aggregate. Note: Poverty estimates Note: Poverty usinginternational estimatesusing povertylines internationalpoverty lineshave havenot notbeen beenpublished publishedyet by yet the by World the Bank World for Bank Vietnam for Vietnamin 2010. in 2010. the Butthe But rate rate of of poverty poverty reduction reduction is slowing, is slowing, linked tolinked to rising rising macro macro instability instability and and slower growth slower growth 1.28 High and sustained rates of economic growth have been a key factor in Vietnam’s success at 1.28 High and sustained rates of economic growth have been a key factor in Vietnam’s success at reducing poverty. But the economy has slowed in recent years. Beginning in late 2007, Vietnam has reducing poverty. But the economy has slowed in recent years. Beginning in late 2007, Vietnam has struggled with struggled economic turbulence with economic turbulence andand inflation, with sharp inflation, with sharp and persistent increases and persistent increases in the prices in the prices of of many basic many commodities. Many basic commodities. workers lost Many workers lost jobs; others received jobs; others received lower lower wages and reduced wages and reduced working working hours due hours due to reduced demand to reduced demandduring the global during the global economic crisis in late economic crisis 2008 and late 2008 early 2009. and early Farmers 2009. Farmers complain that complain thecosts thatthe costsofof agricultural agriculturalinputs inputsare rising, are and rising, profit and margins profit are reduced. margins There are reduced. were again There were again rising food prices and a sharp increase in the costs of electricity and fuel in rising food prices and a sharp increase in the costs of electricity and fuel in 2010, which put additional2010, which put additional pressure on household budgets. Households in urban and peri-urban pressure on household budgets. Households in urban and peri-urban areas have been particularly hard hit areas have been particularly by hard including high inflation, hit by high inflation, including rural-to-urban migrants rural-to-urban who come to the migrants ever come city in who growingto numbers the city into ever seek growing numbers to seek better jobs and higher pay. Migrants send money home to rural areas; the better jobs and higher pay. Migrants send money home to rural areas; the impacts of higher urban prices impacts of higher urban prices are thus also passed on to households living in rural areas through are thus also declining passed on (see, remittances to households living for example in rural VASS areas 2011). through declining Urbanization remittances is increasing at a(see, rapidfor example pace and VASS 2011). Urbanization is increasing at a rapid pace and the face of poverty the face of poverty and sources of vulnerability in urban areas differ in important respects from more and sources of vulnerability in urban areas differ in traditional poverty concerns in rural areas. important respects from more traditional poverty concerns in rural areas. Vietnam has also achieved dramatic progress in improving the non-income dimensions Vietnam of poverty has also achieved and has met dramatic or is progress inlikely to meet improving most of the the non-income Millennium dimensions of poverty Development and Goals has met or is (MDGs) likely to meet most of the Millennium Development Goals (MDGs) 1.29 Table 1.29 1.3 Table documentsprogress documents 1.3 along other progress along dimensions of other dimensions of well-being. well-being.Vietnamese today are Vietnamese today are much much better educated and arguably better prepared to get jobs in industry or services. In 1998, 25 percent better educated and arguably better prepared to get jobs in industry or services. In 1998, 25 percent of of persons aged 15 to 24 did not complete primary school. By 2010, only 12 years later, the percentage had fallen to only 4 percent, and upper secondary enrolments had nearly doubled (60 percent for girls, 54 percent for boys). Moreover, by 2010, there were more girls enrolled in both levels of secondary school than boys; Vietnam scores remarkably well in terms 17of gender parity in education. persons aged 15 to 24 did not complete primary school. By 2010, only 12 years later, the percentage had fallen to only 4 percent, and upper secondary enrolments had nearly doubled (60 percent for girls, 54 percent for boys). Moreover, by 2010, there were more girls enrolled in both levels of secondary school than boys; Vietnam scores remarkably well in terms of gender parity in education. 1.3 Improvements Table1.3 Table in Non-income Improvements in Dimensions of Non-income Dimensions of Poverty, Poverty, 1993-2010 1993-2010 1993 1998 2010 Education % of 15ͲorͲolder who have not completed primary school 35.5 35.7 14.4 % of 15Ͳ24 who have not completed primary school 23.3 25.4 4.1 Primary enrollment rate (net) Female 87.1 90.7 92.8 Male 86.3 92.1 92.5 Lower secondary enrollment rate (net) Female 29.0 62.1 83.2 Male 31.2 61.3 80.2 Upper secondary enrollment rate (net) Female 6.1 27.4 60.1 Male 8.4 30.0 53.9 Health Immunization, DPT1, % of children ages 12Ͳ23 months 91 94 93 Immunization, measles, % of children ages 12Ͳ23 months 93 96 84 Infant mortality (per 1,000 live births) 34 29 14 Incidence of stunting (low height for age), children under 5 51 34 23 Incidence of underweight (low weight for age), children under 5 37 36 12 Life expectancy at birth (years) 68.1 71.0 74.8 % of poor with health insurance n/a 7.8 71.6 Access to infrastructure and durables % using electricity as main source of lighting 48 77 98 % with access to an improved* water source Rural 76 70 87 Urban 89 89 98 % with access to clean** water Rural 17 29 57 Urban 60 75 89 % with sanitary latrine 19 26 69 Rural 10 14 59 Urban 53 68 92 % of households with durable goods TV 22 56 89 Fan 31 68 85 Refrigerator 4 9 43 Car 0 0 1 Motorbike 11 20 76 ** Clean water is defined to include piped water, bottled water, water from deep wells with pumps, and rainwater. * Improved water sources are defined as clean water sources plus handͲdug, reinforced wells and filtered spring sources. Sources: Sources: 2010: 2010: immunization, immunization, malnutrition, malnutrition, and infant and infant mortality mortality statistics comestatistics come from various from of rounds various of the rounds MICS; life expectancy from World Bank World Development Indicators database; all others from World Bank 2000. the MICS; life expectancy from World Bank World Development Indicators database; all others from World Bank 2000. 18 1.30 Vietnamese today are also healthier and live longer than in the 1990s; infant mortality (deaths per 1,000 live births) had fallen to 14 in 2010, which is impressive even by middle-income standards, and life 1.30 Vietnamese today are also healthier and live longer than in the 1990s; infant mortality (deaths per 1,000 live births) had fallen to 14 in 2010, which is impressive even by middle-income standards, and life expectancy had risen to 74.8 years. There was also marked improvement in levels of nutrition, although stunting (low height-for-age) remains a concern in some regions of the country and among minority populations. While immunization coverage looks good on the surface—over 90 percent of children begin the recommended series of childhood immunization (for example, DPT1)—the 2010 Multiple Indicators Cluster Survey (MICS) documents immunization completion rates of only 60 percent (GSO 2011). 1.31 Access to infrastructure and local services improved; the number of households connected to the electricity grid increased from 77 percent in 1998 to nearly universal coverage (98 percent) by 2010. However, many households still do not have access to “improvedâ€? water sources,5 particularly in rural areas, or sanitary latrines. But while challenges in these areas remain, there have been dramatic improvements in coverage since 1998. 1.32 Improvements are also notable in housing quality and ownership of durables. By 2010, 89 percent of Vietnamese households owned TVs (compared to 56 percent in 1998), 85 percent owned an electric fan (compared to 68 percent in 1998), 43 percent owned a refrigerator (compared to 9 percent in 1998), and a substantial 76 percent owned at least one motorbike (compared to 20 percent in 1998). If affluence and quality of life are reflected, at least in part, in the consumer durables that people own and use, then there have been dramatic improvements since the late 1990s. 1.33 According to the most recent national Human Development Report (HDR) for Vietnam (UNDP 2011), the country has achieved or is likely to achieve most of the MDG targets by 2015. However, concerns about clean water and sanitation remain (Goal 10), and Vietnam continues to make slow progress toward environmental goals (Goal 9). Progress is also apparent in composite indicators of well-being 1.34 Recent years have witnessed a greater focus on composite indicators of poverty and deprivation in Vietnam, beginning with the Human Development Index (HDI) in the early 1990s, and more recently the Multi-dimensional Poverty Index (MPI) launched in the 2010 Vietnam HDR.6 The MPI builds on earlier work done to measure nonmonetary poverty, such as the approach to measuring child poverty developed by GSO and MOLISA with support from UNICEF, as well as the multidimensional poverty index used in the 2010 Urban Poverty Survey (UNDP 2011). 1.35 Vietnam has seen steady improvements in human development, evidenced by increases in the HDI over time: the HDI value increased 19 percent between 1992 and 2008. With an HDI of 0.728, Vietnam is now comfortably placed among medium human development countries (table 1.4). Table 1.4 Contribution of HDI Components to HDI Growth, 1992-2008 Year HDI Life Contribution Education Contribution Income Contribution Expectancy of Life Index of Education Index of Income Index Expectancy Index to HDI Index to Index to HDI Growth since Growth since since Previous Previous Previous Period (%) Period (%) Period (%) 1992 0.611 0.670 — 0.776 — 0.386 — 1995 0.639 0.690 18.8 0.808 25.9 0.420 55.3 1999 0.651 0.721 86.1 0.803 -13.9 0.430 27.8 2004 0.701 0.782 40.7 0.826 15.3 0.496 44.0 2008 0.728 0.794 15.2 0.830 5.1 0.559 79.7 Contribution to total change 35.2 N.A. 15.9 N.A. 48.95 in HDI 1992–2008 Sources: 2001 Vietnam HDR; HDI, 1992, 1995, 1999, 2004, 2008. Note: HDI = Human Development Index, N.A. indicates not available. 5 See table 1.3 for definitions of “cleanâ€? and “improvedâ€? water sources.  he Government of Vietnam uses changes in the HDI and in the Gender Development Index as an indicator of progress 6 T toward human development and gender equality. Improvement in the HDI rank and value was also included as a target in the current SEDP 2001–2010. The SEDP 2011–2015 refers to improvements in the HDI as an indication of progress toward development goals, while the 2010 national MDG report cites positive change in the Gender Development Index as a sign of progress toward achieving gender equality and women’s empowerment. 19 1.36 The HDI is a composite index and there have been differences in progress for each of the different HDI sub-indices. Strong economic growth between 1992 and 2008 increased the income index by 45 percent. The life expectancy index also saw significant gains, rising by 19 percent between 1992 and 2008. This reflected steady improvements in average life expectancy from 65.2 years in 1992 to 72.7 years in 2008. The education index, which started from a relatively higher base in 1992, saw a slower rate of increase, rising by only 7 percent by 2008. The contribution of the education index to overall growth in the HDI decreased from around 25.9 percent from 1992 to 1995 to 5.1 percent from 2004 to 2008. Thus, since 1992, rising GDP, together with increased life expectancy, have been the main drivers of improvement in Vietnam’s HDI. Slowing gains in life expectancy are to be expected once years of life expectancy reach higher levels. However, slowing gains in the education index may be cause for concern. 1.37 There is a strong correlation between elements of good governance and higher levels of human development. Of the six dimensions of Vietnam’s Public Administration Performance Index (PAPI), public service delivery is most strongly correlated with the HDI, followed by transparency, participation at local levels, and vertical accountability. Similarly, control of corruption is also highly correlated with the HDI (CECODES, FR, CPP, and UNDP 2012). D. Despite this remarkable progress, the task of poverty reduction is not finished 1.38 Vietnam has made remarkable progress toward its longstanding goal of eradicating poverty. By the end of the 2006–2010 SEDP, only 9.5 percent of households were estimated to live below Vietnam’s official poverty lines, and poverty estimates based on the original GSO-WB basic-needs poverty line suggest similar results. Does this mean that the task of poverty reduction is finished, except for addressing a few remaining pockets of poverty, and a continuing commitment to look after the poorest and most destitute? 1.39 The task may be finished in terms of meeting the most basic food, shelter, and clothing needs of Vietnamese citizens. Vietnam rightly deserves to be recognized for this. But are these the right standards to apply in a rapidly growing, modernizing economy like Vietnam? The remainder of this chapter will discuss why the task of poverty reduction is not finished in Vietnam, and indeed has become more difficult in many respects. 1.40 The task of eradicating poverty is not finished because: â—?â—?  Standards have changed. By the end of the 2006–2010 SEDP, Vietnam’s system for measuring and monitoring poverty no longer adequately captured the living conditions of the population. The original GSO-WB poverty lines were set in the mid-1990s and do not reflect the consumption patterns or broader aspirations of the population today. â—?â—? Many of the erstwhile poor remain vulnerable to slipping back into poverty. Weather  shocks, health shocks, and exposure to other income shocks remain widespread, and in some areas may even be rising. 1.41 Moreover, Vietnam’s rapid pace of development has bred its own challenges. The economy has gone through massive changes since the late 1990s. Workers in their 40s and 50s made schooling and skills training decisions in a much different economy, based on a different set of incentives. Many do not have the skills or training to compete for jobs in today’s rapidly modernizing economy. Even young workers often leave school without adequate training for an expanding skills- based economy. 1.42 The task of eradicating poverty has become more difficult in other important respects. Growth rates have fallen sharply compared to the first half of the 2000s, and growth is expected to remain sluggish in the foreseeable future. In addition, poverty reduction is becoming less responsive to economic growth. The remaining poor are harder to reach; the easy wins due, for example, to land 20 reforms in the early 1990s, rapid expansion in rural areas into cash crop production, and agricultural diversification have for the most part been realized. The remaining poor are more concentrated in isolated regions and among ethnic minority groups, where structural issues linked to assets and location are binding constraints (for example, poorer-quality land, less education and training, and more limited infrastructure and public services). Poverty reduction policies and programs must reflect these changing realities. 1.43 Vietnam’s ongoing structural transformation to a market economy has given rise to trends that suggest new challenges for poverty reduction. â—?â—? Inequality is back on the agenda. There are widespread concerns among Vietnamese citizens from all walks of life about rising inequality. Recent analysis suggests an increase in income inequality between 2004 and 2010, driven predominantly by growing inequality within rural areas. â—?â—? Continuing disparities in human development contribute to income inequalities. While Vietnam has done a good job on coverage of basic services, quality is uneven, and there are large perceived gaps between better-off and poorer households and regions. With the push toward “socializingâ€? health and education services, access has become more closely linked to incomes, and the burden of out-of-pocket spending for health and education is rising. â—?â—? Vietnam’s cities and towns are growing rapidly, due in part to a massive influx of migrants from rural areas of the country. The cost of living in urban areas is rising, due to rising food costs and to rising demand, higher fuel prices, and water and electricity tariffs. The private sector accounts for an increasing share of the urban labor force, and many continue to work in the informal sector without social protection or employment benefits, as was revealed in a number of studies conducted in recent years such as the 2009 Urban Poverty Survey (Haughton et. al. 2010), various rounds of the Vietnam Academy of Social Sciences’ (VASS’s) Rapid Impact Monitoring (RIM) assessments of the global economic crisis (VASS 2009, 2011), and Oxfam-ActionAid’s urban poverty monitoring studies (Oxfam GB/ActionAid 2008, 2011). New forms of vulnerability are developing, in particular among workers in the informal sector and rural migrants in cities like Hanoi and Ho Chi Minh City. Poverty lines used to monitor Vietnam’s progress are low by international standards 1.44 When assessing Vietnam’s performance in recent years, it is important to keep in mind that both official lines and the original GSO-WB poverty line are low by international standards, and, unlike in many other fast-growing economies, the GSO-WB line has not been revised since it was agreed in the mid-1990s. Using a constant standard to assess progress has many advantages. But most countries raise their standards—and their national poverty lines—as they become more affluent and as the aspirations and expectations of citizens change. Figure 1.3 shows the strongly positive relationship in developing and transition countries between national poverty lines (US$ per month, 2005 PPP) and average per capita expenditures (2005 PPP) (Chen and Ravallion 2008). The overall income elasticity of the national poverty line for countries in the sample is .66, with a substantially higher elasticity for the nonfood component of poverty lines (.91) than the food component (.47). Thus, assessed globally, the economic gradient in national poverty lines is driven more by the gradient in nonfood needs, which account for more than 60 percent of the overall elasticity. This is not surprising; food consumption becomes a much smaller share of total consumption as populations become more affluent. In countries like the United States, for example, even the poor spend only 20 to 25 percent of total expenditures on food. 21 ple, even the poor spend only 20 to 25 percent of total expenditures on food. Figure 1.3 National Poverty Lines Rise with Average Per Capita Consumption: Figure 1.3 National Poverty Lines Rise with Average Per Capita Consumption: Developing andTransition Countries (2005 PPP) Developing and Transition Countries (2005 PPP) 3 0 National poverty line ($/month at 2005 PPP) 0 2 0 0 0 0 0 3 4 5 6 7 Log consumption per person at 2005 PPP Note: Fitted values use a lowess smoother with bandwidth=0.8 Source: Chen and Ravallion 2008. Source: Chen and Ravallion 2008. 1.45 The poverty statistics cited in table 1.1 are based on the original GSO-WB poverty line of only US$1.10 per person per month (2005 PPP), which is substantially lower than the US$1.25 per person per day (2005 PPP) “internationalâ€? poverty line calculated by the World Bank and used to measure global progress at reducing poverty. The US$1.25 per person per day international poverty line sets 16 the national poverty lines for the 15 poorest a very low standard; it was constructed by averaging countries in the World Bank’s database of comparator countries7 (Ravallion, Chen, and Sangraula 2008). Higher international poverty lines are typically used for rising middle-income countries. The median poverty line for all developing and transition countries is US$2.00 per person per day (PPP 2005), and the median line for all countries besides the poorest 15 countries is US$2.50 per person per day (PPP 2005). An international poverty line of $4.00 per person per day (PPP 2005) is used for a number of countries in Latin America. Vietnam’s poverty lines are low relative to its rising prosperity and concomitant rising aspirations 1.46 Poverty lines typically increase with economic development because norms change; what was considered an acceptable level of deprivation in the 1990s is no longer acceptable today. Poverty lines also rise because governments have greater capacity and more resources to respond to changing norms. 1.47 Evidence of changing norms is reflected in subjective poverty lines estimated using information reported by households in the 2010 VHLSS on the perceived adequacy of their current levels of consumption. Subjective lines suggest national poverty rates of 20 to 25 percent, substantially higher than current official poverty estimates (Chapter 2). 1.48 Changing norms and higher aspirations are also captured in a number of qualitative field studies and assessments that have been carried out over the past decade. For example, in the 1999 and 2003 Participatory Poverty Assessments (PPAs) carried out by the World Bank in collaboration  alawi, Mali, Ethiopia, Sierra Leone, Niger, Uganda, Gambia, Rwanda, Guinea-Bissau, Tanzania, Tajikistan, Mozambique, 7 M Chad, Nepal, and Ghana. 22 with other donors, international NGOs, and Vietnamese partners, poor respondents defined well- being in terms of adequate food, a stable asset endowment (adequate land, labor, and housing), plus nonmaterial aspects such as community respect and freedom from debt and anxiety (ADB 2003; World Bank 1999). Respondents in the more recent 2008 PPA did not refer to hunger or food security, but instead spoke about risks related to rising food prices, concerns about access to employment, and stable jobs (in the face of emerging impacts from the global financial crisis). 1.49 In research on ethnic minority poverty for this report (Annex 1.1), ethnic minority respondents in three regions were asked about indigenous definitions of success. The most common response was linked to sufficiency of basic needs: enough food to eat year-round, clothes to wear, decent housing, and ability to participate in cultural festivals and customs (such as being able to prepare a pig for the Tet festival). Other respondents realized that ideas of success were changing, pointing to increasing material prosperity and connections to the market economy. One minority official in Muong Khuong district, Lao Cai, said: “In the past it was considered enough to be full and dress warmly (an no, mac am); now people want to eat well and dress beautifully (an ngon, mac dep).â€? Traders mentioned having a larger, cleaner multistoried house as a key indicator of success. Among respondents who have transitioned to trading or other nonagricultural work, the desire for children to be educated and have stable jobs, particularly in the state sector, also formed part of a concept of success. Thus, ideas of well-being, even among poorer Vietnamese, are shifting from satisfaction of basic needs to a higher asset base combined with social status and non-income factors such as health and education. Vietnam increased its official poverty lines in late 2010, and a revision to the GSO- WB line is proposed in this report 1.50 Despite intense internal debate—many policy makers believe Vietnam should set more ambitious goals in the fight against poverty, given its rapid economic growth and vision of itself as a modern industrial society—the new official poverty lines set in 2010 for the 2011–2016 SEDP are still low by international standards. The new urban line is still well below US$2 per person per day (2005 PPP), and the new rural line is only a little above the US$1.25 per person per day lines applied in the world’s poorest countries. 1.51 As noted, the World Bank is working with the GSO and other local partners to update GSO’s poverty monitoring system, through improvements to the VHLSS household survey; more comprehensive welfare aggregates; and a revised GSO-WB poverty line, using an updated food reference basket (from the 2010 VHLSS), a more comprehensive measure of nonfood spending that includes the flow of consumption from household assets (consumer durables and housing), and new spatial cost-of-living indexes. Despite progress, many households remain vulnerable to falling into poverty in Vietnam, and new sources of vulnerability are emerging as a result of external global events and internal instability 1.52 Although tens of millions of Vietnamese households have risen out of poverty over the last decade, many have incomes very near the poverty line and remain vulnerable to falling back into poverty as a result of idiosyncratic shocks, such as job loss, accidents, death or illness of a household member, or economy-wide shocks, for example, effects of climate change on rainfall and temperatures, human and animal influenza pandemics, and impacts of the recent global financial crisis. The combination of large shocks and many small, often local shocks can be difficult to manage for poor, near-poor, and even nonpoor households. The strategies that households use to cope with unanticipated shocks, such as reducing spending on health care, selling off assets like land and livestock, and taking children out of school, can themselves have longer-term adverse consequences. At any point in time, apart from the households we observe living below the poverty line, there may be an additional number of households that face the risk of falling back into poverty—that is, households that remain vulnerable to poverty. 1.53 Some studies have equated vulnerability with the near-poor—households whose incomes lie above but still very close to the poverty line. As noted, Vietnam has defined near-poor poverty lines 23 that are 1.3 times the official poverty line. If a similar approach to defining the near poor is applied to the 2010 GSO-WB poverty line, there were 13 million near-poor households in 2010 in addition to 18 million poor households. The 2008-2010 VASS poverty report (VASS 2011a) used a different methodology to measure vulnerability-to-poverty. The report analyzed poverty dynamics using a panel data set from the 2002, 2004, and 2006 VHLSS and found that one-fourth of those who were poor in 2002 were chronically poor (poor in all three periods), while the remaining three-fourths experienced temporary bouts of poverty and thus were labeled the transient or stochastic poor. The work found a great deal of churning—households moving above and below the poverty line—over the period, including a number of households that escaped poverty. Ethnic minority households were much more likely to be among the chronic poor. Additional 1.54 Additional 1.54 is presented evidence is evidence presented below, below, using using a methodology a methodology initially developed initially developed and and applied in a applied Poverty Assessment for China (World Bank 2009), to assess vulnerability to poverty in a Poverty Assessment for China (World Bank 2009), to assess vulnerability to poverty based based on a panel of 1,800 on households a panel from of 1,800 the 2004, 2006, households from and 2008 VHLSS. the 2004, 2006, andIt constructs 2008 VHLSS. an index of vulnerability-to- It constructs an index of vulnerability-to-poverty, poverty, defined as the sharedefined as the share of the population of the who were population poor who in at least one were poor in year (2004, at least 2006, one or 2008) year (2004, 2006, or 2008) divided by the average poverty rate across all three years. divided by the average poverty rate across all three years. The results summarized in table 1.5 suggest that The results summarized a considerable table 1.5 innumber of suggest households in a that considerable Vietnam that arenumber not be of households poor in a specific in Vietnam that are year nonetheless not poor in a specific year nonetheless remain vulnerable to falling into poverty at some point in remain vulnerable to falling into poverty at some point in time. At the national level, only 7 percent of time. At the national level, only 7 percent of panel households were among the chronic poor (poor panel households were among the chronic poor (poor in all three years), despite an end-period (2008) in all three years), despite an end-period (2008) poverty rate of 13 percent. Vulnerability to poverty poverty was percent. rate of 13high particularly Vulnerability in wealthier areas to of poverty such as high was particularly the country in wealthier the Red areas (where of the country River Delta Hanoi such as the Red River Delta (where Hanoi is located) and the Southeast (where is located) and the Southeast (where Ho Chi Minh City is located). It was also surprisingly Ho Chi Minh City is high located). in provinces in also It was surprisingly the South high Central in provinces Coast in the River and Mekong CentralConsistent South Delta. Coast and Mekong with VASS River Delta. findings, upland Consistent with with regions VASS a high proportion findings, upland of ethnic with regions minorities a highevidenced proportion higher rates of ethnic of chronic minorities poverty. evidenced higher rates of chronic (structural) poverty. Table 1.5 Vulnerability to Poverty Remains High in Vietnam Table 1.5 Vulnerability to Poverty Remains High in Vietnam Consumption poverty (percent) (GSOͲWB) Average VulnerabilityͲ Poor in all 3 Poor in 2 of Poor in 1 of Poor in at Not poor in Headcount, Headcount, Headcount, headcount, toͲpoverty years 3 years 3 years least 1 year any year 2004 2006 2008 2004Ͳ2008 ratio (9) = (4) = [(6)+(7)+(8)] Subgroup (1) (2) (3) (1)+(2)+(3) (5) (6) (7) (8) /3 (10) = (4)/(9) National 7.0 6.7 12.3 26.0 74.0 20.0 13.7 13.0 15.6 1.7 (27) (26) (47) (100) Red River Delta 2.1 5.0 8.5 15.7 84.3 10.9 7.5 6.5 8.3 1.9 (13) (32) (54) (100) East Northern Mountains 10.4 10.3 10.8 31.5 68.5 26.3 17.3 19.0 20.9 1.5 (33) (33) (34) (100) West Northern Mountains 40.5 15.8 16.2 72.5 27.5 59.5 51.4 58.4 56.5 1.3 (56) (22) (22) (100) North Central Coast 10.3 11.5 19.9 41.7 58.3 32.5 25.7 15.6 24.6 1.7 (25) (28) (48) (100) South Central Coast 9.8 8.2 10.0 28.0 72.0 24.0 15.7 16.0 18.6 1.5 (35) (29) (36) (100) Central Highlands 19.1 10.3 3.9 33.3 66.7 31.8 27.9 22.2 27.3 1.2 (57) (31) (12) (100) Southeast 3.1 1.6 6.3 11.0 89.0 8.2 6.2 4.5 6.3 1.8 (28) (14) (57) (100) Mekong River Delta 2.2 4.2 20.0 26.4 73.6 16.9 6.7 11.5 11.7 2.3 (8) (16) (76) (100) Rural 8.8 8.2 14.3 31.3 68.7 24.4 16.6 16.0 19.0 1.6 (28) (26) (46) (100) Urban 0.7 1.6 5.3 7.5 92.5 4.4 3.6 2.5 3.5 2.1 (10) (21) (70) (100) Ethnic minority 34.0 19.4 15.3 68.7 31.3 59.7 49.0 47.5 52.1 1.3 (50) (28) (22) (100) Ethnic majority 2.6 4.6 11.8 19.1 80.9 13.6 8.0 7.4 9.7 2.0 (14) (24) (62) (100) Source: VHLSS tabulations using 2004, 2006, and 2008 panels of households. Source: VHLSS tabulations using 2004, 2006, and 2008 panels of households. 1.55 Vietnam’s rich body of qualitative research on poverty documents continuing concerns about vulnerability. The 1999 Participatory Poverty Assessment (PPA) identified a number of important sources of vulnerability such as crop failures (weather shocks, insects and other pests, landslides), human disasters (severe illness, death of a laborer, alcoholism, drug addiction), other economic shocks (job loss, 24 19 1.55 Vietnam’s rich body of qualitative research on poverty documents continuing concerns about vulnerability. The 1999 Participatory Poverty Assessment (PPA) identified a number of important sources of vulnerability such as crop failures (weather shocks, insects and other pests, landslides), human disasters (severe illness, death of a laborer, alcoholism, drug addiction), other economic shocks (job loss, death of animals, business failures), and material crisis (damage to homes, theft, and violence). (Vietnam-Sweden Mountain Rural Development Program, ActionAid, Save UK, Oxfam GB 1999) 1.56 Risks were also discussed by respondents in the 2003 and 2008 PPAs. The 2008 PPA (see VASS 2009) highlights the fragile balance between opportunities and risks; households must grasp new economic opportunities in order to move out of poverty, but there are risks inherent in grasping new opportunities, and households may be pushed back temporarily into poverty as a result of setbacks, temporary loss of assets, or changes in family circumstances. Many households raised concerns about rising debt and worries about being caught in a “debt spiral.â€? There is widespread evidence that health shocks have pushed some households back into poverty; affected households report selling assets and taking on extra debt in order to cope with health shocks. 1.57 Activities are underway to monitor the impacts of recent shocks on poverty. Oxfam GB and ActionAid8 carried out an annual program of poverty monitoring in 12 sites in Vietnam (nine in rural areas, three in urban areas) between 2007 and 2011, and VASS (with active participation from development partners) carried out several rounds of a Rapid Impact Monitoring (RIM) study beginning in late 2008. (Oxfam GB/ActionAid 2008-2011; VASS 2011b) Results highlight the importance of occasional and often severe individual risks (for example, health related) coupled with more common seasonal risks that are local-context specific (for example, bad weather) in affecting household living conditions. They also document the emerging impacts of “macroâ€? risks such as inflation and global economic crises. Even for the most affected groups, while macro risks worsened existing difficulties (for example, reduced purchasing power), they were found to rarely cause households to relapse into poverty. However, risk and vulnerability were noted as important causal factors in chronic poverty, and were linked to slow poverty reduction among ethnic minority households. Evidence from the RIM and related studies suggests that the 2009 global crisis had a negative but short-lived impact on the living standards of poor households, with particularly adverse effects on Vietnam’s large pool of migrants workers—many of whom work in factories with foreign links (via export production or foreign employers)—and rural households whose livelihoods depend on migrant remittances. 1.58 Three new qualitative field studies were carried out for this report highlight new and old sources of poverty and vulnerability (short summaries are provided in Annex 1.1). Low-income respondents in a study designed to explore “perceptions of inequalityâ€? raised concerns that inflation could widen the gap between the poor and better-off and thereby further reduce opportunities to access education, health care, and other services. Competition for jobs will increase if the economy continues to slow, and good jobs are likely to go to applicants with the right connections or who are willing to pay bribes to potential employers. Concerns about land acquisition have been widely discussed in the press, and were raised again in the perceptions of inequality study as well as a new study carried out jointly by the World Bank and Oxfam to identify the “long-run drivers of poverty reductionâ€? in Vietnam. Erstwhile rural households living in or near urban centers felt vulnerable to having their cultivable land acquired for industrial and other development purposes. Few felt they would be properly compensated for the loss of land, and most saw land acquisition as resulting in an inevitable drop in living standards. A third “positive devianceâ€? study of poverty among ethnic minorities analyzed a range of concerns specifically linked to poverty and progress among ethnic minorities. Minorities depend heavily on earnings from agriculture, both crops and animal products, and were particularly vulnerable to weather shocks and other natural disasters, also to commodity and input price volatility. Ethnic minority respondents were acutely aware of the substantial and persistent gap in living conditions between minority and Kinh households, which they attributed to a number of factors including e.g. gaps in opportunities and differences in treatment. 8. This monitoring was conducted for Oxfam GB and ActionAid by the Ageless Consulting Company. 25 Poverty is increasingly concentrated among Vietnam’s ethnic minority populations, who comprise less than 15 percent of the population but nearly half the remaining poor and two-thirds of the extreme poor. 1.59 Vietnam has 54 officially recognized ethnic groups, of whom the Kinh (Viet) are by the far the most numerous, accounting for nearly 74 million people (85.7 percent of the total population) according to the 2009 Population and Housing Census. In 2009, there were five other ethnic groups (the Tay, Thai, Muong, Khmer, and H’mong) with populations of more than 1 million, and another three (the Nung, Dao, and Hoa) whose populations are between 500,000 and 1 million. There are also a number of ethnic groups with populations of less than 5,000 people. With the exception of the Hoa (Chinese), Khmer, and the Cham, most ethnic minority groups live in highland or upland areas, away from the coastal plains and major cities. The largest minority populations are found in the North-West and North-East and the Central Highland regions, although there are also ethnic population clusters in the North-Central, South-Central, and Mekong regions. Figure 1.4 Kinh and Ethnic Minorities: Average Annual Rates of Real Growth in Per Capita Expenditures, 1998–2010 16 14 Kinh/Hoa Ethnic Minorities 12 10 8 6 4 2 0 Red river East West Northern South Central Southeast Mekong Rural Urtan National Delta Northern Northern Cental Cental Highlans River Delta Moutains Moutains Coast Coast Sources: 1998 VLSS and 2010 VHLSS. 1.60 Despite remarkable progress in reducing overall poverty, including a steady reduction in ethnic minority poverty, there remains a substantial and widening gap in living conditions and poverty rates between the Kinh majority and ethnic minorities. This is illustrated in figure 1.4, which graphs annualized real rates of growth in per capita expenditures (from the 1998 VLSS and 2010 VHLSS) between 1998 and 2010, by region and ethnicity. Since 1998, per capita expenditures have grown at an average annual rate of 9.4 percent for the Kinh and only 7.4 percent for ethnic minorities. Disparities are largest in some of the poorest and least accessible regions of Vietnam. As discussed in Chapter 6, in recent years growth in income has been uneven across minority households, with higher rates of growth among the better-off. Even the fastest-growing minority households are growing more slowly than the average Kinh households. 1.61 Consistent with differential rates of growth, the concentration of minorities among the poor is rising; in 1993, poverty was widespread and minorities comprised only 20 percent of all poor households (figure 1.5). By 1998, the share of minorities among the poor had increased to 29 percent, and by 2010, minorities accounted for 47 percent of the total poor in Vietnam and a 26 resounding 66 percent of individuals in the poorest 10 percent of the population. According to the updated GSO-WB poverty line, 66.3 percent of minorities were poor in 2010 compared to only 12.9 percent of the Kinh. Figure 1.5 Ethnic Minority Poverty Rates and Changing Composition of the Poor, 1993–2010 Composition of Poor by Minority/Majority Poverty Rate for Minority/Majority 100 90 80 80 70 60 60 50 40 40 30 20 21 10 0 0 1993 1998 2004 2006 2008 2010 1993 1998 2004 2006 2008 2010 Ethnic Minorities Kinh/Hoa Ethnic Minorities Kinh/Hoa Sources: 1993, 1998 VLSS; 2004, 2006, 2008, 2010 VHLSS. 1.62 The increasing concentration of minorities among the poor and extreme poor is a serious concern. But not all minorities are poor. There is encouraging evidence of improvements in welfare levels and livelihoods for many minority groups in recent years, and recent analysis of the 2010 VHLSS documents the presence of some better-off ethnic minority households among middle- and upper-income deciles. These issues are explored in greater depth in Chapter 5, which describes encouraging signs of progress in some areas and among some groups, and identifies important pathways for progress. In recent years, growth has favored the better-off, resulting in rising income inequality 1.63 Past work suggests that Vietnam’s development trajectory was one of growth without an appreciable rise in inequality (VASS 2010). The picture has evolved in recent years, however, and there is growing evidence of rising inequality. A new study of citizen perceptions of inequality carried out as background for this report (Annex 1.1) suggests a widespread sentiment that inequality has risen; the sentiment is shared widely across rural and urban populations, and by both rich and poor. 1.64 The annual rate of growth in real household incomes averaged 8 percent between 2004 and 2010, based on successive rounds of the VHLSS. However, growth since the mid-2000s has been uneven across households, with richer households experiencing stronger growth than poorer households. The variation in growth across households is a reflection of a number of powerful, and potentially opposing, changes in the economic fabric: changes in the returns to education and skills in labor markets, sectoral and occupational transitions, and geographic mobility as individuals leave rural areas in search of work. These forces interact with initial differences in human capital and access to services, as well as “proceduralâ€? and institutional inequalities, such as differences in voice and participation among social groups and access to power and influence, to generate differences in living conditions across the population. 1.65 Figure 1.6 presents a growth incidence curve9 using per capita income and shows growth rates by ranked income group between 2004 and 2010. Real income growth rates over the period varied considerably for households at different points in the income distribution, ranging from around 4 percent for households at the bottom of the income distribution to 9 percent for households at the top. Growth was pro-poor, in as much as it contributed to continued progress toward reducing poverty over the period. However, because growth has favored better-off households, both the relative and absolute gap in incomes between the rich and the poor has risen over time. 27 However, because growth has favored better-off households, both the relative and absolute gap in incomes between the rich and the poor has risen over time. Figure 1.6 Growth in Income Per Capita by Income Group, 2004-10 Figure 1.6 Growth in Income Per Capita by Income Group, 2004-10 45000 9 40000 8 VnDong(Jan2010prices) 35000 7 AnnualizedGrowth 30000 6 25000 5 20000 4 15000 3 10000 2 5000 1 0 0 1 2 3 4 5 6 7 8 9 10 RuralIncomeDecile 2004 2010 AnnualizedGrowth2004Ͳ2010 Source: Source: 2004, 2004, 20102010 VHLSS. VHLSS. 1.66 The uneven growth process has contributed to rising inequality and is contributing to concerns 1.66 about increasing The uneven social growth and economic process disparities. has contributed toThe Giniinequality rising index of income and isinequality has risen contributing to concerns modestly from 0.40 to 0.43, adjusted for variations in prices across regions. Inequality in Vietnam about increasing social and economic disparities. The Gini index of income inequality has risen modestly in 2010 was comparable to that in other middle-income countries in the region, such as Indonesia andto from 0.40 although for 0.43, adjusted Thailand, variations it was lower than ininprices China.across regions. This growth has Inequality been accompaniedin Vietnam in 2010 by a shift in was comparable to that the share in other of income from middle-income countries the bottom 60 percent the region, inpopulation of the such to the 40 Indonesia top as percent. The and Thailand, share of income accruing to the top decile increased by 2 percentage points between although it was lower than in China. This growth has been accompanied by a shift in the share of income 2004 and 2010. To place this figure in context—the increase in the share of income going to the top 10 percent was from the bottom 60 percent of the population to the top 40 percent. The share of income accruing to the almost as large as the total share of income going to the bottom 10 percent in Vietnam in 2010. top decile increased Meanwhile, overbythe 2 percentage same period, points between the share 2004 of income and 2010. accruing to theTo place10 bottom this figure percent in context—the decreased increasebyin20the share Focusing percent. of income top to going on the tailthe 10 percent topincome of the was almost distribution, the shareasoflarge income the as of thetotal top 5share of income going to the bottom 10 percent in Vietnam in 2010. Meanwhile, over the same period, the share of percent rose from 20.6 percent to 22.5 percent between 2004 and 2010. In this respect, the patterns income are similarto accruing to the those in China bottom 10 and India, percent where theby decreased top205 percent percent. income earners ofFocusing on theearned top tail20.5 and income of the 21.3 percent of income and consumption, respectively (ADB 2012). 1.67 The trend of rising inequality with economic growth is common across many developing 9 A growth incidence countries curve in the Eastplots Asiathe andannualized rate of Pacific region. growth While between rising two income points in inequality time may befor a specific percentiles manifestation income of the of growthdistribution processes (Ravallion that raise1997). overall income and reduce poverty, and can thus be considered a natural consequence of an economic landscape favoring entrepreneurship, innovation, and economic progress, if left unchecked some types of inequalities can lead to rising social tensions and to undermining social cohesion. The “perceptions 23 of inequalityâ€? study documents “acceptableâ€? and “unacceptableâ€? sources of inequality; wealth is acceptable (and admired) if achieved through hard work, luck, or acquiring more and better education. But wealth obtained through illegal means or misuse of power or influence is not acceptable. As Vietnam continues to grow and basic needs poverty is no longer the primary concern, it will be increasingly important to monitor and promote equitable growth processes that ensure all Vietnamese share in benefits of rapid development.  growth incidence curve plots the annual rate of growth between two points in time for specific percentiles of the income 9 A distribution (Ravallion 1997). 28 acceptable. As Vietnam continues to grow and basic needs poverty is no longer the primary concern, it will be increasingly important to monitor and promote equitable growth processes that ensure all Vietnamese share in benefits of rapid development. Disparities in other aspects of human development remain and in some cases appear to be widening Disparities in other aspects of human development remain and in some cases 1.68 appear to Vietnam be has widening not only succeeded in raising incomes. Progress in human development has been equally impressive. 1.68 Vietnam as in Buthas the not case only of income succeeded raising and ingrowth poverty incomes. reduction, Progress progress in human has been uneven. development has Inequalities beenmay undermine equally growth impressive. Butprocesses as in the ifcase areincome they of by disparities driven growth in circumstances—such and poverty reduction, progress as ethnicity, been and gender, has unequal uneven. opportunities Inequalities for acquiring may undermine a good growth education—that processes if they areultimately prevent some driven by disparities in circumstances—such as ethnicity, gender, and unequal groups from benefiting equally in the gains from high growth and development. opportunities for acquiring a good education—that ultimately prevent some groups from benefiting equally in the gains from high growth 1.69 and development. Consider the example of education. Figure 1.7 depicts the ratio of enrolment rates for majority compared children 1.69 to enrolments Consider for the example ofseveral ethnic education. minority Figure groups. 1.7 depicts theA ratio ratio of of less than enrolment 1 indicates rates that for majority children minority children are participating compared in school to enrolments at a lower for several rate ethnic than the minority majority. groups. Although A ratio there of less than 1 has been indicates that minority children are participating in school considerable progress since 1998, ethnic minority populations continue to have lower enrolments there at a lower rate than the majority. Although than the has been considerable progress since 1998, ethnic minority majority, and these differences are substantial at the upper secondary level.populations continue to have lower enrolments than the majority, and these differences are substantial at the upper secondary level. Figure 1.7 Ratio of Ethnic Minority to Kinh Majoirty Enrolment Rates in Public Schools by Level Figure 1.7 Ratio of Ethnic Minority to Kinh Majoirty Enrolment Rates in Public Schools, of Education, 1998 and 2010 by Level of Education, 1998 and 2010 1 0.9 0.8 NetEnrolmentRate 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1998 2010 1998 2010 1998 2010 Primary LowerSecondary UpperSecondary TayͲThaiͲMuongͲNung KhmerͲCham OtherNorthernUplands CentralHighlands Source: Source: 1998 1998 2010 2010 VLSS, VLSS, VHLSS VHLSS. Incomes matter in terms of access to quality 24 health and education services 1.70 The growing emphasis on “socializationâ€? in the provision of health and education services in Vietnam—which stresses the sharing of social costs and responsibilities between individuals and the state and non-state sectors—means that incomes are beginning to matter more for determining access to basic services. Rising disparities in incomes will contribute to rising social disparities, including disparities in school enrolments (particularly for secondary and higher education) and access to health services. 1.71 A direct consequence of this is that the burden of out-of-pocket health and education expenditures is substantial, particularly for less-well-off households. Analysis based on the VHLSS shows that spending on education rose in real terms between 2004 and 2010 across all levels (figure 1.8), and out-of-pocket costs are higher as students move from primary to lower and upper secondary levels. Compared to the poor, better-off households spend substantially more on education in general and in particular on extra courses and after-school tutoring. Given these advantages, it is not surprising that students from wealthier households perform better in the classroom and on standardized tests, and are more likely to obtain higher degrees and training. 29 the poor, better-off b useholds spen hou nd substantially more on edducation in ge eneral and in particular on extra courses annd after-schoo G ol tutoring. Given these ad dvantages, it is not surprisin nts from weal ng that studen lthier householdds perform be assroom and on etter in the cla o standardize a more likely to obtain higher ed tests, and are h degrees annd training. FiguFigure ure 1.8 Out-of-pocket 1.8 Out-of f-pocket SpendSpending ding per dent, per Stud Student, by by Educa Education ation and Expenditure and Expenditure Quin Quintile, ntile, 2004 d 2010 and 2004 and 2010 12,000 10,000 8,000 6,000 4,000 2,000 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Up pperSecondaryͲ econdaryͲ UpperSe CollegeͲ 2004 4 geͲ 2010 Colleg 2004 010 20 Tuition ntributiontoSch Con hoolorFund ms,Textbooks,St Uniform tationery ExtraCourses Otherexxpenditures ource: So Source: 2004, 2004, 2 VHLSS. 2010 2010 VHLSS R Research Research 1.72 1.72 sugg gests that whil suggests le ill that health while is s more ill health is concen more ntrated among concentrated g the poor, among the theey are poor, less they lik are kely less than the b than the better-off likely to better-off use health toserv usevices (World health Bank B (World services 2012). Moreover, e distribution the M 2012). Moreover, Bank n of public the distribution of public spending spending health decide in the sector in the health sector decidedly edly favors favors the the better-off; e better-off; spending s spendingc on commune on commune hea althhealth centers, centers, utilized by the rural poor, is small compared to spending on government hospitals utilized by the utilized byy the rural po oor, is small compared to sp pending on go overnment ho ospitals utilizeed by the bettter- better-off. Concerns have been raised about the impoverishing effects of catastrophic health costs, including that the poor will forego care when faced with serious illnesses. Most of the poor haveg off. Conce erns have beeen raised abouut the impove erishing effectts of catastropphic health osts, co including that free the poor w forego will health caare when cards, which help face d with the to reduce seriou s illnesses. costs they pay M Most of the pobut for services, or have with free health h concomitant cards, concerns about the quality of care they receive. A number of studies highlight Vietnam’s high out-of-pocket (OPP) health payments; these persist despite improvements in the coverage of the National Health Insurance Scheme as a result of the 2008 Law 25 on Health Insurance. The new law provides fully subsidized health insurance premiums for the poor, and partially subsidized premiums for the near- poor. However health insurance has had a modest impact on reducing out-of-pocket health payments (Lieberman and Wagstaff 2008; Wagstaff 2007) including catastrophic health costs. Households with young children and elderly members have higher exposure to health risks and report higher rates of catastrophic health spending. (Hoang Van Minh et. al. 2012) Urban residents face significant challenges of rising costs and economic instability 1.73 Vietnam has weathered the global economic storm following the financial crisis of 2008–09 better than most countries. Growth hit a decade-low 5.3 percent in 2009, down from a decade-high 8.5 percent just two years before, but in 2010 it bounced back to 6.8 percent. It slipped again to 5.9 percent in 2011, but remained more than 1 point above the average for emerging and developing economies. Growth in 2012 was only 5.7 percent. 1.74 Behind this resilience, however, is a more complicated story of volatility and vulnerability, which plays out in Vietnam’s cities and towns. As export demand fell following the global financial crisis, so did demand for factory labor. Fortunately, the labor market bounced back quite quickly and strongly, in terms of number of working hours and wages in nominal terms. Urban residents were buffeted by inflationary shocks before and after the crisis. In 2008, the GSO reported a price increase of 23 percent overall as Vietnam felt the effects of the global food crisis—with food price inflation registering at 34 percent. Inflation moderated in 2010, but rose again in 2011, to around 18 percent nationally, in both urban and rural areas, with a steeper rise in the price of food and foodstuffs and electricity and fuel. 30 1.75 These events have brought considerable challenges for urban residents, which have been documented in a number of studies and rapid assessments including those by Oxfam/ActionAid, VASS, and the UNDP/GSO cited earlier. For example, 65 percent of households surveyed in the 2009 Urban Poverty Survey reported higher prices for food and essential items as a source of difficulties, making inflation by far the most common factor among job loss, business slowdowns, natural disasters, health shocks, and others (16 percent of households reported job loss or business slowdown as a source of difficulty). On a positive note, a price impact survey undertaken by Oxfam GB and ActionAid in May 2011 found that inflation has not caused families to go hungry or children not to attend school (which may be due to parents giving top priority to their children’s education). Still, serious issues remain. Those living off of savings or fixed incomes, which are not inflation- indexed, such as pensioners, beneficiaries of social protection, and those unable to work due to health issues, are vulnerable to the effects of inflation in obvious ways. 1.76 Combined with employment instability like that introduced by the global recession, inflation also poses especially acute issues for migrants who move to urban areas seeking better work. Migrants already tend to face higher prices for accommodation, electricity, and water than local residents and have difficulty accessing social services; they are therefore especially endangered by instability in their livelihoods. Migrants surveyed in Oxfam/ActionAid’s fourth round of participatory monitoring of urban poverty (Oxfam/ActionAid 2011) reported that wage increases have failed to keep pace with price increases; their average monthly expenditures net of savings and remittances increased 87 percent between 2008 and 2011, while monthly income increased only 66 percent. There have been signs of rising labor tension as a result of this dynamic, and a reduction in remittances to rural areas. Instability in urban livelihoods bears not just on urban poverty, but, via this remittance mechanism, on poverty in rural areas, as well. E. Overview of the report: Vietnam’s old and new poverty reduction challenges 1.77 This report takes the view that despite remarkable progress, the poverty reduction task in Vietnam is not complete. The report aims to do three things. 