(2014) Investir dans les Personnes pour Lutter contre la Pauvreté en Haïti : Réflexions pour l'Élaboration de Politiques Basées sur des Preuves

[page 1] [0] us : 94430 v1 < 2 a © e] 2 A Le S 35 (E TD [0] N E 5 < 2 35 a © e] 2 Q © S E] o __— REPUBLIQUE D'HAITI y ; Observatoire National de la Pauvreté et de l'Exclusion Sociale C) WORLD BANK GROUP [page 2] Investing in people to fight poverty in Haiti Reflections for evidence-based policy making ) WORLD BANK GROUP JSNPES [page 3] © 2014 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.ors Some rights reserved 1234 17161514 This work is a product of the staff of The World Bank with contributions from staff of Observatoire National de la Pauvreté et de l'Exclusion Sociale (ONPES) of the Government of Haïti. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved Rights and Permissions This work is available under the Creative Commons Attribution 30 1GO license (CC BY 30 IGO) http://creativecommonsorg/licenses/by/30/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions Attribution —Please cite the work as follows: World Bank and Observatoire National de la Pauvreté et de l'Exclusion Sociale (ONPES). 2014. Investing in People to Fight Poverty in Haïti, Reflections for Evidence-based Policy Making. Washington, DC: World Bank. License: Creative Commons Attribution CC BY 3.0 1GO Translations—If you create a translation of this work, please add the following disclaimer along With the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation Adaptations—lf you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank and of the Observatoire National de la Pauvreté et de l'Exclusion Sociale (ONPES). Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third- party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail pubrights@worldbank.org, Concept & Design: Manthra Comunicacién Integral / Santiago Calero Cover Design: Manthra Comunicaciôn integral [page 4] Contents Forewords xii Acknowledgments xiv Abbreviations Xvi Overview 1 Introduction 1 Haiti in 2012: Monetary and multidimensional poverty 2 Improvements in monetary and multidimensional poverty 5 Poverty reduction: the importance of transfers and nonagricultural income 9 Conclusions and Priority Areas for Development and Poverty Reduction Policy Action 12 Background and introduction 14 Part l:Poverty and Inequality Diagnostic, 2012 23 Chapter 1: Poverty profile and trends 24 Introduction 24 Poverty and extreme poverty: levels and trends since 2001 25 Poverty profiles 33 Key messages 43 Part Il: Drivers and Constraints for Poverty Reduction 45 Chapter 2: Income generation in rural and urban areas 46 Introduction 46 Income generation in rural areas: opportunities and challenges 49 Income generation in urban areas: opportunities and challenges 63 Internal transfers and remittances: a common strategy for income generation 71 Key messages 76 Chapter 3: Challenges to human capital accumulation 79 Introduction 79 Access to education 82 Access to health care 95 Key messages T4 Chapter 4: Shocks and vulnerability 120 Introduction 120 Shocks, impacts, and household coping mechanisms 123 Vulnerability to natural disasters 134 Key messages 141 [page 5] Chapter 5: Poverty and social protection 145 Introduction 145 Policy framework 146 Social protection needs throughout the life cycle 147 Alignment of social protection, poverty, and risk analysis 150 Key messages 169 Part Ill: Reflections to Promote Evidence-based Policy Making 173 Chapter 6: The way forward: key messages and priority areas of policy actions 174 Urban and rural livelihoods 175 The access to and quality of health and education services 177 Risk management and protection 178 References 212 Appendixes Appendix A. Poverty indicators, disaggregated by department and area of residence, 2012 180 Appendix B. Income Inequality — Lorenz Curves 181 Appendix C. Poverty rate comparisons 182 Appendix D. The methodology for determining the MPI and identifying the categories of the poor, 2012 183 Appendix E. The evolution of the characteristics of households (poor and nonpoor) 185 Appendix F. Poverty correlates 186 Appendix G. Correlates of poverty and food security 190 Appendix H. Definition of concepts 192 Appendix I. Correlates of Labor income, unemployment, underemployment, and informality in urban areas 194 Appendix J. Mincer earnings function and Oaxaca-Blinder decomposition: a methodological clarification 195 Appendix K. Correlates of enrollment and progress in school 199 Appendix L. Descriptive statistics on the shocks reported by households 201 Appendix M. Coping mechanisms 203 Appendix N. Results of the multivariate analysis of shocks 206 Appendix O. incidence maps of weather events 209 [page 6] Boxes Box O.1. À new national poverty line for Haiti 3 Box BI.1. The history of poverty measurement in Haiti 19 Box 1.1. The use of the multidimensional poverty index to identify the chronic poor 31 Box 1.2. Gender inequalities generate great vulnerabilities in Haïti 38 Box 2.1. The correlates of poverty and food security 52 Box 2.2. Estimating correlates of agricultural productivity 57 Box 2.3. The government strategy for rural development 62 Box 2.4. Zooming in on the gender earnings gap using the Oaxaca-Blinder decomposition 66 Box 2.5. Remittances as a return on investment 75 Box 3.1. The intergenerational persistence of education: educational gap analysis 83 Box 3.2. The education system in Haiti 87 Box 3.3. Cholera epidemiological evolution and current policy actions 102 Box 3.4. The health care system in Haïti 106 Box 4.1. Formal and informal mechanisms for risk management: financial inclusion 130 Box 4.2. The disaster risk management strategy in Haïti 138 Box 5.1. Methodology and limitations of ECVMAS data on social protection 154 Box 5.2. Limited access to a national identification document (CIN) can be an obstacle in gaining access to social protection and other services 156 Box 5.3. Kore Fanmi 166 Maps Map 1.1. Moderate and extreme poverty rates, by department, 2012 27 Map 3.1. Literacy rate in Haïti, 2012 86 Map 4.1. The shaking intensity of the 2010 earthquake 139 Map O.1. Flood-prone areas, Haïti 209 Map O.2. Hurricanes, depressions, and tropical storms, by department, 1954-2001 209 Map O.3. Drought-prone areas, Haiti 210 Map O.4. Earthquakes, by magnitude, intensity, and economic damage, Haiti, 1701-2014 210 Map N.5. Soil Liquefaction incidents, February 2010 Pal Map O.6. Landslide incidents during and after the earthquake of January 12, 2010 Pal | [page 7] Figures Figure O.1. GDP per capita in Haiti and in Latin America 2 Figure O.2. Incidence of poverty and number of poor in urban and rural areas 3 Figure O.3. Distribution of household per capita consumption (in Gourdes) 5 Figure O.4. Evolution of extreme poverty in Haiti, 2000-2012 6 Figure O.5. Income inequality in Haiti and in Latin America, circa 2012 7 Figure O.6. Changes in per capita income composition in urban areas per income quintile, 2001-12 10 Figure O.8. Changes in per capita income composition in rural areas per income quintile, 2001-12 11 Figure B1.1. GDP per capita in Haiti and in Latin America 14 Figure B1.2. CDP growth rate in Haiti and Latin America in 1980-2013 15 Figure BI.3. Real and per capita GDP growth in 2001-2013 18 Figure 1.1. Incidence of moderate and extreme poverty in urban and rural areas, 2012. 26 Figure 1.2. Trends in extreme poverty in urban and rural areas, 2000-2012 28 Figure 1.3. Income inequality in Haiti and in Latin America 30 Figure B1.1.1. Poverty decomposition according to the MPI and monetary poverty 32 Figure 1.4. Chronic and transitory poverty, service access deprivation and resilience in Haïti, 2012 32 Figure 1.5. Income composition in urban and rural areas and by poverty status 35 Figure 1.6. Food insecurity in Haiti, 2012. 36 Figure 1.7. Share of the population affected by a climatic shock and poverty level, by department 37 Figure 1.8. Poverty rate by region, economic situation and household head's sector of activity. 42 Figure 2.1. Change in per capita income in urban areas, by income quintile, 2001-2012 47 Figure 2.2. Change in per capita income in rural areas, by income quintile, 2001-2012 48 Figure 2.3. Farm and nonfarm Labor force participation, rural households 50 Figure 2.4. Labor force participation, by type of employment 50 Figure 2.5. Employment, by farm and nonfarm participation 51 Figure 2.6. Economic activity, by poverty Level 51 | [page 8] Figure 2.7. Share of households, by farm activity 54 Figure 2.8. Farm crops grown 55 Figure 2.9. Percentage of households, by Livestock raised 56 Figure B2.4.1. Oaxaca-Blinder decomposition results for different specifications, urban Haiti 67 Figure 2.10. The distribution of hourly Labor income in urban areas, by industry 68 