1.78 First, it proposes revisions to Vietnam’s poverty monitoring system in Chapter 2, including improvements to the VHLSS, more comprehensive welfare aggregates, and a new poverty line, with the aim of bringing these more in line with economic and social conditions in present-day Vietnam. Second, Chapter 3 uses the new methodology to revisit the stylized facts about deprivation and poverty in Vietnam, and develops an updated profile of poverty using data from the 2010 VHLSS and new qualitative field studies. Third, the report selectively analyzes some of the key challenges for poverty reduction in the next decade. Chapter 4 presents new poverty maps based on the 2009 Population and Housing Census and 2010 VHLSS and compares these to earlier poverty maps based on the 1999 census. Chapter 5 focuses on ethnic minority poverty, with the aim of identifying not only the current constraints faced by minority populations but also by documenting important signs of progress. Chapter 6 takes a new look at inequality of outcomes and opportunities, combining analytic work using the VHLSS with findings from the qualitative study of perceptions of inequality. 31 Chapter Annexes Annex 1.1: New qualitative research carried for the 2012 Vietnam Poverty Assessment (1) “Positive devianceâ€? study on ethnic minority poverty This field study, carried out from November 2011 – February 2012, aimed to identify ethnic minority communities that show unusually strong poverty reduction and income growth and identify factors contributing to these positive results. Positive deviance is a methodology that originates in Vietnam, from a 1990s nutrition program led by Save the Children; it has since been applied worldwide by NGOs and researchers (Marsh et al 2004, Ramalingam 2011). Successful families and communities are “positiveâ€? since they escape poverty despite facing the same challenges and obstacles as their neighbors, and “deviantsâ€? (or outliers) because they practice different behaviors from others. The researchers visited ethnic minority communities in Dak Lak province (Ea H’leo district), Tra Vinh province (Chau Thanh and Tra Cu districts) and Lao Cai province (Muong Khuong and Bac Ha districts), conducting semi-structured interviews with over 100 ethnic minority residents and local government officials. Sites were selected using a combination of quantitative analysis and a snowball sample based on expert recommendations and secondary literature. Data from census samples was analyzed to determine rates of poverty reduction (or increase) among ethnic minority respondents only in each province and district over the periods 1999-2006 and 2006-09. Census data was also processed to calculate the mean reported expenditures of ethnic minority respondents (as a proxy for income) by province and district and the percentage of the ethnic minority sample in the top 15 percent of expenditures that resides in each location. A series of qualitative hypotheses was then developed of possible factors leading to poverty reduction and income growth, outlining “provocative propositionsâ€? for qualitative data collection that were explored through interviews and observation in field sites. (2) Identifying Long Run Drivers of Poverty Reduction: The Q-square pilot Oxfam and the World Bank carried out a qualitative pilot study in August, 2011 to identify what have been key long run drivers of poverty reduction over the past two decades in Vietnam. The study was framed around the complementary concepts of opportunities and constraints in assessing income and welfare dynamics at the household and community levels. The longer run aim was to develop a panel data set of households and communities spanning 20 years, drawing on the initial set of communities and households surveyed in the 1992/93 and 1997/98 VLSS. Sites were selected from the 1997/98 VLSS list of districts/communes based on district-level poverty rates and the district-level population of ethnic minorities and Kinh/Hoa. Efforts were made to visit a range of locations, roughly representative of Vietnam’s different regions. In total, the team interviewed 220 households that had been initially surveyed in the VLSS panel, updated household rosters for these households, and held groups discussions with nearly 250 respondents at both village and commune levels. A series of qualitative exercises were carried out including (i) wealth ranking; (ii) time-line exercises are used to explore commune and village histories since 1992 and (iii) card-sorting exercises and mobility diagrams to list and rank opportunities and constraints in the communities over the two decades. Village officials are also asked to discuss their perceptions of how life had changed, what had happened to poverty levels since the early 1990s. Additional life-history interviews and diagrams are conducted with representatives from selected households, focusing on households who had done exceptionally well (and why) or done very poorly (and why). The team also interviewed important ’change agents’ such as local businesses, cooperatives, shops, and projects/programs. (3) Exploring Perceptions of Inequality in Vietnam A field study was carried out in March and April 2012 that aimed to collect and analyze information on perceptions of inequality held by diverse elements of Vietnamese society. The work explored three key areas: (i) perceptions of social and economic disparities within and between different reference 32 groups; (ii) the factors that determine these perceptions, and (iii) the extent to which disaparities have changed over time. Discussions were organized around a number of reference focus groups i.e. better off households, poor households, senior citizens, groups of students as well as working young people, and (in the case of urban areas) rural to urban migrants. Three sentinel groups of sites were selcted -- six locationsn in metropolitan cities, two locations in smaller cities, and seven locations in rural areas. Four overlapping aspects of inequality were higlighted by all groups – inequalities in economic outcomes (incomes, wealth), as well as inquailities in access to education, jobs, and land. Causes of inequality were seen as overlapping and complementary e.g. some rural respondants raised concerns about the poor quality of education in their areas, which contributed to poor skills and unequal access to good jobs. There was strong support for policy measures to ensure equality of opportunities. Many respondents, particularly young, educated people living in urban areas, were tolerant of inequalities in outcomes – for example, ownership of fancy cars, big houses, and the lastest technology – so long as these gains were earned through hard work and legitimate means. Many groups raised concerns about ill-gotten gains, bribery and misuse of power leading to rising inequalities. And there were widespread concerns about ’procedural’ inequalites – the gaps in how systems were supposed to work in principal but failures of systems to work properly in practice e.g. implementation of land compensation policies. 33 References Asian Development Bank. 2003. “Participatory Poverty and Governance Assessment: Central Coast and Highlands Regionâ€?, Hanoi. Asian Development Bank. 2012. Outlook 2012: Confronting Rising Inequality in Asia. Manila: Asian Development Bank. Center for Community Support and Development Studies (CECODES), The Front Review of the Central Committee for the Viet Nam Fatherland Frong (FR), Commission on People’s Petitions of the Standing Committee for the National Assembly of Viet Nam (CPP), and the United Nations Development Program (UNDP). 2012. The Viet Nam Governance and Public Administration Performance Index (PAPI): Measuring Citizen’s Experiences. Hanoi. Chen, Shaohua and Martin Ravallion. 2008. “New Global Poverty Estimates.â€? World Bank Research Digest 3 (1, Fall): 4. Government of Vietnam. 2011. Statistical Handbook. Hanoi. Haughton, J., Nguyen Thi Thanh Loan, and Nguyen Bui Linh. 2010. “Urban Poverty Assessment in Hanoi and HCMC.â€? Joint publication of the UNDP and Vietnam General Statistics Office, Hanoi. Hoang Van Minh, Nguyen Thi Kim Phuong, Priyanka Saksena, and Chris D. James. “Financial Burden of Household Out-of-Pocket Health Expenditure in Vietnam: Findings from the National Living Standards Survey 2002-2010.â€? Social Science and Medicine 30 (2012): 1-6. Hoang, Xuan Thanh, Nguyen Thu Phuong, Vu Van Ngoc, Do Thi Quyen, Nguyen Thi Hoa, Dang Thanh Hoa, and Nguyen Tam Giang. 2012. “Perceptions of Inequality in Vietnam: A Qualitative Study.â€? Background paper prepared for the 2012 Vietnam Poverty Assessment, Hanoi. Iyer, Lakshmi, and Quy-Toan Do. 2008. “Land Titling and Rural Transition in Vietnam.â€? Economic Development and Cultural Change 56 (3). Leiberman, Samuel and Adam Wagstaff. 2008. Health Financing and Delivery in Vietnam: the Short and Medium Term Policy Agenda. Hanoi: World Bank. Marsh, D., D. Schroeder, K. Dearden, J. Sternin, and M. Sternin. 2004. “The Power of Positive Deviance.â€? British Medical Journal 329 (7475): 1177–1179. Nguyen Tam Giang and Hoang Xuan Thanh. 2012. “Long-run Drivers of Poverty Reduction in Vietnam between 1992 and 2011.â€? Background paper prepared for the 2012 Poverty Assessment, Hanoi. Oxfam GB/ActionAid. 2011. “Participatory Monitoring of Urban Poverty in Vietnam: Fourth Round Synthesis Report 2011,â€? Hanoi. Oxfam GB/ActionAid. 2008. “Participatory Monitoring of Urban Poverty in Vietnam: Synthesis Report 2008,â€? Hanoi. Ramalingam, B. 2011. “A Q&A on Positive Deviance, Innovation and Complexity.â€? February 8. Accessed September 3, 2011. http://aidontheedge.info/2011/02/08/a-qa-on-positive-deviance- innovation-and-complexity/. Ravallion, Martin and Shaohua Chen. 1997. “What Can New Survey Data Tell Us about Recent Changes in Distribution and Poverty?â€? World Bank Economic Review, 11(2): 357-382. Ravallion, Martin and Shouhua Chen. 2007. “China’s Uneven Progress against Poverty.â€? Journal of Development Economics 82: 1-42 Ravallion, Martin, Shoahua Chen, and Prem Sangraula. 2008. “Dollar a Day Revisited.â€? World Bank Research Digest 2(4, Summer): 1-16. 34 Turner, Sarah (2011) “’Forever Hmong’: Ethnic Minority Livelihoods and Agrarian Transition in Upland Northern Vietnam.â€? The Professional Geographer. Vietnam Academy of Social Sciences. 2009. “Participatory Poverty Assessment: 2008 Synthesis Report,â€? Hanoi. Vietnam Academy of Social Sciences. 2011a. Poverty Reduction in Vietnam: Achievements and Challenges. Hanoi. Vietnam Academy of Social Sciences. 2011b. “Rapid Impact Assessment - Vietnam in 2011: Synthesis Report,â€? Hanoi. Vietnam-Sweden Mountain Rural Development Programme, ActionAid, Save the Children (UK), and Oxfam (GB). 1999. A Synthesis of Participatory Poverty Assessments from Four Sites in Vietnam: Lao Cai, Ha Tinh, Tra Vinh, and Ho Chi Minh City. Hanoi: World Bank. UNDP. 2001. Doi Moi Processes and Human Development: Vietnam Human Development Report 2001. Hanoi. UNDP. 2011. Social Services for Human Development: Vietnam Human Development Report 2011. Hanoi. Wagstaff, Adam. 2007. “Health Insurance for the Poor: Initial Impacts of Vietnam’s Health Care Fund for the Poor.â€? Policy Research Paper No. WEPS 4134. Washington DC: World Bank. World Bank. 1995. Vietnam: Poverty Assessment and Strategy. Report No. 13442-VN. Washington DC: World Bank. World Bank. 1999. Vietnam Development Report 2000: Attacking Poverty. Washington DC: World Bank. World Bank. 2003. Vietnam Development Report 2003: Poverty. Hanoi: World Bank. World Bank. 2006. Vietnam Development Report 2007: Vietnam Aiming High. Hanoi: World Bank. World Bank. 2009. From Poor Areas to Poor People: China’s Evolving Poverty Reduction Agenda – an Assessment of Inequality and Poverty. Washington, DC: World Bank. World Bank. 2012. Health Equity and Financial Protection Report: Vietnam. Washington DC: World Bank. 35 chapter 2 Updating Vietnam’s Poverty Monitoring System Vietnam’s poverty monitoring system was updated to reflect changing economic conditions since the first Vietnam Living Standards Survey was conducted in 1993. New, comprehensive consumption aggregates were created using data from the 2010 Vietnam Household Living Standards Survey (VHLSS). The GSO-WB poverty line was updated using these aggregates: the new line is 653,000 VND per person per month, yielding a national poverty rate of 20.7 percent. 36 A. Introduction 2.1 Vietnam has a robust system for monitoring changes in poverty, based on a long-running system of nationally representative, comparable Vietnam Household Living Standards Surveys (VHLSS); consistent estimates of household welfare; and a poverty line that has been kept constant in real purchasing power since the mid-1990s, when it was agreed between the General Statistics Office (GSO), the World Bank (WB), and other development partners.10 Consistency in methodology and comparability over time are two of the great strengths of Vietnam’s poverty monitoring system. However, by 2009, it had become clear that key aspects of Vietnam’s poverty monitoring system were outdated. The methods used to measure household welfare and construct the original GSO-WB poverty line were based on economic conditions and the consumption patterns of poor households in the early 1990s. Conditions have changed: Vietnam today is very different from Vietnam in the 1990s. The economy is more diversified and better integrated in the global economy. Connectivity and access to markets have improved, even for households living in more remote rural areas. In addition, the production structure of households has changed: households have access to a much wider array of consumer goods, and they purchase more food from the market rather than producing it at home. Incomes are more diversified, and there has been a rapid shift out of agriculture and into industry and services. These changes affect households across the income distribution. Especially important for determining a poverty line, the consumption patterns of poor households today are substantially different from those of the 1990s. 2.2 This chapter describes revisions and updates to Vietnam’s poverty monitoring system, including improvements to the 2010 VHLSS (and subsequent rounds), revisions to the definition of household welfare to make it a more comprehensive measure of well-being, new indexes to adjust for spatial cost-of-living differences, and an update to the original GSO-WB poverty line. The methodology to construct the new poverty line is consistent with the original GSO-WB methodology, but is based on new information from the 2010 VHLSS.11 The revisions described in this chapter result in a higher estimate of poverty for 2010 than the original GSO-WB poverty line would have yielded, and, particularly for rural areas and areas with high numbers of ethnic minority households, higher poverty estimates compared to official estimates. Reasons for these differences are also discussed. 2.3 The chapter also describes a new methodology for estimating “subjectiveâ€? poverty lines, drawing on experimental questions introduced in the 2010 VHLSS. Poverty estimates based on the subjective poverty line are very similar to those using the updated GSO-WB poverty line. 2.4 The 2010 VHLSS can only give reliable estimates of poverty at the national level, for urban and rural areas and by region. This is due to sample size and design of the sample of the VHLSS, which includes information on both expenditures and incomes. Chapter 3 describes a small-area estimation (poverty mapping) methodology that can be used to estimate poverty at lower levels of disaggregation—in Vietnam’s case, for provinces and districts—and presents new district- and provincial-level poverty maps based on the 2009 Population and Housing Census and 2010 VHLSS. B. Rethinking Poverty and Poverty Measurement in Vietnam 2.5 Poverty is defined as unacceptable deprivation in well-being. But well-being can encompass a multitude of dimensions, and there are many different views about what constitutes an acceptable (or unacceptable) standard of living. In many countries, setting (or revising) the poverty line involves active public debate and a careful balancing of political and scientific considerations. The enormous 10 The original GSO-WB poverty line was prepared as input to the 2000 “Poverty Assessment Attacking Poverty.â€?  similar methodology was used in 2005 by a team of local and international experts, led by the Ministry of Labour 11 A Invalids and Social Affairs (MOLISA), to update Vietnam’s official poverty lines for the 2006–2010 Socio-economic Development Plan and by MOLISA and GSO more recently to construct official poverty lines for the 2011–2015 Socio- economic Development Plan. 37 public response, in India and internationally, to the Indian Planning Commission’s announcement of new poverty estimates and revised urban and rural poverty lines provides a recent example of the challenges inherent in updating poverty lines, with some interesting parallels to current discussions in Vietnam. Many in India feel that the new official poverty lines are far too low (box 2.1). Box 2.1 Do India’s New Official Poverty Lines Measure Up? What are Lessons for Vietnam? The Indian Planning Commission released a new set of poverty estimates and new poverty lines in March 2012. Many observers believe the new poverty lines are much too low—29 rupees per person per day for rural households (just under US$1.25 2005 Purchasing Power Parity [PPP]) and 32 rupees per person per day for urban households (US$1.65 2005 PPP). The Planning Commission’s new estimates showed a 7-percentage-point drop in poverty, the largest drop since the official poverty rate was first calculated in 1962. The announcement caused a furor in the Indian and international press: Indian poverty lines have always been low by international standards, and the new lines were seen as a missed opportunity to rectify this. One important criticism is that the nutrition standards embedded even in India’s new lines continue to be based on the sparse diet that the poor consumed in the 1973–74 National Sample Survey (NSS). Like in Vietnam, consumption patterns in India have changed substantially since these standards were set. Another criticism is that India’s new poverty lines do not “constitute an adequate definition of poverty because they do not take into account malnutrition, sanitation, drinking water, housing and health needsâ€? (Gill 2012). Vietnam’s updated 2010 poverty lines take full account of housing, durables, nutrition, clean water and sanitation, and health needs. If India is using the same methodology it used in the past, why the big controversy now? Over time, the Indian poverty line has increasingly been used as a cut-off to determine eligibility for India’s social welfare schemes and targeted poverty reduction programs. People who fall below the poverty line are eligible for a range of social benefits; states receive funds for some poverty reduction programs (for example, the Public Distribution System, which distributes subsidized rice to poor households) according to the number of residents who fall below the official poverty line. So where the poverty line is set is not just a statistical artifact, but an important policy decision that determines the eligibility of millions of families for public support. The Government of India cannot afford a poverty cut-off that is too high, and—as the controversy continues—it appears that the people of India will not accept a poverty cut-off that is too low. In a recent article in the Hindustan Times, Abhijit Banerjee, Ford Foundation International Professor of Economics at MIT, suggested that the way out of the current muddle is to have “two different poverty lines: an ethical poverty line to describe the standard we should aspire to … and an administrative poverty line which tells us how to best target our limited resources. As [India] gets richer, perhaps the latter will be raised till it is effectively the same as the former. But right now we don’t want to hurt the poorest [by spreading resources too thinly] in the name of being more aggressive about povertyâ€? (Banerjee 2011). Sources: Banerjee 2011; Gill 2012. 2.6 Vietnam’s official poverty lines for the 2011–2015 Socio-economic Development Plan are more akin to Banerjee’s concept of an administrative poverty line: they are designed to help target limited public resources to those most in need, and should be judged by that standard. The updated GSO- WB poverty line better captures what Banerjee refers to as an ethical poverty line; it reflects what Vietnam should aspire to achieve. The good news is that compared to the situation in the 1990s, Vietnam’s administrative and monitoring poverty lines are not very far apart. Moreover, the official poverty lines help to target poverty reduction policies and programs to those most in need, and thus help Vietnam achieve its poverty reduction goals. 38 Capturing Multiple Dimensions of Poverty 2.7 Measuring poverty is a challenging and complicated task, because poverty itself is complex and has many dimensions. This chapter focuses primarily on conventional approaches, based on absolute poverty lines and consumption measures of welfare. While familiar to the public and policy makers in Vietnam, the standard methodology may not fully capture other important dimensions of well-being. For example, households living in large, prosperous cities like Hanoi or Ho Chi Minh City may have access to better-quality schools and health facilities than households in other regions. But students attending higher-quality schools do not necessarily face higher school fees; in fact, households living in areas with poor schools may have to pay more, for instance, on extra tutoring to compensate for quality differences. Poor households that live in areas with low-quality schools but cannot afford to pay more may be at an additional disadvantage not captured in standard poverty analysis. Similarly, two households that look the same in terms of schooling and skills endowments may not earn the same income if one of the households faces discrimination in hiring—due to ethnicity or gender—that limits future prospects. 2.8 A variety of economic and social factors—some subtle and difficult to capture in standard poverty analysis—must be examined to get a full picture of poverty. Conventional poverty measures provide an important starting point for analyzing other dimensions of poverty. For example, the profile of poverty presented in Chapter 3 looks explicitly at other dimensions of poverty, for example, deprivations in education and skills, poor health status, and deprivations in access to basic services such as clean water and sanitation. The aim of multitopic surveys of living conditions (like the VHLSS) is to facilitate the measurement and analysis of poverty in multiple dimensions. The Human Development Index (HDI) described in Chapter 1 is a composite measure of well-being, as is the new Child Poverty Index (used in Chapter 3) and the broader Multidimensional Poverty Index (MPI) proposed by several UN organizations. 2.9 Additional information on other dimensions of deprivation experienced by the poor can be identified by soliciting their perceptions and insights through discussions and open-ended interviews. A number of Participatory Poverty Assessments (PPAs) have been carried out over the years in Vietnam, including three new field studies carried out in preparation for this report (see Chapter 1). Findings from qualitative studies are included throughout the report. These studies let the poor themselves give voice and context to the story that emerges from more conventional statistical analyses—poor men and women in Vietnam highlight concerns about lack of skills and education, access to good jobs and stable employment, and access to land and job security. They also speak about poverty in terms of risks—linked to health shocks, aging, and disability; job loss and uncertain wages; and weather shocks that destroy crops and affect rural incomes. Many of the poor are highly indebted, and risk can undermine new economic initiatives. The importance of social identity is also evident; in rural areas, minority status was often equated with being poor. C. Updating Methods for Measuring Poverty 2.10 Two important decisions are required in order to measure poverty: (a) how to measure welfare—in income or expenditure terms, and (b) what poverty threshold or line to use. Both issues have been the subject of debate in Vietnam, among both local researchers and policy makers and in the international community (box 2.2). 2.11 The GSO-WB approach uses per capita expenditures from the VHLSS as a measure of household welfare. The poverty line is constructed using a standard Cost of Basic Needs (CBN) approach, based on the observed consumption behavior of the poor, as reported in the VHLSS. It includes an allowance for food and nonfood spending. The food allowance (or food poverty line) is based on a single reference food basket for poor households, scaled up or down as needed to meet caloric norms and priced using a vector of national food prices. An additional allowance is added for essential nonfood spending, for example, on fuel, housing, schooling, health care, and clothing based on nonfood spending of households whose food spending is equal to the food poverty line (World Bank 1999). 39 easured? Box 2.2 How is Poverty Me Box 2.2 How is Poverty Measured? The po r headcount ind overty rate (or dex) is definedd as the proport tion of the pop pulation in a sp w pecific period whose onsumption welfare (co per capita) The poverty rate (or falls b below headcount the pove index) erty line is defined as(figure e B2.2.1). the proportion of the population in a specific period whose welfare (consumption per capita) falls below the poverty line (figure B2.2.1). F Figure l Poverty Mea B2.2.1 Conventional asurement Methodology Figure B2.2.1 Conventional Poverty Measurement Methodology Choice of Welfare cator Indic Choice of Welfare Indicator Welfa typically are isWelfare measured isytypically in measuredn terms in termsof of per r capita per consumer capita consumer expenditu expenditures ures or capita or per per ca apita incomes. On a incomes. conceptual On a conceptual l level, income level, e is a measure income is a e of welfare op measure of welfare pportunity—th opportunity—the he level of wel level of ll-being a hous well-being sehold a can aff ford to household can afford to purchase at a particular point in time. Consumption can be thought of as a a particular purchase at a measurepoint in time. of welfare Consumption achievement—the level of be can though well-being ht ofaas that a mea household asure of welfar actually re achievemen achieves at a nt—the level of we point in time. However, ell-being that a household ac incomes are often more variable than ctually achieves at a point in time. Howeve expenditures: for example, er, incomes are farmers e often more va ariable produce more in years when the weather is good than in years with unseasonable temperatures, than expen nditures: for example, droughts, eand flooding. farmmers produce Households smoothm income more in years when variations bythe we saving eather in good isyears good andthan dis-in yearss with unseasonab saving in res, ble temperatur bad droughts, years. Annual a expenditures and flooding. Households sm mooth Htypically reflect variations v incomeconcept a longer-run by saving s of income—that in goodd years is, permanent income—rather than a shorter-run concept of annual income. It is therefore not and dis-sa in bad years. aving surprising Annual that income-based l expenditures s typically poverty statistics can beref flect a longer- very different -runconsumption-based from concept of income—th hat is, permanent income—rath statistics. her In thethan a shor United rter-run States, for concep example, pt of 30 annual percent in ncome. of It is the the income-poor erefore not sur own their own rprising home that inncome- based pove compared erty statistics canto only 15 c be very dif percent of the consumption-poor, fferent from co onsumption-ba and the food share for the income-poor ased statistics. In the United States, for exa is ample, only 24 percent compared to 32 percent of the consumption-poor. It is generally assumed that poor 30 percentt of the income households e-poor own are less the likely eir own to own home their own chome (at least to only compared 15 percent oin high-income t of the consum countries like themption-poor, United and a the food share States) e for the incom and, according me-poor is only to Engel’s law, y 24 percent co will spend a higher proportion ompared to 32 percent of the of expenditures e consumption- on food. -poor. It is gennerally assumed th hat poor house Defining eholds the are less Poverty Lines likely to ow wn their own home h (at least in high-incom me countries lik ke the United Staates) and, accor rding to Engel’ ’s law, will spe end a higher proportion of exp The most commonly used approach to setting poverty lines is the Cost of Basic Needs approach, penditures on food. f which is widely applied in countries throughout the world and described in Ravallion (1994, 1998) Defining the t Poverty Li ine and Bidani (1994). The Cost of Basic Needs approach consists of first defining a and Ravallion basket of food and nonfood items that are adequate for satisfying basic consumption needs of m The most commonly a household, y used approa and then ach to setting calculating the costpoverty lines isConceptually, of this basket. the Cost of fa Basic Cost ofNeeds approach, wh Basic Needs hich is widely app poverty line plied in countri measures the minimum ies throughout the world and income necessary for d described in Ravallion Rhouseholds to (199 purchase a basic R needs 94, 1998) and Ravallion B and Bidani basket of food and other commodities, so that members have sufficient food to remain healthy he Cost (1994). Th andof Bas sic Needs productive appr and have roach consists the means of first fully to participate defini ining a basket society. off food In practical and the terms, nonnfood povertyitems th hat are f satisfying adequate for basic consump line is constructed by firstption needs defining of o a household a reference d, and food basket, then ca reflecting alculating the consumption cost of of this basket. patterns b Conceptua the poor; ally, a Cost of and anchoring it in an agreed nutrition norm f Basic Needs poverty line measures the minimum inc (for example, 2100 calories come necessary per person y for househo olds to per day), and then adding an allowance for nonfood spending on essential goods (health care, purchase a basic needs basket education, of housing, food and d and other durables) co that is ommodities, consistent withsspending o that membe ers of patterns have suffic the poor. cient food to remain r nd productive and healthy an m a have the means to parti icipate fully in p n society. In practical terms, the poverty line is d by first defin constructed e food basket, reflecting cons ning a reference sumption patterns of the poor ng it in r; and anchorin n an agreed nutrition normm (for example, 2100 calories per person per r day), and then lowance for no n adding an all onfood o essential goo spending on re, education, housing, ods (health car h and durables) d s consistent wi that is atterns ith spending pa of the poor r. 40 2.12 Vietnam carried out two Living Standards Surveys in the 1990s—the 1992–93 VLSS and the 1997–98 VLSS—with extensive technical support from international partners. Vietnam then carried out a series of government-financed Vietnam Household Living Standards Surveys (VHLSS) (in 2002, 2004, 2006, and 2008) using a similar approach to the earlier VLSS. The design of the core expenditure and income modules of the VHLSS questionnaires were kept broadly consistent with similar modules of the VLSS modules, with the specific and laudable aim of maintaining comparability over time. As noted, comparability has been one of the great strengths of Vietnam’s poverty data. 2.13 But by 2010, strict comparability was coming at too high a cost. The 2010 VHLSS and related welfare aggregates represent a break with the 2002–2008 VHLSS series in three important respects: (a) the 2010 VHLSS was based on a new master sample based on the 2009 Housing and Population Census, including a new set of communes and enumeration areas; (b) the VHLSS household questionnaire was substantially revised (including revisions to the core consumption module) and reduced in length; and (c) an updated methodology was used to construct a more comprehensive consumption (welfare) aggregate. These improvements are summarized here and described in greater detail in Kozel, Hinsdale, and Nguyen (2013). The VHLSS was Improved and Shortened in 2010 2.14 Sampling. The 2002–08 rounds of the VHLSS used a master sample of communes/urban wards drawn from the 1999 Housing and Population Census. In each round of the VHLSS, half of the enumeration areas (villages) and households within the communes were kept and half replaced, with the aim of ensuring stability in poverty measurement. While good for measurement stability, the 2002–08 master sample was substantially outdated by the end of the period. For example, between 2002 and 2008, there was substantial residential development in erstwhile empty areas (for example, “New Cityâ€? on the outskirts of Hanoi), and residential growth in provincial cities and towns, but these new developments were not included in the master sample used for 2002–08 rounds of the VHLSS. 2.15 A new master sample of communes and wards was developed for the 2010 and subsequent VHLSSs based on the 15 percent sample of the 2009 Housing and Population Census. Analysis suggests that the new sample provides better coverage of smaller households in urban areas, and somewhat better coverage of migrant households, many of whom come to work in urban areas for extended periods. Previous rounds of the VHLSS have been criticized for poor coverage of urban migrants, who are often assumed to belong to rural sending households (Pincus and Sender 2008). A recent study of poverty in Hanoi and Ho Chi Minh City (Haughton et al. 2010) indicates that many unregistered short-term urban migrants—who are likely to be undersampled in the VHLSS—may be vulnerable and have lower living standards than longer-term residents. These issues will be explored more systematically in the future; the 2012 VHLSS includes a special module on migrants, focusing in particular on long- and short-term migration for work purposes. 2.16 The sample of households for the 2012 VHLSS will be drawn from the same communes as the 2010 VHLSS, similar to the design of the 2002–08 sample. For 2014 and subsequent years, GSO is strongly advised to (a) update the master sample through careful relisting of enumeration areas on a regular basis, and (b) add new communes to the VHLSS master sample over time, with particular attention to good coverage in peri-urban areas where new population growth is occurring. GSO is also encouraged to explore alternative approaches to improve coverage of urban migrants, through either a more comprehensive sampling methodology or in-depth surveys of migrant populations. 2.17 Questionnaire Design. The VHLSS has been criticized by some researchers for taking too long to administer in the field, with related concerns about data quality and accuracy. In response to these criticisms, many sections of the 2010 questionnaire were shortened. The consumption modules were redesigned to collect information on food and frequent nonfood spending using a fixed reference period (30 days) rather than a “typical monthâ€? (used in 2002–2008), and a decision was made to administer the VHLSS in four rounds during each survey year.12 Questions designed  he decision to move to a fixed reference period was triggered by difficulties in measuring expenditures and prices during 12 T bouts of high inflation (for example, 2008), and an effort to better capture seasonality in consumption patterns. 41 to collect information on labor earnings also used a fixed reference period (prior month) rather than being based on “typicalâ€? work activities. Additional questions were added to capture Vietnam’s expanding array of social insurance and social assistance programs, and were better measures of remittances and transfers. Improvements were also made to the module on access to poverty programs, including targeting and coverage of benefits from targeted poverty reduction programs such as the National Target Program for Sustainable Poverty Reduction. New, more Comprehensive Consumption Aggregates were constructed 2.18 The first step in estimating a poverty line is to construct a welfare aggregate. The consumption aggregates constructed from the VHLSS follow standard practices well established in the literature (Deaton 1997; Deaton and Zaidi 2002). The consumption aggregates includes (a) food consumption, (b) frequent and infrequent nonfood items (personal care and hygiene, clothing, fuel, household goods), (c) education (tuition, books and uniforms, tutoring, and other fees), (d) health (curative and preventive care, health insurance), and (e) utilities (water, electricity, sanitation and trash collection). Two standard imputations are made in constructing the consumption aggregates, (a) the annual flow of services from durables, and (b) the annual value of housing services/imputed rents. 2.19 The poverty line is defined on the basis of the welfare aggregate. Any changes in the definition of the welfare aggregate will thus require revisions to the poverty line. Different countries use different welfare aggregates for measuring poverty; some countries use income, others use household expenditures. Within countries using household expenditures, there are substantial differences in expenditure aggregates. For example, although many countries include health or education expenditures in the expenditure aggregate, an increasing number of low-income countries in Sub- Saharan Africa do not. If basic health services and primary education services are provided free of charge, they are not captured in household expenditures, however defined, unless imputations are made to value the flow of publically provided services. Instead of trying to value these—which is complicated and controversial—additional analysis can be carried out to measure deprivations in human development, as a complement to income- or expenditure-based measures of deprivation. Many countries, particularly as they become more affluent, include the (imputed) value of durables, housing services, and local amenities in the expenditure aggregate. While broad concepts may be similar—welfare is measured through a household-level expenditure aggregate—the great diversity in actual practice makes it difficult to compare national poverty lines and poverty rates across countries, even when converted into “internationallyâ€? comparable 2005 Purchasing Power Parity (PPP) measures. One reason India’s national poverty line is low in PPP terms is because it is based on a very parsimonious welfare aggregate (box 2.1). 2.20 Two different sets of consumption aggregates have been used for poverty analysis in Vietnam. One set of aggregates (referred to as “temporally comparableâ€?) was designed, as the name suggests, to be strictly comparable with the consumption aggregates initially developed using the 1992–93 VLSS. For example, although new durable goods were added to later rounds of the VHLSS (for example, cell phones, computers), only items available in the 1992–93 VLSS are included in the comparable aggregate. Similarly, estimates of the value of housing services are also based on spending patterns in the 1992–93 VLSS. Because Vietnam’s housing market was very underdeveloped in the 1990s, imputed rents were calculated as a fixed percentage of total nonfood consumption rather than derived using conventional hedonic methods. This same fixed percentage (from 1993) was used to calculate the housing component of the consumption aggregate in all subsequent rounds of the VHLSS through 2008. 2.21 The vast majority of research and analytic work using VHLSS data has used the comparable consumption aggregate. The original GSO-WB poverty line, used extensively in the poverty literature for Vietnam, was constructed using the comparable aggregate, and is based on a reference food basket from the 1992–93 VLSS and related spending on a minimum basket of nonfood items, also based on spending patterns of the poor as reported in the 1992–93 VLSS. 42 2.22 Vietnam today is different from Vietnam in the 1990s, and expenditures, including expenditures of low-income households, are far more diversified. Real estate markets are more developed, particularly in urban areas, and many households put considerable investment into housing and land. Vietnam is similar to other fast-growing economies in this respect. Housing values reported in recent rounds of the VHLSS are more reliable than those collected in earlier rounds. 2.23 A second set of “comprehensiveâ€? consumption aggregates was created for the 2004, 2006, 2008, and 2010 rounds of the VHLSS, which aimed to make optimal use of all the expenditure information in a given year, unencumbered by considerations of strict comparability over time. There are a number of minor and major differences between comparable and comprehensive aggregates (see Annex 2.1 for a detailed description). The comprehensive aggregate includes the imputed value for all durables owned by the household and an imputed flow of services from housing. The latter is a particularly important addition (box 2.3). Box 2.3 How to Value Housing Services in the VHLSS Housing is an important component of household welfare, particularly as countries grow and prosper. Investments in housing are rising rapidly in Vietnam—families purchase new houses, and build or add onto existing dwelling units. Housing expenditures—either actual or imputed— should be fully reflected in the consumption aggregate. In countries where housing markets function well, annual rental payments provide a good measure of the value of housing services. Using information on reported rents, a hedonic for housing can be used to impute the value of housing services (based on characteristics of the dwelling unit and neighborhood characteristics) in cases where information on rents is missing (for example, owner-occupied housing, housing supplied by employers). However, Vietnam is an unusual case; rental markets are still thin and there are not enough renters either in early or more recent rounds of the VHLSS to estimate robust hedonic rent equations. Even the 2010 VHLSS includes only 243 households (out of 9,399) who report spending on rents—around 2.6 percent of total households in the sample. In contrast, the 2009 Housing and Population Census reports that 6.4 percent of all households in Vietnam rent their dwelling unit, including 13.2 percent of households living in urban areas. Prior to 2010, the value of housing services was assumed to be a fixed percentage of nonfood consumption expenditures. Based on shares in 1992–93, the value of housing was set equal to 11.8 percent of nonfood consumption for rural households and 21.4 percent for urban households. In constructing comprehensive aggregates, each household’s annual consumption of housing services is calculated as a fixed share of the reported sales value of the dwelling unit. This fixed share is the same for all households and equals 2.88 percent, which is the median ratio of reported annual rent payments to reported dwelling sales value, among the subsample of households who report renting their dwelling. In essence, this method uses the information collected in the 2010 VHLSS about Vietnam’s rental market to approximate the relationship that prevails in Vietnam between rental and ownership values in housing, and then imputes annual consumption of housing services for all households using this relationship. While this method would not be preferable to hedonic estimation given a more comprehensive survey of Vietnam’s renters, it has the virtue of not assuming that a household’s consumption of housing remains a constant proportion of other nonfood consumption over time, an assumption made in the temporally comparable set of aggregates from 1993 to 2008. Derived directly from the reported value of each household’s dwelling, the measure of housing consumption in the comprehensive aggregates is more sensitive to what each household reports about its living situation. The result is that, in 2010, housing averaged 15 percent of total consumption in the comprehensive aggregates compared to 6 percent in the temporally comparable aggregates (table 2.1). Note, however, that the share of housing is much lower for households in the poorest quintile (7.5 percent) and thus does not have a large impact on 2010 poverty rates. Source: Kozel, Hinsdale, and Nguyen 2013. 43 2.24 Tables 2.1 and 2.2 present comparable and comprehensive consumption aggregates for the last four rounds of the VHLSS.12 By 2010, it was clear that the benefits of maintaining procedural consistency with 1993 consumption aggregates was substantially outweighed by the resulting loss of information; there is a large and growing gap between the temporally comparable and comprehensive aggregates over time. Going forward, it is recommended that the methodology for estimating consumption aggregates and poverty lines be updated on a more frequent basis. How frequently depend on Vietnam’s rate of economic progress and how quickly consumption patterns are changing, will depend on Vietnam’s rate of economic progress and how quickly consumption patterns are particularly changes at the lower end of the income distribution, where there is a trade-off between changing, particularly changes at the lower end of the income distribution, where there is a trade-off depend stability between on and Vietnam’s consistency stability rate of time over and consistency economic and over progress relevance time andof relevance andthehow quickly methodology consumption to contemporary of the methodology patterns are living to contemporary changing, conditions. living particularly Given how quickly conditions. changes Given how at the conditions lower quickly are end conditions of the changing globallyincome are changing distribution, and inglobally Vietnam, where it is and in there Vietnam,that suggested is a trade-off it isthe between methodology suggested that stability be the andin revisited methodology consistency fivebe (or over six) years revisited time and(orrelevance to assess in five whether six) years tothe of it is methodology providing assess accurate whether tois it contemporary estimates.accurate providing living Note, conditions. however, that estimates. Given how quickly despite efforts to ensure procedural consistency, comparisons between the 2010 VHLSS and earlier years Note, however, that conditions despite are effortschanging to ensure globally and procedural in Vietnam, consistency, it is suggested comparisons that the between methodology the 2010 VHLSS be revisited using andin either earlier five (or comparable years orusing six) years either to assess comprehensive comparable whether it or consumption comprehensive is providing accurate aggregates must consumption estimates. Note, be interpreted aggregates however, with must care. that As be interpreted despite described efforts above, with to care. ensure a number As of described procedural consistency, important above, a changes number comparisons were of important introduced in thethe between changes 20102010 VHLSS VHLSS,were introduced suchandasearlier in the years an updated 2010 VHLSS, using sample either frame, comparable asuch shift astoan aorupdated sample comprehensive fixed reference frame, period inathe consumptionshift to a fixed reference aggregates expenditure must beand module, period in the interpreted a revised expenditure with care. As definition of module, and a revised definition of welfare, which make comparisons difficult. The 2010 VHLSS and described welfare, above, which a number make of important comparisons changes difficult. Thewere 2010 introduced VHLSS in thethe and 2010newVHLSS, GSO-WB such poverty as an updated lines the new GSO-WB poverty lines provide a baseline for consistent poverty monitoring going forward, sample frame, provide a baselinea shift to a fixed poverty consistent forand reference period ingoing the expenditure module, and a revised definition of that is, for the 2012 of the VHLSS. forward, that is, for the 2012 and future rounds future rounds monitoring welfare, of which make comparisons difficult. The 2010 VHLSS and the new GSO-WB poverty lines the VHLSS. provide a Table 2.1 baseline for consistent Comprehensive poverty monitoring Consumption going forward, that is, for2004, the 2012 and2008, future rounds Table 2.1 Comprehensive Aggregates forfor Consumption Aggregates thethe VHLSS VHLSS, 2004, 2006,2006, 2008, 2010 2010 of the VHLSS. Mean consumption Average share of total consumption Expenditure component Table 2.1 Comprehensive 2004 Consumption 2006 Aggregates2010 2008 2004 2004,2006 for the VHLSS, 2006, 2008, 2010 2008 2010 Food expenditure 1,753 2,378 2,993 6,515 42 42 38 46 Mean consumption Average share of total consumption Non Ͳfood expenditure Expenditure component 1,050 1,449 2,142 3,220 21 21 22 20 2004 2006 2008 2010 2004 2006 2008 2010 Durables consumption 592 767 1,301 1,972 10 10 12 10 Food expenditure 1,753 2,378 2,993 6,515 42 42 38 46 Education expenditure 261 334 461 769 5 5 5 4 NonͲfood expenditure 1,050 1,449 2,142 3,220 21 21 22 20 Health expenditure 297 339 494 722 6 5 5 4 Durables consumption 592 767 1,301 1,972 10 10 12 10 Utilities and electricity 140 183 233 373 3 3 2 2 Education expenditure 261 334 461 769 5 5 5 4 Housing consumption 1,120 1,390 2,070 3,558 15 15 16 15 Health expenditure 297 339 494 722 6 5 5 4 Total expenditure 5,212 6,840 9,694 17,129 100 100 100 100 Utilities and electricity 140 183 233 373 3 3 2 2 Source: 2004, Housing 2004, Source: 2006, 2006, consumption 2008, 2010 2008, VHLSS. 2010 VHLSS. 1,390 1,120 2,070 3,558 15 15 16 15 Total expenditure 5,212 6,840 9,694 17,129 100 100 100 100 Source: 2004, 2006, 2008, 2010 VHLSS. Table 2.2 Table 2.2 Temporally Temporally Comparable Comparable Consumption Consumption Aggregates Aggregates for VHLSS, 2004, 2006, 2008, 2010 for VHLSS 2004, 2006, 2008, 2010 Source: 2004, Mean consumption 2006, 2008, 2010 VHLSS. Average share of total consumption Expenditure component Table 2.2 Temporally Comparable 2004 2006 2008 Aggregates Consumption 2010 for VHLSS, 2004 2004, 20062006, 2008, 2008 2010 2010 Food expenditure 1,857 2,502 3,153 6,401 49 49 47 54 Mean consumption Average share of total consumption Non Ͳfood expenditure Expenditure component 986 1,396 1,987 2,975 20 21 23 21 2004 2006 2008 2010 2004 2006 2008 2010 Durables consumption 518 638 801 1,268 10 9 9 7 Food expenditure 1,857 2,502 3,153 6,401 49 49 47 54 Education expenditure 246 330 423 732 5 5 5 5 NonͲfood expenditure 986 1,396 1,987 2,975 20 21 23 21 Health expenditure 290 332 465 680 6 5 6 5 Durables consumption 518 638 801 1,268 10 9 9 7 Utilities and electricity 147 191 233 378 3 3 3 3 Education expenditure 246 330 423 732 5 5 5 5 Housing consumption 351 466 622 988 6 6 7 6 Health expenditure 290 332 465 680 6 5 6 5 Total expenditure 4,394 5,855 7,683 13,422 100 100 100 100 Utilities and electricity 147 191 233 378 3 3 3 3 Source: Housing2004, 2006, 2008, 2010 VHLSS. consumption 351 466 622 988 6 6 7 6 Total expenditure 4,394 5,855 7,683 13,422 100 100 100 100 Source: 2004, 2.25 2006, Figure 2008, 2.1 2010 VHLSS. shows overall composition the overall composition of per capita of per expenditures in capita expenditures 2.25 Figure 2.1 shows the the 2010 in the 2010 VHLSS. VHLSS. Spending on food now constitutes less than half of per capita expenditures compared Spending on food now constitutes less than half of per capita expenditures compared to to percent 57 57 in percent 2.25 1998, and in 1998, Figure durables and 2.1 and shows durables andthe housing overall make housing up composition makenearly of a quarter up nearly per of capita aggregate a quarter expenditures of aggregate in the 2010 VHLSS. welfare. welfare. Spending on food now constitutes less than half of per capita expenditures compared to 57 percent in 1998, and durables and housing make up nearly a quarter of aggregate welfare.  hese aggregates are in real terms; they have been adjusted to January terms of the survey year and for regional cost- 13 T of-living differences. 44 44 44 Figure 2.1 Composition of Per Capita Expenditures, 2010 VHLSS Housing 14% Food 46% Other non-food 20% Utilities and electricity 2% Education Durables 4% 10% Health 4% 2.26 Figure 2.2 shows the composition of expenditures, categorized by food, nonfood, durables, housing, and others—broken down by per capita expenditure quintile. Note that the food share falls from 58 percent (in the poorest quintile) to only 32 percent for the wealthiest quintile. In contrast, the poorest individuals spend only 7 percent of their total expenditures on housing and another 7 percent on durables compared to a housing share of 27 percent and a durable share of 12 percent for the wealthiest group of individuals. These gradients are consistent with those of other countries at similar levels of development. Figure 2.2 Composition of Per Capita Expenditures by Per Capita Expenditure Quintile, 2010 VHLSS 100% 7.5 9.9 12.4 15.8 90% 27.1 19.8 80% 20.7 20.3 19.9 70% Housing 17.9 60% 6.6 Other non-food 8.4 50% 9.8 Utilities and electriccity 10.9 40% Education 12.2 Health 30% 58.3 50.3 45.7 41.5 Durables 20% 31.9 Food 10% 0% 1 2 3 4 5 45 Consumption is adjusted for Household Size to Estimate Individual Welfare 2.27 Our objective is to calculate a measure of individual welfare and estimate the number of people who live below the poverty line. But in households, individuals live together, eat together, and often pool their resources. Household surveys like the VHLSS measure expenditures at the household rather than individual level. Different approaches have been used to move from household-level expenditures to individual welfare. One approach is to use equivalence scales and to also adjust for household-level economies of scale. In the absence of a well-defined equivalence scale for Vietnam, and building on past practices, household expenditure is converted into per capita terms by simply dividing by household size. The implications of using alternative measures, adjusting for adult equivalencies and household economies of scale, on the poverty profile are discussed briefly in Chapter 3. Consumption is also Adjusted for Temporal and Spatial Cost Variations 2.28 One of the advantages of the CBN methodology is that it anchors the poverty line at a fixed level of well-being, and consequently allows for consistent poverty comparisons. However, households living in different regions of the country may face different prices for similar consumer goods due to differences in transport, storage, and marketing costs. For example, consumers pay more per kilogram to purchase rice in a market in Ho Chi Minh City than they pay to purchase the same quality of rice in a rural district in the Mekong Delta, where the rice is grown. In contrast, laundry soap may cost more in rural areas than in cities, where it is produced and packaged. Prices also change over time due to inflation and other factors. 2.29 Some countries (for example, Indonesia and Mozambique) account for inflation and spatial cost-of-living differences by constructing poverty lines for different regions, based on region-specific prices and (sometimes) region-specific consumption baskets. In keeping with past practice, a single national GSO-WB poverty line was constructed using information from the 2010 VHLSS. The new GSO-WB poverty line is applied to spatially and temporally adjusted (that is, real) per capita expenditures to calculate poverty rates. 2.30 Temporal adjustments are straightforward; the consumption aggregates described in table 2.1 have been deflated to January of each survey year (for example, 2004, 2006, 2008, 2010) using the GSO’s official Consumer Price Index (CPI) deflators for rice, other foods, and nonfoods. Previous to 2010, spatial adjustments were made using regional CPI deflators provided by the GSO. For 2010, new spatial cost-of-living indexes (SCOLIs) were estimated and are used instead of regional CPI deflators to calculate poverty rates. 2.31 There are three reasons why prices collected for the CPI are not well-suited to measuring spatial differences in the cost of living. First, CPI prices are collected on a frequent basis in outlets where a wide range of consumer goods are available and shopping volumes are high. These are typically located in urban and peri-urban areas. But many of the rural population (including poor households) shop in local markets near where they live. Second, the specification of items whose prices are collected for the CPI is not the same across provinces. Vietnam’s CPI price collection system maintains temporal consistency (prices for the same items are collected over time in each location) but not spatial consistency (the items in the basket may be slightly different in each location). For example, prices of higher-end cotton shirts may be surveyed in large urban areas, while prices for lower-cost polyester shirts are surveyed in smaller towns or rural areas. Regional variations in the specification of items may reflect quality differences rather than only capturing price differences for an identical good. Third, a CPI and SCOLI have different objectives, and the differences make it difficult for the two indexes to rely on the same set of price data. The CPI aims to give equal weight to every Vietnamese dong spent; it is used as a deflator to ensure the real value of currency remains unchanged. Consequently the expenditure patterns of wealthier households have more weight in a CPI because they spend more money, and the CPI price collection system targets outlets with a high volume of purchases. 46 2.32 In contrast, a SCOLI is population-weighted rather VND-weighted; the SCOLI is estimated using the prices paid by the average individual from each area, and prices are aggregated into a population-weighted index that treats everyone equally. In short, compared to the CPI, a SCOLI requires different budget shares for aggregating items into an index, a different set of outlets for price collection, and different weights to aggregate information on individuals to form regional averages. 2.33 Regional adjustments were based on regional CPI indexes in earlier rounds of the VHLSS. However, for 2010, adjustments were made for regional cost-of-living differences using market price data from a SCOLI fielded in conjunction with the second and third rounds of the 2010 VHLSS. The approach is described in Annex 2.2. 2.34 The 2010 SCOLI ranges between 0.7 and 1.0 (table 2.3). The Mekong Delta has the lowest overall cost of living and the Red River Delta (which is also the base region) has the highest cost of living. In all but two of the six regions, the SCOLI shows only a small difference in the cost of living between urban and rural sectors. The two exceptions are the Red River and South East regions, where the urban cost of living is approximately 20 percent higher than the rural cost of living, largely reflecting the higher estimated cost of accommodation services in the metropolitan areas of Hanoi and Ho Chi Minh City. Apart from these two exceptions, the variation in the cost of living is greater across regions than it is between the urban and rural sectors within a region. Table 2.3 Spatial Cost-of-Living Index (SCOLI) for each Region and Sector Region Urban Households Rural Households Red River 1.00 0.79 Midlands & Northern Mountains 0.81 0.79 Northern & Central Coast 0.78 0.71 Central Highlands 0.83 0.78 South East 0.97 0.77 Mekong Delta 0.74 0.70 Note: Calculations are based on a Törnqvist index applied to regional average prices that are pooled over the two rounds of SCOLI data collection, and using person-weighted average budget shares, with housing values based on the hypothetical values reported by all survey respondents. D. Constructing a new GSO-WB Poverty Line 2.35 The poverty line consists of two components, a food poverty line and an additional allocation to account for essential nonfood needs. The food poverty line is estimated in three steps. First, a reference food basket is defined that reflects the consumption patterns of the poor; second, quantities are adjusted to reach an agreed nutrition norm; and third, the cost of purchasing the adjusted reference basket is calculated. An allowance for essential nonfood needs is estimated using an Engel’s curve regression and is then added to the food poverty line in order to construct the total poverty line. Defining the Reference Food Basket 2.36 The reference food basket used to construct the original GSO-WB poverty line is anchored in the food consumption patterns of poor households14 in the 1993 VLSS. The reference food basket for the updated GSO-WB poverty line is anchored in food consumption patterns of poor households in the 2010 VHLSS.  he methodology is described in Annex 2 of the 2000 “Vietnam Development Report: Attacking Poverty.â€? (World Bank 14 T 1999). Food consumption of the 3rd quintile of households, ranked nationally based on per capita expenditures, was used to construct the reference food basket. 47 2.37 Defining the reference basket is an iterative process; we do not know in advance which households are poor (the method is described in Pradhan et al. 2001)15. Households were ranked according to SCOLI-adjusted and temporally adjusted per capita expenditures (henceforth referred to as “realâ€? per capita expenditures) from least well-off to most well-off, and the poor were initially defined as those in the bottom 2.5 percent to 20 percent of the real per capita expenditure distribution. This initial reference basket ultimately became the final reference basket; the 2010 poverty rate, based on an updated GSO-WB poverty line, was close to 20 percent. 2.38 Analyses were carried out to assess the stability of the poverty line food basket across different reference groups; food consumption patterns of the bottom 2.5 to 20 percent (bottom quintile) of individuals were compared with the bottom 2.5 to 10 percent (bottom decile). The initial 2.5 to 20 percent reference group was further divided to compare (a) food baskets for bottom-quintile ethnic minorities and bottom-quintile majorities, and (b) food baskets for bottom-quintile urban and bottom- quintile rural households (Annex table 2.1). 2.39 Food consumption patterns were similar when comparing the poorest 10 percent and the poorest 20 percent of the population. Similarly, the consumption patterns of poor minority households were on average quite similar to consumption patterns of poor majority households. Dietary patterns, however, were different for urban and rural households in the 2.5 to 20 percent reference group: urban poor households consumed less rice and higher-priced calories (meats, oils), and were more likely to consume food and drinks outside the home. Although the GSO-WB poverty line is based on a single national reference basket for poor households, Vietnam’s official poverty lines use different reference baskets for urban and rural households. A number of other countries, including, for example, Indonesia, Mozambique, Papua New Guinea, and Russia, define regional reference baskets that reflect local preferences and tastes. The problem with using different reference baskets, particularly for urban and rural areas, is that the different baskets often reflect diets of different quality, so the poverty line for urban areas (based on consumption patterns of urban households) may give a superior standard of living compared to the poverty line for rural areas (based on consumption patterns of rural households). In 2010, only a small fraction (9 percent) of the poor reference group actually lived in urban areas. Given this, coupled with concerns about avoiding quality differences (that is, setting a higher standard of living for urban households), a single national reference food basket was again used to construct the new GSO-WB poverty line. 2.40 In line with standard CBN practice, food quantities in the reference basket are scaled up to an “acceptableâ€? nutritional norm, holding constant the relative composition of the reference basket (that is, all quantities are scaled up by the same factor). But what constitutes an acceptable norm? International experience shows that countries anchor their poverty lines in very different caloric norms, ranging from a low of 1,800 Kcals for India (GOI 2009) to more than 2,700 Kcals for some countries in Africa. 2.41 The original GSO-WB poverty line was anchored in a caloric norm of 2,100 Kcals per person per day. However, the composition of the Vietnamese population has changed since the early 1990s, when the 2,100 Kcals norm was set. The share of young children in the population (who consume less food) has decreased and the share of adults (who consume more) has increased. A new caloric norm of 2,230 Kcals per person per day was estimated using age- and gender-specific caloric requirements for the Vietnamese population developed by the Nutrition Institute in the Ministry of Health (MOH 2006), and weighted by the relevant age-gender composition of the national population in the 2010 VHLSS. These new norms compare well with international practice (figure 2.3).  e restrict the group to the bottom 2.5 percent to 20 percent to avoid potential problems with outliers and measurement 15 W error. 48 compare well with international practice (figure 2.3). Figure 2.3 Nutrition Norms Used to Anchor Poverty Lines in Different Countries Figure 2.3 Nutrition Norms Used to Anchor Poverty Lines in Different Countries 3500 3000 2500 2000 1500 1000 500 0 Uganda Honduras Indonesia Senegal Egypt India Ecudador ElSalvador Nicaragua Bangladesh Mexico SierraLeone Jordan Chile SriLanka Paraguay Vietnam Malawi Mozambique Guatemala Panama Iraq Cameroon Colombia SouthSudan TimorLeste Sources: UN Statistics Division 2005; World Bank staff estimates. Sources: UN Statistics Division 2005; World Bank staff estimates. 2.42 Table 2.4 compares the calorie and expenditure composition of the 1993 reference food basket 2.42 Table 2.4 to used compares calorie theoriginal estimate the and expenditure GSO-WB composition poverty line with the new food the 1993 of basket use to construct food reference the basket 2010 GSO-WB poverty line. The original food reference basket was heavily dominated by rice (79 used to estimate the original GSO-WB poverty line with the new food basket use to construct the 2010 percent of calories, 46 percent of food spending). The 2010 basket is more diversified; although rice GSO-WB poverty line. continues to The original be important in food reference the food basket consumption was of the poorheavily dominated (66 percent of calories,by30rice (79 of percent percent of calories, 46 percent of foodtheir food spending), spending). Thepatterns consumption 2010 basket is more have become diversified; more diversified although to include, rice continues to be for instance, pork and other meats and seafood, vegetables and fruits, more oils, and more calories from meals important in the food consumption of the poor (66 percent of calories, 30 percent of food spending), their eaten outside the household. Rice calories are very cheap; calories from pork, oils, and seafood consumption patterns have become are more expensive. more The cost diversified of the to include, 2010 reference for be basket will instance, higher thanporktheand other original 1993meats and seafood, vegetables referenceand fruits, basket. In more oils, addition, and there hasmore beencalories from a substantial meals in increase eaten outside the household. the non-quantified share of Rice consumption, that is, food reported under “otherâ€? categories and meals eaten outside the household. calories are very cheap; calories from pork, oils, and seafood are more expensive. The cost of the 2010 More than 95 percent of food consumption was recorded under quantified items in the 1998 VLSS reference basket will be compared higher to less than than 80 the original percent in the 20101993 VHLSS.reference basket. An extended list ofIn foodaddition, items was there has been a included substantial increase the non-quantified in VHLSS, in the 2012 with the aim of share of consumption, getting better (more quantified)that is, food measures reported of food under “otherâ€? consumption (table 2.4). categories and meals eaten outside the household. More than 95 percent of food consumption was recorded under quantified items in the 1998 VLSS compared to less than 80 percent in the 2010 VHLSS. 49 49 An extended list of food items was included in the 2012 VHLSS, with the aim of getting better (more quantified) measures of food consumption (table 2.4). Table 2.4 Composition of the Reference Food Basket, 1993 and 2010 VHLSS Table 2.4 Composition of the Reference Food Basket, 1993 and 2010 VHLSS 1993 2010 Average Average Average Average share of share of total share of share of total total food total food Food item calories expenditure calories expenditure Plain rice (including fragrant and specialty rice) 78.9 46.5 66.4 30.5 Sticky rice 2.7 2.3 4.2 2.5 Maize (in seed equivalent) 1.0 0.4 1.6 0.4 Cassava (in freshͲtype equivalent) 1.9 0.9 1.0 0.3 Potato of various kinds (in freshͲtype equivalent) 1.6 2.5 0.3 0.3 Wheat grains, bread, wheat powder 0.3 0.4 0.3 0.3 Flour noodle, instant rice noodle/porridge 0.3 0.7 1.3 1.6 Fresh rice noodle, dried rice noodle 0.4 0.5 Vermicelli 0.1 0.2 Pork (in equivalent of the pork type with removed fat) 2.4 9.3 4.0 11.1 Beef 0.1 0.8 Buffalo meat 0.0 0.5 0.0 0.2 Chicken meat 0.7 5.1 0.9 5.1 Duck and other poultry meat 0.1 0.7 0.2 1.0 Other types of meat 0.0 0.3 Processed meat 0.1 0.6 Lard, cooking oil 1.8 1.5 4.2 2.5 Fresh shrimp, fish 1.3 8.3 1.4 6.9 Dried and processed shrimps, fish 0.3 1.2 Other aquatic products and seafood (crabs, snails,...) 0.1 0.5 Eggs of chickens, ducks, Muscovy ducks, geese 0.0 0.3 0.7 1.7 Tofu 0.4 0.9 0.6 1.3 Peanuts, sesame 0.7 0.8 0.5 0.4 Beans of various kinds 0.4 0.6 0.3 0.3 Fresh peas of various kinds 0.1 0.4 Morning glory vegetables 0.6 2.2 0.5 1.1 Kohlrabi 0.3 1.0 0.1 0.2 Cabbage 0.2 1.0 0.1 0.4 Tomato 0.1 0.7 0.0 0.4 Other vegetables 0.7 3.3 Orange 0.0 0.2 0.0 0.2 Banana 0.7 1.2 0.6 0.6 Mango 0.0 0.3 0.0 0.2 Other fruits 0.4 1.5 Fish sauce 0.3 2.0 0.2 1.1 Salt 0.0 0.5 0.0 0.3 MSG 0.0 0.8 0.0 0.3 Glutamate 0.0 1.3 Sugar, molasses 1.3 1.3 1.3 1.2 Confectionery 0.6 1.0 Condensed milk, milk powder 0.0 0.1 0.2 0.7 Ice cream, yoghurt 0.0 0.2 Fresh milk 0.1 0.5 Alcohol of various kinds 1.3 1.8 Beer of various kinds 0.8 0.9 0.1 0.3 Bottled, canned, boxed beverages 0.1 0.2 Instant coffee 0.0 0.2 Coffee powder 0.0 0.1 Instant tea powder 0.0 0.1 Other dried tea 1.0 6.3 0.4 1.1 Tobacco 0.0 2.3 Betel leaves, areca nuts, lime, betel pieces 0.0 0.1 Outdoors meals and drinks 3.3 5.9 Other food and drinks 1.0 2.6 50 50 Calculating the Food Poverty Line 2.43 The food portion of the CBN poverty lines is defined as the cost of purchasing the (scaled) reference food basket. There are three sources for food prices that could be used to estimate the food portion of the poverty line: (a) unit values (reported value of food consumption divided by reported quantities) calculated from the 2010 VHLSS survey, (b) food prices collected by the GSO Price Department for the CPI, and (c) food prices collected through the SCOLI survey. 2.44 The original GSO-WB food poverty line was based on CPI food prices provided by the Price Department. However, Vietnam’s new official poverty lines are calculated using unit values from the 2006 VHLSS and adjusted for inflation. Both the SCOLI and CPI prices cover only a subset of food items in the 2010 VHLSS. Unit values (real or imputed in the case of non-quantified consumption) are available for all food items in the VHLSS and, moreover, can be estimated specifically for low-income households, thus reflecting what the poor actually purchase (quality, brand) and what they pay. There are mixed views in the literature (Deaton 1988, 1997; Deaton and Tarozzi 2005) about whether unit values are adequately well specified to be used as prices. Even well-defined items in the household consumption module, such as rice, are available in a range of qualities, and prices vary between urban and rural areas and among regions. Limiting unit values to a group of poor households will help control for quality differences, which are usually linked to income levels (for example, wealthier households tend to purchase higher-quality/more expensive rice). 2.45 Consistent with the methodology used to estimate Vietnam’s official poverty lines, the new GSO-WB food poverty line is calculated using mean unit values for food purchases by poorer households (bottom 2.5 to 20 percent) reported in the 2010 VHLSS. National food poverty lines are estimated for each round of the 2010 VHLSS (June, October, December) using the national reference food basket and food prices (unit values) from each round, and adjusted for inflation and averaged to construct a national food poverty line in January 2010 VND. 2.46 The new GSO-WB food poverty line for 2010 is VND 343,000 per person per month (VND 4,116,000 per person per year). Calculating the Total Poverty Line, including Food and Essential Nonfood Spending 2.47 In addition to food, an allowance must be added for essential nonfood spending such as for fuel, housing, schooling, health care, clothing, and other daily needs. However, estimating the nonfood component of the poverty line is not as straightforward as estimating the food poverty line, because there is no easily defined “normâ€? for nonfood expenditures in the way that caloric norms can be used to define food needs. 2.48 The CBN approach looks to the actual expenditure patterns of the poor in the 2010 VHLSS with the aim of estimating (a) an “austereâ€? allowance for nonfood needs, based on the typical value of nonfood spending by households whose total expenditure just equals the cost of the food poverty line; and (b) “minimal but adequateâ€? allowance for nonfood needs, based on the typical value of nonfood spending by households whose food spending actually reaches the cost of the food poverty line, so that basic food needs are fully met. 2.49 An Engel curve looks at the relationship between the share of spending on food and total per capita expenditures. According to Engel’s law, the food share decreases as expenditures (welfare) The average food rise. The share food average for each sharegroup of households for each can be calculated group of households using an Engel can be calculated using curve regression an Engel curve (Ravallion regression The and The Bidani average average 1994) (Ravallion food food shareasfor and sharefollows: Bidani for 1994) each each group groupas offollows: of households households can can be be calculated calculated an an using using Engel Engel curve curve regression regression (Ravallion (Ravallion and and Bidani Bidani 1994) 1994) as as follows: ‫ݕ‬ follows: ݂ሺ‫ݕ‬ ௜ሻ ௜ ൌ ß™ ൅  ߚଵ Ž‘‰ ቀ ௙ á‰? ൅  ß› ᇱ ൫݀௧ െ  Ý€Ò§ ൯ ൅  ‫݈ܽݑ݀݅Ý?Ý?ݎ‬௜ ‫ݕ‬௜ ݂ሺ‫ݕ‬ ௜ሻ ௜ሻ ݂ሺ‫ݕ‬ ܾ ‫ݕ‬௜ ‫ݕ‬௜ Ò§ ൯ ൅ ௙ሺ௬೔ ሻ ൌൌ ß™ ß™ ൅ ൅ ßš ߚଵ ቀ ቀ ଵ Ž‘‰ Ž‘‰ á‰?௙൅á‰?à³”൅ ௬ ß› ᇱ ൫݀ ß› ᇱ൫݀ ௧െ ௧ െ Ý€Ò§ ൯ ݀൅  ‫݈ܽݑ݀݅Ý?Ý?ݎ‬ ‫݈ܽݑ݀݅Ý?Ý?ݎ‬ ௜ ௜ where is the food budget share, ß™ ܾ ௙ܾቀ ‫ݕ‬௜a national intercept, ‫ݕ‬௜is ೑ á‰? is total (nominal) expenditure divided ௬೔ ௕ ௙ሺ௬೔ሻ ೔ሻ ௙ሺ௬ ௬೔ ௬೔ where by the where where food povertyis is isthethe thefood line, food food and Ý€ budget budget bud is a share, get share, share, vector ß™ ofis αߙ isa a is a national national national demographics intercept, intercept, intercept, with mean ቀ ೑ ቀá‰?Ý€is ೑ . total Ò§ á‰? is total total (nominal) (nominal) expenditure expenditure (nominal) divided divided ௬೔ ௬೔ ௧ ௕ ௕ byby the divided the byfood the food poverty food poverty line, poverty line, and and line, and ݀௧ Ý€ ௧ a is isdat vector is a vector vector of of demographics demographics with of demographics with mean with mean mean Ý€Ò§ .Ý€Ò§ . d. 2.50 In keeping with international practice, we propose to use the upper-bound poverty line (that is, 2.50 with “minimal 2.50 2.50 In In keeping but In with adequateâ€? keeping keeping international with allowance with practice, international international we propose nonfood) for practice, practice, weas wethe to propose use new propose tothe GSO-WB to useupper-bound use the the povertypoverty upper-bound upper-bound line line,poverty which (that poverty is line is, thus line (that (that is, is, with defined as “minimal the but adequateâ€? food poverty line allowance divided for nonfood) by Engel’s as coefficientthe new GSO-WB estimated from poverty line, the regression which (.525) is 16 thus : with with “minimal “minimal butbut adequateâ€? adequateâ€? allowance allowance forfor nonfood) nonfood) as as the the newnew GSO-WB GSO-WB poverty poverty line, line, which which is is thusthus 16 16 defined defined as as thethe food food poverty poverty line line divided divided byby Engel’s coefficient estimated from the regression (.525) : Engel’s ܾ ௙ coefficient estimated from the regression (.525) : ß™ ‫כ‬51 ܾ ௙ܾ ௙ 2.51 The new poverty line assumes the nonfood spending ß™ ‫כ‬of a typical household at the point on the ß™ ‫כ‬ à³” by the food poverty line, and ݀௧ is a vector of demographics with mean Ý€Ò§ . 2.50 In keeping with international practice, we propose to use the upper-bound poverty line (that is, with “minimal but adequateâ€? allowance for nonfood) as the new GSO-WB poverty line, which is thus 16 defined as the defined food as the poverty food line poverty divided line byby divided Engel’s coefficient Engel’s coefficientestimated from the the regression estimatedfrom regression(.525) (.525)16:: ܾ௙ ß™ ‫כ‬ 2.51 The poverty The new new poverty line assumes line assumes the nonfood the nonfood spendingspending of a household of a typical typical household at the at the point point on the on the Engel curve where actual food expenditure is equal to the food poverty line. Engel curve where actual food expenditure is equal to the food poverty line. The new 2.51 The 2.52 GSO-WB poverty new GSO-WB line is poverty line is therefore therefore defined defined as: as: 653,000 VNDVND person/month per per 653,000 ,, person/month which whichequals equalsVND VND343,000 343,000(food poverty (food line) poverty /.525. line) /.525. E. E. New ew N Poverty Poverty Estimates Estimates for 2010: for 2010: GSO-WBGSO-WB and Official PovertyPoverty and Official Methodologies Methodologies 2.53 New poverty estimates based on the new GSO-WB poverty lines and consumption aggregates 2.52 New poverty estimates based on the new GSO-WB poverty lines and consumption aggregates described in this chapter are presented in table 2.5. For purposes of comparison, the table also presents described in this chapter are presented in table 2.5. For purposes of comparison, the table also Vietnam’s presents official household-level Vietnam’s povertyfor poverty estimates official household-level 2010,17 for estimates based 2010,on 17official poverty lines of VND based on official poverty 400,000 person/month lines of VND 400,000 (rural) and VND(rural) person/month 500,000 VND 500,000(urban). andperson/month The GSO-WB person/month poverty (urban). The rates are GSO-WB higher overall—20.7 poverty percent rates are higher compared topercent overall—20.7 14.2 percent—which compared to 14.2 is not surprising because percent—which is notthe GSO-WB surprising because poverty line the (VNDGSO-WB 653,000poverty line (VNDis person/month) 653,000 person/month) higher than is poverty the official higher than the lines. official poverty Comparing the two lines. Comparing the two estimates for 2010, official estimates suggest higher rates of poverty in estimates for 2010, official estimates suggest higher rates of poverty in the North Central and South the North Central and South Central coastal regions compared to GSO-WB estimates, and slightly Central lowercoastal regions rates in compared the Central to GSO-WB Highlands estimates, and Southeast and slightly region. lowerin Differences rates in the poverty Central Highlands estimates for the Southeast primarily andSoutheast region. Differences in poverty reflect the fact that theestimates for the Southeast SCOLI measured a higher primarily reflect cost of living the in the fact that the Southeast compared SCOLI measuredto the higher costregional a CPI-based deflator. of living in the Overall, Southeast GSO-WB to thecompared estimates suggestregional the CPI-based lower poverty deflator. rates in urban areas than official estimates. Overall, the GSO-WB estimates suggest lower poverty rates in urban areas than official estimates. Table 2.5 Poverty Estimates for 2010: Comparing the GSO-WB Methodology Table 2.5 Poverty Estimates for 2010: Comparing the and Official GSO-WB Methodology and Official Methodology Methodology GSOͲWBPovertyRate OfficialPovertyRate Contributionto Contribution to Incidence(%) total(%) Incidence (%) total(%) AllVietnam(national) 20.7 100 14.2 100 16 Urban ଵ 6.0 9 6.9 14 Where ß™ ‫ כ‬is defined as ß™ ‫ כ‬ൌ ß™ ൅  ߚଵ Ž‘‰ ቀ ‫כ‬á‰?. 17 Rural à°ˆ 27.0 91 17.4 86 Official estimates reflect the number of households on the poverty list and not the number of individuals on the poverty list. To the extent that poor households are larger on average than nonpoor households, official estimates RedRiverDelta(Hanoi) 11.4 12 8.4 13 of the share of individuals below the poverty line would be higher than the share of households. EastNorthernMountains 37.7 21 24.2 20 WestNorthernMountains 60.152 9 39.4 9 NorthCentralCoast 28.4 16 24.0 20 SouthCentralCoast 18.1 7 16.9 10 CentralHighlands 32.8 10 22.2 9 Southeast(HCMC) 8.6 7 3.4 4 Mekong Delta 18.7 17 12.6 17 54 Although 16 Where the methodologies α* is are αbroadly defined as α*= α+ β1 log(1/ *). similar (both use a CBN approach based on spending ehavior of the 17  O poor in thereflect fficial estimates VHLSS), the number the new GSO-WB of households on the povertypoverty line list and not the is higher number than of individuals official poverty lines for the on the list. To the extent that poor households are larger on average than nonpoor households, official estimates of the share of llowing reasons: individuals below the poverty line would be higher than the share of households.  ach round of the VHLSS includes around 46,000 households. Detailed information on household income is collected for 18 E all households, but consumption information is collected for only 20 percent of households (three in each enumeration x Official lines were finalized in late 2010, before the 2010 VHLSS data were available and are area), or 9,400 households in total. Only unit record data from the 20 percent sample (income + consumption) are thus based on a food reference basket and consumption behavior of poor households in the 2006 released to the public. VHLSS. As noted, the 2010 VHLSS is different52 from the 2006 VHLSS in a number of important respects, including sampling and design of the questionnaire. 2.53 Although the methodologies are broadly similar (both use a CBN approach based on spending behavior of the poor in the VHLSS), the new GSO-WB poverty line is higher than official lines for the following reasons: â—?â—? Official lines were finalized in late 2010, before the 2010 VHLSS data were available and are thus based on a food reference basket and consumption behavior of poor households in the 2006 VHLSS. As noted, the 2010 VHLSS is different from the 2006 VHLSS in a number of important respects, including sampling and design of the questionnaire. â—?â—? Official poverty lines were estimated using the temporally comparable consumption aggregates rather than comprehensive consumption aggregates. As demonstrated in table 2.1, the comprehensive aggregate is higher due especially to the inclusion of more types of durable goods and, most importantly, a better measure of the value of housing services. But using the new measure of housing services does not in itself lead to a higher poverty rate. We tested a modified comprehensive consumption aggregate that included a value of housing calculated using the original GSO-WB method, and then calculated new poverty lines and poverty rates. The “old housing methodâ€? poverty rate was 21.3 percent, slightly higher than the “new housing methodâ€? poverty rate. â—?â—? Although food poverty lines are similar in the official and GSO-WB approaches, a decision was made to use a lower allocation for essential nonfood spending for the official poverty lines than indicated in the VHLSS data (see discussion in Chapter 1). 2.54 There are other important differences between the two methodologies that might result in different poverty rates in the aggregate and across regions. For example: â—?â—? Official poverty rates for 2010 were calculated on the basis of per capita incomes in the full VHLSS,19 with some adjustments at provincial levels following discussions with MOLISA. As described in box 2.2, income-based poverty estimates are typically different (and yield a different poverty profile) than consumption-based estimates. â—?â—? Income-based poverty rates were adjusted for spatial cost-of-living differences using a CPI-based regional deflator rather than the SCOLI. Consumption-based poverty rates were re-estimated using CPI-based spatial cost-of-living adjustments instead of the SCOLI. The impact was small and worked to raise the poverty rate (to 21.5 percent) rather than lower it. 2.55 Neither set of lines is inherently better than the other. As noted in Chapter 1, they are designed to serve different purposes. The strength of the GSO-WB approach lies in consistent poverty monitoring and its independence from budgetary or political considerations. In contrast, Vietnam’s official poverty lines are primarily intended to help set targets and related resource allocations for targeted poverty reduction programs and policies under Vietnam’s 2011–2015 Socio-Economic Development Plan. In this sense, they are administrative lines, necessarily constrained by resource availability. In response to a new directive on social protection (Resolution 15), MOLISA is developing new measures of average and minimum living standards, which will be used to identify potential beneficiaries of social assistance and social insurance policies and programs. 2.56 Official lines were used in carrying out the 2010 Poverty Census in Vietnam. Local surveys were used to identify poor and near-poor households (using short forms, proxy-means-test scorecards, and short income questionnaires), combined with village-level discussions to determine which households had incomes below the official poverty lines and were eligible to be on the poor list (Prime Minister’s Directive No. 1752/CT-TTg). These lists are being updated annually, again using a mix of survey methods and village-level discussions, often applied differently across the 10,000 or so communes in Vietnam. Analysis suggests that many of those included on the lists are poor, but not all poor households are included on the list (Chapter 3). In short, errors of exclusion are a greater concern than errors of inclusion.  ach round of the VHLSS includes around 46,000 households. Detailed information on household income is collected for 19 E all households, but consumption information is collected for only 20 percent of households (three in each enumeration area), or 9,400 households in total. Only unit record data from the 20 percent sample (income + consumption) are released to the public. 53 F. Are the New GSO-WB Poverty Lines too High? Are They Consistent with Citizens’ Subjective Views? 2.57 An alternative methodology for estimating subjective poverty lines that has received growing attention in the literature (Kapteyn 1994; Ravallion 2012; Ravallion and Lokshin 2002) was also applied in Vietnam based on additional questions added to the 2010 VHLSS to elicit households’ own assessment of whether their consumption of important items, such as foods, foodstuffs, electricity, water, clothing, and housing, was sufficient to meet their needs. (See Annex 2.3 for technical details, and Marra 2012.) For example, the following question was intended to assess adequacy of food (for example, rice, basic food grains, and staples) and foodstuffs (for example, meats, vegetables, condiments): 2.58 The intuition behind subjective poverty lines is straightforward: households whose observed 2.59 2.59 incomes The The are intuition intuition above the behind behind subjective subjective subjective poverty poverty poverty line (that lines lines isisstraightforward: is, marked straightforward: households in red in figurehouseholds 2.4, panel A) whose whose feel they observed observed have incomes incomes are are above above the the subjective subjective poverty poverty line line (that (that is,is, marked marked in in red red in enough or more than enough income to meet their needs, while households with observed incomes in figure figure 2.4, 2.4, panel panel A) A) feelfeel they they have have belowor enough enough or themore more thanenough than subjective enough income income line consider tomeet to their meet their incomes theirneeds, needs, while inadequate whileto meet theirwith households households observed with needs. observed incomes incomes The approach below usedbelow subjective line the subjective here is slightlyline consider theirincomes consider their different and is incomesinadequate based on perceptions inadequate toto of meet themeet their adequacy theirneeds. of needs. The specific The approach items, approach used for here example, used is is here foodstuffs. slightly slightly differentIn the different and andcase is of foodstuffs, isbased based onperceptions on panel perceptions ofof shows Bthe the that, of adequacy adequacy inof2010, specific poorer specific items, households items,for example, for example,(deciles 1 foodstuffs. foodstuffs. Inand the 2) the casewere case of much lesspanel of foodstuffs, foodstuffs, likelyB panel than Bshows better-off shows that, that, in households in2010, 2010, poorer poorerto say their consumption households households (deciles (deciles1 and of 1 and2)foodstuffs were 2) weremuch was muchless less sufficient. likely than better-off households to say their consumption of foodstuffs was sufficient. likely than better-off households to say their consumption of foodstuffs was sufficient. Figure 2.4 Measuring Subjective Poverty Figure Figure2.4 2.4Measuring MeasuringSubjective SubjectivePoverty Poverty Panel A, Stylized Panel A, Case Stylized Case Panel A, Stylized Case Panel Panel B BB Panel Based Based Basedon on the the on 2010 2010 the 2010 VHLSS VHLSS VHLSS 2.59 Overall, responses to these questions suggest that less than 5 percent of the households in 2.60 Overall, the 2010 VHLSS responses felt they tohadthese questions consumed suggest that insufficient less than food 5 percent in the of the households in the 2.60 Overall, responses to these questions suggest amounts that less ofthan 5 percent 30ofdays the preceding households the in the 2010 VHLSS survey. Acutefelt they had hunger is no longer a insufficient consumed major issueamounts for of food Vietnam. in the 30 However, days 11.5 preceding percent the survey. of households 2010 Acute VHLSS felt they had consumed insufficient amounts of food the in percent30 days preceding the survey. hunger is no longer a major indicated insufficient consumption of foodstuffs, and the percentage was significantly higher indicated issue for Vietnam. However, 11.5 of households in rural Acute hunger insufficient is no consumption longer a major foodstuffs, ofpercent issue for and theto Vietnam. percentage However, 11.5 percent of households inindicated than in urban areas—14 compared 5 percentwas significantly (figure higher in rural 2.5). A surprisingly high than urban percentage insufficient areas—14 consumption percent of households (25 compared of foodstuffs, percent in torural 5 percent areas) and the percentage (figure reported 2.5). A were they was significantly surprisingly not able high higher percentage to consume ofrural in households sufficient urban than in (25 electricity areas—14 in the in percent percent daysareas) 30rural compared before reported to the survey. 5 percent they were This (figure not almost able 2.5). to certainlyA surprisingly consume sufficient reflects high percentage electricity supply-side in the problems of 30households with days the before(25 quality percent and the in availability survey. rural areas) This of reported electricity almost certainlythey in 2010 were rather reflects not able to consume than concerns supply-side about problems sufficient electricity affordability; with the quality2010 in and the was 30 days a drought availability before of the survey. electricity in 2010 rather than concerns about affordability; 2010 was a drought year in many parts of of year in many This partsalmost of certainly Vietnam, and reflects supply-side load-shedding and problems brownouts with were the quality widespread. and availability electricity Vietnam, 2010 rather than in load-shedding and and concerns brownouts about were affordability; widespread. 2010 was a drought year in many parts of Vietnam, and load-shedding and brownouts were widespread. 54 e 2.5 Perceived Figure o Consumptio d Sufficiency of a Rural, 2010 on by Urban and Figure 2.5 Perceived Sufficiency of Consumption by Urban and Rural, 2010 100% 90% 80% 70% ble Not applicab 60% More than sufficient 50% Sufficient 40% 30% Insufficient 20% 10% 0% ral Urban Rura Rur l Urban Rural Urban Rural U al Urban Rural U Urban Rural Urban Food odstuff Foo ctricity Elec Wa ater Housing ng Clothin Source: 2010 VHLSS. ource: 2010 VHL So LSS. 2.60 Perceptions of sufficiency also differed across regions. Households in poorer regions (for Perceptions 2.61 example, of sufficiency Northern Mountains,y Central also differe ed across Highlands) were re egions. more Hous likely eholds to report in po oorer insufficient regions levels of s (for example, Mountains, M Northern Concerns consumption. Ce about entral Highla insufficient ands) were electricity more m likely were particularly hight inreport to in ufficient regionsinsu leve the north of els of Vietnam. consumpttion. Concernns about insuufficient elect tricity were particularly p h high in region ns in the norrth of Vietnam. 2.61 The responses to these questions can be used to calculate a subjective poverty line, following an approach proposed in Pradhan and Ravallion (2000). The perceived sufficiency of consumption Tregressed 2.62 isThe responses to these questions can be e used to calcu ulate a subjec ctive poverty line, followi against characteristics of the household such as total consumption, size, gender ing an approach proposedage, composition, in Pradhan n and an educationnd of members. (2000). Ravallion Thee perceived Different regression models sufficiency were used to of consumpti o test for the ion is regressed against sensitivity ofchar racteristics results. Based onoff regression the househo old such results, as to consump otalpoverty subjective ption, lines size, were ender as ge calculated compos the sition, minimum age, and e educationtotal ofexpenditure members. m needed Dif by a fferent household regress sufficient to meet were sion models w used(foodstuff) t for to test consumption needs. the sensitivity esults. of re (Annex 2.3 provides a more detailed description of the derivation of subjective poverty lines.) Based on regression re esults, subjec c ctive poverty lines were calculated as the minimum m total expenditure by needed2.62 household y a Subjective to meet poverty su linesufficient (food for 2010 rangeddstuff) highumption from aconsu needs of VND 888,000 s. per (Annex person 2.3 month to a more 3 provides per a low of VND d detailed description 616,000 per person of the derivation per month n of subjectiv depending ve poverty linon the nes.) exact specification of the regression model. All estimates of subjective poverty lines were higher than Vietnam’s official poverty lines, and Su 2.63 nearlyubjective pov all were verty lines higher for than the 2010GSO-WB r new ranged d from a high poverty h (VND line of VND 888 per per 8,000 653,000 pers person son per mont per month). th to a Most low of VN lines were ND 616,000 per clustered p person pein the range of er month depeVND 700,000 to ending on the VND 800,000. fication of the e exact specifi m e regression model. All estima atesEstimates 2.63 ctive of subjec poverty poverty of subjective lines were lines igher hi suggestthan thatVi etnam’s the offic updated cial poverty GSO-WB li povertyines, and lines near and rly all were high her than related the new poverty n GSO-WB estimates do B poverty indeed line reflect (VND e the 653,000 aspirations perperceptions and persoon perof month the h). Most lines Vietnamese s were population. clustered in the range ofo VND 700,0 8 000 to VND 800,000. 2.64 E Estimates ubjective pov of su verty lines sug ggest that the updated GSOO-WB povert elated ty lines and re stimates do in poverty es ns and perceptions of the Vietnamese ndeed reflect the aspiration V opulation. po 56 55 Annex Table 2. 1 Reference Food Basket for Different Population Groups Table A2.1 Reference Food Basket for Different Population Groups 2.5Ͳ20th 2.5Ͳ10th 2.5Ͳ20th percentile Reference Group: percentile percentile Ethnic Ethnic Subpopulation: (all) (all) minorities majority Urban Rural Food item Plain rice (including fragrant and specialty rice) 66.4 69.1 64.2 68.2 63.1 66.7 Sticky rice 4.2 4.4 7.9 1.1 1.2 4.5 Maize (in seed equivalent) 1.6 2.6 2.7 0.6 1.1 1.6 Cassava (in freshͲtype equivalent) 1.0 1.4 1.9 0.2 0.3 1.0 Potato of various kinds (in freshͲtype equivalent) 0.3 0.2 0.3 0.3 0.3 0.3 Wheat grains, bread, wheat powder 0.3 0.2 0.2 0.4 0.5 0.3 Flour noodle, instant rice noodle/porridge 1.3 1.0 1.1 1.4 1.9 1.2 Fresh rice noodle, dried rice noodle 0.4 0.3 0.3 0.6 0.6 0.4 Vermicelli 0.1 0.1 0.0 0.1 0.1 0.1 Pork (in equivalent of the pork type with removed fat) 4.0 3.6 4.0 4.1 4.3 4.0 Beef 0.1 0.1 0.1 0.1 0.1 0.1 Buffalo meat 0.0 0.0 0.1 0.0 0.0 0.0 Chicken meat 0.9 0.8 1.0 0.8 0.9 0.9 Duck and other poultry meat 0.2 0.1 0.1 0.2 0.1 0.2 Other types of meat 0.0 0.0 0.0 0.0 0.1 0.0 Processed meat 0.1 0.1 0.1 0.1 0.1 0.1 Lard, cooking oil 4.2 3.9 4.0 4.3 4.4 4.1 Fresh shrimp, fish 1.4 1.2 0.8 1.9 1.8 1.4 Dried and processed shrimps, fish 0.3 0.3 0.4 0.3 0.3 0.3 Other aquatic products and seafood (crabs, snails,...) 0.1 0.1 0.1 0.1 0.1 0.1 Eggs of chickens, ducks, Muscovy ducks, geese 0.7 0.6 0.5 0.8 0.8 0.7 Tofu 0.6 0.6 0.6 0.7 0.6 0.6 Peanuts, sesame 0.5 0.4 0.5 0.6 0.5 0.5 Beans of various kinds 0.3 0.2 0.3 0.2 0.3 0.2 Fresh peas of various kinds 0.1 0.1 0.1 0.1 0.1 0.1 Morning glory vegetables 0.5 0.5 0.4 0.7 0.6 0.5 Kohlrabi 0.1 0.1 0.0 0.1 0.1 0.1 Cabbage 0.1 0.1 0.1 0.1 0.2 0.1 Tomato 0.0 0.0 0.0 0.1 0.1 0.0 Other vegetables 0.7 0.6 0.7 0.6 0.8 0.6 Orange 0.0 0.0 0.0 0.0 0.1 0.0 Banana 0.6 0.6 0.6 0.5 0.5 0.6 Mango 0.0 0.0 0.0 0.0 0.0 0.0 Other fruits 0.4 0.3 0.3 0.5 0.6 0.4 Fish sauce 0.2 0.1 0.1 0.2 0.2 0.1 Salt 0.0 0.0 0.0 0.0 0.0 0.0 MSG Glutamate Sugar, molasses 1.3 1.0 0.8 1.7 1.6 1.3 Confectionery 0.6 0.6 0.6 0.7 0.8 0.6 Condensed milk, milk powder 0.2 0.1 0.1 0.2 0.2 0.2 Ice cream, yoghurt 0.0 0.0 0.0 0.0 0.1 0.0 Fresh milk 0.1 0.0 0.0 0.1 0.1 0.1 Alcohol of various kinds 1.3 1.3 1.7 0.9 1.0 1.3 Beer of various kinds 0.1 0.0 0.0 0.1 0.1 0.0 Bottled, canned, boxed beverages 0.1 0.1 0.0 0.1 0.2 0.1 Instant coffee Coffee powder 0.0 0.0 0.0 0.1 0.1 0.0 Instant tea powder Other dried tea 0.4 0.3 0.3 0.4 0.3 0.4 Tobacco Betel leaves, areca nuts, lime, betel pieces Outdoors meals and drinks 3.3 2.1 2.1 4.3 7.6 2.9 Other food and drinks 1.0 0.8 0.8 1.1 1.3 0.9 Source: Source: 2010 2010 VHLSS. VHLSS. 58 56 Chapter Annexes Annex 2.1: Differences between “Temporally Comparableâ€? and Comprehensive Welfare Aggregates Temporally Comparable Comprehensive Food Excludes consumption of tobacco Includes consumption of all 54 and betel nut. Assumes food items food items in VHLSS. Assumes listed in section 5A2 but not listed the only food items consumed in 5A1 were consumed during Tet/ during Tet/holidays were those holidays. Tet/holidays considered listed in section 5A1. Tet/holidays 15.2 days long considered 14 days long. Durables Excludes consumption of certain Includes all types of durables durables: printers, photocopiers, in 2010 VHLSS, but does not mobile phones, microwaves, impute consumption for durables blenders, other transport. Imputes acquired more than 10 years using depreciation rates from prior. Imputes using depreciation 1998 VLSS and real interest rate rates calculated from 2010 of 5 percent. VHLSS data and real interest rate of 5 percent. Housing Imputes housing consumption Imputes housing consumption as 11.8 percent of other nonfood as 2.88 percent of reported consumption for rural households housing values. 2.88 percent is and 21.4 percent for urban the median ratio of rental income households. to housing values for the 2.6 percent of households in the 2010 VHLSS who are renters. Education Equals total expenditures related Also includes supplemental to compulsory school subjects. expenditure on education, e.g., for tutors, typing classes, etc. Health Equals spending on curative Also includes spending on health and preventive care, including insurance. out-of-pocket costs of inpatient and outpatient health services, expenditures for nonprescription medicine, and expenditure on medical tools. Utilities: Electricity, Water, Simple sum of reported spending. Same. Garbage Other nonfood items (e.g., Excludes spending on parties and clothing, fuel, kitchen celebrations, and consumption of items, services, etc.) self-produced daily nonfood items from section 5B1. Temporal deflator GSO’s rice, nonrice food, and Same. nonfood monthly CPI. Spatial deflator GSO’s regional CPI. 2010 SCOLI. 57 Annex Annex 2.2: Spatial 2.2:Spatial Cost-of-living Estimates Cost-of-living Estimates for 2010 for VHLSS 2010 VHLSS A detailed A detailed price survey price survey of 64 of 64 items items was conducted was conducted in thein the main main market market in allin all communes communes inOctober the October in the 2010 2010round roundof the VHLSS of the VHLSSsample (n =(n sample 1049) and in = 1049) andhalf halfcommunes in the in the in December the communes 2010(n = 2010 round the December round 539). The(n 64 = 539). itemsThe 64 items included 45included 45 specifically specifically identified identified foods (including foods (including outdoor outdoor meals), meals), and 19 and another another 19 specially identified nonfoods, including some durable specially identified nonfoods, including some durable goods and services. goods and services. It was was important It important to ensure to ensure consistency consistency over over space space inin thelist the listof of64 items and 64items and to avoid avoid problems problemswith with missing missing observations. observations. SurveyorsSurveyors were given were given detaileddetailed specifications specifications (aided (aided by photographs by photographs to to ensure ensure standardization) and were instructed to take two observations on the price standardization) and were instructed to take two observations on the price of the detailed specification and to of the detailed specification record whether that to record specification and particular whether thatwas particular the most specification common one wasin the the most market.common one in A particular the and size, market. A particular size, and brand name (for packaged goods), was specified brand name (for packaged goods), was specified to avoid variation due to either bulk discounting or qualityto avoid variation due to either bulk discounting. discounting In almost or quality 80 percent of thediscounting. market-item Incombinations, almost 80 percent of the market-item the specification listed incombinations, the questionnaire the specification listed in the questionnaire was indeed the most common; it was available but not was indeed the most common; it was available but not the most common in approximately 5 percent of the most common in approximately 5 percent of markets. To deal with the missing prices problem in markets. To deal with the missing prices problem in the remaining market-item combinations, surveyors also the remaining market-item combinations, surveyors also collected the price of the most commonly collected the price of the most commonly available specification that was not the target specification. The price available specification that was not the target specification. The price of the target specification was the target specification of regressed was regressed against the prices against the of the alternate prices of the specifications alternate (using brand name fixed(using specifications effects,brand or for name fixed effects, or unbranded for unbranded items, items, creating creating quasi-brands quasi-brands by dividing by dividing into intervals based into onintervals their unitbased on and prices) theira unit prices) and a set of regional fixed effects. The regressions were used to impute set of regional fixed effects. The regressions were used to impute the price of the target specification the price of the target specification in about 10 in about 10 percent percent of of markets. markets. District District average or province or provinceprices were prices average used to were used impute tomissing impute the the missing commune-level commune-level pricesprices the remaining inremaining in the few few cases.cases. are aare ThereThere a number number of different of different indexes indexes that used that are are used to adjust to adjust for cost-of-living for cost-of-living differences. The The differences. Consumer Price Index (CPI) is typically based on a Laspayres index. For purposes Consumer Price Index (CPI) is typically based on a Laspayres index. For purposes of the SCOLI, of the SCOLI, new prices were newcombined with prices were regional budget combined sharesbudget with regional from the 2010from shares VHLSS in order the 2010 VHLSSto calculate in order toa Törnqvist calculate aprice index. The Törnqvist Törnqvist index price index. The the geometric isTörnqvist indexaverage of the price is the geometric relativities average of thebetween region i between price relativities and the base region region, i and the weighted by base region, weighted the arithmetic average of by thethe arithmetic budget sharesaverage for the twoof regions. the budget shares for the two regions. à­Ž à­©à­¨ ൅  ୧୨ ୧୨ ܶ ൌ Ý?‫݌ݔ‬ሾà·? ൬ ൰ ÂŽÂ? ቆ ቇሿ Í´ à­©à­¨ ୨ୀଵ where P denotes prices in each region and S is the budget shares. where P denotes prices in each region and S is the budget shares. The Törnqvist index specifically accounts for the fact that consumers will substitute away from items The Törnqvist that are expensiveindex specifically in their own region,accounts relative tothe for the fact basethat consumers region, by using will the substitute awayof budget shares from bothitems the base that are expensive region and the own region when weighting the price relativities. Technically, it closely approximates of in their own region, relative to the base region, by using the budget shares a true both the base region and the own region when weighting the price relativities. Technically, it closely cost-of-living index for any arbitrary utility function, whereas the Laspeyres index (used for the CPI) is an approximates a true cost-of-living index for any arbitrary utility function, whereas the Laspeyres index exact measure of the cost-of-living index only when items are consumed in fixed proportions, without allowing (used for the CPI) is an exact measure of the cost-of-living index only when items are consumed in for substitutions. fixed proportions, without allowing for substitutions. Because only 64 items had prices obtained in the SCOLI survey, while there are over 100 Because only 64 items had prices obtained in the SCOLI survey, while there are over 100 consumption consumption items listed in the VHLSS (including the consumption of housing services and the service flow items listed in the VHLSS (including the consumption of housing services and the service flow from from durables), a mapping on prices to budget shares was formed, where the price relativities for some closely durables), a mapping on prices to budget shares was formed, where the price relativities for some related items closely were items related used as a proxy were usedfor asthe a missing proxy for price the relativities missing pricefor other items. for relativities Two exceptions other items. were Two for utilities, wherewere exceptions the trimmed median for utilities, whereunit value the of electricity trimmed median unittariffs in each value region and of electricity sector tariffs was used in each regionas the proxy to form a price relativity and flow of accommodation services from dwellings. and sector was used as the proxy to form a price relativity and flow of accommodation services For the imputed rents, detailed econometric from dwellings. analysis For of the housing the imputed section of rents, detailed the VHLSS analysis econometric questionnaire of the was undertaken, housing section to of estimate the a hedonic VHLSS house value equation, questionnaire which allowed was undertaken, for regional to estimate differences a hedonic in the house valuecost of constant-quality equation, which allowed housing service. for regional differences in the cost of constant-quality housing service. 59 58 Annex 2.3: Subjective Poverty in Vietnam It is often argued that as countries develop and become less poor, societies’ standards also evolve. Even if the basic point ofAnnex departure measure poverty is toSubjective 2.3: with an Poverty “absoluteâ€? poverty line that is held in Vietnam fixed in real terms over time, societies will need to update this poverty line from time to tome so it It is often remains arguedto relevant that as countries a country’s develop specific and become circumstances. Asless notedpoor, societies’ in chapter standards 2, as countries also growevolve. their Even if the basic point of departure is to measure poverty with an “absoluteâ€? national poverty lines increase over time. Regardless of how carefully an absolute poverty line poverty line that is held is fixed in real terms over time, societies will need to update this poverty line from time to tome so it developed, it is not possible to avoid some degree of arbitrariness. Challenges in setting a poverty line are remains relevant to a country’s specific circumstances. As noted in chapter 2, as countries grow groups by Ravallion (2012) into (i) a referencing problem, including the choice of the reference group and their national poverty lines increase over time. Regardless of how carefully an absolute poverty basket, and (ii) an identification line is developed, problem it is not possible that involves to avoid some degreetranslating households’ Challenges of arbitrariness. utility function into the in setting a measurable expenditure space. poverty line are groups by Ravallion (2012) into (i) a referencing problem, including the choice of the reference group and basket, and (ii) an identification problem that involves translating households’ An alternative utility function method for analyzing into the measurable expenditure poverty space. that has received growing attention builds on subjective welfare questions included in household surveys. A subjective poverty line built up from such An alternative questions method can offer for analyzing an alternative poverty entry point that into has received growing the derivation attention of the poverty builds line, also onhelp subjective with the welfare questions included in household surveys. A subjective poverty line interpretation of the conventionally derived, Cost-of-Basic-Needs (CBN) poverty line. This subjective built up from such questions can offer an alternative entry point into the derivation of the poverty line, also help with the poverty line exercise is particularly interesting in the context of Vietnam given the proposed update to the interpretation of the conventionally derived, Cost-of-Basic-Needs (CBN) poverty line. This subjective 2010 CBN poverty line. poverty line exercise is particularly interesting in the context of Vietnam given the proposed update to the 2010 CBN poverty line. Van Praag (1968) introduced subjective welfare assessment by constructing utility functions based on Van Praag respondents’ (1968) introducedanswers to the question subjective of how welfare much income assessment they regardedutility by constructing as “very bad,â€?, “bad,â€? functions based and so forth, to “very good.â€? A similar method, the Minimum Income on respondents’ answers to the question of how much income they regarded as “very bad,â€?, Question (MIQ), asks about the “bad,â€? and so forth, minimum to “very income thatgood.â€? A similar respondents method, perceive to the necessary Income be Minimum “to make Question (MIQ), ends meetâ€? asks about (Kapteyn the 1994). However, income minimum that respondents applicability perceive to to of the MIQ methodology necessary bethe “to makehas poorest countries ends meetâ€? been (Kapteyn debated (Deaton 1994). and However, Zaidi 2002; applicability Pradhan and the MIQ methodology of Ravallion 2000; Ravallion to the Lokshincountries and poorest has been 2000). Pradhan and debated Ravallion (Deaton (2000) and Zaidi 2002; Pradhan and Ravallion 2000; Ravallion and Lokshin 2000). Pradhan and Ravallion propose an adaptation to Kapteyn’s method by asking households if their consumption of food (and other (2000) propose an adaptation to Kapteyn’s method by asking households if their consumption of things) has been adequate to “meet their needs.â€? The 2010 VHLSS included a set of similar questions, food (and other things) has been adequate to “meet their needs.