Figure 2.11. Composition of occupations in urban areas, by industry 69 Figure 2.12. Education among the self-employed earning Less or more than the average hourly Labor income, urban areas 71 Figure 3.1. Welfare and educational Level in Haiti, 2012 81 Figure B3.1.1. Educational gap among children 10-14 by per capita consumption quintile 83 Figure B3.1.2. Average reduction in education gap given a standard-deviation increase in parent educational level, by per capita consumption quintile 84 Figure 3.2. Educational Level of adults and young adults 84 Figure 3.3. School enrollment for children in Haïti, 2012 87 Figure B3.2.1. The formal education system 87 Figure 3.4. Enrollment rates in primary, secondary, and tertiary education 88 Figure 3.5. School enrollment by area of residence, poverty status, and gender, % 90 Figure 3.6. Number of public and non-public schools, by year 92 Figure 3.7. Educational expenditures by category, children aged 6 to 14 years 93 Figure 3.8. Financing sources for education 94 Figure 3.9. Infant and under-5 mortality rates, by wealth quintile index. 96 Figure 3.10. The maternal mortality ratio, 1990-2013 98 Figure 3.11. Health care service use, Haiti and selected lower-middle-income Latin American countries. 99 Figure 3.12. Share of households encountering problems over the previous 12 months, 2012 101 Figure 3.13. The five most severe shocks among Haitian households, 2012. 101 Figure 3.14. Causes of non-access to health services, by per capita consumption quintile, 2013 104 Figure 3.15. Obstacles in access to health care services, by wealth quintile index 104 Figure 3.16. Coverage of health services 106 Figure B3.4.1. The health service delivery pyramid 107 | [page 9] Figure 3.17. . The density of medical staff: ratio medical staff/poor population 109 Figure 3.18. Incidence of catastrophic health expenditure in Haiti, 2072 112 Figure 3.19. The incidence of catastrophic health expenditures in Africa and Latin America 113 Figure 4.1. Vulnerability to poverty in Haiti, 2012 122 Figure 4.2. Population shares affected by shocks, by department 124 Figure 4.3. Number of shocks by welfare levels 125 Figure B4.1.1. Reasons for not having an account at a financial institution 130 Figure 4.4. Coping strategies, by type of shock 133 Figure 4.5. Climatic shocks and poverty, by department, 2009 134 Figure 4.6. Poverty and vulnerability in Haïti. 135 Figure 4.7. Number of disaster events, by type, Dominican Republic and Haïti, 1980-2010 136 Figure 4.8. Damage among communes as a result of the 2010 earthquake. 139 Figure 4.9. Perceptions of living standards after the earthquake 140 Figure 5.1 Key risks, the life cycle, and social protection in Haïti: a summary 148 Figure 5.2. Access to social security by quintile of per capita consumption 151 Figure 5.3. Coverage of social assistance programs and distribution of beneficiaries. 153 Figure 5.4. Coverage of social assistance programs, by age-group 155 Figure B5.2.1. Availability of national ID among adults 18 years and older 156 Figure 5.5. Incidence of social protection benefits, by quintile of per capita consumption and poverty status 158 Figure 5.6. Benefit amounts and the contribution to the consumption of beneficiaries 159 Figure 5.7. The cost-benefit ratios of various social protection transfers 160 Figure 5.8. Poverty-related spending as a share of GDP 162 Figure 5.9. Main programs under EDE PEP 163 Figure 5.10. Social safety net spending as a share of GDP, Low-income countries 164 Figure 5.11. Coverage of EDE PEP programs, by type and by poverty rate and departmen, 2012-13 167 Figure B.1. Lorenz Curves at National, Urban and Rural levels, 2012 181 Figure J.1. Blinder-Oaxaca decomposition for different specifications, urban areas, Haiti 198 | [page 10] Tables Table O.1. Access to basic services. 6 Table 1.1. Poverty and extreme poverty in Haiti, 2012 25 Table 1.2. Access to basic services. 29 Table 1.3. Basic sociodemographic and socioeconomic characteristics of poor, extreme poor, and nonpoor households. 33 Table 1.4. Poverty incidence, by category of household re] Table 2.1. Land acquisition. 53 Table 2.2. Agricultural inputs. 54 Table 2.3. Activities of agricultural households 55 Table 2.4. Diversity among the crops grown 56 Table 2.5. Livestock inputs. 56 Table 2.6. Correlates of agricultural productivity 58 Table 2.7. Nonfarm activity, by type of household. 