â€? The 2010 VHLSS included a set allowing of similarus a similar us to follow allowing questions, to followmethodology. estimation The exact a similar estimation framing of the methodology. The question, asked of exact framing of the the household head, is the following: question, asked of the household head, is the following: The same question as above is asked about “waterâ€? “electricityâ€? “housingâ€? “clothing & footwearâ€? Out of total respondents to the 2010 VHLSS consumption section, 440 reported insufficient food Out of total respondents to the 2010 VHLSS consumption section, 440 reported insufficient food consumption, 8,218 reported just sufficient food, and 686 indicated that their food consumption consumption, 8,218 reported just sufficient food, and 686 indicated that their food consumption was more was more than sufficient (54 households did not respond). Satisfaction with adequacy of foodstuff (54 households than sufficient (including consumption respond). did not calories higher-cost Satisfaction from with adequacy meat, vegetables, of and oils, foodstuff consumption condiments) was (including higher-cost calories from meat, vegetables, oils, and condiments) was less: 1,079 respondents reported inadequate consumption of foodstuffs, 7,580 indicated less: 1,079 respondents sufficient reported inadequate consumption, consumption and 678 of foodstuffs, claimed their 7,580 consumption indicated was sufficient more than consumption, and 678 claimed sufficient. their consumption was more than sufficient. To calculate a subjective poverty line, we follow Pradhan and Ravallion (2000) in regressing perceived sufficiency of consumption on household expenditure and household (head) characteristics, using the sufficiency of foodstuff as the dependent variable. “Not Applicableâ€? responses were excluded, and 60 the other three categories are subjected to an ordered probit regression including actual household consumption, household size, and characteristics of the household head. Regression coefficients, presented in table A2.1, were also used in calculating a range of subjective poverty lines, including those reported in the chapter. 59 Table A2.1 Subjective Welfare Regression and Variables at Country Means Regression Results Means Variables Coefficient S.E. of Mean S.D. Log total household expenditure 0.717*** 0.029 10.978 0.731 Log household size -0.475*** 0.049 1.435 0.381 Household head is female -0.092** 0.040 0.220 0.414 Household head has a wage job -0.172*** 0.031 0.407 0.491 Household has at least one widow(er) -0.040 0.042 0.186 0.389 Highest grade household head 0.022*** 0.005 7.313 3.683 Household head is registered within the commune 0.046 0.034 0.256 0.437 Household head is of ethnic majority (Kinh) 0.516*** 0.044 0.854 0.353 Share of household < 18 years old 0.206*** 0.078 0.256 0.206 Share of household > 59 years old 0.009 0.093 0.072 0.175 Log land area owned by household 0.029*** 0.005 4.859 3.757 Urban -0.148*** 0.041 1.297 0.457 Cutoff 1 6.264*** 0.277 Cutoff 2 9.327*** 0.289 Number of observations 9,337 Pseudo R2 0.139 Note: The dependent variable is “perceived sufficiency of foodstuff consumptionâ€? with the following answer codes: 1 = insufficient, 2 = sufficient, and 3 = more than sufficient (“not applicableâ€? is recoded as missing). The results are from an ordered probit regression. The natural logarithm is used for the log variables. Significance levels are *** 0.01, **0.05, * 0.1. The means of the variables and the regression are both weighted by population weights. 60 References Banerjee, Abhijit. 2011. “Draw the right line.â€? Hindustan Times, October 24. Accessed May 2012. http://www.hinustantimes.com/StoryPage/Print/761099.aspx. Bertrand, M., and S. Mullainathan. 2001. “Do People Mean What They Say? Implications for Subjective Survey Data.â€? American Economic Review, Papers and Proceedings 91 (2): 67–72. Conti, G., and S. Pudney. 2011. “Survey Design and the Analysis of Satisfaction.â€? Review of Economics and Statistics 93 (3): 1087–1093. Deaton, A., and O. Dupriez. 2011. “Spatial Price Differences within Large Countries.â€? Princeton University Working Papers, Princeton, NJ. Deaton, A., and S. Zaidi. 2002. “A Guide to Aggregating Consumption Expenditures.â€? Living Standards Measurement Study Working Paper No. 135, World Bank, Washington, DC. Deaton, Angus. 1988. “Quality, Quantity, and Spatial Variation in Price.â€? American Economic Review 78 (3): 418–30. Deaton, Angus. 1997. Analysis of Household Surveys: A Microeconometric Approach to Development Policy. Washington, DC: The Johns Hopkins University Press and World Bank. Deaton, Angus, and Alessandro Tarozzi. 2005. “Prices and Poverty in India.â€? In The Great Indian Poverty Debate, ed. Angus Deaton and Valerie Kozel. New Delhi: Macmillan, Chapter 16, pp. 381– 411. Gill, Nikhila. 2012. “Has Poverty Really Dropped in India?â€? New York Times, March 21. Accessed May 2012. http://india.blogs.nytimes.com/2012/03/21/has-poverty-really-dropped-in-india/. Government of India. 2009. “Report of the Expert Group to Review the Methodology for Poverty Estimation.â€? Planning Commission, Government of India, New Delhi. Hansen, H., and T. Nguyen, eds. 2006. Market, Policy, and Poverty Reduction in Vietnam. Hanoi: Vietnam Academy of Social Sciences, Vietnam Cultural Information Publishing House. Haughton, J., Nguyen Thi Thanh Loan, and Nguyen Bui Linh. 2010. Urban Poverty Assessment in Hanoi and HCMC. Hanoi, joint publication of the UNDP and General Statistics Office. Kapteyn, A. 1994. “The Measurement of Household Cost Functions: Revealed Preference versus Subjective Measures.â€? Journal of Population Economics 7 (4): 333–350. Kapteyn, A., and B. Van Praag. 1976. “A New Approach to the Construction of Family Equivalence Scales.â€? European Economic Review 7: 313–335. Kozel, Valerie, Ian Hinsdale, and Nguyen Phong. 2012. “Updated Methodologies for Poverty Monitoring in Vietnam.â€? Background paper prepared for the 2012 Vietnam Poverty Assessment, World Bank, Washington, DC. Krueger, A. B., and D. Schkade. 2008. “The Reliability of Subjective Well-being Measures.â€? Journal of Public Economics 92 (8–9): 1833–1845. Lokshin, M., N. Umapathi, and S. Paternostro. 2006. “Robustness of Subjective Welfare Analysis in a Poor Developing Country: Madagascar 2001.â€? Journal of Development Studies 42 (4): 559–591. Marra, M. 2012. “Estimating Subjective Poverty Lines for Vietnam.â€? Background note prepared for the 2012 Vietnam Poverty Assessment, World Bank, Washington, DC. Ministry of Health. 2006. “Proposed Nutrition Needs for the Vietnamese.â€? Ministry of Health, Hanoi. Phung, D. T. 2005. “Determination of a Consistent Poverty Line for Vietnam.â€? Environmental and Social Department, General Statistics Office, Government of Vietnam. 61 Pincus, J., and J. Sender. 2008. “Quantifying Poverty in Vietnam: Who Counts?â€? Journal of Vietnamese Studies 2 (1) (January): 108–150. Pradhan, M., and M. Ravallion. 2000. “Measuring Poverty Using Qualitative Perceptions of Consumption Adequacy.â€? The Review of Economics and Statistics 82 (3): 462–471. Pradhan, M., M. Suryahadi, S. Sumarto, and L. Pritchettt. 2001. “Eating Like Which Jone’s? An Iterative Solution to the Choice of a Poverty Line Reference Group.â€? The Review of Income and Wealth, Series 47 (4): 473–487. Ravallion, Martin. 1998. “Poverty Lines in Theory and Practice.â€? Living Standards Measurement Study Working Paper 133, World Bank, Washington DC. Ravallion, M. 2012. “Poor, or Just Feeling Poor? On Using Subjective Data in Measuring Poverty.â€? World Bank Policy Research Working Paper 5968, World Bank, Washington, DC. Ravallion, M., and B. Bidani. 1994. “How Robust Is a Poverty Profile?â€? The World Bank Economic Review 8 (1) (January): 75–102. Ravallion, M., and M., Lokshin. 2002. “Self-Rated Economic Welfare in Russia.â€? European Economic Review 46 (8) (September): 1453–1473. Taylor, M. P. 2006. “Tell Me Why I Don’t Like Mondays: Investigating Day of the Week Effects on Job Satisfaction and Psychological Well-being.â€? Journal of the Royal Statistical Society: Series A (Statistics in Society) 169 (1): 127–142. United Nations Statistics Division. 2005. Handbook on Poverty Statistics: Concepts, Methods, and Policy Use, Special Project on Poverty Statistics. New York: United Nations. Van Praag, B., and M. Warnaar. 1997. “The Cost of Children and the Use of Demographic Variables in Consumer Demand.â€? In Handbook of Population and Family Economics, ed. Mark Rosenzweig and Oded Stard. Amsterdam: North-Holland, Chapter 6, pp: 241–273. World Bank. 1999. Vietnam Development Report 2000: Attacking Poverty. Washington, DC: World Bank. 62 chapter 3 Poverty Profile: Establishing the Facts about Poverty and the Poor in Vietnam A new poverty profile is presented that characterizes the poor and the extreme poor and compares them with the rest of society along a number of key dimensions including geographic location, ethnicity, sector of employment, income sources, educational attainment, ownership of durable goods, landholdings, household amenities, child poverty, and coverage under social protection and poverty reduction programs and policies. Statistical analysis is complemented by a rich body of qualitative research. The poor in Vietnam today are similar in important respects to the poor in the late 1990s. Among other factors, poverty is linked to rural and upland locations, agricultural livelihood, ethnic identity, low educational attainment, exposure to risk and rising vulnerability. 63 A. Introduction 3.1 Poverty reduction remains a challenge in Vietnam, albeit one that has changed dramatically in scope and nature over the last two decades. This chapter revisits the basic facts about poverty and the poor in Vietnam. It takes stock of what we know about poverty today and draws comparisons with the situation of the poor in the late 1990s, with the aim of highlighting both important areas of progress and remaining and new challenges. The chapter presents a new profile of the poor, using the 2010 General Statistics Office-World Bank (GSO-WB) poverty line and more comprehensive measures of household welfare proposed in Chapter 2. The analysis is primarily based on the 2010 Vietnam Household Living Standards Survey (VHLSS), but also draws selectively on earlier rounds of the Vietnam Living Standards Survey (VLSS), (particularly the 1998 VLSS), and other sources, such as recent Participatory Poverty Assessments and qualitative field studies, 2009 poverty maps, and other supplementary data sets. 3.2 A poverty line only discriminates between poor and non-poor households. It ignores the fact that not all poor people are the same; some have incomes or consumption very close to the poverty line, while others live in much poorer conditions. Nor are the non-poor homogeneous; some live near the poverty line (referred to as the “near-poorâ€? in Vietnam) while others are much more prosperous. The analysis presented in this chapter recognizes the broad economic diversity among poor and non-poor households in Vietnam. At the lower end of the welfare distribution, we distinguish between the “extreme poorâ€? (per-capita expenditures below two-thirds of the poverty line) and “poorâ€? (per- capita expenditures below the poverty line). The remainder of the population is analyzed on the basis of per-capita expenditure quintiles and deciles. Specifically: â—?â—? Individuals are ranked by per-capita expenditures from least well-off to most well-off, then divided into five equally-sized population groups (for quintiles) and ten equally sized population groups (for deciles). Quintile 1 comprises the poorest 20 percent of the population, and quintile 5 comprises the wealthiest 20 percent. Similarly, decile 1 comprises the poorest 10 percent of the population and decile 10 the wealthiest 10 percent. â—?â—? Individuals are also categorized into expanded per-capita expenditure quintiles, where the poor are classified into two groups (all poor and extreme poor) and the non-poor are classified by the standard per-capita expenditure quintiles. Expanded quintiles thus comprise six groups: o The extreme poor: individuals whose per-capita expenditures are less than  two-thirds of the poverty line (poorest 8 percent of the population) o All poor: individuals whose per-capita expenditures are below the poverty  line (poorest 20.7 percent of the population) o And quintiles 2 through 5 (as above). 3.3 In the context of the 2006-2010 Socio-Economic Development Plan (SEDP), the Ministry of Labour, Invalids and Social Affairs (MOLISA) introduced a “near-poorâ€? classification, which includes households whose per-capita income lies between the poverty line and 1.3 times the poverty line. If this definition is applied to the 2010 GSO-WB poverty line, roughly three-quarters of individuals in quintile 2 would fall into the near-poor group. 3.4 As a follow-on to the Millennium Development Goals, the World Bank is proposing to launch a new global initiative designed to accelerate the rate of poverty reduction among the poorest and most destitute and to promote shared prosperity over the next decade. Research from countries throughout the world shows that the poorest and most destitute are more difficult to reach than those living close to the poverty line; they face a structural barriers and specific constraints, and better policies and programs are needed to address these specific challenges. In many countries, including Vietnam, the extreme and destitute poor are falling further behind. This chapter develops profiles of the extreme poor as well as the total poor, and recognizes that many of the near-poor (quintile 2) remain vulnerable to falling (back) into poverty. 64 3.5 In constructing the poverty profile, households and individuals are also categorized by socioeconomic group (ethnic minority, Kinh majority), sector (urban, rural), and economic region. The Government of Vietnam has identified eight economic regions encompassing 63 provinces, more than 680 districts, and two major urban areas (Hanoi and Ho Chi Minh City). Annex 3.1 provides a description of the eight economic regions including the North East region, North West region, the Red River Delta (which houses Hanoi), the North Central Coast, the South Central Coast, the Central Highlands, the South East (which houses HCMC), and the Mekong River Delta. The North East and North West are mountainous regions where the majority of Vietnam’s ethnic minorities reside. Ethnic minorities also live in upland areas of central and southern regions, particularly the Central Highlands. The two deltas (Red River, Mekong) are major rice growing regions, and the majority of Vietnam’s rice exports come from the Mekong River Delta. The Stylized Facts about Poverty and Poor Households at the End of the 1990s 3.6 The Vietnam Development Report 2000: Attacking Poverty (World Bank 1999) described the key characteristics of poor households at the end of the 1990s, drawing on the 1993 and 1998 VLSS combined with a series of Participatory Poverty Assessments (PPAs) carried out in 1999. These early PPAs stressed core poverty concerns like hunger; lack of productive assets; high exposure to adverse shocks like drought, flooding, and illnesses; and concerns about social marginalization and isolation (particularly for ethnic minority groups). Many poor households struggled to feed and educate large families, and child poverty was widespread. Landlessness was rising, and there were limited options for off-farm employment (box 3.1). Box 3.1 Defining Characteristics of Poor Households at the end of the 1990s By the end of the 1990s, the key defining characteristics of poor households included: â—?â—? The poor lived in rural areas and were predominantly farmers with low levels of educational attainment, limited access to information, and low function skills. In 1998, nearly four-fifths of the poor were agriculture households. â—?â—? Poor households had small landholdings, and landlessness was increasing, especially in the Mekong Delta. Households that were unable to make a living from the land found few opportunities for stable off-farm income generation. There was an urgent need for reforms to stimulate demand for off-farm employment. â—?â—? Households with many children or few laborers were disproportionately poor and were particularly vulnerable to rising and variable health and education costs. Newly formed households went through an initial phase of poverty, often aggravated by limited access to land. Poor households were also frequently caught in a debt trap. â—?â—? Poor households were vulnerable to seasonal hardship and household-specific and communitywide shocks and some were socially and physically isolated. â—?â—? Poverty among ethnic minority groups had declined, but not as rapidly as for the majority population. Ethnic minorities faced many specific disadvantages that could best be addressed through an Ethnic Minority Development Program. â—?â—? Migrants to urban areas who were poor and who had not secured permanent registration faced difficulties accessing public services and some felt socially marginalized. Further work was needed to identify the best way to help these groups. â—?â—? Children were overrepresented in the poor population; they were less able to attend school and were trapped in a cycle of inherited poverty. Many felt insecure and uncertain about their future. Source: World Bank 1999. 65 Many of these Stylized Facts are still True Today 3.7 Although poverty has fallen dramatically, many of the factors that characterized the poor in the 1990s still characterize the poor today: low education and skills, heavy dependency on subsistence agriculture, physical and social isolation, specific disadvantages linked to ethnic identity, and exposure to natural disasters and risks. Those that moved out of poverty acquired more schooling and job skills, diversified out of agriculture and into manufacturing and services, and reduced exposure to seasonal hardships and shocks through income diversification and migration. But some of the stylized facts have changed. For example, issues such as ethnic minority poverty that were only emerging as concerns in the late 1990s are much greater concerns today. Other issues, like poverty The Poor in Vietnam still Predominately Live in Rural Areas and are Increasingly and vulnerability among migrants in urban areas, have become lesser concerns. Although income poverty remains very low in Vietnam’s cities and towns, there is evidence that new forms of poverty Concentrated in Upland Regions are arising: urban households are particularly vulnerable to sharp bouts of inflation and a rising cost of living. Risk remains an important feature of the rural economy as well, including weather-related risks and the emerging impacts of climate change for agriculture. As shown in table 3.1, an estimated 20.7 percent of the population was poor in 2010 and 8 nt was extremely B. The Poor Poverty poor. in Vietnam remains a rural phenomenon still Predominately in Vietnam; Live in Rural Areas andmore are than 90 percent o oor and 94 Increasingly percent of the Concentrated extreme poor live in Regions in Upland rural areas. The poor in urban areas for the most par n smaller cities 3.8 As and showntowns (Section in table 20.7 percent qualitative G). However, 3.1, an estimated studies of the population was poorcomplete in 2010 and for 8 this report and percent was extremely poor. Poverty remains a rural phenomenon in Vietnam; more than 90 percent on t research of urban the poverty poor and 94 percent(Haughton et.poor of the extreme al. live 2010) suggest in rural that areas. The urban poor low-income in urban areas for the households are most part live in smaller cities and towns (Section G). However, qualitative studies complete for this cted by other (non-income) dimensions of poverty, such as poor sanitation, lack of adequate housing report and recent research on urban poverty (Haughton et. al. 2010) suggest that urban low-income ed coverage of social households insurance, are impacted by increasing exposure other (non-income) to risk, dimensions and continuing of poverty, such as poor vulnerability sanitation, to poverty. lack of adequate housing, limited coverage of social insurance, increasing exposure to risk, and continuing vulnerability to poverty. Table 3.1 2010 Table Poverty 3.1 2010 Headcount Poverty andComposition, Headcount and Composition, by Region by Region and Sector and Sector Poverty ExtremePoverty Shareof Contribution Contribution totalpop Index(%) tototal(%) Index(%) tototal(%) (%) National 20.7 100.0 8.0 100.0 100.0 RedRiverDelta 11.4 12.3 2.8 7.8 22.3 EastNorthernMountains 37.7 20.8 17.9 25.8 11.5 WestNorthernMountains 60.1 9.1 36.5 14.4 3.2 NorthCentralCoast 28.4 16.5 9.7 14.6 12.0 SouthCentralCoast 18.1 7.4 5.9 6.3 8.5 CentralHighlands 32.8 9.5 17.0 12.9 6.0 Southeast 8.6 7.2 3.1 6.9 17.5 MekongRiverDelta 18.7 17.1 4.8 11.4 19.0 Rural 27.0 91.4 10.7 94.4 70.3 Urban 6.0 8.6 1.5 5.6 29.7 Source: Source: 2010 2010 VHLSS. VHLSS. 3.9 The spatial distribution of poverty has changed over time. In the 1990s, poverty was widespread The spatial distribution in Vietnam. of poverty Although poverty has rates were changed higher in some over regionstime. widespread In the than others, example, poverty (for1990s, in sparsely was settled provinces in the Northern Mountains and Central Highlands), the majority of the poor lived in the etnam. Although poverty rates were higher in some regions than others, (for example, in sparsely d provinces in the Northern Mountains and Central Highlands), the majority of the poor lived in the 66 densely settled Delta regions (figure 3.1). Poverty fell throughout Vietnam between 1998 and 2010 fell more rapidly in fast-growing regions around Hanoi and Ho Chi Minh City (that is, the Red more densely settled Delta regions (figure 3.1). Poverty fell throughout Vietnam between 1998 and 2010, Figure 3. 1 it but Level and fell more Composition rapidly regions around Hanoi Figure of Poverty in fast-growing 3.2 Minh and Ho Chi Level and City Composition (that is, the Red Riverof Poverty Delta and by Region, the 1998 Southeast). byin Uneven progress has resulted in substantial changes Region, the spatial2010 distribution of poverty, with the remaining poor becoming more concentrated in the upland areas in the north of Vietnam and in the Central Highlands (figure 3.2). Chapter 4 uses poverty mapping methods to look at 80 80 the spatial distribution of poverty at lower levels of spatial disaggregation (provinces and districts). 70 70 Figure Figure 3. 1 3. 1 Level Figure and Level 3.1 andand Composition Level Composition WB ͲGSO Poverty of Composition ofrate poverty Poverty of Figure Figure 3.2 Figure 3.2Level 3.2 Level Leveland Composition andand Composition of Poverty of Poverty Composition of WBͲGSO poverty rate 60 by by Region, Region, 19981998 Contribu on to total Poverty by Region, 1998 60 by by Region, Region, 2010 Poverty by Region, 20102010 Contribu on to total 80 80 80 80 50 Na onal WBͲGSO 50 poverty rate: 70 70 70 70 37.4 40 WBͲGSO WBͲGSO poverty rate poverty rate WBͲGSO WBͲGSO poverty rate poverty rate 40 to total on to total Contribu onContribu to total on to total Contribu onContribu 60 60 60 60 30 30 Na onal WBͲGSO 50 50 50 50 Na onal WBNaͲGSO onal WBͲGSO poverty rate: poverty rate: poverty rate: 20.7 20 37.4 37.4 20 40 40 40 40 10 30 30 30 10 30 Na onal WBNa onal WBͲGSO ͲGSO poverty rate: poverty rate: 20.7 20.7 20 20 20 20 0 0 Red River Delta East Northern West Northern North Central South Central Central Southeast Mekong River Red River Delta East Northern West Northern North Central South Central Central Southeast Mekong River Mountains Mountains Coast Coast Highlands Delta Mountains Mountains Coast Coast Highlands Delta 10 10 10 10 Source: 1998 VLSS. 0 0 0 Source: 2010 VHLSS. 0 Red East Red River Delta RiverNorthern Delta East West Northern Northern West North Northern Central South North Central Central South Central Southeast Southeast Central Mekong River Mekong River Red River Delta Red East River Delta East Northern West Northern Northern North West Northern Central North South Central Central Central South Central Southeast Southeast Central Mekong River Mekong River Mountains Mountains Mountains Mountains Coast Coast Coast Coast Highlands Highlands Delta Delta Mountains Mountains Mountains Mountains Coast Coast Coast Highlands Coast Highlands Delta Delta Source: Source: 1998 1998 1998 VLSS. VLSS.VLSS. Source: 2010 VHLSS. Source: Source: VHLSS. 2010 2010 VHLSS. C. Many Source: of the Poor are Farmers Whose Livelihoods are Primarily Linked to Agriculture C. Many of the Poor are Farmers Whose Livelihoods are Primarily Linked 3.10 C. The poorof C. Many Many in the to Agriculture the Poor of Poor Vietnam are are are Farmers Farmers still Whose Whose predominately Livelihoods Livelihoods farmers; are are Primarily Primarily 32.9 percent Linked ofLinked to Agriculture to Agriculture agricultural households live 19 the 3.10 below 3.10 poverty poor The The line, poor in Vietnam in Vietnam which areare is still are still nearly predominatelythree predominately times farmers; farmers; higher 32.9 32.9 percent than percent the national of agricultural of agricultural poverty households households rate, and live live 3.10 The poor in Vietnam19 19 still predominately farmers; 32.9 percent of agricultural households agriculturalbelow below households livethe the the poverty below make poverty up line, line, poverty 65 which line, 20 percent which is nearly which is of is nearly the three nearly poor three times three and higher times times higher73 higher percent than thanthan the national the the nationalthe extreme ofnational poverty poverty poverty rate, rate,poor andand compared rate, and agricultural with a agricultural households agricultural population households households share make of make makeup 65 only up up41 65 percent 65 percent percentof of percent of the the the poor poor (table poor and and 73 and 3.2). 73 percent 73 percent percent of the of Agriculturalthe of extreme the poor extreme extreme poor households compared poor compared compared also contribute with withwith a population a population a population share share of of share ofonly only 41 only41percent percent 41 percent(table (table 3.2). 3.2). (table Agricultural Agricultural 3.2). households Agricultural households also alsocontribute households contribute also contribute disproportionately to the poverty disproportionately gap and to the poverty gappoverty and povertyseverity. severity. disproportionately disproportionately topoverty to the the poverty gap and gap poverty severity. and poverty severity. Table 3.2 Poverty Headcount and Composition in 2010, by Sector of Employment Table 3.2 Poverty Table Table 3.2 Headcount Poverty 3.2 Poverty and Headcount Composition Headcount and and Composition Composition in in 2010,by by in 2010, Sector 2010, of Household Head Sector by Sector ofEmployment of of Employment Employment of Household of Household of Household HeadHead Head Poverty Poverty Poverty Extreme Poverty Extreme Extreme Poverty Poverty of Share ShareShare of of Contribution  Contribution  Contribution  total  pop Contribution total pop Contribution Contribution totalpop IndexIndex (%) (%) total(%) tototal to(%) IndexIndex (%) (%) tototal total(%)(%) (%) to(%) Index(%) tototal(%) Index(%) tototal(%) (%) National National 20.7 20.7 100.0100.0 8.0 8.0 100.0100.0 100.0100.0 National 20.7 100.0 8.0 100.0 100.0 Employment Employment ofhousehold ofhousehold head:head: Not employed Notemployed 13.2 13.2 9.1 9.1 5.3 5.3 9.6 9.6 14.4 14.4 Employmentofhouseholdhead: Agriculture Agriculture 32.9 32.9 64.8 64.8 14.1 14.1 72.5 72.5 40.9 40.9 employed Not FamilyFamily businessbusiness 13.2 5.9 5.9 9.1 4.4 4.4 5.3 1.2 1.2 9.6 2.3 2.3 15.4 15.4 14.4 Agriculture Employed Employed forwages forwages in: in: 32.9 64.8 14.1 72.5 40.9 Family business Industry Industry &manufacturing &manufacturing 5.913.2 13.2 4.0 4.4 4.0 1.2 2.7 2.7 2.1 2.1 2.3 6.3 6.3 15.4  forConstruction Construction Employed wagesin: 19.3 19.3 7.7 7.7 5.1 5.1 5.3 5.3 8.3 8.3 Services  Industry&Services manufacturing 14.0 13.214.0 10.0 10.0 4.0 4.4 4.4 2.7 8.2 8.2 2.114.9 14.9 6.3 Source: Source: Construction2010 2010 VHLSS. VHLSS. 19.3 7.7 5.1 5.3 8.3 Services 14.0 10.0 4.4 8.2 14.9 3.11 3.11The2010 Source: The level Source: VHLSS. level and 2010 composition and VHLSS. compositionof household of household income income across the expanded across the expandedper-capita expenditure per-capita expenditure quintiles quintiles is described is described in figure in figure The The 3.3. 3.3. height height of each bar reflects of each the average bar reflects the average levellevel of per-capita income of per-capita income 20 Defined as households where the head’s main job is in agriculture. for each for each group.group. Figure Figure 3.4 looks in greater 3.4 looks detail in greater detail at the the composition at composition of income of income for each for each group, broken group, broken 3.11 The level and composition of household income across the expanded per-capita expenditure quintiles described in figure 3.3. The height of each is19 bar reflects the average level of per-capita income 19 Defined Defined as households as households wherewhere the head’s the head’s mainmain job is job is 67 in agriculture. in agriculture. for each group. Figure 3.4 looks in greater detail at the composition of income for each group, broken 71 71 3.11 The level and composition of household income across the expanded per-capita expenditure down by by quintiles down income is from described income fromagriculture in agriculture figure 3.3. sources The sources (crop height ofcultivation, (crop each cultivation, livestock, bar reflects forestry, the average livestock, forestry, aquaculture, level of per-capita aquaculture, andand income agriculture wages), group. Figure for eachnonfarm family 3.4 looks in greater enterprises, detail at the composition non-agriculture wages, social of income transfers, fordomestic each group, and agriculture wages), nonfarm family enterprises, non-agriculture wages, social transfers, domestic and broken overseas down by income remittances, andand otherfrom agriculture sources. sources According to(crop figure cultivation, 3.4, poor livestock, households forestry, derive aquaculture, roughly half overseas remittances, and agriculture wages), other nonfarmsources.family According to enterprises, figure 3.4, poor non-agriculture households wages, social derive roughly transfers, half their income their income from agricultural from agricultural activities, activities,including including agricultural agricultural wages. wages. However, However, whatwhatdifferentiates differentiates the domestic and overseas remittances, and other sources. According to figure 3.4, poor households the incomes incomes deriveof the poor of the roughly from poor half wealthier from their wealthier income households households from is not agricultural the is not level the of including level activities, income of income fromfrom agricultural agricultural agricultural wages.activities; activities; However, crop crop incomes what are surprisingly differentiates the equal incomes across of the wealth poor from quintiles, wealthier reflecting households incomes are surprisingly equal across wealth quintiles, reflecting Vietnam’s broadly egalitarian Vietnam’s is not the broadly level of egalitarian income from agricultural distribution distribution activities; of agriculture of agriculturecrop land. incomes What land. What are surprisingly differentiates differentiates equal incomes thethe across incomes the of ofwealth poor the poorquintiles, fromfrom reflecting wealthier wealthier Vietnam’s households households is,is, broadly instead, the egalitarian extent to distribution which households of agriculture have land. successfully What differentiates diversified into the incomes off-farm of the activities. poor from Progress inin instead, the extent to which households have successfully diversified into off-farm activities. Progress wealthier households is, instead, the extent to which households have successfully diversified into 1990s the the 1990swaswas driven by by driven on-farm on-farm diversification, forfor instance into cash crops, livestock, and (in some off-farm activities. Progress in the diversification, 1990s was driven by instance on-farm into cash crops, diversification, livestock, for instance and into(in cashsome parts of parts the of country) the country)fish and fish shrimp and shrimp farming farming (World (World Bank Bank 1999). 1999).But progress But crops, livestock, and (in some parts of the country) fish and shrimp farming (World Bank 1999). But progress in recent in recent years yearshashasbeen been driven by driven progressdiversification by diversification in recent years into business into has business and been driven trading and trading and, even by diversification more more and, even into importantly, importantly, business by salaried by salaried and trading employment and,employment even more inin industry andand importantly, industry by salaried and manufacturing manufacturing andjobs employment in the jobs in service industry inthe sector. serviceand Even manufacturing sector. Eventhetheextreme and jobs extreme poor inhave poor the have income service income sources sector. sources Even outside the extreme agriculture, poor although have as income shown in sources the next outside section, agriculture, this differsalthough outside agriculture, although as shown in the next section, this differs for poor minority households for poor as shown minority in the next households section, this differs for poor minority households compared to poor minorities. compared compared to poor minorities. to poor minorities. Figure 3.3 Household Income by Expanded Figure 3.4 Composition of Income by Figure 3.3 3.3 Figure Household Income Household Quintile,Incomeby Expanded 2010 by Expanded Figure 3.43.4 Figure Composition Composition Expanded Income of of Quintile, byby Income 2010 Expanded Expanded Quintile, Quintile, Level of household 20102010 million VND incomes, Composition ofQuintile, 2010 Quintile, household 2010 income (percent) Level of household Level of household (January incomes, million incomes, 2010) VND million (January VND 2010) (January 2010) Composition of household Composition income of household (percent) income (percent) Source: 2010 VHLSS. Source: 2010 VHLSS. Source: 20102010 Source: VHLSS. VHLSS. Source: 2010 VHLSS. D. D. Ethnic Ethnic Identity Identity D. Ethnic Matters Matters Identity even even more Matters forfor more even Poverty Poverty more Today Today for Poverty Today 3.12 Although Vietnam’s 53 ethnic minority groups make up only 15 percent of the total population, 3.12 Although 3.12 AlthoughVietnam’s Vietnam’s 5353ethnic ethnicminority minoritygroups makeup groupsmake up only 15 percent only 15 of the percent of thetotal totalpopulation, population, they account for nearly half (47 percent) of the total poor and 68 percent of the extreme poor in Vietnam. they account for nearly they account half (47 for nearly halfpercent) of the of (47 percent) total thepoor and total 68and poor percent of the extreme 68 percent poor in Vietnam. of the extreme poor in (Figure 3.5). Although living conditions for many minorities have improved since the late 1990s, the (Figure 3.5). Although Vietnam. living conditions for (Figure 3.5). Although living many minorities conditions for many have have since improved minorities improvedthe late since 1990s, the the late concentration 1990s, the of minoritiesof among the poor has nonetheless increased dramatically —by 25 percentage concentration ofconcentration minorities among minorities the pooramong the poor has has nonetheless nonetheless increased increased dramatically dramatically —by —by 25 percentage 25 percentage points for the extreme poor (from 43 percent in 1998 to 68 percent in 2010) and 19 1998 to 47 percent in 2010). percentage points for the poor (from 28 percent in72 72 68 Ethnic minori es Majority Figure 3.5 Composition of Poor and Better-off Households in 2010, 100% 90% by Ethnicity Figure 3.5 Composition of Poor and Better-off Households in 2010, by Ethnicity 80% 70% Ethnic minori es Majority 100% 60% 90% 50% 80% 40% 70% 30% 60% 20% 50% 10% 40% 0% 1998 2010 1998 2010 1998 2010 1998 2010 1998 2010 1998 2010 30% Extreme poor All poor Quin le 2 Quin le 3 Quin le 4 Quin le 5 20% 10% Sources: 1998 VLSS and 2010 VHLSS. 0% 1998 2010 1998 2010 1998 2010 1998 2010 1998 2010 1998 2010 3.13 Extreme as Despite progress, poorshown in the Table All poor Quin le 23.3, 66.3 percent Quin le 3 of le Quin minorities Quin still 4 le 5 lived below 37.4 1998 poverty line andSources: percent 2010below lived VLSS and VHLSS. the extreme poverty line in 2010. In comparison, only Sources: percent of the Kinh 1998 VLSS majority and 2010 VHLSS. population was still poor and 2.9 percent lived below the extreme po 3.13 Despite progress, as shown in the Table 3.3, 66.3 percent of minorities still lived below the poverty3.4) ine in 2010. (Table line andBecause Kinh thelived 37.4 percent make below up a much the extreme povertylarger shareInof line in 2010. the population comparison, only in Vie 12.9 percent of the Kinh majority population was still poor and 2.9 percent lived below the extreme 13 still hey account Despite line over for just progress, poverty inas half (53 shown 2010. (Table inpercent) the Table 3.4) Because thethe of Kinhtotal 3.3, 66.3 make poor up a in Vietnam. percent much ofshare larger minorities still lived belo of the population overty line and in Vietnam, they still account for just over half (53 percent) of the total poor in Vietnam. 37.4 percent lived below the extreme poverty line in 2010. In comparison, onl rcent of the Kinh Table majority 3.3Table Ethnic 3.3 Ethnic population Minority Poverty: Minority was Poverty: still poor Headcount Headcount and andand 2.9 percent Composition Composition in in 2010,lived 2010, Regionbelow Region the and and Sector extreme Sector p ne in 2010. (Table 3.4) Because the Kinh Poverty make up a muchExtreme share of the population in Vie larger Poverty Share of ey still account for just over half (53 percent) of the total poor in Vietnam. Contribution Contribution total pop Index (%) to total (%) Index (%) to total (%) (%) Ethnic Minority Poverty: Headcount Table 3.3 National 66.3 and Composition 100.0 37.4 in 2010, 100.0 Region and Sector 100.0 Poverty Extreme Poverty Red River Delta 13.1 0.2 0.0 0.0 1.0 Share of East Northern Mountains 64.8 35.4 34.9 33.9 36.2 West Northern Mountains 72.8 Contribution 18.9 45.5 Contribution 20.9 total pop 17.2 North Central Coast 71.2 to total (%) Index (%) 34.8 to total 12.1 14.0 Index (%) (%) (%)13.0 South Central Coast 78.4 5.3 50.7 6.1 4.5 National Central Highlands 66.3 76.6 100.0 15.2 37.4 50.4 100.0 17.7 100.0 13.1 Southeast 46.4 3.5 22.2 3.0 5.0 Red DeltaDelta River River Mekong 13.1 50.4 0.2 7.6 0.0 23.3 0.0 6.2 1.0 10.0 East Northern Mountains 64.8 35.4 34.9 33.9 36.2 West RuralNorthern Mountains 72.8 68.9 18.9 95.5 45.5 39.3 20.9 96.8 17.2 91.9 North Urban Central Coast 71.2 36.5 14.0 4.5 34.8 14.8 12.1 3.2 13.0 8.1 South Central Coast 78.4 5.3 50.7 6.1 4.5 Source: 2010 VHLSS. Source: Central 2010 VHLSS. Highlands 76.6 15.2 50.4 17.7 13.1 Southeast 46.4 3.5 22.2 3.0 5.0 Mekong River Delta 50.4 7.6 23.3 6.2 10.0 69 Rural 68.9 73 95.5 39.3 96.8 91.9 Urban 36.5 4.5 14.8 3.2 8.1 able 3.4Table Kinh 3.4 Table Kinh Majority Majority Poverty: Poverty: 3.4 Kinh Majority Headcount and Headcount Poverty: and Composition Headcount and Composition Composition in 2010, in in 2010, by by Region 2010, by Region Region and and Sector Sector and Sector Poverty Extreme Poverty Poverty Extreme Poverty Share of Contribution Contribution total popShare of to total (%) Index (%) Index (%) Contribution to total (%) Contribution (%) total pop National Index (%)12.9 to total (%) 100.0 Index (%) to 2.9 total (%) 100.0 (%) 100.0 National Red River Delta 12.9 11.4 100.0 22.9 2.8 2.9 24.7 100.0 26.0 100.0 East Northern Mountains 14.4 8.0 3.3 8.2 7.2 Red RiverWest DeltaNorthern Mountains 10.7 11.4 0.6 22.9 1.3 2.8 0.3 24.7 0.8 26.0 North Central Coast 20.4 18.6 4.9 19.8 11.9 East Northern Mountains South Central Coast 14.4 13.0 8.0 9.2 2.1 3.3 6.5 8.2 9.2 7.2 Central Mountains West Northern Highlands 10.7 12.4 0.6 4.6 1.5 1.3 2.4 0.3 4.8 0.8 North Central Coast Southeast 20.4 6.9 18.6 10.5 2.3 4.9 15.3 19.8 19.7 11.9 Mekong South Central River Delta Coast 16.1 13.0 25.5 9.2 3.3 2.1 22.7 6.5 20.5 9.2 Central Highlands Rural 12.4 17.0 4.6 87.7 3.9 1.5 89.1 2.4 66.6 4.8 SoutheastUrban 6.9 4.8 10.5 12.3 1.0 2.3 10.9 15.3 33.4 19.7 Mekong River Delta Source: 2010 VHLSS. 16.1 25.5 3.3 22.7 20.5 Rural 17.0 87.7 3.9 89.1 66.6 3.14 Urbanbeyond Looking the headcount, 4.8 the 12.3 poverty conditions 1.0 experienced 10.9 by ethnic 33.4 poor are minority more severe Source: than the Source: 20102010 VHLSS. experienced by poor Kinh households. Minorities are more heavily conditions VHLSS. concentrated among 3.14 the Looking extreme beyond poor, asthe the headcount, illustrated in table poverty conditions 3.5, andbyboth experienced ethnicthe depth minority poorand are severity of substantially poverty are more severe than higher for minorities. the conditions experienced These differences by poor are illustrated Kinh households. Minorities aregraphically more heavilyin figure 3.6: ooking beyond the concentrated headcount, among the extreme the poverty poor, conditions as illustrated experienced in table 3.5, and both the depth and the distribution of welfare (per-capita expenditures) for minorities who fall below the poverty line is by ethnic severity of minority poo poverty are substantially higher for minorities. These differences are illustrated graphically in figure ere than skewed the to the 3.6: conditions left theand ofexperienced the overall distribution distribution welfare has (per-capitaby a poor Kinh much thinner expenditures) households. “tailâ€? for minorities than Minorities distribution thebelow who fall the poverty are more of welfare for he Kinhamong ted majorities.the In line is skewedcontrast, extreme poor poor, to the left Kinh and the as have illustrated overall welfarehas distribution levels in table a muchmuch closer 3.5, thinner and to both “tailâ€? thanthe poverty of than poor line the depth the distribution and severi welfare for Kinh majorities. In contrast, poor Kinh have welfare levels much closer to the poverty line ethnic minorities. re substantially higher than poor for minorities. These differences are illustrated graphically in figure ethnic minorities. welfare bution of Table (per-capita Table 3.5 expenditures) Poverty Headcount, for in Gap, and Severity minorities who 2010, Kinh and fall Ethnic below the poverty li Minorities 3.5 Poverty Headcount, Gap, and Severity in 2010, Kinh and Ethnic Minorities o the left and the overall distribution has a much thinner “tailâ€? than the distribution of welfar Headcount PovertyGap PovertySeverity orities. In contrast, poor Kinh Contributio have welfare levels much closer to the poverty line than Contributio norities. ntototal ntototal Contribution Index(%) (%) Index(%) (%) Index(%) tototal(%) Poor: Kinh/Hoa 12.9 53.3 2.7 39.7 0.9 31.1 Table 3.5 Poverty Headcount, Ethnicminorities 66.3 Gap, and 46.7 Severity 24.3 in 2010, Kinh 60.3 and 11.3 Ethnic Minorities 68.9 Headcount PovertyGap PovertySeverity Extremepoor: Kinh/Hoa Contributio 2.9 31.5 0.5 Contributio 21.5 0.1 15.1 Ethnicminorities ntototal 37.4  68.5 9.7 nto78.5 total 3.7 Contribution 84.9  Source: 2010 Source: 2010 Index VHLSS. (%) VHLSS. (%) Index(%) (%) Index(%) tototal(%) Poor: Kinh/Hoa 12.9 53.3 2.7 39.7 0.9 31.1 Ethnicminorities 66.3 46.7 24.3 60.3 11.3 68.9 Extremepoor: Kinh/Hoa 2.9 31.5 70 0.5 21.5 0.1 15.1 Ethnicminorities 37.4 68.5 9.7 78.5 3.7 84.9 Source: 2010 VHLSS. Figur re Figur 3.6 Figure re 3.6 Distrib 3.6bution Distrib bution of of Welf Distribution of fareKinh Welf fare for Welfarefor Kinh for and a and a Ethnic Kinh Ethnic and M M Minorities, Minorities, Minorities, Ethnic 0 2010 20100 2010 Figur bution of Welf re 3.6 Distrib a Ethnic Minorities, fare for Kinh and M 0 2010 3.15 T 3.15 There T are There 3.15 impo There are ortantareimpo ortant nces differen important differen innces thein differences spthe in patial sp patial the distribut distribut tion spatial tionof of Kinh Kinh distribution aand a of and ethnic Kinh mi ethnic miinority inority and popula popula ethnic ations ations minority 3.15 T There are impo ortant differen nces in the sp patial distribut tion of Kinh and a ethnic mi inority popula ations in Vietnamm.inpopulations Vietnam Minority m. Minority populations populations in Vietnam. r Minority remain r heavil remain heavil ly concentrate ly concentrate populations remain ed ed in inthe heavily the t tand East East and West West in concentrated No orthern No orthern the East Moun Mounntains, and ntains, West in Vietnam m. Minority in populations r Highlands, remain heavil ly concentrate ed in the East and t in West Noorthern Moun ntains,In in the Ce inNorthern entral the Highla Ce Mountains, entral ands, Highla ands, and the (tooand Central some(too some exten exten nt) nt) the in and in the N(toNorth NorthN some extent) Central Central l l Coast.the Coast. North In contrast, In c Central ccontrast, theCoast. theKinh Kinh in the Ce contrast, entral the KinhHighla ands, and population is (to o some exten concentrated nt) in in the large North N (including cities Central l Hanoi Coast. and In contrast, cHo Chi Minh the KinhCity), population npopulation is concentr n israted concentr inrated large e large incities e cities (inclu (inclu uding uding Hanoi Hanoi a and and a Ho Ho ChiMinh Chi M Minh M City), City), thetthe t Red Red Rive er er Rive and and population the Red n is concentr rated in deltas, large e cities (inclu Hanoi and udingelevations a along Ho Chi Minh M City), t inland the Rive Red areas. er and dMekong Mekong deltas, andd River deltas, in andin and n lower nMekong lower elev vationselev vations and along alongin lower the the coast coast annd an nd inlandare inland eas.coast the are eas. The spat The and spattial tial distributi ion distributi ion The of of Mekong spatial d deltas, distribution and of inn lower elev poverty vations tends to along the follow the spatial coast an nd inland are distribution eas. of Therespective spat tial distributi their ion of populations: poverty te poverty ends te ends to follow wto follow the w the spatial spatial distributiondistribution o their of of o theirrespecrespec ctivective populat populat ions: poor ions: poor Ki Kiinh inhhousehold household ds ds areare poverty poor Kinhte ends to follow households ware spatial distribution the concentrated in the odeltas of their respec and in ctive populations: provinces along poor the Ki inh household North Centralds are Coast. concentra concentra ated in the delated in the delltas ltas and in pro and in pro ovinces alongovinces along g the North C entral Coast. g the North Central Coast. In contrast, most In contrast, most mm poor min poor min nority nority In contrast, concentra most ated in the del poor minority ltas and in pro households ovinces along live in upland areas, with g the North Central Coast. In contrast, mostthe West Northern m poor min Mountain nority household dshousehold live in ds uplive pland in up pland areas, areas, with the with West Wthe WestWNorthern Northern n Mountain region n Mountain region and and Central CCentral C High hlands High hlands region householdand live in Highlands ds Central uppland areas, accounting with the for W a somewhat West Northern nhigher Mountain share of poor region and C minorities ethnic Central High than hlands accountinngaccountin for their ng ashare some for ewhat in athesome ewhatr higher higher population. share r share of po Notably, of oor po acrossoor all ethnic ethnicmi mi inorities locationsinorities (with than thannntheir the their share of share exception in the inRedthe popul lation. popul River lation. Delta, accountin ng for a some ewhat higher r share of po oor ethnic mi inorities than n their share in the popul lation. Notably, Notably, acrosswhere all across loc very all ethnic cations few loccations (with (with the minorities the exceptio exceptio n of reside), nRed povertyof RedRiv Riv ver rates verDelta, amongDelta, whwh here minorities here ethnic very few very w w ethnic few ethnic mino orities mino orities Notably, across all loc cations (with the exceptio n of Red Riv ver Delta, wh here very few w average ethnic mino between orities reside), po reside), overty four and po a rates overty rates among seven times a higher among ethnic thanminorities ethnic minorities av poverty verage av verage rates betwe among betwe enthe en four four and and s seven Kinh (figures s 3.7 seven times and higher times h than h Majorities higher 3.8). thanpooverty po overty living reside), po overty rates among a ethnic minorities av verage between four and seven s h times higher than po overty rates amoongratestheamo ong Kinh the Kinh (figures (figures 3.7 and 3.7 and 3.8). 3.8). Ma Ma ajoritiesajoritieslivin livin g g in in minority minority in minority areas have substantially better living conditions on average than the minorities living y y areas areas have have s substantially s substantially better better in rates amo ong the Kinh (figures 3.7 and 3.8). Ma ajorities living in minority s y areas have substantially better living con living nditions these con onnditions sameavverage on av areas. verage than th hethan thhe minorities minorities living l l in these living in these e same e same areas. areas. living con nditions on av verage than th he minorities livingl in these e same areas. ure Figure Figu 3.7Figu 3.7 ure Level aLevel 3.7 andLevel and a Composition and Composit Composit ofy Poverty y ion of Poverty ion of Poverty Figure FFigure igure 3.8 3.8 3.8 Lev Level Lev vel vel and and and Compo CompoComposition osition ositionofof Pove of erty Pove erty Figuure by Regi 3.7 ion, by Level Regi for and a Composit ion, for byRegion, Kinh/H Hoa forKinh/H ion Hoa Kinh/Hoa of Poverty y F igure byby Poverty 3.8 Regio by Lev Regio on,vel on,for Region,and Compo Ethnic for Ethnic osition Minorities Minorities for Ethnic of Pove erty Minorities by Regi Hoa ion, for Kinh/H by Regioon, for Ethnic Minorities 80.0 80.0 0 80.0 0 80.0 80.0 0 80.0 70.0 70.0 0 70.0 0 70.0 70.0 Incidence Incidence 0 70.0 Incidence Incidence 60.0 60.0 0 60.0 0 60.0 Contribu on total on to total to Incidence Contribu on to on Contribu Contribu Incidence to total total 60.0 0 60.0 Contribu on to total Contribu on to total 50.0 50.0 0 50.0 0 50.0 50.0 0 50.0 40.0 40.0 0 40.0 0 40.0 40.0 0 40.0 30.0 30.0 0 30.0 0 30.0 30.0 0 30.0 20.0 20.0 0 20.0 0 20.0 20.0 0 20.0 10.0 10.0 0 10.0 0 10.0 10.0 0 10.0 0.0 0.0 0 0.0 0 0.0 ver Red Riv East Red Riv West East North ver West South North Central South Southeast Central M Mekong Southeast M Mekong Red River Red River EastEast West West North North South South CentralSoutheast Central Southeast Mekong M Mekong M Delta 0.0 Delta Northern Northern Central Northern Northern Central Central Highlands Central Highlands River Delta River Delta 0.0 0 Delta Northern Delta NorthernNorthern Northern Central Central Central Central Highlands l l Highlands River Delta River Delta Red Riv Mountains Mountainsver s East Mountains Coast West Mountainss Coast North Coast Coast South Central Southeast M Mekong Red River Mountains East Mo Mountains ountains Mo West Coast ountains North Coast South Coast Coast Central M Southeast Mekong Delta Northern Northern Central Central Highlands River Delta Delta Northern Northern Central Centrall Highlands River Delta Mountains Mountainss Coast Coast Mountains Mo ountains Coast Coast Source: 20 010 Source: Source: VHLSS. 2010 20 010 VHLSS. VHLSS. Source: Sou urce: Sou 2010 urce: 2010 VHLSS. VHL 2010 LSS. VHLLSS. Source: 20 010 VHLSS. Souurce: 2010 VHL LSS. 3.16 Maps 3.1 and 3.2 illustrate the strong spatial segregation between poor minority and poor 75 75 majority households in Vietnam. Poor minorities75 are heavily concentrated in the East and West Northern Mountains, upland areas in the North Central Coast, and the Central Highlands. In contrast, 71 3.16 Maps 3.1 and 3.2 illustrate the strong spatial segregation between poor minority and poor majority households in Vietnam. Poor minorities are heavily concentrated in the East and West Northern Mountains, upland areas in the North Central Coast, and the Central Highlands. In contrast, poor people from the majority population are concentrated in the Red River Delta, along coastal regions, and in the Mekong Delta. poor people from the majority population are concentrated in the Red River Delta, along coastal regions, and in the Mekong Delta. 3.1 Spatial Map Map Distribution 3.1 Spatial of Poor Distribution of Minorities Poor Minorities Map Map 3.2 3.2 Spatial Spatial Distributionof Distribution Poor Kinh ofPoor Kinh Sources: Cuong et al. 2012. Sources: Cuong et al. 2012. 3.17 There are important differences in livelihood strategies and employment patterns between 3.17 poorThere are important majority differences and minority households livelihood in(Figure 3.9). strategies and employment Poor minorities of between patterns earn three-quarters their total poor majority and minority income households from agriculture (Figure and allied 3.9). including activities, Poor minorities earn three-quarters wage employment of their in agriculture. total income In contrast, poor majority households earn only 42 percent from agriculture and allied activities from agriculture and allied activities, including wage employment in agriculture. In contrast, and a much higher poor share from off-farm activities, both salaried non-farm employment and family enterprises. Forestry majority households earn only 42 percent from agriculture and allied activities and a much higher share is important for minorities, but much less so for poor majorities, in large part reflecting differences from off-farm activities, both salaried non-farm employment and family enterprises. Forestry is important in residential patterns. Notably, the composition of income is similar between ethnic minorities and majorities but for minorities, much in the less so wealthiest for poor majorities, in large part reflecting differences in residential quintile. patterns. Notably, the composition of income is similar between ethnic minorities and majorities in the wealthiest quintile. 76 72 Figure 3.9 Composition of Income for Extreme Poor, Poor, and Top Quintile in 2010: Figure 3.9 Composition of Income for Extreme Poor, Poor, and Top Quintile in 2010: Comparing Comparing Kinh/Hoa Kinh/Hoa andand Ethnic Ethnic Minority Households Minority Households Source: 2010 VHLSS. Source: 2010 VHLSS. E. Poverty is Still Linked to Low Education Attainment E. Poverty is Still Linked to Low Education Attainment 3.18 3.18 Vietnamese Vietnamese today are today farare far better better educated than educated they were than they a were decade ago. Primary completion a decade rates ago. Primary completion were high already by the end of the 1990s, as evidenced in the first panel of Figure 3.10. Since then, the rates were high already by the end of the 1990s, as evidenced in the first panel of Figure 3.10. Since other panels illustrate the rapid increase in enrolments at lower and upper secondary levels, leading to an then, the other panels illustrate the rapid increase in enrolments at lower and upper secondary levels, increase in the number of students who attend colleges and universities. However, lack of education leading to an increase continues to bein an the number important of students determinate of poverty, who attend and this colleges was highlighted and universities. by respondents in both However, lack of education continues urban and rural areas asto be an a cause important of rising inequalitydeterminate (Chapter 6). of poverty, and this was highlighted by respondents in both urban and rural areas as a cause of rising inequality (Chapter 6). Figure 3.10 Schooling Achievement by Age Cohort, 1998 and 2010 Completed Primary Completed Lower Secondary 100 1998 100 1998 90 90 2010 2010 80 80 70 70 60 60 Percent Percent 50 50 40 77 40 30 30 20 20 10 10 0 0 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ Age Age Completed Upper Secondary Completed University 100 100 1998 1998 90 90 2010 2010 80 80 70 70 60 60 Percent Percent 50 50 40 40 30 30 20 20 10 10 0 0 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61+ Age Age Source: 1998 VLSS, 2010 VHLSS. 73 3.19 As shown in Table 3.6, individuals living in households whose head did not complete primary school have the highest poverty rate in 2010 (nearly 40 percent or twice the national average) as well as the highest extreme poverty rate (nearly 19 percent or two-and-a-half times the national average). The inverse relationship between education and poverty has become stronger over time: in 1998, households whose heads had completed primary or less schooling accounted for 55 percent of the total poor. By 2010, they accounted for 75 percent of the poor. Rising levels of education coupled with rapid income diversification has been a powerful force for poverty reduction in Vietnam since the late 1990s. Table 3.6 3.6 Table Poverty Headcount Poverty andComposition and Headcount in2010, 2010, Composition in by by Education Education of Household of Household Head Head Poverty ExtremePoverty Shareof Contribution Contribution totalpop Index(%) tototal(%) Index(%) tototal(%) (%) National 20.7 100.0 8.0 100.0 100.0 Householdhead's  highesteducational qualification: None 39.8 46.1 19.3 58.1 24.0 Primary 23.5 28.5 7.9 25.0 25.1 Lowersecondary 15.3 18.4 4.2 13.2 24.9 Uppersecondary 8.7 4.2 2.1 2.6 9.9 Vocational 5.8 2.6 0.8 0.9 9.4 Highereducation 0.7 0.2 0.1 0.1 6.6 Source: 2010 VHLSS. Source: 2010 VHLSS. 3.20 Table 3.7 describes the distribution of education for persons 21 years and older across 3.20 Table 3.7 describes the distribution of education for persons 21 years and older across expanded expanded per-capita expenditure quintiles, illustrating in yet another way the strong relationship per-capita expenditure between risingquintiles, illustrating levels of education another in yet wealth and increasing way the in Vietnam. strong By 2010, 40 relationship between rising percent of persons 21 yearsand evels of education increasing and older wealth in the richest in Vietnam. quintile By a had completed 2010, percent 40 degree; university inof persons contrast, less21 years than 2 and older percent in the poorest quintile were university graduates. In fact, more than a quarter of those in the n the richest quintile had completed a university degree; in contrast, less than 2 percent in the poorest poorest quintile had not even completed primary school by 2010. quintile were university graduates. In fact, more than a quarter of those in the poorest quintile had not 3.21 Table 3.7 also highlights the gaps in education between ethnic minorities and Kinh majorities. ven completed primary school by 2010. Even among the poor, minorities are substantially less educated than their Kinh economic peers: for example, 39 percent of poor minorities had not completed primary school compared to only 16 percent Table of poor Kinh of 3.7 Distribution majorities. CompletedAchievement gaps Education are in in part 2010, by due to a historical Ethnicity legacy of lower and Expanded Quintiles education achievement among many minority populations, but also reflect lower (albeit increasing) (persons age 21 and older) current enrolment rates. Figure 3.11 illustrates the relationship between education and total per- Lower in Table capita expenditures for Kinh and minorities documented 3.7. Upper Higher None Primary secondary secondary Vocational education National Extreme Poor 37.1 28.3 23.4 9.3 1.2 0.7 All Poor 26.7 29.7 28.7 12.3 1.3 1.4 Quintile 2 12.4 26.6 34.7 20.7 3.4 2.3 Quintile 3 6.6 21.6 31.8 27.0 6.1 6.9 Quintile 4 4.7 14.2 23.1 30.3 9.8 17.8 Quintile 5 2.0 7.7 15.6 25.6 9.2 40.0 Rural 13.1 23.1 30.6 21.9 4.7 6.7 Urban 4.7 12.5 17.6 25.9 9.0 30.3 National 10.6 20.0 74 26.7 23.1 5.9 13.7 Majority ven completed primary school by 2010. Distribution Table 3.7 Table of Completed 3.7 Distribution Education of Completed 2010, in2010, Education in byby Ethnicity Ethnicity and Expanded and Expanded QuintilesQuintiles (persons age 21 and older) (persons age 21 and older) Lower Upper Higher None Primary secondary secondary Vocational education National Extreme Poor 37.1 28.3 23.4 9.3 1.2 0.7 All Poor 26.7 29.7 28.7 12.3 1.3 1.4 Quintile 2 12.4 26.6 34.7 20.7 3.4 2.3 Quintile 3 6.6 21.6 31.8 27.0 6.1 6.9 Quintile 4 4.7 14.2 23.1 30.3 9.8 17.8 Quintile 5 2.0 7.7 15.6 25.6 9.2 40.0 Rural 13.1 23.1 30.6 21.9 4.7 6.7 Urban 4.7 12.5 17.6 25.9 9.0 30.3 National 10.6 20.0 26.7 23.1 5.9 13.7 Majority Extreme Poor 21.7 25.1 33.6 16.1 2.5 1.0 All Poor 16.4 31.2 34.5 14.2 1.8 2.0 Quintile 2 10.7 26.2 36.0 21.2 3.3 2.6 Quintile 3 6.3 21.6 32.2 27.0 6.0 6.9 Quintile 4 4.5 14.6 23.4 30.3 9.8 17.4 Quintile 5 2.0 7.8 15.7 25.6 9.0 39.9 Ethnic minorities 3.21 Table 3.21 Table3.7 Extreme 3.7 Poor also also highlights highlights 44.2 the gaps the gaps in 29.8 in education education between between 18.7 ethnic minorities 6.1 minorities ethnic and Kinh 0.6and Kinh 0.6 majorities. majorities. Even Evenamong among the Allthe poor, Poorpoor, minorities minorities 38.6 are are substantially substantially 28.0 less less 21.9 educated educated than 10.1 than their their KinhKinh 0.9 peers:peers: economic economic 0.6 for for example, 39 example, 39 percentof percent Quintile 2 poor minorities of poor 23.3 had minorities had28.5not not completed 25.8 completed primary primary 17.5school school compared compared 3.9 only to 16 to only 0.9 16 percent percent of of Quintile 3 12.2 21.5 25.3 26.1 8.2 6.8 poor Kinh poor Kinh majorities. Achievement majorities. Achievement gaps gaps7.2 are in are in partpart due to a historical due to a historical legacy legacy of of lower lower education education Quintile 4 9.3 18.3 29.0 10.0 26.3 achievement achievement among among Quintile 5 many many minority minority 4.2 populations, populations, 1.7 butbut 9.2 also also reflect reflect 23.0 lowerlower (albeit (albeit 17.1 increasing) increasing) 45.0 currentcurrent enrolmentrates. enrolment rates. Figure 3.11 illustrates the therelationship relationship between education total total and and per-capita Source: VHLSS. illustrates 2010 Figure Source: 2010 VHLSS. between education per-capita expenditures for Kinh and minorities documented in Table expenditures for Kinh and minorities documented in Table 3.7. 3.7. Figure 3.11 Education Achievements by Expanded Quintiles (persons age 21 and older) Figure 3.11 Figure 3.11 Education Achievements Education Achievements by by Expanded Expanded Quintiles Quintiles (persons (persons age age 21 21older) and older) and Kinh/Hoa Kinh/Hoa Kinh/Hoa 79 Ethnic Minorities Ethnic Ethnic Minorities Minorities Source: Source: 2010 2010 VHLSS. VHLSS. Source: 2010 VHLSS. 3.22 High 3.22 Highlevels of current levels of enrolmentsindicate current enrolments indicatethat future that generations future of workers generations will be of workers better will be better prepared to participate in Vietnam’s modernizing prepared to participate in Vietnam’s modernizing economy than previous generations. However,in economy than previous generations. However, gaps gaps in enrolments between children from poor and better-off 75 households have persisted (Table 3.8), enrolments between children from poor and better-off households have persisted (Table 3.8), includingincluding gaps gaps between between enrolments for enrolments for Kinh andethnic Kinh and ethnicminority minoritychildren. (Table children. 3.9)3.9) (Table Most primary-school-aged Most primary-school-aged and poor, children—richand children—rich minority and poor, minority and majority—are majority—are enrolled in school. enrolled enrolments But But in school. among enrolments (poor)(poor) among 3.22 High levels of current enrolments indicate that future generations of workers will be better prepared to participate in Vietnam’s modernizing economy than previous generations. However, gaps in enrolments between children from poor and better-off households have persisted (Table 3.8), including gaps between enrolments for Kinh and ethnic minority children. (Table 3.9) Most primary-school-aged children—rich and poor, minority and majority—are enrolled in school. But enrolments among (poor) minorities drop off at the lower secondary level, and children from lower- income households are much less likely to be enrolled in upper secondary schools than children from better-off households. Chapter 6 analyzes the links between education and rising inequality, including the role of inequality in opportunities (especially education) in perpetuating poverty across generations. Table 3.8 School Enrolment Rates (net) for Boys and Girls in 2010, Table 3.8 School Enrolment Rates (net) and Girls for BoysQuintiles by Expanded in and 2010, by Expanded Quintiles and Region Region Primary Lower Secondary Upper Secondary Male Female Total Male Female Total Male Female Total Extreme Poor 91.6 88.8 90.2 62.2 70.8 66.6 16.4 28.1 22.9 All Poor 90.2 90.2 90.2 68.6 75.6 72.2 28.1 36.1 32.4 Quintile 2 93.7 92.6 93.2 77.5 82.6 79.9 50.0 56.5 53.0 Quintile 3 94.1 92.9 93.5 84.9 85.5 85.2 58.1 62.5 60.3 Quintile 4 92.5 93.7 93.1 90.5 90.4 90.5 66.0 73.6 69.5 Quintile 5 93.3 97.6 95.3 86.1 90.3 88.0 76.2 85.6 80.9 Red River Delta 95.0 93.5 94.3 89.6 91.9 90.6 69.2 67.2 68.2 East Northern Mtns 93.0 90.9 91.9 85.2 83.0 84.1 56.0 60.7 58.3 West Northern Mtns 93.3 93.9 93.6 80.9 65.5 74.2 47.4 38.8 42.7 North Central Coast 90.9 91.1 91.0 83.8 87.6 85.8 54.7 58.9 56.8 South Central Coast 92.1 90.7 91.4 89.5 86.4 88.1 58.4 69.6 64.0 Central Highlands 95.4 87.7 91.9 67.3 78.2 73.1 45.6 52.5 49.3 Southeast 90.3 97.9 94.1 76.1 81.8 78.4 52.8 63.1 58.0 Mekong Delta 91.4 92.7 92.0 66.1 76.5 71.2 39.2 50.5 44.1 Rural 92.4 91.9 92.2 78.9 82.8 80.7 49.3 54.5 51.8 Urban 92.9 95.2 94.1 83.5 85.0 84.2 68.8 76.2 72.5 National 92.5 92.8 92.6 80.0 83.3 81.5 53.9 60.1 57.0 Source: Source: 2010 2010 VHLSS. VHLSS. 3.23 Gender gaps in minority school enrolments have received a lot of attention in Vietnam. These Table 3.9 Net gaps School have Enrolment closed Rates at the primary Kinh/Hoa forpersist level but and Ethnic at the secondary Minority level Boys and above. and reverse Girls in 2010, However, by Expanded Quintile gender gaps—substantially higher enrolments for girls compared to boys at the secondary level— have started to emerge at the secondary level, particularly among children from poor (majority) Primary Lower Secondary Upper Secondary households and in the Central Highlands, the Southeast, and the Mekong Delta. Concerns have Male Female Total Male Female Total Male Female Total been raised that boys from poor households are leaving school earlier than girls to take up jobs in Majority the service sector and manufacturing, “pushedâ€? by poverty and economic imperatives and “pulledâ€? Extreme Poor employment by expanding 96.4 92.4opportunities 94.5 69.7 and94.1 in nearby cities 81.8 leaving towns. While 27.6 school48.5 after six39.9 or All Pooryears of education eight 88.3 may make94.2sense 91.0 71.9 incentives, given short-run 85.8 79.5 34.2 made education choices 46.4 today40.8 Quintile 93.2 2 children for the will follow 92.1 rest of 92.7 These young their lives. 84.2 may 75.7 workers 79.6 not have50.7 57.7 and the education 54.0 Quintile 3 get good jobs 93.8 skills to 93.4 93.0 as Vietnam’s in the future 85.2 economy 85.7 continues85.4 58.1modernize, 63.3 and to grow and 60.7 Vietnam’s Quintile 4 economic development 92.4 94.6will be constrained 93.5 91.0by the lack of90.7 90.5 an educated 66.7 75.4 labor and skilled 70.7 force. Quintile 5 93.2 97.5 95.3 86.0 90.2 87.9 76.8 85.3 81.0 Ethnic minorities Extreme Poor 91.4 86.1 88.7 59.4 62.5 61.0 12.4 19.2 16.1 All Poor 92.5 86.5 89.3 65.5 63.1 64.4 22.4 26.3 24.5 Quintile 2 97.4 96.1 96.8 76 90.1 72.2 81.6 46.1 48.3 47.1 Quintile 3 100.0 90.5 95.4 78.0 82.1 80.3 57.9 43.4 53.1 Quintile 4 94.5 74.9 85.5 80.1 88.9 84.4 58.4 41.2 52.3 National 92.5 92.8 92.6 80.0 83.3 81.5 53.9 60.1 57.0 Source: 2010 VHLSS. Table 3.9 Net 3.9 Table School Enrolment Net School Rates Rates Enrolment for Kinh/Hoa andand for Kinh/Hoa Ethnic Minority Ethnic Boys Minority and Boys andGirls Girls in 2010, by Expanded Quintile in 2010, by Expanded Quintile Primary Lower Secondary Upper Secondary Male Female Total Male Female Total Male Female Total Majority Extreme Poor 92.4 96.4 94.5 69.7 94.1 81.8 27.6 48.5 39.9 All Poor 88.3 94.2 91.0 71.9 85.8 79.5 34.2 46.4 40.8 Quintile 2 93.2 92.1 92.7 75.7 84.2 79.6 50.7 57.7 54.0 Quintile 3 93.8 93.0 93.4 85.2 85.7 85.4 58.1 63.3 60.7 Quintile 4 92.4 94.6 93.5 91.0 90.5 90.7 66.7 75.4 70.7 Quintile 5 93.2 97.5 95.3 86.0 90.2 87.9 76.8 85.3 81.0 Ethnic minorities Extreme Poor 91.4 86.1 88.7 59.4 62.5 61.0 12.4 19.2 16.1 All Poor 92.5 86.5 89.3 65.5 63.1 64.4 22.4 26.3 24.5 Quintile 2 97.4 96.1 96.8 90.1 72.2 81.6 46.1 48.3 47.1 Quintile 3 100.0 90.5 95.4 78.0 82.1 80.3 57.9 43.4 53.1 Quintile 4 94.5 74.9 85.5 80.1 88.9 84.4 58.4 41.2 52.3 Quintile 5 100.0 100.0 100.0 100.0 100.0 100.0 25.7 100.0 75.1 Source: Source: 2010 2010 VHLSS. VHLSS. 3.24 There are many reasons why children from poor and ethnic minority households do not stay in 3.23 Gender Highin gaps school. minority costs out-of-pocket schoolareenrolments have received one factor (Chapter 1). Location lot a is of attention another. In uplandin Vietnam. These regions, gaps have closed at in particularly the the Northern level primary persist but upper Mountains, at the secondary secondary levellocated schools are often and above. at someHowever, remove reverse from rural communities, and students are forced to board rather than commute to school each day gender gaps—substantially higher enrolments for girls compared to boys at the secondary level—have from their homes. Background qualitative studies carried out for this report also highlight widespread started to emerge concerns the secondary at about level, the poor quality particularly of schools among in some children from poor (majority) households and rural areas. in the Central Highlands, the Southeast, and the Mekong Delta. Concerns have been raised that boys from Vietnamese Farmers have Small Landholdings and Landlessness is Rising poor households are leaving school earlier than girls to take up jobs in the service sector and 3.25 “pushedâ€? manufacturing, An early andby strong poverty and economic commitment imperatives by the government to and “pulledâ€? distribute by rights expanding land use employment equitably among farmers in Vietnam has resulted in a pattern of land distribution that remains remarkably opportunities in nearby cities and towns. While leaving school after six or eight years of education may equitable by international standards. Rural growth and on-farm diversification were the driving forces for poverty reduction in the 1990s. Most rural households continue to have small landholdings and, in recent years, few households were able to substantially improve their living conditions through 81 expanded cultivation of annual crops. A high percentage of Vietnamese farmers continue to grow rice, in part driven by state restrictions on the use of land. Land use restrictions are primarily in place for rice production, and affect land in the Mekong and Red River Deltas (Markussen, Tarp, and van den Broeck 2009). Except in the Mekong Delta, rice is grown primarily for own consumption rather than as a source of cash income. 72 percent of poor households in Vietnam grew rice according to the 2008 VHLSS; 90 percent of this rice consumed at home, and only 18 percent of poor households were net sellers of rice. Instead, rising wealth among rural households is linked to on-farm diversification into cash crops, and even more important, diversification into off-farm activities. The last decade is notable for rapidly expanding opportunities for stable off-farm income generation, including in industrial centers and nearby towns. 3.26 Less-well-off rural households cultivated, on average, more land than better-off rural households in 2010 (Table 3.10). However, these statistics should be interpreted with care; much of the land cultivated by ethnic minorities is in upland regions and often of lower quality due to sloping and rocky terrain and lack of dependable irrigation. Better-off households cultivate more perennial cropland, which is used for commercial activities (including coffee, an important cash crop). 77 cultivated by ethnic minorities is in upland regions and often of lower quality due to sloping and rocky terrain and lack of dependable irrigation. Better-off households cultivate more perennial cropland, which is used for commercial activities (including coffee, an important cash crop). Table 3.10 Table 3.10 Average Landholdings for Average Landholdings Households in Rural Households for Rural 2010, by in 2010, by Consumption ConsumptionQuintile Quintile Quintile 1 2 3 4 5 3.27 The proportion of landless rural households has risen in all regions since the late 1990s (Table All land (sq. m.) 8235 6049 5901 5723 5608 3.11). However, with the exception of the Mekong Delta, landlessness is not associated with higher of which: 3.27 The proportion of landless rural households has risen in all regions since the late 1990s (Table poverty. In fact, Annual initial crop analysis land suggests a positive relationship 3765 3322 between 2927 rural landlessness 2826 and wealth in 2302 3.11). However, with the exception of the Mekong Delta, landlessness is not associated with higher most regions in the north Perennial of Vietnam. 698 land analysis (Table 3.12). 1031But 54 percent 1145 of the rural poor living 2463 in the poverty. In fact, initial suggests a positive relationship between rural landlessness and1640 wealth in Southeast Source:region Source: 2010 and VHLSS. 2010 VHLSS. 48 percent of the rural poor living in the Mekong Delta are landless (landless rates most regions in the north of Vietnam. (Table 3.12). But 54 percent of the rural poor living in the among extreme poor are similar). Concerns have been raised over the years about the links between 3.27 Southeast region and of The proportion 48 landless percent of the ruralrural poor living in households has Mekong the risen inDelta are landless all regions since(landless rates the late 1990s (Table landlessness and poverty. Some were concerned that legislation allowing the opening up of land markets 3.11). extremewith However, among are exception poor the of the similar). Concerns Mekong have Delta, been raised landlessness over the years about notlinks is the associated between with higher late-1990s in thepoverty. In fact, would encourage initial analysis poor a suggests farmers positiveto for quick sell land between relationship ruralprofits, leaving landlessness and themwealthwithout landlessness and poverty. Some were concerned that legislation allowing the opening up of land markets adequate means of livelihoods; others argued that land markets would promote greater efficiency. in most regions in the north of Vietnam. (Table 3.12). But 54 percent of the rural poor living in the in the late-1990s Southeast region andwould 48 encourage percent poor of the rural to farmers sell living poor land forinquick the profits, leaving Mekong Deltathem are without landless (landless (Ravallion and Van der Walle 2008a, 2008b) The picture 82 have is mixed. Respondents living in Tra Vinh amongmeans ratesadequate extreme poor are similar). Concerns been raised over of livelihoods; others argued that land markets would promote greater efficiency. the years about the links province in the Mekong Delta interviewed for the positive deviance between landlessness and poverty. Some were concerned that legislation allowing the opening study (Chapter 1) noted expanding up (Ravallion and Van der Walle 2008a, 2008b) The picture is mixed. Respondents living in Tra Vinh opportunities for “land-poorâ€? of land markets households in the late-1990s would in the Mekong encourage poor and Southeast farmers to sell tolanddiversify for quick into higher profits, paid off- leaving themprovince without farm activities. the Mekong inadequate However, Delta interviewed means off-farm of for the others livelihoods; diversification positive argued studyland deviance that requires relevant (Chapter 1) noted markets education and expanding would promote skills. Althoughgreater young opportunities efficiency. for “land-poorâ€? (Ravallion and Van households in the Mekong der Walle 2008a, and Southeastpicture 2008b) The to diversify into higher paid off- is mixed. Respondents living in workers can acquire Tra Vinh these provinceHowever, skills, the in the Mekong situation is more Delta interviewed complicated for households with older workers. farm activities. off-farm diversification requires for the positive relevant deviance education and study (Chapter skills. Although young 1) noted work is needed More expanding to understand opportunities the complex for “land-poorâ€? households linksin between the Mekong landlessness and Southeast and poverty in Vietnam’s to diversify into higherworkers can acquire paid off-farm these skills, the situation activities. However, is morediversification off-farm complicated for households with oldereducation requires relevant workers. and skills. southern provinces. More work Although young is needed workers to understand can acquire the complex links between these skills, landlessness the situation and poverty is more in Vietnam’s complicated for households with older workers. southern provinces. More work is needed to understand the complex links between landlessness and a,b poverty Table 3.11 Percentage in Vietnam’s of Rural Households without Allocated or Swidden Land southern provinces. 1993 1998a,b 2010 Table Table 3.11 3.11 Percentage Percentage Northern Mountains ofof Rural Rural Households Householdswithout without Allocated 2.0 or Swidden Land Allocated or 3.7 Swidden Land, 8.1 Red River Delta 1993 3.2 1998 4.5 2010 13.4 North Central Northern Coast Mountains 2.0 3.8 3.7 7.7 8.1 15.5 South Central Red River DeltaCoast 3.2 10.7 4.5 5.1 13.4 19.7 Central Highlands North Central Coast 3.8 3.9 7.7 2.6 15.5 17.3 Southeast South Central Coast 10.7 21.3 5.1 28.7 19.7 58.9 Mekong Delta Central Highlands 3.9 16.9 2.6 21.3 17.3 33.6 Southeast 21.3 28.7 58.9 National Mekong Delta 16.9 8.2 21.3 10.1 33.6 22.5 Source: 1993 and 1998 figures taken from the World Bank 2000 Vietnam Development Report, table 2.4. 2010 figures are World Bank estimates from 8.2 National 2010 VHLSS. 10.1 22.5 a. Swidden Source: 1993 Source: and 1998 land1993 figuresis and land 1998 taken figures cleared from thetaken for from the World cultivation World Bank by 2000 Bank 2000 Vietnam cutting Vietnam and Development burning Development the Reporttable , table 2.4. 2010 figures are vegetation. Report, World Bank estimates b. “Landâ€? 2010 2.4.from figures includes are Worldcropland, 2010 VHLSS. annual Bank estimates from 2010 VHLSS. perennial cropland, forestry land, water surface, and shifting- Note: Swidden land is land cleared for cultivation by cutting and burning the vegetation. a. Swidden cultivation farmland. land land is It cultivation ponds, cleared forgardens, excludes by cutting and burning and the vegetation. land classified as “other.â€? “Landâ€? includes annual cropland, perennial cropland, forestry land, water surface, and shifting-cultivation farmland. It b. “Landâ€? includes annual cropland, excludes gardens, ponds, and land classified as “other.â€? perennial cropland, forestry land, water surface, and shifting- cultivation farmland. It excludes gardens, ponds, and land classified as “other.â€? Table 3.12 Table 3.12Percent Percentof ofRural RuralHouseholds without Allocated Households without orSweden Sweden Allocatedor Land Land inin 2010, 2010, by Per-Capita by Region and Region and Expenditure Quintile Quintile Table 3.12 Percent of Rural Households without Allocated or Sweden Land in 2010, Quintile by Region and Per-Capita Expenditure Quintile Extremepoor 1 2 3 4 5 RedRiverDelta 2.2 4.6 Quintile 4.8 7.9 14.6 30.5 EastNorthernMountains Extremepoor 0.7 1 2.2 2 4.8 3 4 9.6 5 20.9 31.4 RedRiver WestNorthern Delta Mountains 2.2 0.5 4.6 0.6 4.8 5.37.9 14.6 5.5 30.5 38.7 56.9 EastNorthern NorthCentral Coast Mountains 0.7 7.9 2.2 7.9 4.8 9.99.6 20.9 14.9 31.4 22.6 52.0 West SouthCentral Coast NorthernMountains 0.5 2.5 0.610.6 5.3 14.65.5 38.7 16.7 56.9 21.7 25.3 North CentralHighlands Central  Coast 7.9 13.2 7.9 9.6 9.9 17.0 14.9 22.6 27.6 52.0 21.1 23.9 Southeast SouthCentralCoast 2.5 43.4 10.653.9 14.6 43.4 16.7 21.7 53.6 25.3 56.5 68.5 Central Mekong River Highlands Delta 13.2 50.3 9.647.5 17.0 29.0 27.6 21.1 29.7 23.9 30.6 34.9 Southeast Source: 2010 VHLSS. 43.4 53.9 43.4 53.6 56.5 68.5 Source: 2010 VHLSS. MekongRiverDelta 50.3 47.5 29.0 29.7 30.6 34.9 Source: 2010 VHLSS. 78 83 means toward achieving better living standards. For example, access to sanitation interacts with health care, good nutrition, and water supply to influence the health of individuals. Homes built with more durable building materials provide safer shelter and reduce labor costs for repairs and new construction. 3.29 Vietnam F.  has achieved Housing and Local widespread improvements Infrastructure have Improvedin the quality of housing Substantially since andthe access to infrastructure in recent Late 1990syears. These are evident in recent rounds of the VHLSS, and were also reported in supporting field 3.28 studies. Housing For example, conditions respondents are an important in the measure long-run of quality drivers of life, both asof poverty ends reduction study in themselves andas (Nguyen and Hoang means 2012) toward describe achieving substantial better livingimprovements in rural infrastructure standards. For example, access to sanitationsince the early 1990s interacts with health care, good nutrition, and water supply to influence the health and increased access to associated social and economic services, markets, and information. These of individuals. Homes built include with more durable building materials provide safer shelter and reduce labor costs for repairs and new better road and bridge access for rural communes and remote villages, new irrigation facilities, and a construction. rapid expansion of media services and technologies into rural areas. Associated with this, many 3.29 Vietnam has achieved widespread improvements in the quality of housing and access to households have invested infrastructure in recent in years. new These of assets types are evident that improve in recent rounds mobility and access of the VHLSS, and were to also information, motorbikes, includingreported TVs, field in supporting mobile phones, studies. and even For example, computers respondents in long-run in the urban areas. drivers These of povertywidespread reduction study (Nguyen and Hoang 2012) describe substantial improvements improvements in economic and social infrastructure have resulted from the combined efforts of many in rural infrastructure since the early 1990s and increased access to associated social and economic services, markets, and government infrastructure information. investment These include better programs across road and bridge the different access infrastructure for rural communes sectors, and remote and provide a villages, good foundation for growth of the rural economy and continued reductions in rural poverty in the coming new irrigation facilities, and a rapid expansion of media services and technologies into rural areas. years. Associated with this, many households have invested in new types of assets that improve mobility and access to information, including motorbikes, TVs, mobile phones, and even computers in urban 3.30 areas. These Although poor still improvements thewidespread in economic own fewer durable goods social andthan infrastructure better-off have resulted the from households, comparative the combined efforts of many government infrastructure investment programs across the different in table 3.13sectors, statistics infrastructure indicate andsubstantial increases provide a good in durable foundation goods for growth ownership of the since rural economy 1998. and For example, continued 51 percent in 2010, reductions in of the rural owned poor in poverty a motorbike the coming years. compared to 2 percent in 1998; 74 percent owned a TV compared to 30 percent 3.30 Although in still the poor 1998; ownand 46durable fewer percent owned goods a better-off than rice cooker or electric the stove households, compared to 1 comparative percent in 1998, and statistics 37 3.13 in table percent owned indicate a mobile substantial phone.in increases The extreme durable goods poor ownedsince ownership very1998. little in For1998, but example, in 2010, 51 percent of the poor owned a motorbike compared to 2 percent by 2010, 40 percent owned a motorbike, 61 percent owned a TV, 28 percent owned a rice cooker or stove, in 1998; 74 percent owned a TV compared to 30 percent in 1998; and 46 percent owned a rice cooker or electric and 24 percent owned ato stove compared mobile phone. 1 percent Wider in 1998, andaccess to transport, 37 percent owned a TVs, mobile and mobile phone. The phones extreme haspoorimproved the spread of information owned very little in and 1998, but bythe helped 40to poor 2010, become percent lessasocially owned motorbike,isolated and more 61 percent owned integrated a TV, with 28 percent owned a rice cooker or stove, and 24 percent owned a mobile phone. Wider access to the wider economy. transport, TVs, and mobile phones has improved the spread of information and helped the poor to become less socially isolated and more integrated with the wider economy. Table 3.13 Household Ownership Rates of Durables in 1998 and 2010 (Percent) Table 3.13 Household Ownership Rates of Durables in 1998 and 2010 (Percent) National Poor ExtremePoor 1998 2010 1998 2010 1998 2010 Car 0.2 1.3 0.0 0.0 0.0 0.0 Motorbike 20.3 75.9 2.4 50.9 0.4 39.6 Mobilephone ͲͲ 69.8 ͲͲ 37.1 ͲͲ 24.2 TV 55.7 89.3 30.2 73.6 11.9 61.3 Computer 0.7 16.8 0.0 0.3 0.0 0.4 Refrigeratororfreezer 9.0 42.6 0.0 5.3 0.0 2.2 Airconditioner 0.7 8.2 0.0 0.1 0.0 0.2 Electricfan 68.4 85.2 45.9 65.2 26.3 49.4 Ricecookerorelectricstove 19.3 77.6 1.1 45.6 0.0 28.3 Source: VHLSS. 2010VHLSS. Source: 2010 3.31 Despite improvements, many of the poor still do not have access to clean water (36 percent of households in the bottom quintile, 14 percent in the second quintile) or adequate sanitation (21 percent of households in the bottom quintile and 8 percent in the second quintile do not have flush or 84 semi-flush toilets). Although Vietnam has done a remarkable job at making electricity widely available (more than 95 percent of households are connected to the grid) and improving the reliability of supply, 11 percent of households in the bottom quintile are still not connected to the electricity grid. Many of the households without access to clean water, adequate sanitation, and electricity are ethnic 79 minorities living in less accessible upland regions of Vietnam (Table 3.14). As described in Chapter 1, these households are deprived not only in income terms, but also in terms of access to public goods and services. Table 3.14 Percentage of Households with Access to Housing and Neighborhood Amenities in 2010, by Quintile Quintile Total 1 2 3 4 5 Tap water 7.5 13.3 21.7 32.8 59.2 26.9 Clean (nontap) water 56.4 72.8 71.2 62.3 39.7 60.5 Flush toilet 12.8 31.2 48.4 67.6 88.7 49.7 Semi-flush toilet 66.0 61.3 46.8 30.7 10.9 43.1 Solid house 12.0 19.7 26.9 34.5 62.5 31.1 Semisolid house 64.9 66.2 64.7 60.7 36.3 58.6 Household with electricity 89.0 97.9 99.4 99.3 99.6 97.0 Source: VHLSS 2010. G. Urban Poverty is Low According to GSO-WB Estimates, and Concentrated in Smaller Cities and Towns 3.32 The poverty rate in urban areas is only 6 percent compared to 27 percent in rural areas. Because only 30 percent of the Vietnamese population lives in urban areas, the urban poor comprise only 8.6 percent of the total poor in Vietnam. 3.33 Although poverty in Vietnam is primarily a rural phenomenon, understanding and addressing urban poverty is increasingly important. Vietnam is urbanizing rapidly; the urban population grew by 3.4 percent per year between 1999 and 2009 compared to an annual population growth rate of only 0.4 percent in rural areas. The urban population is forecast21 to reach 45 percent of the total population by 2020—a major increase over the 30 percent registered in the 2009 Housing and Population Census. In light of this rapid change, it is vital to better understand the factors that influence the living conditions of low-income urban households, including how poverty is distributed across urban areas. 3.34 City size is one important correlate of poverty. The sample size of the 2010 VHLSS is too small to estimate poverty rates for different types of cities. Instead, the poverty mapping methods described in Chapter 4 were used to estimate poverty rates by city size, ranging from very large “special citiesâ€? (for example, Hanoi and Ho Chi Minh City) to small Class 5 cities, which include district towns with 4,000 or fewer inhabitants. Table 3.15 presents poverty statistics by city size ranging from extra- large (that is, Hanoi and Ho Chi Minh City) to extra-small Class 4 and 5 towns. 3.35 Poverty levels decrease with city size; if measured by the 2010 GSO-WB poverty line,22 only 1.9 percent of the population in the largest cities is poor, while the poverty rate in the smallest cities is 11.2 percent. Poverty depth (the poverty gap) and poverty severity (the squared poverty gap) also decrease with city size. The urban poor are overwhelmingly concentrated in small cities and towns; small and extra small cities account for only 43 percent of the urban population but over 70 percent of the urban poor. Conversely, 32 percent of Vietnam’s urban population lives in Hanoi and Ho Chi Minh City, but only 11 percent of the urban poor live in these two cities. 21 Ministry of Construction plan, as part of Decree 10/1998/QD-TTg, 1998.  everal of Vietnam’s largest cities have developed their own poverty lines; for instance, Hanoi recently announced a new 22 S poverty line of 750,000 VND per person per month for the 2011–2015 Socio-Economic Development Plan, and the poverty line used by Ho Chi Minh City is 1,000,000 VND per person per month. 80 fewer inhabitants. Table 3.15 presents poverty statistics by city size ranging from extra-large (that is, Hanoi and Ho Chi Minh City) to extra-small Class 4 and 5 towns. Table 3.15 Table Poverty 3.15 Poverty by by City City Size Size Extra- Extra- Large Medium Small Rural Large Small Special City class Class 1 Class 2 Class 3 Class 4, 5 City Number of cities in category 2 7 14 45 634 Average population (000) 4,075 467 225 86 11 % of total population 9.5 3.8 3.7 4.5 8.1 70.4 % of urban population 32.1 12.9 12.4 15.3 27.3 Poverty rate (%) 1.9 3.8 4.2 5.8 11.2 25.6 Poverty gap (%) 0.4 0.6 0.7 1.1 2.4 6.8 Share of urban poor (%) 11.0 8.8 9.2 5.9 55.0 Source: Source: World World Bank Bank estimates. estimates.  3.36 Smaller cities can be thought of as more “ruralâ€? than larger cities; urban poverty is concentrated 3.35in the Poverty levels decrease with city size; if measured by the 2010 GSO-WB poverty line,21 only 1.9 more “rural-likeâ€? urban areas. This is consistent with the stylized facts presented earlier in the of percent the population chapter; the poorin inthe cities is poor, while largest overwhelmingly Vietnam therural poverty live in rateAnd areas. in the smallest indeed, citiescities smaller is 11.2 percent. Poverty are more depth than rural-like (the poverty gap)inand larger cities morepoverty severity aspects (the than just squared poverty population. gap) provides also decrease Table 3.16 an city size. of withoverview housing The and local urban poor services, also education are overwhelmingly concentratedlevels of urban in small residents, cities and towns; categorized by small and extra city size and for rural areas. Although access to electricity is universal across all small cities account for only 43 percent of the urban population but over 70 percent of the urban poor.city types, smaller cities lag the larger ones in most other basic services. Use of gas for cooking is less common, Conversely, 32 percent use of firewood for of Vietnam’s cooking is moreurban and lives population common, access Hanoi in to and piped Ho Chi water Minh is much City, less but only common in11 percent of the smaller urban cities poor and live in towns. In these fact, atwo cities. group of smaller cities report having no access to piped water at all. Similarly, fewer households in small cities have flush toilets and substantial numbers use 3.36 Smaller cities can be thought of as more “ruralâ€? than larger cities; urban poverty is concentrated firewood instead of gas for cooking. Smaller cities and towns also lag larger cities in the education in the more level “rural-likeâ€? of the household urban head.areas. This is consistent with the stylized facts presented earlier in the chapter; the poor in Vietnam overwhelmingly live in rural areas. And indeed, smaller cities are more rural-like than larger Table cities 3.16 in more Percent aspects of than just Households population. with Specific Characteristics, by City Size 3.37 Table 3.16 provides an overviewExtra Large and of housing local Medium services, also Large Smalleducation levels Rural Extra Small of urban residents, categorized by city size and for rural areas. Although access to electricity is universal across all Primary city types, cities lag the larger ones in 20.2 education smaller most other 21.8 20.7 Use basic services. 23.7 26.2 of gas for cooking30.0 is less Secondary education 19.0 21.0 20.5 20.1 22.6 27.0 common, use of firewood for cooking is more common, and access to piped water is much less common Tertiary education 49.7 41.7 46.5 40.1 30.6 14.9 in smaller cities and towns. In fact, a group of smaller cities report having no access to piped water at all. Dwelling walls of solid material 98.2 90.6 92.4 86.7 79.9 69.5 Similarly, fewer Dwelling households walls in small of semisolid cities have material toilets and 1.2 flush 4.5 5.0 8.4 numbers substantial 11.9 use firewood 16.0 instead of gas for cooking. Smaller cities Dwelling walls of temporary material and towns 0.6 also lag 4.9 larger cities 2.6 in the 4.9 education 8.2 level of the 14.5 household head. Dwelling roof of solid material 35.1 21.5 25.2 19.5 17.9 13.4 Dwelling roof of semisolid material 6.0 11.5 18.1 20.7 26.6 39.6 Dwelling roof of temporary material 58.8 67.0 56.8 59.8 55.5 47.1 Has flush toilet 99.3 89.6 92.7 82.9 69.6 38.8 Has other kind of toilet 0.5 9.9 5.0 14.6 24.9 50.4 Has no toilet 0.2 0.5 2.3 2.5 5.5 10.9 21 Drinks water from pipe 74.2 74.3 75.5 57.2 33.6 8.0 Several of Vietnam’s largest Drinks water from well cities have developed 25.3 their own 15.9 poverty 21.3 lines; for 35.6 instance, 52.2 Hanoi recently 58.3 announced a new poverty Drinks water other source line of 750,000 VND per 0.6 person per 9.9 month 3.2 for the 2011–2015 7.2 Socio-Economic 14.2 33.8 Development Plan, and Uses electricity the for poverty line used by Ho lighting Chi Minh 99.7 City is 1,000,000 99.7 99.8 VND per person 99.6 per month. 99.0 94.1 Uses electricity for cooking 2.1 1.4 1.1 1.9 1.8 1.5 86 Uses gas for cooking 89.3 70.7 75.5 66.9 55.6 22.9 Uses firewood for cooking 0.7 12.0 7.2 15.7 32.2 64.6 Source: World Bank estimates from 2009 Population Census. Note: Education level is highest attainment of the household head. 81 H. Poverty has Become Less Correlated with Demographic Factors, although Aging is Emerging as an Issue and Child Poverty Remains a Concern 3.37 Compared to the 1990s, demographic factors such as high dependency ratios and female headship have become less linked to poverty. Comparisons between 1999 and 2009 population “pyramidsâ€? for Vietnam (GSO 2010) highlight the sharp reduction in the proportion of children in the population and an increase in the proportion of older adults. Recent qualitative studies (e.g. the long- run drivers of poverty reduction study; Nguyen and Hoang 2012) identify important links between changing household structures and the dynamics of income and well-being. The nationwide family planning campaign, active since the late 1980s, were widely acknowledged at all field sites as having made an important contribution to poverty reduction. Most couples (nearly 80 percent according to the 2010 VHLSS) now have only two children, which helps reduce household spending on basic services like education and health and allows for more “qualityâ€? spending on children. 3.38 The long-run drivers study, with its two-decade reference period, also identified several positive impacts for families that had more children. The Vietnamese economy has been expanding and creating new jobs. Although poor rural households struggled to raise and educate children born in the 1980s and early 1990s, these children are now grown, and many are working in off-farm activities or have migrated to work in urban areas. Rather than being a burden, they contribute to supporting their parents and younger siblings who stay home. Figure 3.12 Figure for Vietnam: 3.12 Population Pyramids for Vietnam: 1999 1999 and and 2009 2009 2009 85+ 80-84 1999 75-79 70-74 65-69 60-64 55-59 50-54 45-49 40-44 35-39 30-34 25-29 20-24 15-19 10-14 5-9 0-4 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 Male % Female Source: Source: GSO GSO 2010. 2010. 3.40 Female-headed households with children were identified in a number of sites as more vulnerable to and at risk for poverty, in large part because they were dependent primarily on the earnings of the 82 female household head. Many respondents felt that two parents are required to work to support a family in Vietnam. Moreover, men in rural areas are better paid than most women because they take on different 3.39 Female-headed households with children were identified in a number of sites as more vulnerable to and at risk for poverty, in large part because they were dependent primarily on the earnings of the female household head. Many respondents felt that two parents are required to work to support a family in Vietnam. Moreover, men in rural areas are better paid than most women because they take on different (heavier and more dangerous) tasks. Single mothers struggle with the lack of adequate daycare facilities, particularly in rural areas, and may not receive support from extended family. 3.40 Aging is another important source of vulnerability. Vietnam has a high proportion of widows; according to the 2010 VHLSS, 19 percent of households include a widow, and 12.5 percent are currently headed by a widow. The proportion of widows in an age cohort rises sharply with age: 47.6 percent of women aged 66-70 are widowed compared to only 9.7 of men in the same age cohort; 67.6 percent of women aged 76-80 are widowed compared to 24.5 percent of men in the cohort. Participatory Poverty Assessments (PPAs) and recent qualitative studies carried out, for instance, by Oxfam, highlight the vulnerability of households headed by elderly persons, and in particular widows, in part linked to the limited coverage of social insurance and pensions for Vietnam’s aging population (UNFPA 2011). Vulnerability linked to aging is a growing challenge in Vietnam, and additional research on the links between poverty, vulnerability, and aging is needed. Aging and Economies of Scale in Consumption Aging and Economies of Scale in Consumption 3.42 New work on aging and household economies of scale and composition was carried for this 3.41 New work on aging and household economies of scale and composition was carried for this report to address the concern that conventional poverty profiles based on per-capita consumption tend to report to address the concern that conventional poverty profiles based on per-capita consumption underreport poverty among small households (particularly those with only elderly members) and over- tend to underreport poverty among small households (particularly those with only elderly members) report and poverty among over-report large households poverty among large (including those households with many (including children). those The study with many explores children). The different study methods to adjust explores for different economies methods of scale to adjust (size) in household for economies welfare of scale (size) (measured in household in terms welfare of per-capita (measured in consumption). While some terms of per-capita types of consumption consumption). typesas While some such of food are moresuch consumption directly a function as food of household are more directly function of sizea(although household young size children (although eat young less than othereat children adults), less types than like adults), other electricity types like and housing electricity are fixed costs andandlesshousing directlyare costs fixed to linked and lesssize. household directly To linked adjustto household for economies ofTo size. adjust scale, for economies individual of is welfare scale, redefined asindividual welfare is redefined as Ü» ‫ כݕ‬ൌ ሺܰሻà°? 3.43WhereWhere Y is total Y is total household household expenditures, expenditures, N is Nthe the number isnumber of household of household members, members, andand θ isÈ™ais a scale scale parameter, which parameter, ranges which from ranges 1 to from 0.0. 1 to When WhenÈ™ θ= 1, individual = 1, individual welfare welfare is is equal equal to per-capita to per-capita expenditures expenditures economies (no (no of scale). Engel curve analysis undertaken as part of the study suggest that economies of scale). Engel curve analysis undertaken as part of the study suggest that moderate moderate scale economies hold for (that hold for Vietnam scale economies = .68). is, È™(that Vietnam is, θ = .68). 3.44 Table 3.17 3.42 Table presents poverty 3.17 presents poverty rates for different for different demographic demographic groups groups and and different different household household demographic demographic compositions compositions using conventional using conventional per-capita per-capitaexpenditure expenditure measures measures (È™ (θ = 1)= 1)and andmoderate moderate (È™ = 0.8)(θ = 0.8) and more and more substantial substantial (θ = 0.6) adjustments (È™ = 0.6) adjustments for economiesfor economies of scale. of scale. Using Using conventional conventional measures, we see measures, the standardwe see the results: standard higher poverty results: higher than the poverty national than for average theminority nationalhouseholds average for minority and for large households and for large households with more dependents (two or more children). Households households with more dependents (two or more children). Households with three or more children with three or more children (around 10 percent of households in 2010) are more likely to be poor (around 10 percent of households in 2010) are more likely to be poor even after adjusting for economies even after adjusting for economies of scale. Child poverty, therefore, remains an important concern. of scale. poverty, low Child although In addition, therefore, remains in absolute an important numbers concern. at present, smallIn addition, although households with elderlylowmembers in absolute numbers emerge at present, as a new small group households with elderly of vulnerable/poor as wemembers emerge as a new adjust progressively group of vulnerable/poor for economies of scale. The as adjust progressively we number for economies of scale. The number of these households of these households is likely to increase as the population ages and Vietnam is likely increase as tobecomes population the more ages and urbanized. Vietnam Ongoing becomes efforts more to develop a urbanized. modern socialOngoing efforts protection to develop system a modern for Vietnam social should keep (single) protection system elderly should keephouseholds and widow/widower for Vietnam in sight (single) elderly andaswidow/widower target populations deservingin households special sight as attention. target populations deserving special attention. 83 Table 3.17 Demographic Characteristics and Scale Economies for the Poor Percent Poor % Household Population size θ = 1 θ = 0.8 θ = 0.6 All households 100.0 4.5 20.7 21.2 No widow 81.0 4.4 20.3 20.5 With widow 19.0 4.8 23.6 24.1 25.2 Female-headed 24.8 4.0 14.9 16.5 18.2 Male-headed 75.2 4.6 22.6 22.5 23.0 Widow-headed 12.5 4.1 21.5 23.2 26.0 Ethnicity = Kinh 82.2 4.4 13.2 13.4 14.3 Ethnicity = not Kinh 17.8 5.1 62.2 63.0 62.9 Household Composition Single adult 0.7 1.0 4.0 11.3 19.9 Single elderly/widow/ widower 0.7 1.0 14.9 29.6 51.1 2 adults 3.8 2.0 6.8 10.9 16.9 Single parent 0.6 2.0 21.4 26.7 34.5 2 elderly 1.2 2.0 22.3 31.9 46.0 Other 2-member household 1.2 2.0 17.0 23.6 34.3 Nuclear 1 child 6.5 3.0 14.0 16.8 19.3 Nuclear 2 children 14.0 4.0 25.1 26.8 28.3 Nuclear 3+ children 5.3 5.3 47.3 45.1 42.9 Extended family no children 20.4 3.9 8.7 9.7 11.1 Extended family 1 child 19.9 4.8 15.0 14.8 15.1 Extended family 2 children 12.0 5.6 26.2 24.0 22.2 Extended family 3+ children 4.7 7.5 56.3 52.4 46.7 Joint family no elderly 6.0 5.7 29.9 26.4 24.0 Joint family with elderly 3.0 6.0 20.9 18.4 17.0 Source: World Bank estimates. Child Poverty Rates Remain High, and Children Face Multiple Deprivations that could Impact their Long-term Development 23 3.43 Children face a higher risk of poverty than adults, and poverty affects them differently. They have different dietary requirements, for example, and the role of education is vital at this stage of life. A child-specific approach to measuring poverty can highlight and emphasize those needs that are especially crucial for children and their development, and enable more effective poverty reduction objectives, strategies, and policies. 3.44 The most common approach to measuring child poverty examines income and/or expenditures at a household level. According to the 1998 VLSS, 47.2 percent—nearly half—of all children lived below the original GSO-WB poverty line. By 2010, this figure had fallen to 29.2 percent. Extreme child poverty fell more slowly—from 16.8 percent in 1998 to 12.5 percent in 2010. Furthermore, in households with three or more children, child poverty remains high, as noted in the previous section. But the monetary approach to measuring child poverty reflects only one dimension of well-being, and does not capture the intra-household distribution of resources. The conventional methodology has thus been extended to assess child poverty along additional dimensions. 23 Information in this section was provided by UNICEF/Hanoi. 84 3.47 In 2008, MOLISA and UNICEF developed a Vietnam-specific multidimensional poverty measurement approach, based on the Convention on the Rights of the Child. The approach incorporates eight poverty domains, including deprivations in education, nutrition, health, shelter, water and sanitation, child labor, leisure, and social inclusion and protection. Poverty prevalence can be calculated for any one 3.45 In 2008, MOLISA and UNICEF developed a Vietnam-specific multidimensional poverty of these domains, and a measurement multidimensional approach, based on the child poverty Convention rate on the (MDCP) Rights constructed of the Child. to measure the The approach children who percentage ofincorporates eight are poor poverty in at least domains, domains. in twodeprivations including This methodology education, has been nutrition, health, applied to the shelter, water and sanitation, child labor, leisure, and social inclusion and protection. Poverty prevalence 2006, 2008, and 2010. can be calculated for any one of these domains, and a multidimensional child poverty rate (MDCP) constructed to measure the percentage of children who are poor in at least two domains. This 3.48 UNICEF’s monetary child poverty rate (MCP) measures the proportion of children living in methodology has been applied to the 2006, 2008, and 2010. households whose welfare levels fall below the GSO-WB poverty line. In contrast, the MDCP identifies 3.46 UNICEF’s monetary child poverty rate (MCP) measures the proportion of children living in the proportion of children suffering from deprivation in at least two of the eight selected domains. The households whose welfare levels fall below the GSO-WB poverty line. In contrast, the MDCP identifies MDCP is systematically the proportion ofhigher childrenthan the from suffering MCP, indicating deprivation around that two in at least of the one-third ofdomains. eight selected children living in Vietnam—or Thean MDCP is systematically estimated 7 million higher than the MCP, children—are indicating that considered around one-third of children multidimensionally poor, living in contrast to in Vietnam—or an estimated 7 million children—are considered multidimensionally poor, in contrast five around one into who around oneare poor in five according who to conventional are poor according income to conventional incomeor or expenditure criteria. (figure 3.13) expenditure criteria. (figure 3.13) Figure Monetary 3.13 3.13 Figure and Monetary Multidimensional and Multidimensional Child Poverty Child Poverty inin Vietnam, Vietnam, 2006-10 2006-10 Source: Source: 2006, 2006, 2008, 2008, 2010 2010 VHLSS. VHLSS. 3.47 A deeper analysis of the degree of overlap between the MCP and the MDCP reveals that the 3.49 A deeper methodsanalysis degree of the groups identify different of of overlap children. While between the are some children MCP and as identified the MDCP poor reveals according to that the both methods, there is also a group that is only identified as poor by the multidimensional approach, methods identify different groups of children. While some children are identified as poor according to and likewise for the monetary approach. Using the 2006 VHLSS data, GSO and MOLISA estimate that both methods, 18there also is of percent a group children that is exclusively are captured only identified as poor by the MDCP andby thenot would multidimensional have been consideredapproach, and poor by the MCP. This result underlines the stark difference between child likewise for the monetary approach. Using the 2006 VHLSS data, GSO and MOLISA estimate that 18 and overall poverty and the importance of a multidimensional measure to complement the standard monetary measurement percent of children are captured exclusively by the MDCP and would not have been considered poor by of poverty. 3.48 Figure 3.14 indicates the disparities that exist among subgroups of the national population. The MDCP declined for both ethnic categories from 2006 to 2010, but children from ethnic minority households are still almost three times more likely to be multi-dimensionally poor than their Kinh/Hoa 91 peers. The figures also provide evidence of a significant urban-rural divide; children in rural areas are twice as likely to be multi-dimensionally poor than children in urban areas. While child poverty in rural areas has shown some decline in recent years, the MDCP indicates that urban poverty is rising. 85 wice as likely to be multi-dimensionally poor than children in urban areas. While child poverty in rural reas has shown some decline in recent years, the MDCP indicates that urban poverty is rising. Figure Figure 3.14 Multidimensional 3.14 Multidimensional Child Child Poverty in Poverty in Vietnam Vietnam by by Selected Selected Sociodemographic Sociodemographic 2006-2010 2006-2010 Variables, Variables, Source: 2008, 2006,2008, Source: 2006, 2010 2010 VHLSS. VHLSS. 3.49 Figure 3.15 provides a breakdown by domain of the MDCP for 2010. Health, water and .51 sanitation, and leisure Figure 3.14 indicates aredisparities clearly the domains the of most that exist concern. among These figures subgroups indicate of the that more national population. The than one in three children aged 2 to 4 (36.7 percent) was not fully immunized and had not visited a health for MDCP declined both facility in theethnic prior 12 categories from months (health); 2006 almost to of2010, two out but five aged 0 tochildren from lived 15 (39.2 percent) ethnic minority ouseholds are still almost in dwellings without three times hygienic more sanitation orlikely multi-dimensionally to bewater safe drinking poor (water and sanitation); thanthan their Kinh/Hoa and more two out of three children aged 0 to 4 did not have any toys or books (leisure). eers. The figures also provide evidence of a significant urban-rural divide; children in rural areas are Figure 3.15 wice as likely to be multi-dimensionally Figure Child 3.15 poorPoverty Child than Rate by children Poverty Rate Domain, byin urban Domain,2010 areas. While child poverty in rural 2010 reas has shown some decline in recent years, the MDCP indicates that urban poverty is rising. .52 Figure 3.15 provides a breakdown by domain of the MDCP for 2010. Health, water and anitation, and leisure are clearly the domains of most concern. These figures indicate that more than one n three children aged 2 to 4 (36.7 percent) was not fully immunized and had not visited a health facility n the prior 12 months (health); almost two out of five aged 0 to 15 (39.2 percent) lived in dwellings without hygienic sanitation or safe drinking water (water and sanitation); and more than two out of three hildren aged 0 to 4 did not have any toys or books (leisure). Source: 2006, 2006, 2008, 2010 VHLSS. Source: 2008, 2010 VHLSS. 92 I. Poor Households are Still Vulnerable to Weather Shocks 3.53 Located in one of the earth’s five typhoon86 centers, Vietnam is prone to natural disasters, including frequent tropical storms and flooding. The 2008 VHLSS collected information on whether households had experienced weather shocks between 2003 and 2008 and the types of shocks. Results are 53 Located in one of the earth’s five typhoon centers, Vietnam is prone to natural dis cluding frequent tropical storms and flooding. The 2008 VHLSS collected information on w ouseholds had experienced weather shocks between 2003 and 2008 and the types of shocks. Resu I. Poor Households are Still Vulnerable to Weather Shocks esented in Table 3.18. Households in rural areas are much more likely to experience weather 3.50 Located in one of the earth’s five typhoon centers, Vietnam is prone to natural disasters, including frequent and an their urban counterparts, the tropical poor storms andare more exposed flooding. The 2008 VHLSS shocksinformation to collected than theon nonpoor. whether Househo households had experienced weather shocks between 2003 and 2008 and the types of shocks. Results are e Central Highlands more likely are presented in Tablethan those in any 3.18. Households other in rural are muchto areas region experience more droughts, while likely to experience weather shocks than their urban counterparts, and the poor are more exposed to shocks than the the Central Coastal nonpoor. regions Householdsareinmost likely the Central to experience Highlands storms are more likely or flooding. than those in any other (Le, Nguyen, region to and 012). experience droughts, while those in the Central Coastal regions are most likely to experience storms or flooding. (Le, Nguyen, and Phung 2012). TablePercent Table 3.18 of HouseholdsExperiencing of Households 3.18 Percent Natural Experiencing Natural Disasters, Disasters, 2003-082003-08 Other forms Flood, of extreme Drought storm Landslide weather National 6.7 12.9 0.7 15.2 Rural 8.6 15.5 0.9 19.4 Urban 1.8 6.3 0.1 4.3 Red River Delta 2.6 10.3 0.4 28.6 East Northern Mountains 9.4 7.0 1.7 23.0 West Northern Mountains 8.1 14.3 1.3 22.6 North Central Coast 15.8 29.3 1.1 30.3 South Central Coast 7.3 25.9 0.4 7.4 Central Highlands 19.2 10.9 0.4 4.9 Southeast 2.9 5.1 0.3 1.3 Mekong River Delta 3.5 10.2 0.5 1.4 Poor 14.2 17.9 1.2 22.9 NonͲpoor 5.6 12.2 0.6 14.1 Source: 2008 VHLSS. Source: 2008 VHLSS. 