61 Table 2.8. Household participation in non farm activities, by industry. 61 Table 2.9. Household enterprise profile 62 Table 2.10. Labor market indicators geographically disaggregated. 63 Table 2.11. Labor market indicators in urban settings, by poverty level. 65 Table 2.12. Gender, poverty and Labor income in urban areas, by industry 69 Table 2.14. Remittances and other income percent unless otherwise indicated 74 Table 2.15. Uses of transfers in rural areas 74 Table 2.16. Uses of transfers in urban areas 74 Table 3.1. Basic health indicators 82 Table 3.2. The average students completes primary school at nearly 16 years of age 89 Table 3.3. Health outcomes among children, by wealth quintile index, 2005-06 and 2012 97 Table 3.4. Maternal and child health service utilization, by wealth quintile index, 2005-06 and 2012 98 Table 3,5. Children's health outcomes and service utilization, by educational attainment of the mothers 100 | [page 11] Table 3.6. Proportion of households that consider sickness and cholera the most severe problems, by poverty line, residence, and gender. 103 Table 3.7. Health care providers, by the location and poverty Level of the population served. 110 Table 3.8. Per capita annual out-of-pocket health expenditures, by poverty Line. 110 Table 3.9. Per capita out-of-pocket health expenditures, by gender and location 11 Table 3.10. Household out-of-pocket health expenditures, by service type. (N = 4,929) 11 Table 4.1. The prevalence of types of shocks faced by households, by poverty status 127 Table 4.2. The prevalence of types of shocks, by household type 128 Table 4.3. The economic impact of shocks, by household poverty status. 129 Table 4.4. Disasters in the Dominican Republic and Haïti compared, 1980-2010 136 Table 4.5. Triggers and consequences of hazards in Haiti 138 Table B5.1.1. Sample and population sizes for social protection variables in ECVMAS 2012 154 Table 5.1. Alignment of EDE PEP programs With risks and vulnerabilities across the life cycle 168 Table A.1. Poverty indicator, disaggregated by department and area of residence, 2012 180 Table C.1. Poverty rates based on different poverty lines and welfare measures, 2000-12 182 Table E.1. Characteristics of poor households, 2001 and 2012 185 Table F1. Linear regressions to identify poverty correlates, by area of residence 186 Table G.1. Correlates of poverty and food security 190 Table 1.1. Correlates of Labor income, unemployment, underemployment, and informality in urban areas, Haïti 194 Table J.1. Mincer equation results, urban areas, Haïti 196 Table J.2. Average hourly Labor income, urban areas, Haïti 197 Table J.3. Gender earnings differentials, Oaxaca- Blinder decomposition, urban areas, Haiti 197 Table K.1. Correlates of enrollment and progress in school 199 Table L.1. Idiosyncratic economic shocks affecting households 201 Table L.2. Prevalence of types of shocks faced by households, by Location 202 [page 12] Table L.3. Impact of three main types of shocks, by household poverty status 202 Table M.1. Shocks: main coping mechanisms 203 Table M.2. Coping mechanisms to address the most important shocks, by type of shock 204 Table M.3. Coping mechanisms for the most important shocks, households in extreme poverty 204 Table M.4. Coping mechanisms for the most important shocks, resilient households 205 Table N.1. Correlations of the main shocks experienced by households 206 [page 13] : Investing in People to Fight Poverty in Haïti Foreword The following is a new study on how poverty and vulnerability manifest themselves in Haïti. Some may question the value and relevance of such an undertaking and wonder if it was really necessary to engage in a new study of these phenomena, already so scrutinized and publicized in Haiti and around the world. What is actually new about poverty and vulnerability to justify this study? What have we truly lear- ned about poverty and vulnerability that can help us reduce their adverse effects and promote Haitis development? There are a multitude of reports, academic pa- pers and documentaries on the reality of poverty in the country, covering nearly every aspect in detail. In recent years, the fight against poverty has been an important part of government action. Thus, starting in 2004, the government developed the interim framework for poverty reduction, which became the National Strategy Document