93 Poverty Reduction and Social J. Limited Coverage is Provided by Existing Protection Programs 3.51 This report focuses on diagnostics. Follow-up work on policy and program implications is planned, including on the design and targeting of social protection and poverty reduction policies and programs. Access to poverty reduction programs and policies is an important aspect of well-being for low-income households. But concerns have been raised about both the targeting and coverage of Vietnam’s existing poverty reduction programs. These issues are examined briefly using information collected in the 2010 VHLSS: each round of the survey includes information on whether households have been formally classified as poor—that is, whether they are on the official MOLISA poverty list—and thus eligible for benefits under existing government programs, most notably the National Targeted Program for Sustainable Poverty Reduction (NTP-SPR). Each round of the VHLSS also includes information on whether the household received program benefits. This information can be used to assess coverage and targeting of Vietnam’s poverty programs. 87 coverage and targeting of Vietnam’s poverty programs. 3.55 Analysis suggests that coverage is problematic (a substantial number of households that should be on the poverty list are not) but targeting is less of a concern (most households on the list are from the 3.52 groups). poorest Analysis Note, however, suggests that the 2010 that coverage VHLSS (a is problematic data were collected substantial numberbefore the government of households that should be the implemented poverty on the census poverty fornot) list are 2011–2015 thebut Socio-Economic targeting is less of a concernDevelopment Plan and (most households used on the listthis are from the poorest groups). Note, however, that the 2010 VHLSS data were collected information to update the official poverty list. Thus, while the official poverty rate for 2010 is 14.2 before the government implemented the poverty census for the 2011–2015 Socio-Economic Development Plan percent, only this and used 10.6information update the surveyed percent oftohouseholds in the official poverty 2010 list. VHLSS Thus, reported while the official being povertyon thefor rate (old) MOLISA 2010 is poverty list. only 10.6 percent of households surveyed in the 2010 VHLSS reported being 14.2 percent, on the (old) MOLISA poverty list. 3.56 Table 3.19 shows the percentage of households (by expanded expenditure quintile) that reported 3.53 Table 3.19 shows the percentage of households (by expanded expenditure quintile) that being classified as poor by commune authorities, and are thus on the official MOLISA poverty list. 8 reported being classified as poor by commune authorities, and are thus on the official MOLISA percent of individuals poverty in the list. 8 percent of 2010 VHLSS individuals are 2010 in the classified VHLSS as are extreme pooras classified byextreme the updated GSO-WB poor by the updated GSO-WB poverty line. However, only 52 percent these households poverty line. However, only 52 percent these households said they were on the official poverty said they were on thelist. official poverty list. Similarly, 20.7 percent of individuals were classified as poor using the updated Similarly, 20.7 percent of individuals were classified as poor using the updated GSO-WB poverty line, GSO-WB poverty line, but only 36 percent of these households said they were on the official poverty but only 36 percent list. Thus of these coverage households is low, said but leakage they of were to benefits thenon-poor onthe official poverty is modest; Thus list.only coverage 12.2 isof percent low, but leakage householdsof benefits to the non-poor in the second is 6.3 quintile and modest; only percent of12.2 percent in households households ofthe in the third quintile second said quintile they were on the official poverty list. and 6.3 percent of households in the third quintile said they were on the official poverty list. Table 3.19 Percentage of Households Officially Classified as Poor, Officially Quintile, by Expanded Table 3.19 Percentage of Households Classified2010 as Poor, by Expanded Quintile, 2010 2010 Extreme poor 52.0 All poor 36.0 Quintile 2 12.2 Quintile 3 6.3 Quintile 4 2.6 Quintile 5 0.4 and rural elites. 3.54 This Figure 3.16 said, being there describes in are clearly greater detailproblems with program how households coverage, on the poverty including list are distributed coverage of 3.57across the 3.16 Figure describes welfare in greater distribution. The detail how households on great majority—nearly 70 the poverty list percent—of are distributed households across are also the poorest households. the welfare distribution. classified as poorDeeper The the using analysis great GSO-WB of poverty coverage majority—nearly Only targeting and 70 percent—of line. atofthe households 11.5 percent regional are also those classified officially level indicates that as poor classified coverage lower is as using in high-poverty poor are the GSO-WB in the upper halfprovinces, poverty line.of the welfare Only such as in the distribution. 11.5 percent of thoseWhileNorth there is officially room and West classified as North for improvement, poor are East, and these in the upperhigher in targeting results are better than in many other countries, where program benefits are frequently some better-off provinces half captured of the welfare urbanWhile and households distribution. by better-off areas. and MOLISA there is elites. rural room may for This face pressure improvement, being said, there to these targeting are spread program results are clearly problems better more benefits with equitably across than programprovinces; in many coverage, the given where other countries, including increasing program coverage of the concentration benefits poorest are frequently households. of Deeper the poor captured in better-off by analysis high-poverty of coverage regions, this households and targeting at the regional would lead to reduced program coverage. level indicates that coverage is lower in high-poverty provinces, such as in the North West and North East, and higher in some better-off provinces and urban areas. MOLISA may face pressure to spread program benefits more94 equitably across provinces; given the increasing concentration of the poor in high-poverty regions, this would lead to reduced program coverage. Figure 3.16 Distribution of Population on the Official Poverty List by Expanded Per-Capita Expenditure Figure 3.16 Distribution of Population Quintile, on the 2010 List by Expanded Per-Capita Official Poverty 80 70 60 50 Percent 40 30 20 10 0 88 3.58 Table 3.20 looks in detail at the coverage of Vietnam’s various social protection and poverty Expenditure Quintile, 2010 3.55 Table 3.20 looks in detail at the coverage of Vietnam’s various social protection and poverty reduction policies for households classified by expanded expenditure quintile (Nguyen and Vu 2012). Coverage rates are low in general and social insurance programs are not well targeted to the poor. Few households reported receiving vocational training in 2010. Analysis of the coverage of social assistance measures presents a more nuanced picture. Many of the policies included under the National Target Program for Sustainable Poverty Reduction are well-targeted toward the poor (for example, education fee reductions and subsidies, production support, food support) but, consistent with the analysis above, the coverage of these programs is very low. In general, less than a third of the extreme poor were covered by these poverty reduction policies in 2010. Health coverage (free health cards) is better, but benefits accrue to households across the welfare distribution. Table 3.20 Coverage of Social Protection and Poverty Reduction Policies by Expanded Quintiles Percentage of People in Households Extreme All Quintile Quintile Quintile Quintile Receiving: Total Poor Poor 2 3 4 5 All transfers and programs 72.6 88.8 77.2 68.1 67.8 70.6 74.5 All social insurance 32.1 11.2 14.3 20.4 28.0 41.1 58.1 Employment subsidy 1.5 1.2 0.8 1.3 1.6 1.8 1.7 Pension 9.2 2.9 2.2 5.4 7.0 11.6 19.5 Having social insurance 26.7 7.5 11.9 15.6 23.4 34.1 50.0 Vocational training 0.1 0.2 0.3 0.2 0.0 0.0 0.0 All social assistance 56.6 87.4 72.0 60.6 54.7 47.9 41.0 Allowances for veterans, merit households 4.0 2.9 2.8 5.2 4.8 4.6 2.6 Allowances for policy households 4.9 11.8 8.8 5.0 4.1 3.3 1.6 Health subsidy allowances 32.7 29.6 31.3 34.3 34.9 29.8 33.7 Education subsidy allowances 8.3 36.0 15.0 7.6 4.0 4.2 2.3 Allowance for recovery from disaster, fire 4.9 7.4 6.7 7.4 5.7 3.8 1.0 Loan from Vietnam Bank for Social Policies 13.1 33.7 25.6 14.2 10.3 8.6 3.2 Health program 12.0 54.7 29.3 11.9 5.2 2.3 0.7 Education fee reduction and exemption 5.5 25.8 14.9 5.4 1.9 0.7 0.1 Housing program 1.1 4.4 2.9 1.3 0.4 0.2 0.0 Cultivation land for ethnic minorities 0.1 0.1 0.5 0.3 0.0 0.0 0.0 Agricultural extension 7.8 25.5 14.4 7.3 6.1 4.7 1.9 Clean water 1.9 9.1 4.5 2.1 0.6 0.5 0.2 Food supports 5.2 24.9 10.4 5.6 2.0 1.9 0.2 Production support 9.0 27.9 14.5 9.0 8.0 5.6 2.1 Source: Nguyen and Vu 2012. 3.56 Table 3.21 presents similar estimates stratifying for urban versus rural households, also for Kinh majorities and ethnic minorities. Minorities report substantially lower coverage of social insurance programs, albeit greater access to NTP-SPR support, and greater access to social assistance 89 programs more generally. Higher coverage is not surprising given the very high poverty rates among ethnic minorities. Table 3. 21 Coverage of Social Protection and Poverty Reduction Policies by Urban/Rural and Ethnicity Percentage of People in Households Receiving: Total Urban Rural Kinh/Hoa Ethnic Minorities All transfers and programs 72.6 75.3 71.5 70.3 86.1 All social insurance 32.1 56.2 22.0 35.2 14.0 Employment subsidy 1.5 2.0 1.3 1.6 0.8 Pension 9.2 17.9 5.5 10.1 4.0 Having social insurance 26.7 48.9 17.3 29.3 11.0 Vocational training 0.1 0.0 0.1 0.0 0.6 All social assistance 56.6 44.0 61.9 52.2 82.0 Allowances for veterans, merit households 4.0 2.6 4.6 4.2 2.4 Allowances for policy households 4.9 2.3 5.9 4.1 9.4 Health subsidy allowances 32.7 31.9 33.0 33.0 30.7 Education subsidy allowances 8.3 3.5 10.3 4.1 32.7 Allowance for recovery from disaster, Fire 4.9 1.3 6.4 4.8 5.6 Loan from Vietnam Bank for Social Policies 13.1 6.8 15.8 9.7 33.2 Health program 12.0 3.4 15.6 6.4 44.1 Education fee reduction and exemption 5.5 1.8 7.1 3.2 18.8 Housing program 1.1 0.2 1.5 0.4 4.8 Cultivation land for ethnic minorities 0.1 0.0 0.2 0.0 0.8 Agricultural extension 7.8 1.1 10.6 4.7 25.9 Clean water 1.9 0.2 2.7 0.6 9.7 Food supports 5.2 1.4 6.8 2.8 19.1 Production support 9.0 1.4 12.1 6.0 26.2 Source: VHLSS 2010. Notes: Program coverage is the portion of population in each group that receives the transfer. Specifically, coverage is (number of individuals in the group who live in a household where at least one member receives the transfer) / (number of individuals in the group). Program coverage is calculated setting as the expansion factor the household expansion factor multiplied by the household size. Source: Nguyen and Vu 2012. 90 Chapter Annexes Annex 3. 1 Overview of Vietnam’s Eight Economic Regions Vietnam’s eight regions include the North East, the North West, the Red River Delta, the North Central Coast, the South Central Coast, the Central Highlands, the South East, and the Mekong River Delta. The North East lies to the north of the Red River Delta. It includes nine provinces, with a population of 8.2 million. The Viet (Kinh) people make up the majority, with the exception of where a number of minority groups reside. Economic development in the region is mainly based on mining, especially coal and various minerals, forestry, perennial crops, vegetables, and tourism at sites like Ba Be lake, Tam Dao, and Ha Long Bay. The North West is in the mountainous northwestern part of the country, bordering China and Laos. It covers six provinces, with a population of 4.2 million. The Thai people make up the majority, but more than 20 other ethnic groups live in North West region. High mountains make communications difficult. The economy is based on agriculture and industrial crops such as tea and maize. The soil contains various minerals that have not yet been exploited. The Red River Delta’s population is 18.8 million inhabitants, a majority of which (96.2 percent) are Viet people who live in 10 provinces. The region is the economic, political, and cultural center of the country, with the capital Hanoi and the port of Haiphong. The economic engines are industrial production and services. It is also the second- largest rice producer of the country. The North Central Coast has about 10.1 million inhabitants consisting of 25 ethnic groups the majority of which are Viet people. The region is located between the Lao border and a long coastal line. It offers good conditions for overseas trading and tourism. The South Central Coast encompasses eight provinces with a combined population of 8.9 million. The majority of the population are Viet people, but other minorities include Bana, Cham, and RaGlai. Economic development is mainly based on industrial production, especially in Da Nang and Khanh Hoa provinces, and in new industrial centers, namely the Chu Lai economic zone and the Dung Quat economic zone (with the Dung Quat refinery). The long coastline offers good potential for the development of the marine economy in the region. The Central Highlands region has a population of 5.3 million that is ethnically dominated by the Bana, Coh, Ede, and Giarai. It shares a border with Cambodia and Laos and covers the poorest areas of the country, with sluggish economic development and weak infrastructure. Its fertile soil is good for industrial crops such as coffee, pepper, and rubber. The South East consists of seven provinces and 14.9 million people, of which Viet people are the majority and Cham and Kh’mer are the main ethnic minorities. This region is the most economically developed and also the most urbanized region in Vietnam, with the economic hub Ho Chi Minh City. Other provinces of the region such as Binh Duong, Dong Nai, and Ba Ria-Vung Tau are industrialized and contribute significantly to economic development in the region. The Mekong River Delta includes 13 provinces and 17.3 million people of which Viet is the main group and Hoa and Khmer the minorities. It is the largest rice-growing area and produces half of Vietnam’s total rice production. In addition, the region is home to a large aquacultural industry of catfish and shrimp and a variety of fruits. 91 References GSO (General Statistics Office of Vietnam). 1998. “Decision to Approve the Orientations of the Master Plan for the Development of Vietnam’s Urban Centers till 2020.â€? Decree 10/1998/QD-TTg. Hanoi, January 23, 1998. GSO (General Statistics Office of Vietnam). 2011. “Report on Multidimensional Child Poverty in Vietnam.â€? Prepared jointly by UNICEF and GSO, Hanoi, September. GSO (General Statistics Office of Vietnam), 2010. “Migration and Urbanization in Vietnam: Patterns, Trends and Differntials.â€? Prepared with support from UNFPA based on 2009 Housing and Population Census, 15% sample. Hanoi. Haughton, J., Nguyen Thi Thanh Loan, and Nguyen Bui Linh. 2010. “Urban Poverty Assessment in Hanoi and HCMC.â€? Joint publication of the UNDP and Vietnam General Statistics Office, Hanoi. Le, T. D., C. V. Nguyen, and T. D. Phung. 2012. “Natural Shocks, Vulnerability to Poverty in Vietnam.â€? Background paper for the 2012 Vietnam Poverty Assessment, Hanoi. Markussen, T., Finn Tarp, and K. van den Broeck. 2009. “The Forgotten Property Rights: Restrictions on Land Use in Vietnam.â€? Discussion Paper No. 09-21, Department of Economics, University of Copenhagen, Copenhagen. Nguyen, Cuong Viet and Linh Vu. 2012. “Poverty Targeting and Social Protection Strategies in Vietnamâ€?. Background paper prepared for the 2012 Vietnam Poverty Assessment, Hanoi. Nguyen Tam Giang and Hoang Xuan Thanh. 2012. “Long-run Drivers of Poverty Reduction in Vietnam between 1992 and 2011.â€? Background paper prepared for the 2012 Vietnam Poverty Assessment, Hanoi. Ravallion, Martin, and Dominique van de Walle. 2008a. Land in Transition: Reform and Poverty in Vietnam. New York: Palgrave Macmillan; Washington, DC: World Bank. Ravalliion, Martin and Dominiqe van de Walle. 2008b. “Land and Poverty in Reforming East Asiaâ€?. Finance and Development 45(3): 38-41. UNFPA (United Nations Population Fund). 2011. “The Aging Population in Vietnam: Current Status, Prognosis, and Possible Policy Responses.â€? United Nations Population Fund, Hanoi. World Bank. 1999. Vietnam Development Report 2000: Attacking Poverty. Washington DC: World Bank. 92 chapter 4 Spatial Dimensions of Poverty: 1999 and 2009 Poverty Maps New poverty and inequality maps were created using Vietnam’s 2009 Population and Housing Census in combination with the 2010 Vietnam Household Living Standards Survey. Poverty rates are highest in rural, inland, upland areas, and especially for ethnic minorities. Regions with high poverty are also characterized by high inequality. Poverty is becoming more spatially concentrated over time. 93 A. Introduction 4.1 Household surveys are an important source of information on poverty and living conditions. But there is also widespread demand for information on poverty at a more disaggregated level, such as districts, communes, and villages, than is typically available through national household surveys. Knowing where poor people live is important information for designing effective poverty reduction policies and programs, including targeted poverty reduction programs and policies to promote infrastructure investment and improve access to public goods and services in poor areas. 4.2 Spatial targeting requires reliable information on poverty outcomes at the local level. The Ministry of Labor, Invalids and Social Affairs’ (MOLISA’s) system for determining eligibility for support under the National Target Program for Sustainable Poverty Reduction and other social programs uses a bottom-up process of local surveys combined with village-level discussions to produce poverty estimates at the commune level. But analysis suggests that coverage is uneven and there is a need to improve information on poverty outcomes at the local level (Nguyen et al. 2012). Estimation of poverty for small geographical units (for example, districts and communes) is data intensive. While household surveys like the Vietnam Household Living Standard Survey (VHLSS) collect detailed information on household incomes and expenditures, the sample sizes are too small to yield reliable estimates of poverty at the district or commune level. In contrast, Vietnam’s decennial Population and Housing Censuses do not suffer from small-sample problems; they cover the whole population. Censuses also collect valuable information on individual and household characteristics that provide insights into living standards. But the Census does not collect the detailed information on income or expenditures needed to directly measure poverty. 4.3 Small area estimation techniques (often referred to as poverty mapping methods) have been developed to estimate poverty at the small-area level. One popular approach, introduced by Elbers, Lanjouw, and Lanjouw (2002, 2003), combines household survey data and census data at the unit record level. The approach exploits a census’s coverage of the entire population and the household survey’s detailed information on income and expenditure. First, an expenditure (or income) model is estimated using the household survey data. The dependent variable is expenditure (or income), and the explanatory variables are a set of household and community characteristics that are comparable and that are available in both the household survey and the census. Subsequently, the parameter estimates from the expenditure model are applied to the census data in order to predict expenditure for all households in the population. From there it is a straightforward procedure to estimate poverty measures in small areas such as communes and districts. 4.4 The small area estimation method has been applied in a large number of countries to produce maps not only of poverty measures but also of other welfare indicators (see Bedi, Coudouel, and Simler [2007] for review of applications). In Vietnam, a number of poverty maps have been developed in the past using the Elbers, Lanjouw, and Lanjouw small area estimation method. Minot, Baulch, and Epprecht (2003) combined the 1993 Vietnam Living Standard Survey (VLSS) and the 1994 Agricultural Census to estimate poverty at the local level in rural areas of Vietnam. Minot, Baulch, and Epprecht (2003) constructed a poverty map using the 1998 VLSS and a 33 percent sample of the 1999 Population and Housing Census. Nguyen (2009) applied the 2002 VHLSS to the 33 percent sample of the 1999 Population and Housing Census to produce a poverty map for 2002. Nguyen et al. (2010) further updated the rural poverty map for 2006 using the 2006 VHLSS and the 2006 Rural Agriculture and Fishery Census. 4.5 The General Statistics Office (GSO) completed a new census of the population in 2009 and a new round of the Vietnam Household Living Standards Survey in 2010. These datasets were used to construct new poverty and inequality maps for Vietnam. This chapter documents these new estimates of poverty at the province and district level 24 of Vietnam, using the updated 2010 poverty line and t is not feasible to produce reliable commune-level poverty estimates using the 15 percent sample of the 2009 24 I Population and Housing Census. These will be done at a later date if GSO makes the unit record data available for the full 2009 census. 94 comprehensive consumption aggregates described in Chapter 2. The estimates are based on the 15-percent sample of the 2009 Population and Housing Census. In addition, poverty is estimated at the provincial and district level for different groups including rural, urban, Kinh/Hoa, and ethnic minority subpopulations. Estimates of provincial- and district-level inequality are also presented, as is a complementary set of “wealth maps,â€? that is, maps that show which provinces and districts account for the wealthiest 15 percent of the Vietnamese population. 4.6 The chapter then turns to an assessment of spatial changes in poverty based on the 1999 and 2009 poverty maps. Although poverty at the national level has fallen substantially over this period, the rate of progress has not been uniform across all localities. Against a background of substantial aggregate growth and poverty reduction, poverty today has become more concentrated in certain regions of the country and within certain socioeconomic groups. Building on these findings, the mapping methodology is used to assess whether the 62 “poorest districtsâ€? identified under Program 30A are indeed among the poorest in Vietnam. Initial findings from policy simulations to assess the gains from spatial targeting in 2010 compared to 1999 are also briefly described. The policy message emerging from both exercises is that spatially targeted poverty reduction policies, including, for example, area-based schemes, will continue to play an important role in Vietnam. B. 2009 Poverty Maps 4.7 Small area estimation methods are used to construct per capita expenditure-based poverty rates for regions, provinces, and districts in Vietnam. Table 4.1 presents regional estimates of the poverty rate and per capita expenditure that are computed directly using per capita expenditure data of the 2010 VHLSS and those estimated from the poverty mapping method. The 2012 VHLSS is representative at the regional level, and the regional poverty rate directly estimated from expenditure data can be thus regarded as the benchmark against which to compare the poverty map estimates. Table 4.1shows that estimates of the poverty rate are quite similar across the two approaches. Table 4.1 Per Capita Expenditure and Poverty Indexes stimates from the 2010 VHLSS E Estimates from Small Area Estimation Method Per Capita Per Capita Expenditure P0 P1 P2 Expenditure P0 P1 P2 (thousand VND) (thousand VND) Northern 10,927.1 44.87 0.1558 0.0701 10,826.4 43.85 0.1483 0.0679 Mountain (250.2) (1.54) (0.0069) (0.0042) (340.9) (1.76) (0.0082) (0.0046) Red River 21,546.0 11.95 0.0265 0.0088 20,515.2 10.65 0.0203 0.0060 Delta (605.6) (0.85) (0.0025) (0.0010) (592.2) (1.02) (0.0025) (0.0009) Central 14,222.6 23.73 0.0635 0.0251 14,002.1 22.48 0.0520 0.0180 Coast (267.3) (1.33) (0.0051) (0.0028) (268.7) (1.05) (0.0031) (0.0013) Central 13,069.0 32.74 0.1149 0.0542 12,931.0 33.29 0.1146 0.0536 Highlands (490.9) (2.75) (0.0128) (0.0077) (351.8) (1.25) (0.0056) (0.0032) South 24,297.4 7.02 0.0172 0.0064 23,350.9 7.07 0.0139 0.0043 East (935.9) (0.96) (0.0036) (0.0018) (844.9) (0.84) (0.0020) (0.0007) Mekong 14,858.2 18.71 0.0425 0.0143 14,497.9 17.45 0.0359 0.0112 River Delta (265.8) (1.10) (0.0033) (0.0015) (280.7) (1.08) (0.0029) (0.0011) Source: Estimation based on the 2009 Vietnam Population and Housing Census and the 2010 VHLSS Note: Standard errors are in parentheses. P0 is the poverty headcount, P1 is the depth of poverty, P2 is the severity of poverty. 95 4.8 Table 4.2 presents estimates using poverty-mapping methods of the mean of per-capita expenditure and the estimated poverty rate, and the absolute number of poor people and the contribution to national poverty for all 63 provinces in Vietnam. Lai Chau, Ha Giang, and Dien Bien are the three poorest provinces, with a poverty rate of more than 70 percent. As expected, Hanoi and Ho Chi Minh City are the least-poor cities, followed by Da Nang, Hai Phong, Quang Ninh, Binh Duong, and Ba Ria-Vung Tau. Similar estimates were made for Vietnam’s 668 districts and, along with provincial estimates, are presented in the figures and maps that follow (Nguyen et al. 2012). Table 4.2 Per-Capita Expenditure and Poverty Rate of Provinces Province Number Share Per Capita Poverty Rate Number Share of People in Expenditure (%) of Poor in Total Total (thousand VND) People Poverty Pop. (%) Mean Std. Err. Mean Std. Err. Northern Mountain Ha Giang 724,352 0.84 7422.7 448.1 71.46 2.99 517,586 3.07 Cao Bang 510,884 0.60 9,325.7 515.1 53.11 3.26 271,348 1.61 Bac Kan 294,660 0.34 10,136.1 792.0 45.97 5.32 135,448 0.80 Tuyen Quang 725,467 0.85 11,238.3 917.9 39.95 5.41 289,798 1.72 Lao Cai 613,074 0.71 9,711.5 817.8 56.77 3.90 348,018 2.06 Dien Bien 491,046 0.57 7,625.9 611.7 71.06 3.65 348,953 2.07 Lai Chau 370,134 0.43 6,809.2 465.3 76.41 2.99 282,805 1.68 Son La 1,080,641 1.26 8,326.0 590.3 63.60 4.02 687,305 4.08 Yen Bai 740,904 0.86 10,621.9 794.5 45.33 4.72 335,860 1.99 Hoa Binh 786,963 0.92 10,439.0 675.5 47.31 4.23 372,330 2.21 Thai Nguyen 1,124,785 1.31 14,170.5 1,117.1 21.99 3.42 247,386 1.47 Lang Son 731,886 0.85 10,292.1 715.1 45.69 4.29 334,364 1.98 Bac Giang 1,555,720 1.81 12,823.4 889.4 23.83 4.33 370,722 2.20 Phu Tho 1,313,926 1.53 13,535.9 806.9 23.62 3.20 310,380 1.84 Red River Delta Ha Noi 6,448,837 7.52 29,344.6 1,375.7 4.94 0.89 318,488 1.89 Quang Ninh 1,144,381 1.33 18,538.0 1,243.9 12.12 1.81 138,656 0.82 Vinh Phuc 1,000,838 1.17 15,743.1 869.0 11.99 2.83 119,989 0.71 Bac Ninh 1,024,151 1.19 17,590.4 1,145.4 10.19 2.37 104,327 0.62 Hai Duong 1,703,492 1.99 15,261.3 827.5 14.84 2.73 252,716 1.50 Hai Phong 1,837,302 2.14 20,316.9 1,140.2 7.93 1.62 145,625 0.86 Hung Yên 1,128,702 1.32 16,063.4 812.6 12.78 2.36 144,273 0.86 Thai Bình 1,780,953 2.08 13,578.2 873.7 18.95 3.86 337,435 2.00 Ha Nam 785,057 0.92 14,269.8 1,011.8 16.56 4.07 130,009 0.77 Nam Dinh 1,825,770 2.13 14,866.4 814.6 14.04 2.70 256,321 1.52 Ninh Bình 898,458 1.05 14,955.3 878.3 15.28 3.33 137,314 0.81 Central Coast Thanh Hoa 3,400,238 3.96 13,118.2 474.9 26.48 2.09 900,393 5.34 Nghe An 2,913,054 3.40 13,356.4 576.6 26.74 2.57 778,900 4.62 Ha Tinh 1,227,554 1.43 13,222.9 578.5 21.55 2.97 264,499 1.57 Quang Binh 846,924 0.99 13,847.2 798.8 23.20 4.14 196,475 1.17 Quang Tri 597,984 0.70 12,567.1 621.0 29.55 3.15 176,710 1.05 96 Province Number Share Per Capita Poverty Rate Number Share of People in Expenditure (%) of Poor in Total Total (thousand VND) People Poverty Pop. (%) Mean Std. Err. Mean Std. Err. Thua Thiên Hue 1,087,578 1.27 14,453.7 955.1 19.43 3.03 211,283 1.25 Da Nang 887,068 1.03 23,087.9 1,311.7 2.39 1.05 21,218 0.13 Quang Nam 1,419,502 1.65 12,703.2 528.7 23.47 2.73 333,146 1.98 Quang Ngãi 1,217,159 1.42 12,955.1 573.2 23.65 2.80 287,827 1.71 Binh Dinh 1,485,943 1.73 14,498.9 834.9 16.68 3.16 247,882 1.47 Phú Yên 861,993 1.00 13,377.2 793.1 22.08 3.47 190,348 1.13 Khanh Hoa 1,156,902 1.35 16,778.1 1,244.5 15.51 2.87 179,462 1.06 Ninh Thuan 564,128 0.66 11,626.1 799.1 34.52 4.36 194,759 1.16 Binh Thuan 1,169,450 1.36 13,428.5 693.8 21.44 3.04 250,692 1.49 Central Highlands Kon Tum 430,036 0.50 11,112.5 796.7 47.58 3.37 204,624 1.21 Gia Lai 1,272,791 1.48 11,222.1 439.8 43.34 2.07 551,632 3.27 Dak Lak 1,728,380 2.01 13,445.5 639.8 30.32 2.03 524,104 3.11 Dak Nong 489,441 0.57 11,719.4 500.0 32.50 2.83 159,063 0.94 Lâm Dong 1,186,786 1.38 15,173.1 687.8 21.96 1.97 260,629 1.55 South East Binh Phuoc 874,961 1.02 14,370.4 849.9 17.20 3.58 150,477 0.89 Tay Ninh 1,066,402 1.24 15,459.4 737.6 11.78 2.51 125,615 0.75 Binh Duong 1,482,635 1.73 18,378.5 1,168.5 7.82 2.10 115,901 0.69 Dong Nai 2,483,210 2.89 17,293.1 1,129.8 11.73 2.21 291,223 1.73 Ba Ria - Vung Tau 994,836 1.16 18,704.2 1,336.3 9.97 2.22 99,206 0.59 Ho Chí Minh 7,123,340 8.30 29,431.0 1,342.5 2.94 0.51 209,427 1.24 Mekong River Delta Long An 1,436,913 1.67 16,334.8 703.5 10.97 1.64 157,596 0.93 Tien Giang 1,670,215 1.95 16,578.6 875.9 9.53 2.14 159,215 0.94 Ben Tre 1,254,588 1.46 16,022.7 745.8 10.00 2.00 125,506 0.74 Tra Vinh 1,000,932 1.17 13,507.1 688.8 22.28 3.09 222,988 1.32 Vinh Long 1,028,365 1.20 16,038.5 887.7 11.76 2.26 120,947 0.72 Dong Thap 1,665,420 1.94 13,820.8 605.6 15.58 2.42 259,532 1.54 An Giang 2,144,772 2.50 13,739.4 595.5 18.22 2.50 390,808 2.32 Kiên Giang 1,683,149 1.96 13,057.1 580.7 24.02 2.62 404,319 2.40 Can Tho 1,187,088 1.38 17,911.6 1,029.2 11.70 1.97 138,868 0.82 Hau Giang 756,625 0.88 13,369.3 690.7 19.68 3.41 148,915 0.88 Soc Trang 1,289,441 1.50 12,561.6 604.5 27.28 3.10 351,709 2.09 Bac Liêu 856,249 1.00 12,533.0 670.7 23.30 3.74 199,528 1.18 Ca Mau 1,205,107 1.40 12,456.9 682.5 26.36 3.48 317,609 1.88 Sources: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. 97 4.9 Map 4.1 shows the spatial distribution of poverty by provinces and districts in 2009. Poverty rates are highest in the mountainous Northern areas and lowest in the Mekong and Red River Deltas. Disaggregating down to the district level reveals a greater degree of heterogeneity in terms of both pockets of extreme poverty and pockets with particularly low levels of poverty. As discussed later in the chapters, such heterogeneity across sub-national localities translates into gains from spatial targeting of resources for poverty reduction.  Map 4.1 Predicted Poverty Rates of Provinces and Districts, 2009 Map 4.1 Predicted Poverty Rates of Provinces and Districts, 2009 Panel A Province Panel B District Panel A Province Panel B District Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. 4.10 Map 4.2 4.10 Map 4.2 graphs graphs the thedensity density of ofthe thepoor poor across the across country. the Because of their country. Because large of their populations, large the populations, Mekong and Red the Mekong and River Red RiverDeltaDelta regions still account regions significant for afor still account number a significant numberof poor people of poor living people in living Vietnam. in Vietnam. However, However, as shown below as shown (map (map below 4.10), 4.10), the picture in 2009 in the picture 2009 less is much accentuated is much than at the less accentuated time thanofatthe thepreceding time of the census, and as census, preceding such indicates and as a clear such attenuation indicates a of the pattern clear described attenuation in pattern earlier of the studies of poverty described in earlierin studies Vietnamof (see Minot, poverty in Baulch, Vietnam and (seeEpprecht 2003) where Minot, Baulch, the distribution 2003) of and Epprecht the where the distribution number of the was of poor people number of poor inversely people with correlated was inversely the spatialcorrelated distributionwith the spatial of poverty distribution rates. In the lateof poverty rates. In the late 1990s, the incidence of poverty was highest in 1990s, the incidence of poverty was highest in more sparsely populated localities and these thus more sparsely populated localities and accounted these for only a thus fraction for accounted modest of only a modest the poor. fraction Today, of thepoverty although poor. Today, rates although poverty remain spatially rates remain concentrated spatially the concentrated, distribution the distribution of poor people of poorspread is more evenly people is more across theevenly spread country. across the Consequently country. Consequently Vietnam’s poorest communities now account Vietnam’s poorest communities now account for a larger share of the poor population. for a larger share of the poor population. 108 98 Map Map4.2 4.2 Densityof Density ofPoverty ( Number Poverty ( ofPoor PoorPeople), Number of People), 2009 2009 Source: Estimation based Source:Estimation based on on the the 2009 2009 Population Housing andHousing Population and Censusand Census the andthe 2010 2010 VHLSS. VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Inequality is higher in poorer regions 4.11 In Vietnam, there is a positive relationship between poverty and inequality (measured by the is higher InequalityGini poorer in more index). A regions equal distribution in well-being is associated with a lower poverty rate (figure 4.1) at the district and province level, while regions with high poverty rates tend to be more unequal. 4.11 Vietnam, In This there result is is a in large positive part relationship driven by between persistent gaps poverty in well-being and inequality between (measured ethnic minorities and Kinhby the Gini index). majorities (see below, also Chapter 5). However, there remains a great deal of heterogeneity A more equal distribution in well-being is associated with a lower poverty rate (figure 4.1) at the in inequality outcomes, particularly when results are disaggregated to the district level. district and province level, while regions with high poverty rates tend to be more unequal. This result is in 8 large part driven by persistent gaps in welfare between ethnic minorities and Kinh majorities (see below, also Chapter 5). However, there remains a great deal of heterogeneity in inequality outcomes, particularly when results are disaggregated to the district level. 99 Figure 4.1 Relationship between the Poverty Rate and Gini Index Figure 4.1 Relationship between the Poverty Rate and Gini Index Panel A: Provinces Figure Panel A:4.1 Relationship between the Poverty Rate Provinces PanelPanel and B: B: Index Gini Districts Districts Panel A: Provinces Panel B: Districts 100 80 100 80 80 60 80 60 60 rate Poverty rate Poverty Poverty rate 40 60 rate 20 Poverty40 40 20 40 20 20 0 0 .25 .3 .35 .4 .45 .2 .25 .3 .35 .4 .45 Gini index Gini index 0 0 .25 .3 .35 .4 .45 .2 .25 .3 .35 .4 .45 Source: Estimation Source: based on Estimation based the2009 onthe Gini 2009 index Population Populationand Housing and Census Housing andand Census the 2010 VHLSS the 2010 . Gini index VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Relationship between Poverty and other Characteristics Relationship between Poverty and other Characteristics Relationship 4.13 Although between Poverty Vietnam rural Characteristics andaother remains country, urbanization has been accelerating in recent years. 4.12 Although Vietnam remains a rural country, urbanization has been accelerating in recent About 30 percent of people now reside in urban areas (GSO 2011). Overall, urban areas tend to have Although 4.13 About years. Vietnam of 30 percent remains people a now country, rural reside in urbanization urban areas has been2011). (GSO accelerating Overall,in recent urban years. areas lower poverty, and poverty tends to decrease as the urban population share increases (Ravallion, Chen, tend About have to30 lower percent ofpoverty, and poverty people now reside in tends urban to decrease areas (GSOas the urban 2011). population Overall, share urban areas increases tend to have and Sangraula 2007). Figure 4.2 shows that poverty is negatively correlated with the urban population (Ravallion, lower poverty, and and Chen, Sangraula poverty tends to 2007). decreaseFigure as the4.2 shows urban that poverty population shareisincreases negatively correlated (Ravallion, with Chen, share the at the provincial and district level but, again, with considerable variability across space. and urban population Sangraula 2007). share Figureat the 4.2 that and provincial shows district poverty level but, again, is negatively withwith considerable correlated the urbangeographic population variability. share at the provincial and district level but, again, with considerable variability across space. Figure 4.2 Poverty Rate and Proportion of Urban Population Figure Panel A Figure Poverty 4.2 4.2 Provinces Rate Poverty Rateand Proportion and Proportion ofof Urban Urban Population Panel B Districts Population Panel Panel A Provinces A Provinces Panel B Districts Panel B Districts 100 80 100 80 80 60 80 60 Poverty rate Poverty rate 60 rate 40 60 Poverty rate 40 Poverty 40 20 40 20 20 20 0 0 0 20 40 60 80 0 20 40 60 80 100 Percentage of urban population Percentage of urban population 0 0 0 20 40 60 80 0 20 40 60 80 100 Source: Estimation based on the Percentage of 2009 Population and Housing Census and the 2010 VHLSS. urban population Percentage of urban population Source: Source: Estimation Estimation on the based on based 2009 Population the 2009 andHousing Populationand HousingCensus Censusand and 2010 thethe VHLSS. 2010 VHLSS. 4.13 Despite the ongoing urbanization process, poverty in Vietnam is still largely a rural phenomenon; 110 consistent with the updated poverty profile presented in Chapter 3, results using the poverty mapping live in rural areas. Map 4.3 compares poverty rates approaches confirms that 95 percent of the poor 110 in urban and rural areas both at province and district levels. Urban poverty is found to be uniformly lower, and there are substantial differences in poverty rates between urban and rural areas within a given province or district. As discussed in Chapter 3, 70 percent of the urban poor live in smaller cities and towns, rather than Vietnam’s large (special, Class 1 and 2) cities. 100 province or district. As discussed in Chapter 3, 70 percent of the urban poor live in smaller cities and towns, rather than Vietnam’s large (special, Class 1 and 2) cities. Map 4.3 Urban and Rural Poverty Rates Map 4.3 Urban and Rural Poverty Rates Panel Panel A Urban A Urban and and Provinces Provinces Districts Districts Panel Panel B Rural B Rural Provinces Provinces andand Districts Districts 4.14 Analysis based on mapping methods also confirms that poverty has become increasingly 111 there is a strong correlation between the share concentrated among ethnic minority populations, and of ethnic minorities in the population and the poverty rate, at both the province and district levels (figure 4.3).25  he mapping methodology may underestimate ethnic minority poverty, because it assumes that minorities receive the 25 T same returns to their endowments as the Kinh majority. Studies suggest that minorities not only have lower levels of assets, but also receive lower returns on their assets (Baulch and Dat 2012). Estimates presented here and in Chapter 3 provide lower bound estimates of geographically disaggregated poverty levels. 101 4.15 Analysis based on mapping methods also confirms that poverty has become increasingly concentrated among ethnic minority populations, and there is a strong correlation between the share of ethnic minorities in the population and the poverty rate, at both the province and district levels (figure 4.3).24 Figure 4.3 Poverty Figure Rate 4.3 Poverty and Rate Proportion and ofEthnic Proportion of Ethnic Minorities Minorities Panel A Panel AProvinces Provinces Panel Panel B Districts B Districts 100 80 80 60 60 Poverty rate Poverty rate 40 40 20 20 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Percentage of ethnic minority population Percentage of ethnic minority population Source: Source: Estimationbased Estimation on the 2009 based on Populationand 2009 Population and Housing Housing Census Census and and thethe 2010 2010 VHLSS. VHLSS. 4.15 Consistent with Chapter 3, Vietnam’s poor are increasingly concentrated in the Northern 4.16 MountainsConsistent with Chapter and Central 3, Vietnam’s Highlands, where there poorare increasingly are high concentrated proportions in the in Northern of minorities local Mountains and Central Highlands, where there are high proportions of minorities in local populations. populations. Map 4.4 The Map Rate Rate 4.4 The Poverty Poverty of Kinh/Hoa and of Kinh/Hoa and Ethnic Minority Ethnic People People Minority Panel A A Panel Kinh/Hoa Kinh/Hoa People People Panel B Panel B Ethnic Ethnic Minority Minority People People 24 The mapping methodology may underestimate ethnic minority poverty, because it assumes that minorities receive the same returns to their endowments as the Kinh majority. Studies suggest that minorities not only have lower levels of assets, but also receive lower returns on their assets (Baulch and Dat 2012). Estimates presented here and in Chapter 3 provide lower bound estimates of geographically disaggregated poverty levels. 112 102 113 C. Inequality and Wealth Maps 4.16 We employ two measures of inequality, the Gini index and the ratio of the 90th-to-10th expenditure percentile (a measure of “absoluteâ€? inequality). Provincial results are presented in table 4.3. Provincial- and district-level estimates are presented in the figures and maps that follow, and elsewhere (Nguyen et al. 2012). 4.17 Consistent with table 4.3, maps 4.5 and 4.6 illustrate that inequality of expenditures tends to be higher in provinces and districts with low average expenditures. Districts with high poverty rates in the Northern Mountains (these also have a high percentage of minorities) have higher expenditure inequality than other regions. This finding is noteworthy in light of the common (often implicit) view in Vietnam that everyone in poor communities is similarly poor. But the finding also resonates with other empirical studies of inequality (see Elbers et al. 2004). While there may be poor localities where everyone is similarly poor, more in-depth analysis at the commune level (see targeting simulations described in Annex 4.1) suggests there is still substantial inequality at low levels of geographic disaggregation. Communes in Vietnam typically consist of four to six villages; empirical work suggests that villages tend to be more ethnically and economically homogeneous than communes. Table 4.3 Inequality and Wealth Measures for Provinces Provinces Gini Index Ratio of 90th to 10th Percentage of People Expenditure Percentile in the Richest 20% Mean Std. Err. Mean Std. Err. Mean Std. Err. Northern Mountain Ha Giang 0.374 0.018 4.93 0.35 3.55 0.89 Cao Bang 0.351 0.016 5.10 0.40 4.73 1.14 Bac Kan 0.321 0.018 4.21 0.32 5.31 1.62 Tuyen Quang 0.329 0.021 4.38 0.37 7.54 2.13 Lao Cai 0.397 0.019 6.12 0.53 7.38 1.99 Dien Bien 0.404 0.023 5.82 0.56 4.51 1.29 Lai Chau 0.376 0.017 4.82 0.29 2.99 0.80 Son La 0.360 0.013 4.82 0.27 4.20 1.02 Yen Bai 0.354 0.019 5.20 0.46 7.24 1.91 Hoa Binh 0.345 0.018 4.70 0.35 6.83 1.57 Thai Nguyen 0.308 0.021 4.11 0.42 13.33 3.44 Lang Son 0.325 0.018 4.31 0.32 5.77 1.69 Bac Giang 0.281 0.012 3.60 0.22 8.55 2.29 Phu Tho 0.305 0.013 4.01 0.26 11.30 2.21 Red River Delta Ha Noi 0.382 0.013 6.02 0.40 49.03 2.16 Quang Ninh 0.324 0.015 4.50 0.34 25.76 3.65 Vinh Phuc 0.275 0.012 3.47 0.19 15.81 2.73 Bac Ninh 0.297 0.014 3.85 0.26 22.08 3.55 Hai Duong 0.289 0.013 3.63 0.18 14.49 2.33 Hai Phong 0.322 0.014 4.32 0.28 30.29 3.26 Hung Yên 0.290 0.012 3.68 0.21 16.96 2.49 Thai Bình 0.271 0.014 3.36 0.19 9.40 2.33 Ha Nam 0.273 0.015 3.41 0.23 11.33 2.95 Nam Dinh 0.271 0.014 3.40 0.19 12.97 2.50 Ninh Bình 0.283 0.016 3.57 0.24 13.63 2.55 Central Coast Thanh Hoa 0.316 0.011 3.95 0.15 10.11 1.15 Nghe An 0.328 0.016 4.15 0.21 10.88 1.33 103 Provinces Gini Index Ratio of 90th to 10th Percentage of People Expenditure Percentile in the Richest 20% Mean Std. Err. Mean Std. Err. Mean Std. Err. Quang Binh 0.322 0.017 3.99 0.26 11.75 1.81 Quang Tri 0.323 0.012 4.42 0.25 9.45 1.51 Thua Thiên Hue 0.305 0.016 3.90 0.29 13.22 2.80 Da Nang 0.283 0.011 3.63 0.21 40.11 4.16 Quang Nam 0.281 0.009 3.55 0.17 8.04 1.42 Quang Ngãi 0.290 0.012 3.76 0.20 8.72 1.58 Binh Dinh 0.293 0.015 3.57 0.23 12.42 2.28 Phú Yên 0.297 0.015 3.60 0.22 9.69 2.02 Khanh Hoa 0.325 0.017 4.44 0.35 20.18 3.50 Ninh Thuan 0.313 0.015 4.19 0.30 7.28 1.92 Binh Thuan 0.287 0.012 3.64 0.19 10.02 1.91 Central Highlands Kon Tum 0.414 0.011 7.60 0.47 9.97 2.04 Gia Lai 0.374 0.008 6.18 0.24 8.87 1.16 Dak Lak 0.356 0.011 5.34 0.25 12.50 1.70 Dak Nong 0.307 0.007 4.44 0.15 7.03 1.19 Lâm Dong 0.337 0.010 4.98 0.23 16.80 2.00 South East Binh Phuoc 0.294 0.009 3.53 0.16 11.53 1.91 Tay Ninh 0.287 0.008 3.35 0.14 13.49 1.79 Binh Duong 0.300 0.008 3.62 0.15 22.47 3.65 Dong Nai 0.319 0.014 3.93 0.27 19.47 3.27 Ba Ria - Vung Tau 0.331 0.015 4.14 0.28 23.46 3.70 Ho Chí Minh 0.357 0.009 4.73 0.18 51.17 2.87 Mekong River Delta Long An 0.285 0.009 3.57 0.13 17.55 2.15 Tien Giang 0.277 0.010 3.46 0.14 18.18 2.72 Ben Tre 0.269 0.009 3.36 0.13 16.29 2.33 Tra Vinh 0.294 0.009 3.76 0.15 10.49 1.80 Vinh Long 0.284 0.011 3.58 0.17 16.81 2.66 Dong Thap 0.261 0.007 3.18 0.10 9.59 1.60 An Giang 0.278 0.009 3.39 0.13 9.98 1.49 Kiên Giang 0.293 0.010 3.72 0.14 9.43 1.48 Can Tho 0.328 0.017 4.29 0.33 22.59 2.76 Hau Giang 0.271 0.008 3.39 0.12 9.22 1.70 Soc Trang 0.298 0.011 3.79 0.16 8.44 1.46 Bac Liêu 0.271 0.010 3.32 0.13 7.25 1.56 Ca Mau 0.288 0.012 3.58 0.17 7.76 1.63 Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. 104 Map Map 55 4.4. Expenditure Gini Indices Map 4. 5Expenditure ExpenditureGini GiniIndices Indices Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts Source: Estimation basedEstimation Source: on the 2009 on the 2009and Population based Housing Population and Census and theand 2010 Housing Census the VHLSS. 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. th th Map Map 4.6 4.6 Ratio Ratio of the of the 90Expenditure 90th Expenditure Percentile Percentile toto the10th the 10 Expenditure Expenditure Percentile Percentile Map 4.6 Ratio of the 90th Expenditure Percentile to the 10th Expenditure Percentile Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. 116 116 105 4.20 Map 4.7 shows the locations of the wealthiest 20 percent of households in Vietnam—the so- 4.18 Map 4.7 shows the locations of the wealthiest 20 percent of households in Vietnam—the so- called middle class and rich. As expected, individuals in the top quintile of the per-capita expenditure called middle class and rich. As expected, individuals in the top quintile of the per-capita expenditure distribution are spatially concentrated in the Delta regions, especially in Hanoi and Ho Chi Minh City and distribution are spatially concentrated in the Delta regions, especially in Hanoi and Ho Chi Minh City the in inand immediate surrounding the immediate areas. areas. surrounding Map 4.7 Proportion of People in the Richest Expenditure Quintile (%) Map 4.7 Proportion of People in the Richest Expenditure Quintile (%) Panel Panel AA Provinces Provinces Districts Panel B Panel B Districts basedEstimation Source: Source: Estimation based on the 2009 on the 2009 Population Population and Housing and Housing Census Census and the 2010 and the 2010 VHLSS. VHLSS. D. D. TheThe Evolution Evolution ofPoverty, of Spatial Spatial to 2009 1999 Poverty, 1999 to 2009 4.19 Chapter 4.21 Chapter 1 documentsVietnam’s 1 documents rapid reduction Vietnam’s rapid reductionininpoverty sincethe povertysince the early early 1990s 1990s based based on on a range a range of poverty lines applied to successive rounds of the VHLSS. However, the VHLSS of poverty lines applied to successive rounds of the VHLSS. However, the VHLSS is only representative is only representative at higher levels of spatial aggregation, that is, by region and urban and rural sector. at higher levels of spatial aggregation, that is, by region and urban and rural sector. The 2009 poverty The 2009 poverty maps can be compared with 1999 poverty maps to measure progress at the maps can be compared with 1999 poverty maps to measure progress at the provincial and districts levels, provincial and districts levels, also to look at changes in the spatial distribution of poverty over time. also tosection This look at in the patterns changes spatial describes of poverty, of spatial distribution poverty albeit over leaving for time. futureThis indepthdescribes worksection spatial analysis of patterns of poverty, the causal mechanisms albeit leaving that for underpin future these work indepth analysis of the causal mechanisms that patterns. underpin these patterns. 4.20 Comparisons of maps 4.8 and 4.9 show that poverty fell most rapidly between 1999 and 2009 in the Comparisons 4.22 provinces and of maps 4.8 districts 4.9 Deltas. andtwo in the show that poverty and fell districts Provinces most rapidly between in the Northern1999 and 2009 in Mountains the andprovinces and districts Central Highlands in the two experienced Deltas. Provinces substantially and lower rates districtsreduction. of poverty in the Northern Mountains maps and District-level highlight the variation within provinces, such as in the Central Highlands. Central Highlands experienced substantially lower rates of poverty reduction. District-level maps highlight the variation within provinces, such as in the Central Highlands. 117 106 Map Map4.8 ProvincialPoverty Provincial 4.8 Poverty Rates Rates Map 4.8 Provincial Poverty Rates A 1999 PanelPanel A 1999 Panel Panel B 2009B 2009 Panel A 1999 Panel B 2009 Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Note : the Source: 1999 poverty Estimation rates based are 2009 on the obtained from Minot, Population Baulch, Census and Housing and Epprecht (2003). and the 2010 VHLSS. Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Map 4.9 District Poverty Rates Map 4.9 District Poverty Rates Map 4.9 District Poverty Rates Panel A 1999 Panel B 2009 Panel A Panel A 1999 1999 Panel B 2009 Panel B 2009 Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS. Note: rates Note: the 1999 poverty are obtained the 1999 fromare poverty rates Minot, Baulch, obtained and from Epprecht Minot, (2003). Baulch, and Epprecht (2003). 118 118 107 4.23 Areas with high incidence of poverty are not necessarily the areas with the highest numbers of poor people. 4.21 AreasFor example, with many provinces high incidence of poverty the not inare Northern Mountains necessarily have the areas a high with the incidence of poverty highest numbers ofhave but poorlow population people. densities, For example, thus account andprovinces many a small share forNorthern in the of thehave Mountains total a highin poor Vietnam.of incidence Map poverty but have low population densities, and thus account for a small share of the total 4.10 shows the density of the poor across the country in 1999 and 2009. In 1999, the poor were highly poor in Vietnam. Map concentrated 4.10 in the Red Riverthe shows density Delta of the poor and Mekong Riveracross the Delta; country these areasin 1999 had and 2009. moderate In 1999, poverty rates but the poor were highly concentrated in the Red River Delta and Mekong River Delta; high population densities. By 2009, however, poverty had become less spatially concentrated. these areas had The moderate poverty rates but high population densities. By 2009, however, poverty had become less number of poor decreased remarkably in the two Delta regions, but much less markedly in the Northern spatially concentrated. The number of poor decreased remarkably in the two Delta regions, but much Mountains and Central Highlands. less in the Northern Mountains and Central Highlands. Map 4.10 Map Poverty 4.10 Density Poverty Density (Number ofPoor) Poor) (Number of Panel Panel A 1999 A 1999 Panel PanelBB2009 2009 Source: Estimation based on the 2009 Population and Housing Census and the 2010 VHLSS Note: the 1999 poverty rates Source: are obtained Estimation basedfrom Minot, on the 2009Baulch, Epprecht and and Population (2003). Housing Census and the 2010 VHLSS Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). 4.24 Nearly all provinces and districts experienced a decline in poverty between 1999 and 2009 (figure 4.22 Nearly all provinces and districts experienced a decline in poverty between 1999 and 2009 4.4). But the rate of progress was slower in areas that had very high or very low rates of poverty in 1999, (figure 4.4). But the rate of progress was slower in areas that had very high or very low rates of and much faster in areas that started the period in the middle ranges (that is, with a headcount of 25 to 55 poverty in 1999, and much faster in areas that started the period in the middle ranges (that is, with a percent) of 4.5). (figure headcount 25 to 55 percent) (figure 4.5). 4.23 Provinces with lower levels of inequality in 1999 also in general achieved a larger reduction in poverty. This largely reflects the growing gap between Kinh and ethnic minority households; high inequality areas typically had a high proportion of ethnic minorities (figure 4.6). 119 108 Figure 4.4 Poverty Rates, 1999 and 2009 Figure 4.4 Poverty Rates, 1999 and 2009 Figure 4.4 Poverty Rates, 1999 and 2009 Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts Panel A Provinces Panel B Districts 100 100 100 100 80 80 80 80 raterate raterate Poverty Poverty 60 60 60 60 Poverty Poverty 2009 2009 40 40 40 40 2009 2009 The The The The 20 20 20 20 0 0 0 0 0 20 40 60 80 100 0 20 40 60 80 100 0 20 The 401999 Poverty rate 60 80 100 0 20 The 401999 Poverty rate 60 80 100 The 1999 Poverty rate The 1999 Poverty rate Source: Source:Estimation based Estimation on based onthe the2009 2009Population Population andHousing and HousingCensus Census and the 2010 and the 2010 VHLSS. VHLSS. Source: Note: the Estimation based 1999 poverty on rates 2009 Population theobtained are andBaulch, from Minot, Housing Census and and the Epprecht 2010 VHLSS. (2003). Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Figure 4.5 Poverty Reduction, 1999-2009, and Poverty Rate, 1999 Figure 4.5 Figure Poverty Reduction, 4.5 Poverty Reduction, 1999-2009, 1999-2009, and and Poverty Rate, 1999 Poverty Rate, 1999 Panel A Provinces Panel B Districts Panel A Provinces Panel A Provinces Panel B Districts Panel B Districts 40 60 40 60 points) points) (percentagepoints) (percentagepoints) 30 40 30 40 reduction (percentage reduction (percentage 20 20 20 20 Poverty reduction Poverty reduction 10 0 0 -20 -20 0 0 20 20 40 40 60 60 80 80 0 20 40 60 80 100 The 0 20 40 60 80 100 1999 Poverty The 1999 Poverty rate rate The 1999 1999 Poverty The rate Poverty rate Source: Source: Source:Estimation based Estimation Estimation on based based on onthe the the2009 2009 2009Population Population Population andHousing and and HousingCensus Housing Census and Census and the and 2010 VHLSS. the 2010 the VHLSS. VHLSS. Note: Note:the 1999 the poverty 1999 rates poverty are rates obtained are obtainedfrom fromMinot, Minot,Baulch, Baulch,and Epprecht and (2003). Epprecht (2003). Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). 4.25 4.25 Figure 4.6with Provinces Provinces with lower lowerin Change levels levels of of inequality Poverty, inequality in 1999 in 1999-2009, 1999 also also in Compared general in to general achieved the Initial Gini a achieved larger Index, larger1999 a reduction reduction in in Figure 4.6 Change in Poverty, 1999-2009, Compared to the Initial Gini Index, 1999 poverty. poverty. This largely This largely reflects reflects the growing gap between Kinh the growing gap between Kinh Panel and ethnic and ethnic minority households; high minority households; high Panel A Provinces B Districts Panel A Provinces Panel B Districts inequality areas typically inequality areas had a typically had a high high proportion proportion of ethnic minorities of ethnic (figure 4.6). minorities (figure 4.6). 60 40 Poverty reduction (percentage points) Poverty reduction (percentage points) 30 40 20 20 10 0 120 120 0 -20 .24 .26 .28 .3 .32 .34 .2 .25 .3 .35 .4 The 1999 Gini index The 1999 Gini index Source: Estimation Source: Estimation based basedononthe 2009 the Population 2009 andand Housing Population Census Housing and the Census and2010 VHLSS the 2010 VHLSS Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). Note: the 1999 poverty rates are obtained from Minot, Baulch, and Epprecht (2003). 109 Reduction Contribution of the Rural Nonfarm Sector to Poverty 4.26 A number of factors are responsible for differential rates of progress across provinces and Contribution of the Rural Nonfarm Sector to Poverty Reduction 4.24 A number of factors are responsible for differential rates of progress across provinces and districts in Vietnam, and new work is underway to better understand some of the key drivers of progress over the last decade. Income and employment diversification has been a strong force for growth and poverty reduction. Much attention has been paid to diversification linked to rural-to-urban migration and the role of remittances. In a number of other countries, the expansion of the rural nonfarm sector has been shown to play a beneficial role in rural development and improving the lives of the poor. The rural nonfarm sector can help absorb excess agricultural labor, provide insurance against agricultural shocks, reduce rural-to-urban migration and, more generally, promote a more equitable distribution of income (see, for example, Ferreira and Lanjouw 2001; Lanjouw and Lanjouw 2000; Oseni and Winters 2009). 4.25 Between 1999 and 2009, a major shift occurred in rural occupations in Vietnam. While in 1999, more than 81 percent of the working population worked in agriculture, by 2009, this has dropped to about 71 percent. The growth of the rural nonfarm sector has been primarily due to expansion in the number of of low-skilled blue collar occupations in the construction, manufacturing, trade, and food preparation sectors. More than half of the increase in fast-growing blue collar nonfarm industries in rural Vietnam is the result of an expanding construction sector (table 4.4). Table 4. 4 Rural Employment and Percent of the Working Population in Sector Description 1999 (%) 2009 (%) Farm All agriculture and forestry and fishing 81.4 71.2 Nonfarm Self-employed nonfarm, nonfarm wage 18.6 28.8 labor, rural-urban commuters White-collar nonfarm Finance, consulting, science, 5.9 5.8 government, television, healthcare, education, Communist party Blue-collar nonfarm Mining, processing, construction, 12.6 23.0 reparation, trading, food preparation, transportation, cleaning Construction All construction, site preparation, building 1.6 7.5 activities Other blue-collar nonfarm All other blue-collar nonfarm jobs 11.0 15.5 Source: 1999 and 2009 Vietnam Population and Housing Censuses. 4.26 Results from the district-level poverty maps, augmented with data from the 1999 and 2009 Population and Housing Censuses, were used to explore the determinates of rural nonfarm diversification and its contribution to poverty reduction. Proximity to an urban center was found to stimulate rural nonfarm employment, in particular, proximity to large cities (Lanjouw and Marra, 2013). In terms of economic significance, the nonfarm sector of rural districts that are on average 10 kilometers further removed from the nearest city grew 1.63 percentage points more slowly between 1999 and 2009. Although the absolute magnitude may seem small, providing jobs for around 2 percent of the working population for every 10 kilometers of urban proximity is substantial. In addition, analysis suggests that growth in the rural nonfarm sector did indeed contribute to poverty reduction between 1999 and 2009; the poverty headcount was reduced by .0186 (1.86 percent) for a 10- percentage-point increase in the growth in the nonfarm sector. A similar picture emerges when we consider reductions in the severity of poverty (P1), and even the poorest of the poor, captured in reductions in the squared poverty gap (P2), were found to benefit from an expanding nonfarm sector. These findings stand in contrast to Hoang et al. (2012), whose findings suggest that the very poor do not benefit from expansion in the rural nonfarm sector because they lack the education and skills to access nonfarm jobs. It is clearly important to look beyond the household level to understand the potential indirect labor market effects of an expanding nonfarm sector. 110 E. In what other Ways can Mapping Methods Inform Policy Design and Evaluation? 4.27 This chapter has documented changing patterns in the spatial distribution of poverty between 1999 and 2009. But what do these imply for the design of policy? A series of simulations were carried out to assess how much the spatial disaggregation provided by poverty maps can help to improve area-based targeting schemes in Vietnam (details provided in Annex 4.1). The simulations are based on a hypothetical transfer scheme that aims to minimize poverty at the national level (focusing on the squared poverty gap, or severity of poverty) by using spatial targeting at different levels of geographic disaggregation, that is, province, district, and commune. The initial results clearly show that in both 1999 and 2009 there were potentially large gains in targeting performance by disaggregating to the local level. An important corollary of these findings is that the benefits from spatial targeting become increasingly evident as more and more disaggregated data on poverty are considered. The simulations show that a given impact on poverty can be achieved at considerably less expense with detailed spatial targeting than with a uniform transfer. 4.28 A second key finding is that the benefits from spatial targeting, at any level of geographic disaggregation, are more clearly evident in 2009 than 1999. This finding follows directly from the evidence presented in the previous section on the changing spatial distribution of poverty in Vietnam over time. As Vietnam has prospered, moderately poor households living in relatively well-off areas in 1999 (for example, Red River Delta) were able to cross the poverty line, so that by 2009 such relatively well-off areas no longer contributed as much to overall poverty. Poverty has become more concentrated in poor districts. For policy makers, this is an important finding, because it indicates that there may be a stronger rationale for using area-based targeting to reach the poor today than was previously the case. 4.29 But these findings should be viewed as illustrative only. They do not take account of important practical and political considerations such as how the hypothetical transfers would be financed, the costs of administering such a scheme, possible behavioral responses of households, and the possibility of local capture linked to power and influence. The anticipated albeit hypothetical gains from targeting must be juxtaposed against the potential costs and political-economy considerations, and should be scrutinized against other possible policy objectives. In practice, a combination of geographic targeting between villages and means-tested targeting on poor households within villages is likely to be the best way forward for Vietnam. 4.30 We close this chapter with a brief assessment of the targeting performance of Program 30A, one of MOLISA’s newer area-based targeted poverty reduction programs. A welfare ranking of districts is drawn up, based on criteria developed by MOLISA (incorporating information on income, as opposed to expenditures, and other indicators of well-being), and the poorest 62 districts on the list are singled out for specific policy interventions (box 4.1). Mapping methods were used to see whether the 62 poorest districts identified by MOLISA’s criteria are also the poorest as measured by the per-capita expenditure criteria underpinning the Vietnam poverty map for 2009. Figure 4.7 illustrates the close correlation between the two approaches; the districts targeted by MOLISA are also among the poorest identified by the independent mapping methodology. Spatial targeting in Vietnam is not only warranted on empirical and conceptual grounds, but appears administratively and logistically feasible, as evidenced by one well-established program. 111 Box 4.1 Box Overviewof 4.1Overview Program30A ofProgram 30A Program30A, Program 30A,named namedafter Prime after Minister Prime Decision Minister 30A30A Decision in 2008, is a comprehensive in 2008, poverty is a comprehensive poverty reduction program targeted at 61 (now 62) of the country’s poorest districts through 2020. These lie in reduction program targeted at 61 (now 62) of the country’s poorest districts through 2020. These districts districts lie in 20 provinces throughout the country, but most of the districts are located in the 20 provinces throughout the country, but most of the districts are located in the northeastern mountainous northeastern mountainous region. The program focuses on four primary areas: (a) increasing region. The program focuses on four primary areas: (a) increasing income through production, job creation, and income through production, job creation, and labor exports; (b) improving education standards; labor exports; (b) improving education standards; (c) improving the quality of local administrators; and (d) (c) improving the quality of local administrators; and (d) investing in infrastructure. investing in infrastructure. Funding commitments for the different components are made in three-year tranches. According toFunding MOLISA, commitments state budget for the different funding was VNDare components for 2009–11 8.5made inFor trillion. three-year 2012–15, tranches. funding According is VND to MOLISA, 7.2 state budget trillion. funding for A substantial 2009–11 portion wasfunding of the VND 8.5 trillion. has gone For toward2012–15, funding boosting is VND incomes by 7.2 trillion. A paying citizens substantial portion of thespecified to protect areas funding has of forest. gone toward boostingas However, with Program incomes 135-II, by paying the vast citizens majorityspecified of to protect areas offunds forest.are invested However, infrastructure. inwith as Thus far, Program 135-II, no attempt the vast majority of been has fundsmade to evaluate are invested the impact Thus in infrastructure. of this program. far, no attempt has been made to evaluate the impact of this program. The 62 districts selected under Program 30A do not receive support directly only through 30A. The 62 districts selected under Program 30A do not receive support directly only through 30A. Their Their designation as particularly needy districts also makes them eligible for other targeted designation as particularly programs. in order also needy districts For example, makes them to improve cadre eligible forProgram quality, other targeted 30A isprograms. linked toFor theexample, 600 in order to improve cadre quality, Program 30A is linked to the 600 Deputy Chairman Deputy Chairman Program, which is run by the Ho Chi Minh Youth League and the Ministry Program, which is run by Chi the Ho of Minh Home YouthThis Affairs. League and the program, Ministry initiated of Home in 2011, Affairs. targeted 600This program,in communes initiated in 2011, an the 62 districts targeted 600 communes in the 62 districts an additional (trained) person additional (trained) person to support the People’s Committee. to support the People’s Committee. Figure District Poverty: 4.7 District Figure 4.7 MOLISA compared Poverty: MOLISA to Poverty compared to Map Estimates Poverty Map Estimates 124 112 Annex 4. 1 The Spatial Distribution of Poverty and the Gains from Spatial Targeting Chapter 4 documents changing patterns in the spatial distribution of poverty between 1999 and 2009. But what do these patterns imply for the design of policy? A series of simulations was carried out to assess how much the spatial disaggregation provided by poverty maps can help to improve area- based targeting schemes in Vietnam.26 We consider here the distribution of a hypothetical budget to the population of Vietnam. We assume that we have no information about the poverty status of this population other than the geographic location of residence and the level of poverty in each location. As a benchmark case, we make the extreme assumption of no knowledge whatsoever about the spatial distribution of poverty, in which case our given budget is distributed uniformly to the entire population. We set up a series of comparisons to this benchmark, where we assume knowledge about poverty levels in progressively smaller subpopulations. For a given level of disaggregation, we ask how knowledge about poverty outcomes across localities can be incorporated into the design of a transfer scheme so as to improve the overall targeting performance relative to the benchmark case. In light of the observations made above, concerning the evolving spatial distribution of poverty in Vietnam, we ask whether and how our conclusions differ between 1999 and 2009. We consider a transfer scheme that makes use of our knowledge of the spatial distribution of poverty in such a way that poverty is minimized at the national level. We consider the gains from spatial targeting at alternative levels of disaggregation. We focus on the squared poverty gap, a measure of poverty that is particularly sensitive to the distance between a poor person’s income level and the poverty line.27 We specify a poverty line that accords with a poverty rate of around 20 percent nationally, in each respective year, and we consider a modest hypothetical budget that would be insufficient, in and of itself, to eliminate all poverty, even if it were perfectly targeted at the household level. The results from this exercise show clearly, first, that in both 1999 and 2009, there are potentially large gains in targeting performance from disaggregating to the local level. These benefits are clearly seen when we examine the squared poverty gap as our poverty measure of choice. The impact on the headcount rate is, unsurprisingly, more muted, given that we do not “optimizeâ€? our transfer scheme with respect to this poverty measure. An important corollary of these findings is that the benefits from spatial targeting become increasingly evident as more and more disaggregated data on poverty are used. We show that a given impact on poverty can be achieved at considerably less expense with detailed spatial targeting than with a uniform transfer. The results from this exercise also show that the benefits from spatial targeting, at any level of disaggregation, are more clearly evident in 2009 than in 1999. This finding follows directly from the evidence presented in the earlier section on the changing spatial distribution of poverty in Vietnam over time. As Vietnam has prospered, moderately poor households living in relatively well-off areas in 1999 were able to traverse the poverty line, so that by 2009, such relatively well-off areas no longer contributed as much to overall poverty levels. Poverty has become more spatially concentrated. For policy makers, this is an important finding, because it indicates that there may be an even stronger rationale for spatial targeting of resources today than was the case a decade earlier.  e build on an earlier analysis in Ravallion (1993), who finds that spatial disaggregation to the broad regional level in 26 W Indonesia, the lowest level at which household survey data provide reliable estimates of poverty, improves targeting, but only to a modest extent. In contrast, Elbers et al. (2007) find that fine geographic targeting offers significant benefits over broad targeting.  e focus on the squared poverty gap because of its appealing properties from both a conceptual and technical point of 27 W view. The basic approach explored here would also work for other poverty measures, particularly Foster-Greer-Thorbecke measures with values of parameter α greater than 1. However, with the headcount measure (the FGT measure with α=0) welfare, “optimizationâ€? is not well defined and the approach taken here is thus less obviously applicable (see, for example, Ray [1998, 254–55]). 113 Transfer Scheme We postulate that the government has a budget, S, available for distribution and wishes to transfer this budget in such a way as to reduce poverty. We specify a baseline case in which the government is assumed to have no knowledge of who the poor are or where they are located. It is therefore unable to distribute its budget in any manner other than a lump-sum transfer to the entire population of size N. We thus calculate the impact of transferring S/N to the entire population. Kanbur (1987) shows that to minimize poverty summarized by the Foster-Greer-Thorbecke (FGT) class of poverty measures with parameter value α>1, the group with the highest FGT(α-1) should be targeted on the margin.28 Hence, to minimize the squared poverty gap (equal to a poverty measure from the FGT class with α=2), target populations should be ranked by the poverty gap (FGT with α=1) and lump-sum transfers made until the poverty gap of the poorest locality becomes equal to that in the next poorest one, and so on, until the budget is exhausted. Theand Budget Poverty results Linesalso show that the benefits from spatial targeting, at any level of from this exercise disaggregation, are more clearly evident in 2009 than in 1999. This finding follows directly from the We assume evidence thatin presented the budget the earlier The available section results on from for thethis changing distribution exercise spatial also show has been distribution that the of exogenously poverty benefits in Vietnam from set.over As is at spatial targeting, intuitively clear, any level of the time.potential As Vietnam benefits from targeting has prospered, moderately disaggregation, are more will poor vary with the households clearly evident in overall living 2009 inthan size relatively in of budget. well-off 1999. This areasIn finding in the 1999 followslimit, asfrom directly the the budget were able goes to traverse to infinity, the there poverty is no evidence line, need presentedso that for in by the 2009, section targeting, earlier relatively such as even on well-off a uniform the changing areas transfer spatial no longer contributed distributionwill eliminate of poverty poverty. in Vietnam over As a as much to overall benchmark, we poverty time. levels. identify Asthe Vietnam Poverty has has per-capita become prospered, more spatially moderately consumption poor value concentrated. of the living households 25thForinpolicy relatively percentilemakers, well-off of the areas in 1999 consumption this is an important finding, because were able it indicates to traverse that there the poverty may line, so thatbeby an even 2009, stronger such rationale relatively well-offfor spatial areas no longer contributed distribution. We scale this consumption value by the total population. Our benchmark budget is set 29 Gains from targeting targeting also vary of resources today than as muchwith towasthethechoice overall case poverty of poverty a decade levels. earlier. Poverty hasline. The more become higher the concentrated. spatially poverty line, the makers, For policy to equal 5 percentthis this of is total value. an important finding, because it indicates that there may be an even stronger rationale for spatial to theofnonpoor s need for targeting, as leakage targeting diminishes to zero. In this study, we select as the resources today than was the case a decade earlier. Transfer SchemeGains from targeting also vary with Gains nchmark a poverty from that targeting line yields also a povertyvary with rate the 20 the choice of exactly choice of poverty percent inof Theline. poverty line. both 1999 andThe higher higher the 2009. the poverty poverty line, line, the the less need less We targeting, for need for postulate targeting, thatas the leakage Transfer as leakage government Scheme to the has to the S, nonpoor nonpoor a budget, available diminishes diminishes for to to zero. distribution andzero. In this wishesIntothis study, study, transfer we we select select asas the the benchmark this budget benchmark in such a way a poverty a poverty as to line line that reduce that yields a We poverty. yields a poverty poverty rate exactly of exactly a baseline specifyrate of case in20 20 percent which percent in in both the government 1999 both is andand 1999 2009. 2009. mulating the Impact of assumed to Uniform Transfers have no knowledge We postulate that the government has a budget, S, available for distribution and wishes to transfer of who the poor are or where they are located. It is therefore unable to this budget in such a way as to reduce poverty. We specify a baseline case in which the government is distribute its budget in any manner other than a lump-sum transfer to the entire population of size N. We Simulating the Impact Uniform of Uniform Transfers Simulating Our policy simulation thus calculate thethe in Impact assumed the impact ofcase of have to of transferring no knowledge uniform S/N to the Transfers of who the is poor transfers entire calculated population. in they are or where are located. a very It is therefore unable to straightforward distribute its budget in any manner other than a lump-sum transfer to the entire population of size N. We nner. Budget S is divided Our policy Kanbur by Our total policy simulation (1987)thus population in simulation calculate shows the to impact the thatcase N. The inuniform of the of minimize resulting case poverty of transferring S/N transfer uniform to the transfers summarizedisentireathe added is Foster-Greer-Thorbecke transfers is in by population. calculated ato each calculated very inpredicted a very straightforward straightforward manner. (FGT) manner. Budget ispoverty S of class Budget divided Sby yis ( r )divided by total population N. The resulting transfer a is added to each predicted measures with parameter total populationvalue N.Ä®>1, Thethe group with the resulting highest FGT( transfer a isÄ®-1) shouldto each predicted added Kanbur (1987) shows that to minimize poverty summarized by the Foster-Greer-Thorbecke penditure in our bedatabase, to margin.27 targeted on the yield ch +a. to Hence, For each the minimize replication squared r we poverty estimate gap posttransfer national >1, the to Ä®(equal a poverty measure ( r) expenditure in (FGT) class our of poverty database, to measures yield ywith ch +a. parameter For value each group replication r with the we highest FGT(Ä®-1) should estimate posttransfer national verty. The average across expenditure from the the in our FGT class r be with replications database, Ä® =2), target targeted on the of to the estimated yield populations margin.27 Hence, be should . posttransfer For by to ranked minimize each the poverty replication poverty thesquared gap (FGT poverty rates r we gap with yields estimate Ä®=1) (equal our post-transfer to a poverty measure ratelump-sum and pected poverty national poverty. associated The transfers withaverage made from the FGTacross theuntil the benchmark, class the Ä® poverty with r=2), replications gap of untargeted target of locality the populations poorest the lump-sum estimated should becomes transfer be rankedposttransfer equal by to poverty scheme. the that inpoverty the This new gap (FGT rates yields with Ä® =1) our poverty. next poorest one, and expected The so poverty on, and average until the rate lump-sum across budget associated transfers the exhausted. is made with r replications the until benchmark, the of the poverty gap untargeted of the estimated poorest locality posttransfer lump-sum transfer becomes equal topoverty scheme. therates that in This new rate can mated povertyyields ourbe compared expected to the next poverty original rate national-level associated with poverty is the estimate from the poverty benchmark, estimated poverty poorest one, rate can andcompared be so on, until the to budget the national-level untargeted exhausted. original poverty estimate lump-sum from the transfer poverty p to gauge the impact scheme. the of This Budget and transfer. new Poverty Linesestimated poverty rate can be compared to the original national-level poverty map to gauge the impact of the transfer. estimate from the Budget povertyand Poverty map Lines the impact of the transfer. to gauge We assume that the budget available for distribution has been exogenously set. As is intuitively mulating the Impact “Optimalâ€? ofpotential clear, the Simulating benefits Geographic the Impact assume We targeting from Targeting that will of “Optimalâ€? the budget vary with available the overall Geographic for distribution size of budget. Targeting has been exogenously In the set. As is intuitively limit, as the Simulating the Impact of “Optimalâ€? Geographic Targeting budget goes to infinity,clear, there the is potential no need benefits for from targeting, targeting as even awill vary uniform with the transfer overall will size eliminateof budget. poverty. In the limit, as the Simulating the of impact Simulating the budget goes “optimalâ€? to infinity, targeting there scheme is no need for is targeting, slightly as even complicated. more a uniform will eliminate poverty. transfer Following As a benchmark, we identify the impact of the “optimalâ€? targeting scheme is slightly more complicated. Following the per-capita consumption value of the 25th percentile of the consumption Simulating distribution.28the We impact scaleAsthisbenchmark, a of we identify the “optimalâ€? consumption value targeting by the per-capita the scheme total consumption population. Ourvalue is slightly of the complicated. more benchmark 25th percentile of Following the consumption Kanbur nbur (1987), we want to equalize Kanbur (1987), the we following want distribution.28 toWe expression equalize scale the this across following consumption the expression value poorest by across the total thebudget locations population. is set country: of abenchmark poorest Our to locations of a budget iscountry: set to (1987), we want equal 5 percent to of this value. the following expression across the poorest locations of a country: equalize total equal 5 percent of this total value. z z ( z  y G acc() ) ³ ac dF ( z  y  ac )  dFc ( y ) , (7) (7) Gc (ac(7) ) ³ 0 0 y) , c( 27 27and Thorbecke (1984), the FGT class of poverty measures takes the following form: Following Foster, Greer Following and Thorbecke Foster, Greer(1984), (1984), FGT the of poverty measures class ofmeasures 28 Following Foster, Greer and Thorbecke the FGT class poverty takestakes the following the following form: form: which is z times the poverty gap in location c, after every person in the location has received a transfer ac. 1 ich is z times the poverty 1 Fc(y) gap in location is the FGT (D ) average c, after of ( the w Revery )¦ FGTwi (person simulated1 ( D) ( xi( ))Dthe / zin expenditure)¦ w location has ))D received  ( xi / z i (1distributions of c. The a transfer function ac. (x)+ gives the y) is the average of the R partâ€? simulated ¦ expenditure distributions ¦ wi (x)+=x, if x of function 0. c. The otherwise (x)+ givesac the i “positive of its argument, that is, is positive, Transfers (which must be where xi is xiper capita expenditure where for xi is those per individuals capita expenditure for weight with individuals thosewith wi who are below with the weight wpoverty i who line are and below zero the for lineline poverty and for those zerofor ositive partâ€? of its w argument, here is per nonnegative) that is, capita add (x)+=x, expenditure up to a givenif forx is positive, those budget S: otherwise individuals weight 0.wiTransfers who are ac below (which the poverty must be and zero line and ¦ i is total population ¦ i size. takes a value of 0 for the Headcount Index, 1 for the those above, z is the povertythose line above,and z is w the is total poverty linepopulation and wsize. is D takes total a value population of size. 0 Dfor the takes aHeadcount value of 0 for the Headcount to a given nnegative) add up above, budget z is the poverty S: Index, 1 for Poverty the (8) Gap Poverty and 2 for 28 Index, Gap and the 1 for 2 Squared ¦ the for Poverty the N a Squared Poverty c c Gap Sand Gap., Poverty 2 for Gap. For the Squared For further furtherPoverty discussion, Gap. discussion, For see seefurther Ravallion discussion, Ravallion (1994).see Ravallion (1994). (1994). ¦ 28 29 The Theconsumption consumption The consumption distribution is constructed distribution c distribution on the basis is constructed on isthe constructed of basis the average, ofon the the basis of across average, the average, r replications, across across r replications, of household-level r replications, of household-level of household-level predicted (8) predicted per-capita N S c ac consumption per-capita consumption , in the predicted per-capita in consumption the population population census. census. in the population census. c where Nc is the population size of location c. After transfers, 126 there is a group of locations all sharing the 126 same (maximum) poverty gap rate in the country. These are the only locations receiving transfers. We ere Nc is the population size of location c. After transfers, there is a group of locations all sharing the describe below how this problem is solved given that we are working with a database of incomes for me (maximum) poverty gap rate in the country. These are the114 only locations receiving transfers. We every household in the 15 percent sample population census. cribe below how this problem is solved given that we are working with a database of incomes for ry household in the 15 percent sample population census. z ³ (z  y  a )  (7) Gc ( a c ) c dFc ( y ) , 0 the location son in which has received is z times the poverty gap in ac. a transfer location c, after every person in the location has received a transfer ac. which is z times the poverty gap in location c, after every person in the location has received a butions is The of c. Fc(y) function the average transferof (x)+ ac.thegives Fc(y) the R simulated is the average expenditure of the R distributions of c. The simulated expenditure function (x)+ distributions gives of c. the The function (x)+ e, otherwise 0. Transfers “positive partâ€? givesacthe of its (which argument, must that “positive be is, (x)+=x, partâ€? that is,otherwise if x is positive, of its argument, x Transfers (x)+=x, if0. (which must acotherwise is positive, be 0. Transfers ac nonnegative) add up to a given budget S: (which must be nonnegative) add up to a given budget S: (8) (8) ¦N a c c c S, is a group there where of Nc is locations the all sharing population size of the location c. After transfers, there is a group locations of is where Nc is the population size of location c. After transfers, there a groupall of sharing the locations all sharing e the only same locations (maximum)receiving povertytransfers. gap rateWe in the country. These are the only locations receiving transfers. We the same (maximum) poverty gap rate in the country. These are the only locations receiving transfers. re working describewith a database below Wehow of incomes this describe problem for is solved below how given that this problem we are is solved working given that we with a database are working of with a incomes databaseforof incomes every household in the 15 percent sample population census. for every household in the 15 percent sample population census. Solving Solving the Problem the Problem – “Optimalâ€? – “Optimalâ€? Geographic Geographic Targeting Targeting nterest in minimizing As describedAs described the FGT2, in Elbers Elbers inoptimal et al. al. (2007), et (2007), givengiven our interest in ourinterest in minimizing minimizingthe the FGT2, FGT2, optimal optimal geographic targeting implies that after transfers there is a group of locations all sharing the same (maximum) a group locations ofknow geographic we how all compute targeting to sharing implies the thatsame after transfers there is aS(G) group of locations all sharing the same poverty gap inS(G), the country.we simply adjust We determine G until the level of equals transfers thegoingoriginally to each given locationbudget for solving by first el of transfers (maximum) going to poverty transfers S. Toaimplementeach gaplocation in the by first country. this scheme, We we must determine the level of transfers going to each location by first we know how different problem. compute tominimum S(G), we Following simply thesolve adjust notation G the following equation until introduced S(G) equals above, for ac: the minimum budget S(G) consider the originally budget S(G) given budget for d above, consider solving a the different needed to bring down all locations’ poverty gaps to at most level G/z. minimum problem. budgetFollowing S(G) the notation introduced above, consider the This amounts to transferring an transfers G/z. levelwe needed S. This Toamounts to how bringimplement this to transferring scheme, z an we must solve the following equation for ac: know todown amountcompute all G ac locations’ (G) ( zwe S(G), ³ to poverty y  ac )  adjust locations simply gaps dFcwith (yto)Gatuntilmost level before-transfer S(G) G/z. poverty equals This amounts the originally gaps to transferring above given G/z, for budget an such that (A.1) G ( a ( G )) G 0 . G ( a (G )) G abovetransfers G/z, amount suchS.that ac To implement howto (G) c locations c G with thisz scheme, . before-transfer Once we must poverty solve the following adjust G gaps until above equation G/z, thethat such for ac: c c . Once we know (A.1) transfers InS.what ToOnce to compute follows implementwe know we S(G), this 0 (z drop how ³  the zscheme, ywe a location to compute c ) dFc ( y ) simply we must index S(G), . solve c for we theease simply S(G) equals of following notation. adjust G until equation Using S(G) originally given budget for integration equals ac: by parts it the originally for can budget given be shown that for transfers In(A.1) what follows we drop the G S. To 0 ³  y  ac ) this ( zimplement z location index c 127 dFcscheme, ( y) for we must solve the following equation for ac: . ease of notation. Using integration by parts it can ³  be shown that (A.1) . G ( z z y  ac ) dFc ( y) z a (A.1) .) ease of In what follows we (A.2) G( a) 0 the(location drop 0 ³z ya )  dFc index (y for ³ 0 ( y )dy. Using integration by parts it can Fnotation. be shown that In what follows wez drop the location index c a ease of notation. Using integration by parts it can be z for In what (A.2) In be shown that other follows shown words, G( that a) drop(the we we need0 ³z to y  a)  dF z location compute index ( y )c for ease the ³ 0 surface z a F( ofy )notation. under dy. the Using integration by parts it can expenditure distribution between ³ ³  (A.2) expenditure levels y=0 and y=z-t, G ( a ) ( z  y  a ) dF ( y ) F ( y ) dy for values of t up to z. Instead of computing G(t) exactly, we use a . In(A.2) other words, we need to 0 compute the surface 0under the expenditure distribution between z z a simple approximation. For this to work  we y split the(interval [0,zF] in n equal segments and assume that the expenditure levels y=0 andG “poverty (A.2) In mappingâ€? other In words, othersoftware y=z-t, we need words, (a) for has we 0 to ³ (z generated values compute need to a) t dF of expected compute up to the surface ³ y ) z. Instead headcounts0 under the surface (y of for )dy the . computing poverty under expenditure thelines G(t) exactly, we use a z k/n distribution expenditure , where k=0, between distribution …,n. simple approximation. For this to work we split the interval [0,z] in n equal segments and assume that the between In other words expenditure levelswe y=0 haveand expenditure a table y=z-t, levels ofy=0 F (z and for k/n values). Using y=z-t, of fortthe up table valuesto z.of we t approximate Instead up to of computing z. Instead F(yof ) by linear G(t) computing interpolation exactly,G(t) we usefor exactly, awe use a In other software “poverty mappingâ€? words, we hasneed generated to compute expected the surface under headcounts for poverty lines z k/n,distribution the expenditure where k=0,between …,n. y between simple table simple approximation. values. approximation. ForWith thisthe For to approximated work we this to work split we split[distribution, expenditure the interval the 0,z]interval in n equal [0,z] it is in n equal easy segments to solvesegments and assume that assume and for transfers theas that other words we In expenditure levels the have y=0 a table “poverty and of F(z k/n y=z-t, mappingâ€? for ).values softwareUsinghas thet generated of table up to we z. approximate Instead expected F(y) by linear computing of headcounts G(t) for interpolation exactly, poverty we for 29 lines zuse a k/n, where a function “poverty of G (see mappingâ€? below). has software In practice, generated we find thatheadcounts expected n = 20 gives forsufficiently poverty lines precisez k/n results. , where k=0, …,n. y between simple approximation. table values. k=0, …,n. WithFor In thethisapproximated other to work we words we havesplit the expenditure a tableintervalof F(z [0,z distribution, k/n). n equal ] in Using it is the segments easy to solve table weand for assume transfers approximate that asthe F(y) In other words we have a table of F(z k/n). Using the table we approximate F(y) by linear interpolation 29 forby linear a function “poverty of G (see mappingâ€? below). software In practice, has generated we find that expected n = 20 gives headcounts sufficiently interpolation for y between table values. With the approximated expenditure distribution, it is. easy tofor poverty precise lines z k/n results., where k=0, …,n y between Thetablecomputational values. With set-up is as follows expenditure the approximated (note that the numbering it distribution, we is adopt easy to meanssolvegoing from z in for transfers as In other wordssolve we have a table ofas for transfers F(a z k/n ). Using function of G the table (see we approximate below). In practice, F( y) by we find linear that n interpolation for = 20 gives sufficiently athe direction function of G of(see 0 rather below). than Inthe other way practice, we findaround). that n Define = 20 gives b0=0,sufficiently and for k=1,...,n, precise bk 29 as the surface under results. y between precise The computational table values. With results. set-up the is approximated as follows (note that the numbering expenditure distribution, we it adopt is easy means to solvegoing for from z in as transfers the (approximated) expenditure distribution between z-kz/n and z-(k-1)z/n, divided by z: 29 theadirection function of 0 G rather (see than below). The computational the In other practice, way set-up around). we find is as (note Define that follows n = b 20 (note 0 =0, givesand for sufficiently that the numbering k=1,...,n, b precisek as the surface results. under The computational set-up is as follows that the numbering we adoptwe means adopt means going from going z in from z in the (approximated) the expenditure direction of distribution 0 rather than between the other z-kz/n way and around). z- ( k-1 )z Define/n , divided b0=0, by and zfor : k=1,...,n, bk as the surface the direction of 0 rather than the1 other way around). Define b0=0, and for k=1,...,n, bk as the surface under 30 The computational bk set-up is under the (approximated) F( asz kz / n)(note follows expenditure that z  (the  F (distribution k numbering 1) z / n) we between adopt z-kz/n and means from z in going divided z-(k-1)z/n, by z: the (approximated) expenditure2 distribution between z-kz/n and z-(k-1)z/n, divided by z: (A.3)of 0 rather than1 the direction thenother way around). Define b0=0, and .for k=1,...,n, bk as the surface under the (approximated) (A.3) bk expenditure F ( z  kz / n) distribution  F (z  between (k  1 z-kz/n ) z / z- and n) (k-1)z/n, divided by z: (A.3) 2 n 1 . Let g0 be the original bk poverty F ( z gap,  kzor / nin) terms F (z  of(the 1) z / n) above, g0=G(0)/z. Fork=1,...n, put k discussion (A.3) Let g0 be the original 2n 1 gap, poverty gap, or in terms of the discussion . above, g0=G(0)/z. Fork=1,...n, put Let g0 be the original bk poverty gk 2 gn F (z  or kz in/n terms )  Fof ( zthe (k  1) z / n) discussion above, g0=G(0)/z. Fork=1,...n, put (A.3) (A.4) (A.4) k 1  bk . . . Let g0 be the original poverty gap, or in terms of the discussion above, g0=G(0)/z. Fork=1,...n, put (A.4) g k g k 1  bk . Let g The be g0k30 are O the the ther original poverty interpolation poverty gaps ofgap, schemes the are or in approximated terms For possible. the discussion of expenditure instance, if the above, distribution poverty g0=G gap for ( successively is given0)/zat. For tablek=1,...n lower values put an even ,zk/n, lines z-kz/n poverty(A.4) . Letg simpler k beg computation a k the  bk . itself. presents k 1 per-capita transfer Often, the poverty needed to mapping bring down software the will giveline poverty percentiles to z-kz/n of :the expenditure The gk are the povertyThese distribution. gaps can of the also approximated be used for interpolation, expenditure but the distribution formulas are for more successively cumbersome, lower since the percentiles poverty lines(A.4) are not equallyg z-kz/n. Let ak be the per-capita k g spaced. k  1  b k . transfer needed to bring down the poverty line to z-kz/n: The gk are the poverty (A.5) a k kz gaps / n of the approximated expenditure distribution for successively lower . poverty lines z-kz/n. Let ak be the per-capita transfer needed to bring 115 down the poverty line to z-kz/n: The gk are the a poverty kz /gaps n of the approximated expenditure distribution for successively lower (A.5) k . poverty lines We can z-kz/n now solve. Let for ak be the per-capita per-capita transfers transfer as a function needed ofto bring the down poverty intended the poverty gap lineg coefficient 1 0.75 Inequality< 0.5 0.25 0 Ͳ0.25 Source: Source: World World Bank Bank estimates estimates fromfrom a Shorrock’s a Shorrock’s decomposition decomposition by income income source. by source. Note: A relative concentration coefficient greater than 1 suggests that the income source is inequality increasing, and a Note: A relative concentration coefficient greater than 1 suggests that the income source value less than 1 suggests that it is inequality decreasing (that is, it is not disproportionately amongincreasing, is inequality concentrated richer and a value households). less than 1 suggests that it is inequality decreasing (that is, it is not disproportionately concentrated among richer households). 6.51 Income from the agricultural sector, notably income from crop activities, agricultural wage Figure 6.12 Contribution of different Income Sources to the Gini, 2010 labor, and livestock and aquaculture, is inequality decreasing. Agricultural wage labor and cropping activities are among the most equalizing income components.52 A rise in the relative concentration 0.45 coefficient of agriculture between 2004 and 2010 implies that the extent to which agriculture was Manufacturing Services Pensions ShareofGiniCoefficientof 0.4 equalizing declined over time. Relative to its share of income, however, the contribution of the 0.35 agricultural sector to overall inequality is low; the agricultural sector (including agricultural wages) 0.3 contributed approximately 29 percent of total income but accounted for only 15 percent of inequality. inequality In0.25 rural areas, agricultural sideline activities were a relatively equalizing source of income in 2004; in 0.2they had become mildly disequalizing, a change that reflects the faster growth in these sources 2010 0.15 of income among richer rural households. 0.1 0.05 The distribution of remittance incomes has become more equalizing over time in both rural and 6.52 urban 0 areas. In 2004, the share from remittances in the richest quintile was double that in the poorest quintile; by 2010 the shares of remittances were similar. The change in the distributional impact on 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 remittances appears to be predominantly driven by changes in migration patterns among richer households. The quantitative and perceptions studies both suggest a declining role for higher-paid Agriculture Livestock Non Agricultural NonͲ Remittances Scholarships Other and richer international migration among households; Agricultural  Wages the share of remittances Agricultural  coming from international migration has declined from 35 percent of remittances to 30 percent over time. Income from Source: 2010 VHLSS. remittances dropped in absolute terms in the top quintile, and the share of international remittances declined from 47 percent of remittance income to 42 percent among the richest 20 percent of the The distribution of remittance incomes has become more equalizing over time in both rural and 6.52 population. urban areas. In 2004, the share from remittances in the richest quintile was double that in the poorest quintile; by 2010 the shares of remittances were similar. The change in the distributional impact on remittances appears to be predominantly driven by changes in migration patterns among richer households. The quantitative and perceptions studies both suggest a declining role for higher-paid  gricultural sidelines activity, notably livestock, aquaculture, and agricultural services, are the least equalizing of all 52 A agricultural sources and contribute more to income inequality than crop income. This is corroborated when examining the 181 continue to be an important source of income for both rich structure of incomes across income quintiles; sideline activities and poor households. 160 Figure Figure of of 6.12 Contribution 6.12 Contribution different Income different Income Sources to the Sources toGini, the2010 Gini, 2010 0.45 ShareofGiniCoefficientofinequality Manufacturing Services Pensions 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 Agriculture Livestockand Non Agricultural NonͲ Remittances Scholarships Other Acquculture Agricultural Wages Agricultural Businesses Wages Source: 2010 VHLSS. Source: 2010 VHLSS. 6.53 Households working in the nonagricultural sector earn more than those working in the agricultural 6.52 sector, and The distribution their incomes of remittance hasgrown have incomes become at a faster more pace. equalizing Figure over time in6.13 bothshows per-capita rural and incomes urban conditional areas. In 2004, the sharethe upon fromsector of employment remittances of quintile in the richest the household was double head. that Incomes of households in the poorest with a household quintile; by 2010 the head sharesemployed in white-collar of remittances were similar. The change in occupations in the nonagricultural the distributional sector on are highest impact in both urban remittances andto appears rural areas, followed be predominantly the by by driven incomes changes of self-employed in migration nonagricultural patterns among richer workers. In rural areas, households. Thehouseholds quantitativewhose head works and perceptions in both studies suggest have agriculture the lowest a declining role forincomes higher-paidin both periods and the lowest international average migration richerNote growth. among that the households; thedifference between coming share of remittances these households and agricultural from international households was relatively small in 2004 but has grown over time. migration has declined from 35 percent of remittances to 30 percent over time. Income from remittances dropped in absolute terms in the top quintile, and the share of international remittances declined from 47 Figure percent of6.13 income Income Per-capita remittance per among to 42 percent Year by theOccupation of the richest 20 percent Household of the population. Head in Rural and 2004 and 2010 Urban Areas, 2004 and 2010 6.53 Households working in the nonagricultural sector earn more than those working in the 40,000 agricultural sector, and their incomes have grown at a faster pace. Figure 6.13 shows per-capita incomes conditional upon the sector of employment of the household head. Incomes of households with a 35,000 household head 30,000 employed in white-collar occupations in the nonagricultural sector are highest in both ThousandJan.2010VND urban and rural areas, followed by the incomes of self-employed nonagricultural workers. In rural areas, 25,000 households whose head works in agriculture have the lowest incomes in both periods and the lowest growth. Note that the difference between these households and agricultural households was average20,000 relatively small in 2004 but has grown over time. 15,000 2004 10,000 2010 5,000 0 19 WhiteͲCollar BlueͲCollar SelfͲEmployed Agriculture WhiteͲCollar BlueͲCollar SelfͲEmployed Employee Employee Employee Employee Rural Urban OccupationofHouseholdHead Source: Source: 2004, 2004, 2010 2010 VHLSS. VHLSS. 161 an individual works in the agricultural or 6.54 Education is an important determinant of whether nonagricultural sector, and the type of nonagricultural work conducted. The relationship between education and employment type can be readily seen for more recent labor market entrants who have 6.54 Education is an important determinant of whether an individual works in the agricultural or nonagricultural sector, and the type of nonagricultural work conducted. The relationship between education and employment type can be readily seen for more recent labor market entrants who have completed their schooling. Figure 6.14 shows the structure of employment for workers aged 25 to 30 in 1998 and 2010. Having an upper secondary education or above is a significant determinant of having nonagricultural employment, and those with a college education are the most likely to be found in more attractive, higher-skilled employment.53 Figure 6.14 Workers Aged 25-30 by Education Level and Job Type Figure 6.14 Workers Aged 25-30 by Education Level and Job Type 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1998 2010 1998 2010 1998 2010 1998 2010 PrimaryͲ Rural PrimaryͲ Urban UpperSecondary UpperSecondary andAboveͲ Rural andAboveͲ Urban ManualWork LowerͲSkilled HighSkilled Source: Source:2010 VHLSS. 2010 VHLSS. Note: High-skilled workers are professional/office workers. These positions are usually classified as white-collar work. Note: High-skilled Lower-skilled workers workers are workersare in professional/office workers. the service sector, sales, These machine positions operators, and are usually skilled classified asworkers. manual/handicraft white-collar work. Lower-skilled Manual workers workers include are workers agricultural in the laborers andservice unskilledsector, manualsales, machine operators, and skilled manual/handicraft workers. workers. Manual workers include agricultural laborers and unskilled manual workers. Returns 6.55 6.55 Returnsto education to education haveincreased have increased over over the the2000s, 2000s, with with larger larger substantially substantially increases increases for for workers in urban areas (figure 6.15). Empirical work carried out for this workers in urban areas (figure 6.15). Empirical work carried out for this report finds evidence of risingreport finds evidence of rising returnsreturns to education in the wage labor market during the 2000s; for non-agricultural jobs, the hourly to education in the wage labor market during the 2000s; for non-agricultural jobs, the hourly wage wage return to a year of schooling increased from 5.3 percent in 2004 to 5.8 percent in 2010. The return to a year labor incomeof schooling increased return to education fromon (based 5.3 percent total earnings) in 2004 to 5.8 is greater percent than the wagein 2010. returnThe (basedlaboron income return to education (based on total earnings) is greater than the wage return (based on hourly earnings) to education, since more-educated individuals work longer hours in the wage labor market than less- educated individuals. An additional year of education is estimated to have raised labor incomes by 9.7 percent in 2010 compared to a labor income return of 8.9 percent in 2004. Returns to education are higher for workersThose 53  in urban areas with upper in ruraland than education secondary areas aboveandare have risen still likely to be faster over found doing time. In unskilled workurban in ruralareas, an additional areas, either in the agricultural sector or as an unskilled manual laborer in the nonagricultural sector. In the qualitative assessment, focus year of schooling was groups in rural associated areas 7.6 percent with a where discussed instances individualsincrease in hourly who had obtained wages, were higher education while in rural unable areas it was to find skilled with associatedwork a 4.1 (either percent lower- increase. or higher-skilled Within work), and hence rural areas, returned returns to farming. to attributed They education amongobservation ethnic minorities this worrying to differences in the quality of education between urban and rural areas, and to students choosing fields of study, such as than those are lower pedagogy, accrued for which by the labor market majority, demand and appeared to decline between 2004 and 2010. The lower is limited. returns for ethnic minority workers reflect the fact that minorities tend to work in lower-paid occupations, including wage employment in the agriculture sector. 162 assessment, focus groups in rural areas discussed instances where individuals who had obtained higher education were unable to find skilled work (either lower- or higher-skilled work), and hence returned to farming. They hourly earnings) to education, since more-educated individuals work longer hours in the wage labor market than less-educated individuals. An additional year of education is estimated to have raised labor incomes by 9.7 percent in 2010 compared to a labor income return of 8.9 percent in 2004. Returns to education are higher for workers in urban areas than in rural areas and have risen faster over time. In urban areas, an additional year of schooling was associated with a 7.6 percent increase in hourly wages, while in rural areas it was associated with a 4.1 percent increase. Within rural areas, returns to education among ethnic minorities are lower than those accrued by the majority, and appeared to decline between 2004 and 2010. The lower returns for ethnic minority workers reflect the fact that minorities tend to work in lower-paid occupations, including wage employment in the agriculture sector. Figure 6.15 Hourly Wage and Labor Income Returns to Schooling Figure 6.15 Hourly Wage and Labor Income Returns to Schooling 12% Returntoanadditionalyearofschooling 10% 8% 6% 4% 2% 0% 2004 2010 2004 2010 2004 2010 2004 2010 2004 2010 Ͳ2% Ͳ4% National Urban Rural Rural,Minority Rural,Majority WageReturns LaborIncomeReturns Source: 2004, 2010 VHLSS. Source: 2004, 2010 VHLSS. 6.56 The increase in returns over time has increased the gap between the wages and incomes of The increase 6.56 individuals in returns with higher over levels and lower time has increased of education the gap (World Bankbetween the wages staff estimates). 54 and incomes of Since education 53 individuals with higher and lower levels of education (World Bank staff estimates). is unequally distributed across the working-age population and adjusts only slowly overeducation Since time, someis unequally peopledistributed across will benefit more the working-age from population nonagricultural growthand adjusts and higheronly slowly returns to over time, some education people than others. Therefore, nonagricultural growth and rising returns to education are associated with will benefit more from nonagricultural growth and higher returns to education than others. Therefore,rising inequality in income. growth and rising returns to education are associated with rising inequality in income. nonagricultural 6.57 6.57 TheThelinklink between between education education andand rising rising income income inequality inequality cancan be explored be explored through through examining examining the the relative gap between the incomes of more and less educated households, relative gap between the incomes of more and less educated households, which rose between 2004 and which rose between 2010.2004 and 2010. In 2004, In 2004, households households with at least one with at least one working-age working-age individual with individual with a college a collegeearned education education1.3 earned 1.3 times more income than those with an upper-secondary-educated individual, and 2.5 times more income than those with an upper-secondary-educated individual, and 2.5 times more than times more than households with no education. By 2010, the college-educated households earned households with no education. By 2010, the college-educated households earned 1.7 and 3 times more, respectively. Figure 6.16 shows income in urban and rural households, by education level. More educated households earn more than less educated households, and the incomes of the most educated households grew faster than all other education categories between 2004 and 2010 in both rural and urban areas. Although urban 54  There hashouseholds continued been a substantial toreturns rise in the earn more in every to education education over time, although category in the majority of2010, this riseas they beendid has in driven by urban areas. Assessments of the average wage earned by individuals with different 2004, the ratio of incomes of rural households to urban households at education levels above lower levels of education find low rates of return in the early 1990s. In 1993, the return to education using a basic Mincerian earnings equation was found to secondary be has fallen over approximately This suggests time. (Gallup 4 percent 2002; Glewwethat and the Patrios decline in mean 1999). Returns incomes between in the 1990s were low rural and urban by international standards, although they were similar to rates of returns found in China in the early 1990s (Psacharopoulos 1994). 53 There has been a substantial rise in the returns to education over time, although the majority of this rise has been driven by urban areas. Assessments of the average wage earned by individuals with different levels of education find low rates of return in the early 1990s. In 1993, the return to education using a basic Mincerian earnings 163 equation was found to be approximately 4 percent (Gallup 2002; Glewwe and Patrios 1999). Returns in the 1990s were low by international standards, although they were similar to rates of returns found in China in the early 1990s (Psacharopoulos 1994). 1.7 and 3 times more, respectively. Figure 6.16 shows income in urban and rural households, by education level. More educated households earn more than less educated households, and the incomes of the most educated households grew faster than all other education categories between 2004 and 2010 in both rural and urban areas. Although urban households continued to earn more in every education category in 2010, as they did in 2004, the ratio of incomes of rural households to urban households at education levels above lower secondary has fallen over time. This suggests areas that the due to the is decline relatively in mean richer, incomes more educated between individuals rural and in rural urban areas areas is due catching to the up to relatively their urban richer, more peers, rather than to catch-up at the bottom end of the income distributions. educated individuals in rural areas catching up to their urban peers, rather than to catch-up at the bottom end of the income distributions. Figure 6.16 Per-capita Income per Year by Education of most Educated Working-age Household Member, Figure 6.16 Per-capita Income per Urban and Year Rural by Education Households, 2004 of andmost 2010 Educated Working-age Household Member, Urban and Rural Households, 2004 and 2010 35,000 PerͲcapitaIncomeinThousandJan.2010VND 30,000 25,000 20,000 15,000 10,000 5,000 0 Noeducation Primary Lowersecondary Uppersecondary Collegeor vocationalTraining Urban2004 Urban2010 Rural2004 Rural2010 Source: VHLSS. 2010 VHLSS. Source: 2010 E. Inequalities I. Inequalities E. in Opportunities that inOpportunities Perpetuate that PerpetuateIncome Differences Income across Generations Differences across Generations 6.58 The analysis of opportunities is predominantly focused on education. This choice of focus was driven in part by the perceptions study; education and employment were central concerns in many focus 6.58 The analysis of opportunities is predominantly focused on education. This choice of focus groups. was focus Thisin driven part was also by the motivated by perceptions the empirics, study; education which and suggest employmentan increasingly important were central role of concerns in education as a determinant of income inequality. It is recognized that the focus on education many focus groups. This focus was also motivated by the empirics, which suggest an increasingly comes at the exclusion of important other role importantas of education opportunities that of a determinant drive inequality, income however, inequality. in particular It is recognized access that to health the focus on 54 care and basic public services. education comes at the exclusion of other important opportunities that drive inequality, however, in particular 6.59 access Growth into thehealth demandcare and for basic public educated labor andservices. 55 increases in the return to education in urban areas that education imply Growth 6.59 in the is an increasingly demand important—and for educated labor and dividing—asset increases in the inreturn Vietnam. levels Education in to education urbanin the labor market and in households are rising as more educated younger cohorts join the areas imply that education is an increasingly important—and dividing—asset in Vietnam. Education labor market and less educated levels oldermarket in the labor retire. cohortsand However, the in households stock are of as rising education among the more educated working-age younger population cohorts join the changes labor slowly market and response in less to changing educated returns; therefore, older cohorts initial differences retire. However, in education the stock of education endowments among the can translate into large differences in incomes as returns to education rise and the demand for skilled labor in the nonagricultural sector grows. 54 For 55  Foran anexcellent oninequalities discussionon excellent discussion inequalitiesin in these these other other important important dimensions, dimensions, seebackground see the the background paper paper for thefor the 2008– 2008–2010 Vietnam 2010 Vietnam PovertyPoverty Assessment Assessment by Hoang by etHoang et al. (2010). al. (2010). 185 164 working-age population changes slowly in response to changing returns; therefore, initial differences in education endowments can translate into large differences in incomes as returns to education rise and the demand for skilled labor in the nonagricultural sector grows. 6.60 Whether income inequality and disparities will perpetuate across generations depends on whether investments in human capital among younger generations are responding to changes in income generation opportunities, or whether they reflect inequalities in opportunities linked to their circumstances of birth, such as where a child was born, the characteristics of their parents, or ethnicity. The evidence suggests that inequalities in education are likely to be transmitted to future generations, implying that deprivations continue to be perpetuated across generations and require decisive action. 6.61 The transmission of deprivations across generations was reflected in multiple focus group discussions, where groups highlighted that children born to poorer households were likely to drop out of school earlier than those born to richer households, and to work in less-skilled occupations. Many participants recognized that gaps in education enrolment have narrowed between better-off and worse-off households at lower levels of education, but suggest that gaps remain at higher levels of education, and quality gaps arise at all ages, implying that poverty perpetuates across generations. As one member of a lower-educated migrant group expressed it, “Education is an important cause of inequality. Without education, I work as an unskilled worker and send my children to lower-quality schools. With a good education and income, I could send my children to good schools. It is a vicious cycle.â€? (lower-educated migrant group, Ho Chi Minh City) 6.62 Substantial progress has been made in equalizing enrolments and completion rates at the primary level. Between 1998 and 2010, differences in enrolments at the primary and secondary level have narrowed across the rich and the poor and in rural and urban areas, as can be seen in figure 6.17. At the primary level, educational enrolment is close to universal for all groups, although important differences remain between ethnic minorities and the majority, and across minority groups, as discussed in Chapter 5. Figure 6.17 Figure 6.17 Ratio Ratio Enrolments ofof in Primary, Enrolments Lower in Primary, Secondary, Lower and Upper Secondary, andSecondary School by Various Upper Secondary School Groups, 1998 and 2010 by Various Groups, 1998 and 2010 1998 VLSS, Source: 1998 Source: 2010 VHLSS. VLSS, 2010 VHLSS. 6.65 The characteristics of a child’s parents and household wealth continue to be significant predictors of whether a child is enrolled in lower secondary or upper secondary school, although their impact on enrolment diminished between 1998 and 2010. Educational enrolment at the secondary level is affected by income, which can be considered a short-term liquidity constraint, and is linked to longer-term, or permanent, factors such as parental education (World Bank 2011).55 The evidence also suggests that the 165 impact of income on education decisions is twice as large for ethnic minorities as for the Kinh/Hoa majority (World Bank 2011). 6.63 Educational investment continues to be unequally distributed at higher levels, an inequality that will feed into inequalities in outcomes later in life. Gaps in enrolment at an upper secondary level continued to be high in 2010, and a child’s background plays a large role in determining their educational attainment at a higher level. Upper secondary enrolment for children in rural areas is still only 70 percent of enrolment rates for children in urban areas, and ethnic minority enrolment is only half that of the majority. Only four poor students are enrolled in upper secondary school for every 10 richer students enrolled. Since many of those richer students will continue on to college or university, the final education difference between students residing in the top and bottom income quintiles will be wider than it is for upper secondary education. 6.64 The characteristics of a child’s parents and household wealth continue to be significant predictors of whether a child is enrolled in lower secondary or upper secondary school, although their impact on enrolment diminished between 1998 and 2010. Educational enrolment at the secondary level is affected by income, which can be considered a short-term liquidity constraint, and is linked to longer-term, or permanent, factors such as parental education (World Bank 2011).56 The evidence also suggests that the impact of income on education decisions is twice as large for ethnic minorities as for the Kinh/Hoa majority (World Bank 2011). 6.65 Beyond family background, the quality of schooling is an important factor that influences the skills that a child acquires in school. At the primary level, the characteristics of teachers, schools, and classrooms are statistically significantly related to student achievement in math and science, and these inputs have been found to be unequally distributed across schools in Vietnam (World Bank 2011). 6.66 Evidence from the Young Lives data suggests that children from poorer households perform worse on math tests prior to entering primary school, and continue to perform worse than children from richer households throughout primary and lower secondary school. Figure 6.18 shows the average rank of children in math tests at ages 5, 8, 12, and 15 by household wealth quantile. At age 5, prior to entering school, the average math scores of children increase with wealth quantiles, so that children from the poorest 25 percent of households have lower scores, on average, than children from other wealth quantiles. 6.67 Most worrisome, the circumstances that a child is born into appear to be a more important determinant of success than a child’s potential when entering school. Figure 6.19 shows the score trajectories of children who had math scores in the top and bottom 20 percent at age 5. Trajectories are divided by the wealth status of their households at age 8. We can see that high-scoring children from poor households perform poorly relative to their high-scoring peers from rich households. Similarly low-scoring children from rich households make more substantial gains in their scores over time than low-scoring children from poorer households. 6.68 The perceptions study indicates that parents perceive significant variation in the quality of education across rural and urban areas at all levels of education. A frequently raised concern is that teachers in rural areas at higher levels appeared to be less qualified than those in urban areas, and that the poor were unable to afford to send their children to the same quality schools as rich children. ncome is also likely to be related to unobserved correlates such as local returns to education, which are also likely 56 I to positively influence education decisions. Furthermore, income is unlikely to reflect a true liquidity constraint since households also have access to savings and formal and informal credit institutions. 166 0.45 Average 0.4 3rdQuantile 0.35 TopWealthQuantile 0.3 Figure Figure 6.18Average 6.18 Rankin AverageRank inMath Math Test, Test, by by Wealth Wealth Quantile, Quantile, at at Ages5, Ages 8, and 5,8, and 15 15 Years Years 5 8 12 15 0.7 AverageRankinMathTest 0.65 Age 0.6 Source: World 0.55 Bank staff estimates using Young Lives data. BottomWealthQuantile 0.5 6.68 Most worrisome, the circumstances that a child is born into2nd appear to be a more importa Quantile 0.45 determinant of success than a child’s potential when entering school. Figure 6.18 shows the sco 0.4 3rdQuantile trajectories of children who had math scores in the top and bottom 20 percent at age 5. Trajectories a 0.35 TopWealthQuantile divided by the wealth status of their households at age 8. We can see that high-scoring children from po 0.3 households perform poorly 5 relative to 8 their high-scoring 12 15 peers from rich households. Similarly lo scoring children from rich households make more substantial gains in their scores over time than lo Age scoring children from poorer households. Source: World Bank staff estimates using Young Lives data. Source: World Bank staff estimates using Young Lives data. 6.68 Figure Most worrisome, Figure Average 6.196.19 the Rank in inMath circumstances Average Rank Math Test, a that Test, childby by isInitial born Initial into TestTest Score appear Score and to and WealthbeWealth a more important determinant of success than a child’s potential when entering school. Figure 6.18 shows the score 1 trajectories of children who had math scores in the top and bottom 20 percent age 5. Trajectories at Children are BetterOff with 0.9 divided by the wealth status of their households at age 8. We can see thatHigh high-scoring children from poor AverageRankinMathTest TestScores (n=126) 0.8 households perform poorly relative to their high-scoring peers from rich households. Similarly low- 0.7 from rich households make more substantial gains in their scores over time than low- scoring children 0.6from poorer households. scoring children PoorerChildrenwith 0.5 HighTestScores(n=44) 0.4 Figure 6.19 Average Rank in Math Test, by Initial Test Score and Wealth 0.3 BetterOffChildrenwith 0.2 1 Low Better Test Off Scores Children (n=44) with  0.9 AverageRankinMathTest 0.1 High Test Scores (n=126) 0.8 0 0.7 PoorerChildrenwith PoorerChildrenwith 0.6 5 8 Low High Test Test Scores Scores (n=108) (n=44) 0.5 Age 0.4 Source: World Bank staff estimates using Young Lives data. Source: World0.3 Bank staff estimates using Young Lives data. BetterOffChildrenwith 6.69 A striking 0.2perceived inequality in education quality is found between Low richer TestScores and poorer (n=44) households in 0.1urban areas, where the rich children can go to high-quality schools, attend extra 6.69 The perceptions classes, and pay study indicates that parents perceive significant variation in the quality 0 private tuition, including for English and computer courses. Meanwhile, poor education across rural children and attend urban average areas schools 5 atfew with all levels extra 8 of In classes. education. A was the past, there Poorerfrequently differentiation little with Children  raised concern in is th LowTest the quality of education services, but now such differentiation in the cities Scores(n=108) in Vietnam is perceived to teachers in rural areas at higher levels appeared to be less qualified than those in urban areas, and that t be very big, and the rich are viewed asAge having the capability to invest in better-quality education for poor were unable afford toSource: their children. Forto send example, World their children a student Bank staff from estimates to the Ward using same in 26Lives Young Ho Chiquality data. schools Minh City as rich children. reports that: “As early as the child is still in preschool, the rich families will start to seek their way into 6.69 The perceptions study indicates that parents perceive significant variation in the quality of good primary schools, the poorer families just want their children to be literate, so they education across rural and which urban areastheir at all 188 of education. A frequently raised concern is that levels don’t care about school children are going to. Previously, there was a small teachers in rural numberareas at higher levels of international appeared schools for the to be rich less qualified families to choosethan those from, bothin urban rich andareas, poor and that the poor were unable to would students affordattend to send thetheir same school,to children thethere now same quality are more schools schools as rich children. providing a wider range of services, the rich-poor gap also gets widened.â€? 188 167 6.70 Unequal education quality is perceived to start from an early age, with children from poorer households sending their children to lower-quality kindergartens. Some poorer households in An San ward, Tam Ky city, Quang Nam, reported not being able to afford to send their children to kindergarten. Others who were able to do so expressed concerns about quality differences between the preschools attended by their children and those attended by children from wealthier backgrounds: “The disparity can be found right from the preschool level. The poor households, who try their best, can send their kids to school[s] that cost 500,000 VND per month. The better-off households, on the contrary, send their kids to key schools that ask for fees of 700,000 to 900,000 VND per month. The diet and care services among these schools are different.â€? 6.71 Although empirical evidence on quality differences at higher levels of education is limited, looking at the composition of education expenditures across households can give insight into why quality differences may emerge. As noted in Chapter 1, spending on inputs like extra courses is substantially higher among richer and urban households at the lower and upper secondary level, and the amount spent on these courses has increased over time among the richest households. These trends are strongest in urban areas, but can also be seen in rural areas. If children from richer households are able to benefit from extracurricular activities and additional training through tutoring and foreign language studies, they are likely to receive a higher-quality and more rounded education than children from poorer households. 6.72 There is evidence of inequality of opportunities in Vietnam beyond education, and that circumstances beyond the control of an individual contribute substantially to these inequalities in access to basic services. Attitudes toward inequality, and whether it is perceived as unjust, unnecessary, and undesirable, depend on the processes that form it. An important factor is whether inequalities are perceived to be driven by differences in factors for which the individual can be held accountable (“effortsâ€?) or are due to circumstances that fall beyond an individual’s responsibility (“circumstancesâ€?) (Roemer 1998). Factors beyond an individual’s control that lead them to have different levels of well-being can thus be considered inequalities of opportunity (Paes de Barros et al. 2009). 6.73 The Human Opportunity Index (HOI), developed by Paes de Barros et al. (2009), captures inequality of opportunity by examining the extent to which the circumstances that children are born into, such as gender, parental education, and ethnicity, affect the likelihood of their access to basic building blocks of human capital, such as education and health services. The index captures two moments of access to basic services. It captures absolute levels of access, and then calculates how different the access rate is across gender, location, parental background, income, and other indicators capturing circumstances. The degree of inequality is measured by the D-index, which captures the dissimilarity in access rates due to differences in circumstance. Differences is the degree of inequality of opportunity and can be interpreted as the fraction of a given inequality that needs to be redistributed in order to achieve equality. The D-index measure of inequality of opportunity is used to scale down the average national access rate of a service to the given HOI. 6.74 The HOI in Vietnam was examined between 2004 and 2010 in a background paper for the poverty assessment led by researchers from the Vietnamese Academy of Social Sciences, with inputs from the World Bank (VASS 2012). Opportunities for access to basic building blocks were examined in three domains—education, health, and housing infrastructure—and the paper investigates whether access to these basic foundational blocks is evenly spread across children in the population or circumscribed by inherent characteristics beyond an individual’s control. The circumstances examined include a number of individual and household characteristics, including gender, parental education and well-being (expenditures), location, and ethnicity. 168 6.75 In international comparisons with countries in Africa and Latin America and the Caribbean, Vietnam fares well on some dimensions, such as access to electricity and school attendance, and poorer on others, such as access to piped water and flush toilets. Specifically, the HOI for school attendance is higher than that of most African countries and several countries in the Latin America and the Caribbean region, while the HOI for access to electricity is higher than all African countries and only slightly lower than most Latin American and Caribbean countries. The international comparison is, however, less favorable in other dimensions. Vietnam’s HOI for access to piped water is higher than only some African countries, and it is lower than all Latin American and Caribbean countries. The HOI for flush toilets is in the middle of the whole range of African and Latin American and Caribbean countries. However, Vietnam falls considerably behind top-performing countries in both of these basic services. 6.76 Although equality of access is high for education “quantityâ€? in 2010, the HOI suggests that the quality of education is more divergent across the population. Among 7-to-11 year-olds, both the coverage rate and HOI are high, suggesting that there are low inequalities in accessing primary education, and access overall is high. At the lower secondary level, however, although the coverage rate is high, the evidence suggests that there are some inequalities in access. The education of the household head is the most important characteristic determining whether a child attends lower school between ages 12 and 15, followed by household well-being (expenditure). These two circumstances account for more than 50 percent of the dissimilarity. Although ethnic minorities have lower education outcomes, ethnicity alone plays a smaller role than well-being and education of the household head, a finding that suggests that differences in other circumstances contribute substantially to and reinforce inequalities across ethnicities. 6.77 The quality of schooling received by a child is measured by his or her ability to advance independently to lower secondary school without help when he or she is in the last grade of primary school. Only 62 percent of pupils in grade 5 would be able to advance to the lower secondary school without help. The considerable difference between the HOI of the quantity and quality dimensions of education suggests that a greater emphasis needs to be placed on raising quality in the education system, in general, and primary school, in particular. Household well-being and education are the two most important circumstances determining the quality of education received. 6.78 Although the HOI for access to electricity and improved water sources is high, the coverage of access to improved sanitation facilities is lower and less evenly distributed than the other infrastructure measures. Although there was significant progress during 2002–08, and further improvement in 2010, the coverage rate was approximately 64 percent in 2010, suggesting that more can be done to improve access to this basic service.57 Furthermore, a substantial gap between the coverage rate and HOI indicates a remarkable inequality in access to this service. The region a household is located in plays the biggest role in determining access to clean water and sanitation, followed by a household’s well-being, ethnicity, and the education of the household head. 6.79 The HOI is high for some indicators of health and low for others. Notably, the index suggests that Vietnam is doing well on the fraction of women receiving prenatal care, assistance at delivery, and child immunization against measles; 92 percent of children aged 1 to 5 were vaccinated against measles in 2010. Immunization against polio, however, displays a lower coverage rate.  ue to changes in the sampling frame between 2008 and 2010, it is not possible to compare the progress achieved 57 D between 2002 and 2008 to that achieved between 2008 and 2010. Therefore, access to improved sanitation facilities is analyzed separately in 2010. 169 Vietnam is doing well on the fraction of women receiving prenatal care, assistance at delivery, and child immunization against measles; 92 percent of children aged 1 to 5 were vaccinated against measles in 2010. Immunization against polio, however, displays a lower coverage rate. 6.83 6.80 Household well-being is a leading determinant for opportunities in the health domain. Figure 6.20 Household well-being is a leading determinant for opportunities in the health domain. Figure shows6.20 relative the shows importance of circumstances the relative importance for key of circumstances health for indicators key health in 2010, indicators decomposed in 2010, decomposed into the fraction intoattributable the fractionto different circumstances. attributable Ethnicity is to different circumstances. the most Ethnicity important is the circumstance most important for access to circumstance formothers, care for access toandcareaccounted for mothers, one-quarter forand accounted dissimilarities offor one-quarter of receiving prenatal indissimilarities care and in receiving assistance prenatal at delivery. care and Among children, assistance household at delivery. Among well-being, children, region of residence, household well-being, and theofeducation residence,of the region and the education of the household head account for 65 percent household head account for 65 percent or more of the dissimilarity in opportunities. or more of the dissimilarity in opportunities. Figure Figure 6.20 Relative 6.20 Importance of RelativeImportance of Circumstances forHealth Circumstances for HealthOpportunities Opportunities 100 23 26 27 31 24 80 18 19 60 23 30 Percent 21 16 12 40 21 17 19 25 27 20 15 12 12 12 9 9 8 8 0 Receivingantenatal Skilledattendantat Immunization Complete Completion carebyskilled delivery againstMeasles Immunization Immunization person Opportunity againstPolio Childgender HHcomposition Location Ethnicity Source: Source: VASS VASS 2012. 2012. 6.81 An analysis of the HOI at the region level suggests that there is substantial heterogeneity An analysis 6.84 across of the regions with HOI regard toat the region access level suggests to improved sanitationthat there facilities in is substantial both heterogeneity the initial year examined,across 2002, and in improvements between 2002 and 2008, and in 2010. The South East shows regions with regard to access to improved sanitation facilities in both the initial year examined, the largest 2002, and and most stable increase, while the North West had a very low HOI in 2002, which improved in a slow in improvements between 2002 and 2008, and in 2010. The South East shows the largest and most stable and unstable manner. increase, while the North West had a very low HOI in 2002, which improved in a slow and unstable manner. F. Inequalities in Connections, Voice, and Influence 56 Due 6.82 Qualitative to changes and quantitati in the sampling frameve evidence between 2008suggests that and 2010, it inequality in Vietnam is not possible reflects to compare processes the progress achieved that may be more socially and economically damaging, such as inequalities in social between 2002 and 2008 to that achieved between 2008 and 2010. Therefore, access to improved sanitationand political capital, facilities which manifest is analyzed in 2010. through inequalities driven by influence, connections, and uneven themselves separately voice. Inequalities of these forms were raised in many focus groups, urban and rural, rich and poor 191identified as having risen in recent years. alike, as an important driver of inequality, and were 58 6.83 Corruption has been recognized in previous work as a systemic problem in Vietnam, and the qualitative evidence reflects many of the issues raised in previous analyses of corruption and transparency in the country (Anderson et al. 2009; Cecodes, FR, CPP, and UNDP 2012; World Bank 2010; World Bank, Embassy of Sweden, and Embassy of Denmark 2011), but does so through the lens of rich-poor differences and inequality, therefore shedding light on how inequalities in socioeconomic outcomes interact with, are magnified by, and are perpetuated by inequalities in power and connections. Inequality of treatment by public authorities was raised with respect to a  uantitative evidence suggests mixed trends in reported corruption, as would be expected (World Bank 2010). Surveys 58 Q of firms suggest that corruption is less of an obstacle for their operations, but the same surveys show that the magnitude of bribes, as a percentage of revenues, has not declined. Individual reports from household surveys suggest that, while citizens do not find that corruption has worsened, they do not report that the situation has improved (World Bank 2010). 170 number of things, from land conversion practices that favor investors over landholders to the uneven quality of public service delivery in hospitals and public notaries that led to frustration among poorer and less-well-connected individuals. 6.84 Rural respondents were concerned about increasing disparities in employment opportunities in the public sector, and cited the need to pay bribes or have connections to obtain jobs as teachers, doctors, in state-owned enterprises, and as public officials.59 These concerns were widespread and expressed by individuals from all backgrounds, including commune officials. Evidence from the nationally representative Provincial Administrative Procedural Index study suggests that 29 percent of individuals agree that bribes are required to obtain jobs in the public sector, and nearly half of all respondents believe that connections are important in obtaining various types of state employment (Cecodes, FR, CPP, and UNDP 2012). Moreover, these views are shared in urban and rural areas. 6.85 Unfair recruitment mechanisms in the public sector are linked to concerns about youth unemployment following substantial investment in higher levels of education. Focus groups of youth, in particular, voiced frustration with perceived procedural inequalities that affected their ability to translate their education into good jobs, such as the unfair roles of power and relationships to get public sector employment. In their words: “Money is not enough. Money without connections can’t get you a job in the public sector. I know some cases where the workers quit their job in pursuit of higher education but after graduation, they returned to work in the previous position as if they had never attended such courses.â€? (better-off group, Cam Hung commune, Hai Duong) “In my place, there are some guys who have to work as simple workers after completing university just because their families do not have 50 billion VND to 70 billion VND to bribe their way into an agency just to work as an administrative assistant. Many with poor academic performance somehow passed university entrance exams and were placed [in] a job after graduation. This is irrational but unlikely to abate in the future.â€? (senior citizen, Cam Hung commune, Hai Duong) 6.86 In peri-urban areas undergoing conversion of agricultural land into nonagricultural land for industrial zones, inequalities in outcomes related to land were seen as an unfair source of disparities, whereby people with connections and information gain from land speculation while those without are unable to convert their land into income. Focus group participants perceived that the current land conversion policies and processes favored commercial investors, and that local landowners did not secure their rights to proper compensation and resettlement, effective vocational training, occupation replacement, and employment generation. As one group expressed it: “Many owners of bogus projects have exploited loopholes under Decree 64 to appropriate land from local farmers with false claims of using it [the land] for public utilities.â€? (poor group, Me Tri, Ha Noi) 6.87 Focus group participants raised concerns suggesting that corruption in land management is regressive since it involves a transfer of land at lower-than-market prices from poorer households to relatively well-off investors. People with connections and access to information were reported n 2010, the public sector (including state-owned enterprises and civil servants) accounted for only 4 percent of 59 I nonagricultural work and 15 percent of wage or salaried jobs, but for 52 percent of high-skilled jobs in rural areas. In urban areas, the data suggest that public sector jobs account for 9 percent of all nonagricultural work, 28 percent of wage or salaried jobs, and 42 percent of high-skilled jobs. Ho Chi Minh City stands out as having the highest private sector opportunities in the nonagricultural sector, while the North West mountains regions have the lowest private sector opportunities for highly skilled wage or salaried work. 171 to have made substantial profits from land speculation and trade, while those who lost land in the process have to struggle for their basic necessities after land conversion. A key concern here is speculative behavior, wherein land was bought at a low price and resold shortly after at a higher price, as reported by youth in Me Tri, Ha Noi: “People in [the] land sector they know in advance the information so that they can advise others to buy land when the price is low and then sell it out at much higher prices.â€? 6.88 Unequal access to public services was another major source of concern across focus groups, with differences in treatment noted between those who “do politicsâ€? and ordinary people. Concerns about access to quality public services are widespread and cover multiple forms of public services, from lengthy administrative procedures such as registering a marriage to the length of wait and quality of treatment given by doctors and hospital staff in public hospitals. In addition, concerns were raised in multiple settings regarding who receives the benefits from public social assistance programs targeted at the poor. 6.89 It is perceived that those who have been officials of government agencies are often given priority when they go through administrative procedures. In particular, a commonly voiced concern was that richer people use bribes to better access education or health care services. Participants expressed concern over the predominance of valuing money over traditional ethical values on the part of employees in public services as outcome inequalities widen. As one person put it: “For example, when it comes to doing paperwork at the ward people’s committee, if you had been with the state before you retired, you will still be given priority over other ordinary people. Even if you have to queue up, you will still be quicker to have the paperwork done than the others. Likewise in hospital, if you are an average person, you will not get the same treatment as the privileged.â€? (youth group, Ho Chi Minh City) 6.90 The use of power, connections, and corrupt means to get ahead in life and acquire better public services and employment opportunities was seen as unacceptable by many focus group participants, and was a key source of frustration. The evidence suggests that whether inequality in outcomes is viewed as acceptable or not appears to depend more on the process by which the inequality is generated than on the level of disparity. A key concern among focus group participants in both urban and rural areas was whether existing inequalities in outcomes were generated through fair or unfair means, such as corruption, misuse of power, and dishonest business practices. Unfair use of political capital and corruption were perceived to have affected well-being through multiple routes, from employment opportunities and land conversion to the ability to access high-quality public services and education. 172 6.91 If left uncurbed, inequalities in voice and connections that manifest themselves in a myriad of forms, from uneven land conversion practices to poor public service delivery, are likely to be damaging for social cohesion, economic progress, and growth. In the perceptions study, these inequalities provoked the most concern and frustration among participants, and were the focus of lengthy discussions. Inequalities in voice and connections are likely to play a role in determining whether individuals tolerate rising inequality in the future, directly through a sense of injustice and indirectly through their revised expectations of growth. There are suggestions that this may already be occurring via a reduction in the perceived return to education in rural areas, where focus group participants have suggested that their inability to translate education into employment opportunities, in part due to a lack of transparent recruitment mechanisms, has diminished their perception of the value of education for future generations. Box 6.1 discusses policy recommendations for dealing with inequality. Box 6.1 Emerging Policy Recommendations: Inequality Three key messages emerge for policy makers in Vietnam from the analysis of inequality. First, income inequality has risen in Vietnam, indicating that growth processes have been less favorable to poorer households and that poorer households are being left behind. Ethnic minority households have experienced slower growth on average than Kinh majority households, although there is substantial variation among minority households depending on endowments and sources of income. There is evidence of regional variation in growth rates, which has contributed to the rise in inequality. In addition, households characterized by lower average education levels are less likely to benefit from growth processes and to transition into the nonagricultural sector than more educated households. These patterns suggest an active role for policy to help households overcome the structural constraints facing poorer households that limit their growth potential. Second, inequality of outcomes affects the opportunity of children to fulfill their potential, and circumstances overtake potential early in life in Vietnam. Evidence presented in this chapter suggests that children who show promise at age 5 are unable to sustain that promise by age 8 to the same degree as their peers from better-off households. Inequality in opportunities of this form are likely to dampen growth and progress in Vietnam, since they imply that the full potential and talent of Vietnamese children are not being fully achieved. Moreover, it contributes to social tensions. Closing the gap in early childhood development and education quality in Vietnam is, therefore, desirable in terms of both equity and efficiency. Finally, there is widespread concern that inequality in connections, influence, and voice is affecting many aspects of Vietnamese peoples’ lives, from the ability of individual’s to attain public sector employment to obtaining access to good-quality public services and administration. These inequalities in political and social capital are not acceptable to Vietnamese citizens from all backgrounds, and inequality in income and spending that is due to unfair processes is less tolerated than inequality that arises through talent and hard work. Promoting transparent processes in Vietnam is necessary to ensure equitable growth—growth that is viewed as fair by its population. 173   Chapter Annexes Annex 6. 1 Why doâ€? Perceptions of Inequalityâ€? Diverge from Empirical Measures of Inequality? The empirical measurement of inequality includes four components (Cowell 2011). Perceptions of inequality may differ from empirical measures of inequality due to the following considerations: (a) the factor examined, (b) the unit of analysis, that is, whether a household or individual; (c) the reference group, that is, the universe of comparison, such as inequality at the national, regional, rural, or urban level; and (d) the inequality thermometer, or the tool used to capture changes in inequality, such as the Gini or Theil index. This section examines why perceptions may vary from empirical measures of inequality. First, it may be that our measures of inequality focus disproportionately on easily measured dimensions of inequality, such as outcomes, while Vietnamese people focus on other dimensions of inequality, such as the quality of education they receive or whether there is perceived unfairness in society. This chapter discussed modalities of inequality as they were perceived through the eyes of Vietnamese people. Not all modalities of inequality were discussed in each focus group, and the emphasis on different modalities of inequality varied substantially by group. For example, young working people often discussed employment inequalities in greater detail; ethnic minorities paid more attention to livelihood-related modalities of inequality in terms of access to market, credit, and technical services; and students and senior groups talked more about education and the unfair roles of power and connections in employment.60 Second, perceptions may differ from empirical measures because the frames of reference used in empirical analysis differ from that used by individuals when thinking about inequality. In contrast to most empirical measures of inequality, which capture inequalities at the national, regional, rural, or urban level, perceptions of inequality are often rooted in direct life experiences and have a narrower focus. Groups often discussed disparities within their communities, and then conceptualized a step up from their income levels to compare themselves with people in more favorable places or higher positions. For example, in contrast to the decline in inequality attributable to differences between rural and urban areas, rural respondents perceive inequality between rural and urban areas to have risen. However, in contrast to the empirical measure of inequality that compares the average level of welfare within urban areas to the average levels of welfare within rural areas, participants in the focus groups compared their own rural communities to nearby urban centers in the region. Since the empirical measures of inequality and perceptions of inequality are taking place at different levels of aggregation, it may be that, at a more local level, perceptions of inequality and measures of inequality converge.61 An empirical examination of inequality at a lower level of aggregation than normally used in a quantitative assessment may help to bridge the gap between empirical measures and perceptions of inequality. Figure 6A.1 shows inequality at a district level in 1999 and 2009, where a district is a lower 60 Another concern is that the incomes or expenditures of the rich are underreported and undercaptured in household surveys. Therefore, empirical measures of inequality may be downward biased (Cowell 2011; VASS 2011). 61 It may also be that people do not compare mean levels of welfare, but instead compare the richest people in urban areas with the richest, or poorest, in rural areas. District-level inequality was computed using small area estimation techniques. See Benjamin et al. (2009) for more 62  details. 174 unit of analysis than normally used when empirically examining inequality.62 District-level inequality 61 unit of rose analysis than in previously normally used low-inequality when empirically districts examining inequality. and fell in higher-inequality districts.District-level inequality While this gets closer rose to previously in unit the low-inequality of analysis used by districts our focusand fell in group higher-inequality respondents, districts. since Whileof the frames this gets closer reference to used ofvary the unit to appear analysis used by our substantially focus across group respondents, individuals, it remains since the frames of reference used appear to an approximation. vary substantially across individuals, it remains an approximation. Figure 6A.1 District-level Expenditure Figure 6A.2 District-level Expenditure Inequality, 1999 and 2009 Inequality, 1999 and 2009 Absolute Gini Figure 6A.1 District-level Expenditure Inequality, Figure 6A.1 District-level Expenditure Inequality, 1999 and 2009 1999 and 2009Coefficients Absolute Gini Coefficients 0.45 0.8 0.7 0.4 0.6 A b so lu te expen ditu re gini District level gini, 2009 0.5 0.35 0.4 0.3 0.3 0.2 0.25 0.1 0 0.2 1998 2004 2006 2008 2010 0.2 0.25 0.3 0.35 0.4 0.45 District level gini, 1999 Income Consump on The most commonly used measures of inequality—the Gini Coefficient, the class of generalized entropy measures including the Theil index, and ratios of outcomes of people at different percentiles 6.95 The most commonly used measures of inequality—the Gini Coefficient, the class of generalized of the outcome distribution—capture inequality in relative terms. However, individuals may view entropy measures including the Theil index, and ratios of outcomes of people at different percentiles of inequality in absolute terms (Amiel and Cowell 1999; Ravallion 2004). For example, if everyone’s the outcome distribution—capture inequality in relative terms. However, individuals may view inequality income rises by 7 percent, then relative measures of inequality will not register a rise in inequality in absolute terms (Amiel and Cowell 1999; Ravallion 2004). For example, if everyone’s income rises by even though the absolute gap has grown. Evidence from a developed country setting suggests that 7 percent, then relative measures of inequality will not register a rise in inequality even though the approximately 40 percent of individuals in a study on concepts of inequality thought of inequality in gap has absolute terms absolute grown. rather than Evidence from a relative terms developed (Amiel country and Cowell setting 1999). suggests There that approximately is evidence in Vietnam that 40 percent of individuals in a study on concepts of inequality thought of inequality absolute inequality has been rising. Figure 6A.2 shows that the absolute Gini has risen in Vietnam in absolute terms rather 1998. terms (Amiel and Cowell 1999). There is evidence in Vietnam that absolute inequality has than relative since been rising. Figure 6A.2 shows that the absolute Gini has risen in Vietnam since 1998. Whether individuals view inequality in relative or absolute terms is very difficult to capture, and 6.96 Whether individuals view inequality in relative or absolute terms is very difficult to capture, and there are only hints of this in the qualitative assessment. The suggestive evidence indicates that, there are in Vietnam, only hints there this in oflikely are to the qualitative be some assessment. individuals The think who also suggestive aboutevidence inequality indicates that, in in an absolute Vietnam, there are likely to be some individuals who also think about inequality in an sense, and others who think of it in a relative sense. Therefore, even if relative measures of inequality absolute sense, and others who remain think they constant, of it may in a perceive relative sense. to be even Therefore, inequality if For relative rising. measures example, of inequality the first commentremainbelow constant, they may perceive inequality to be rising. For example, the suggests one focus group was discussing inequality in absolute terms, while the second first comment below suggests one comment focus group suggests thatwas discussing another focus inequality group was in absolute discussingterms, while the inequality in second relative comment suggests Vietnamese terms. Whether that another focus group people was discussing conceptualize inequality inequality in relative in absolute terms.terms or relative Whether Vietnamese will be examined people further conceptualize in follow-up inequality work in absolute or relative terms will be examined further in follow-up work underway. underway. “The group “The claimed that group claimed the government’s that the government’s move move toto increase increasethe salary the base salary baseat at times of inflation times of only broadened the income gap between the better-off and the poor. Justifying inflation only broadened the income gap between the better-off and the poor. Justifying the irrationality of raising the salaryof the irrationality base in percentage raising basethey terms, the salary cited an example in percentage terms, they the where increase cited is 20 percent an example and the poor with the lower salary will get just some dozens of thousand where the increase is 20 percent and the poor with the lower salary will get just some VND while the better- dozens of thousand VND while the better-incomed with the often higher salary base will receive additional millions of VND to their pay.â€? Site Report from Phuc Xa Ward, Hanoi 61 District-level inequality (better-off was computed using small area estimation techniques. See Benjamin et al. (2009) for residents) more details. “The students claimed that the rich-poor gap over the past five years has been 197 increasingly widened due to the increasing relative gap: the rich develop faster than the poor.â€? Site report from Linh Xuan Ward, Ho Chi Minh City (student group).  175 References Adams, Richard H. 1994. “Non-farm Income and Inequality in Rural Pakistan: A Decomposition Analysis.â€? The Journal of Development Studies 31 (1): 110–33. Amiel, Yoran, and Frank Cowell. 1999. Thinking about Inequality. Cambridge UK: Cambridge University Press. Anderson, James H., Alcaide Garrido, Maria Delfina, and Tuyet Thi Phung. 2009. “Vietnam Development Report 2010: Modern Institutions.â€? World Bank, Washington DC. Asian Development Bank. 2012. “Outlook 2012: Confronting Rising Inequality in Asia.â€? Asian Development Bank, Manila. Banerjee, Abhijit V. and Esther Duflo. 2003. “Inequality And Growth: What Can The Data Say?â€? Journal of Economic Growth 8 (3) (September): 267–99. Benjamin, Dwayne, and Loren Brandt. 2002. “Property Rights, Labour Markets, and Efficiency in a Transition Economy: The Case of Rural China.â€? Canadian Journal of Economics 35 (4) (November): 689–716. Benjamin, Dwayne, Loren Brandt, and Brian McCaig. 2009. “The Evolution of Income Inequality in Vietnam between 1993 and 2006.â€? University of Toronto, Toronto. Benjamin, Dwayne, Loren Brandy, and John Giles. 2005. “The Evolution of Income Inequality in Rural China.â€? Economic Development and Cultural Change 53 (4) (July): 769–824. Benjamin, Dwayne, Loren Brandt, John Giles, and Sangui Wang. 2007. “Inequality and Poverty in China during Reform.â€? Poverty Monitoring, Measurement and Analysis Working Paper 2007, No 7. Partnership for Economic Policy–Poverty Monitoring, Measurement and Analysis. Bourguignon, F. 2004. “The Poverty-Growth-Inequality Triangle.â€? World Bank, Washington, DC. CECODES, FR, CPP, and UNDP. 2012. “The Viet Nam Governance and Public Administration Performance Index (PAPI): Measuring Citizens’ Experiences.â€? A Joint Policy Research Paper by the Centre for Community Support and Development Studies (CECODES), The Front Review of the Central Committee for the Viet Nam Fatherland Front (FR), the Commission on People’s Petitions of the Standing Committee for the National Assembly of Viet Nam (CPP), and the United Nations Development Programme (UNDP), Hanoi. Cowell, F. A. 2011. Measuring Inequality (third edition) . Oxford: Oxford University Press. Elbers, Chris, Peter Lanjouw, Johan Mistiaen, and Berk Özler. 2008. “Reinterpreting Between-group Inequality.â€? Journal of Economic Inequality Springer 6 (3) (September): 231–45. Gallup, J. 2002. “The Wage Labour Market and Inequality in Vietnam in the 1990s.â€? World Bank, Washington, DC. Glewwe, P., and H. Patrinos. 1999. “The Role of the Private Sector in Education in Vietnam: Evidence from the Vietnam Living Standards Survey.â€? World Development 27 (5): 887–902. GSO (General Statistics Office of Vietnam). 2009. “Vietnam Population and Housing Census 2009: Migration and Urbanization in Vietnam: Patterns, Trends and Differentials.â€? Ministry of Planning and Investment, General Statistics Office, Government of Vietnam, Hanoi. Hirschman, Albert O., and Michael Rothschild. 1973. “The Changing Tolerance for Income Inequality in the Course of Economic Development; with a Mathematical Appendix.â€? Quarterly Journal of Economics 87 (4): 544–66. Hoang, Thanh Houng, Le Dang Trung, Pham Thi Anh Tuyet, Pham Thai Hung, and To Trung Thanh. 2010. “Preserving Equitable Growth in Vietnam.â€? Background paper for the 2008–2010 Vietnam Poverty Assessment. Vietnamese Academy of Social Sciences, Hanoi. Hoang, Xuan Thanh, Nguyen Thu Phuong, Vu Van Ngoc, Do Thi Quyen, Nguyen Thi Hoa, Dang 176 Thanh Hoa, and Nguyen Tam Giang. 2012. “Perceptions of Inequality in Vietnam: a Qualitative Study.â€? Background paper for the 2012 Vietnam Poverty Assessment. Hanoi. McCaig, Brian, Dwayne Benjamin, and Loren Brandt. 2009. “The Evolution of Income Inequality in Vietnam between 1993 and 2006.â€? University of Toronto, Toronto. McKay, Andy, and Finn Tarp. No date. “Welfare Dynamics in Rural Vietnam, 2006 to 2010.â€? Policy Brief No. 3 of 2012, Central Institute for Economic Management, Hanoi, Vietnam. Paes de Barros, R., F. Ferreira, J. Molinas Vega, and J. Saavedra Chanduvi. 2009. “Measuring Inequality of Opportunities in Latin America and the Caribbean.â€? World Bank, Washington, DC. Psacharopoulos, G. 1994. “Returns to Investment in Education: A Global Update further Update.â€? World Development 22 (9): 1325–1343. Ravallion, Martin. 2004. “Competing Concepts of Inequality in the Globalization Debate.â€? In Brookings Trade Forum 2004, ed. S. Collins and C. Graham. Washington, DC: Brookings Institution, 1–38. Ravallion, Martin, and Dominique van de Walle. 2008. “Does Rising Landlessness Signal Success or Failure for Vietnam’s Agrarian Transition?â€? Journal of Development Economics, Elsevier 87 (2) (October): 191–209. Ravallion, Martin, and Shouhua Chen. 2007. “China’s Uneven Progress against Poverty.â€? Journal of Development Economics 82 (1): 1–42. Roemer, John 1998. Equality of Opportunity. Cambridge, MA: Harvard University Press. Roemer, John E. 2006. “Economic Development as Opportunity Equalization.â€? Cowles Foundation Discussion Papers 1583, Cowles Foundation for Research in Economics, Yale University, New Haven, Connecticut. Roemer, John. 2011. “Equality of Opportunity as Opportunity Equalization.â€? Department of Political Science Discussion Paper, Yale University, New Haven. Stark, O., J. E. Taylor, and S. Yitzhaki. 1986. “Remittances and Inequality.â€? Economic Journal 96 (383): 722–740. VASS (Vietnamese Academy of Social Sciences). 2008. “Participatory Poverty Assessment 2008.â€? Viet Nam Academy of Social Sciences, Hanoi. VASS (Vietnamese Academy of Social Sciences). 2011. Poverty Reduction in Viet Nam: Achievements and Challenges. Hanoi: The World Publisher. VASS (Vietnamese Academy of Social Sciences). 2012. “Opportunities for Children in Vietnam.â€? Background paper for the 2012 Programmatic Poverty Assessment, World Bank, Washington, DC. Wells-Dang, Andrew. 2012. “Ethnic Minority Development in Vietnam: What Leads to Success?â€? Background paper for the 2012 Poverty Assessment, World Bank, Washington, DC. World Bank. 1999. Vietnam Development Report 2000: Attacking Poverty. Washington DC: World Bank. World Bank. 2004. Vietnam Development Report 2003: Poverty. Washington DC: World Bank. World Bank. 2006. World Development Report: Equity. Washington DC: World Bank. World Bank. 2009. From Poor Areas to Poor People: China’s Evolving Poverty Reduction Agenda – an Assessment of Inequality and Poverty. Washington, DC: World Bank. World Bank. 2010. “Assessing and Monitoring Governance in the Land Sector: The Land Governance Assessment Framework.â€? World Bank, Washington DC. World Bank. 2011. Vietnam – High-quality Education for All by 2020. World Bank, Washington DC. World Bank, Embassy of Denmark, and Embassy of Sweden. 2011. “Recognizing and Reducing Corruption Risks in Land Management in Vietnam.â€? National Political Publishing House, Hanoi. 177 Publishing licence number No:742/QÄ?LK-LÄ? and Ä?KKHXB CXB No: 279-2012/CXB/21-69/LÄ? Issued on 28 December 2012 Design and printed by Golden Sky Co., Ltd.