for Growth and Poverty Reduction (DSNCRP) in 2007, and the Action Plan for National Recovery and Development of Haiti (PARDH) in 2010 after the earthquake, and finally the Strate- gic Plan of Development of Haïti (PSDH) in 2012, accompanied by the first Triennial Investment Program (PTI) 2014-2016. In each case, the government has sought to link economic growth with the struggle for poverty reduction. Contrary to what has been produced in the past, this report provides an updated picture of poverty, taking into account the living conditions of the people after the 2010 earthquake. It also includes the new national poverty lines derived from the post-earthquake living conditions survey, Enquête sur Les Conditions de Vie des Mé- nages Après le Séisme (ECVMAS), from which an analysis of the causes and effects of the endemic poverty in the country were produced. This report does not simply address poverty and vulnerability in and of themselves. On the contrary, it helps better identify challenges and opportunities, while also proposing ways to improve the current situation. Michel Présumé Secretary of State for Planning Ministry of Planning and External Cooperation of the Republic of Haiti | [page 14] WorldBank - ONPES | Foreword Despite numerous challenges, Haiti has made marked progress over the last de- cade. The percentage of its people living in extreme poverty has fallen from 31 to 24 percent between 2000 and 2012. Living conditions have broadly improved. There is better access to education, health, and housing services than a decade ago. All of this is welcome news. When we started this report, we knew that the people of Haiti faced multifaceted difficulties across sectors. What we did not know was their magnitude, their geo- graphic distribution, or their effects on different groups of the population. Thanks to the joint efforts of the government of Haïti and its partners, including the World Bank, to collect the Enquête Sur les Conditions de Vie des Ménages Après le Séis- me (ECVMAS), develop the new national official poverty line, and produce this study, we now have à much clearer picture of the obstacles facing the country, and a precise diagnostic on which we can base policy priorities going forward. We now know that poverty is particularly high and persistent in rural areas, with nearly 75 percent of the rural population remaining poor. We also know that the fight against income inequality has not advanced, and this high inequality has actually increased in rural areas. This study is able to document constraints and opportunities to set the country on a sustainable path of poverty and inequality reduction. Alongside sustained economic growth, and strengthening of governan- ce and institutions, we have identified three priority areas for action. First, invest in people by improving access to education, health and basic services. Second, boost income generation prospects, especially in agriculture and among the ur- ban self-employed. Third, in the face of high vulnerability to shocks and natural disasters in particular, shield the Less well-off from losing their gains through bet- ter social protection and risk management. While there is no silver bullet or perfect recipe to guarantee an end to poverty in Haiti, this study can serve as an indispensable tool for policy discussions based on solid evidence, and for program design guided by robust information. We hope it can be used as a building block to construct a better future for Haiti. Mary Barton-Dock World Bank Special Envoy for Haiti | [page 15] : Investing in People to Fight Poverty in Haïti Acknowledgments This report is the result of a joint effort by the World Bank and the National Observatory for Poverty and Social Exclusion (ONPES) of the Ministry of Planning and External Cooperation (MPCE). The team at the World Bank was led by Federica Marzo (Economist) and Facundo Cuevas (Senior Economist) and comprised Natalia Garbiras Diaz and Thiago Scot, under the overall supervision of Louise Cord (Practice Manager), Mary Barton-Dock (Haïti Country Director), and Raju Jan Singh (Haiti Program Leader). The cross-sectorial Poverty Assessment team that authored the background papers included Aude-Sophie Rodella; Bernard Atuesta Montes; Alan Fuchs and Prospère Backiny-Yetna; Ghemisola Oseni; Tanya Savrimootoo, Eli Weiss, and Barbara Coello; Javier Sanchez Reaza and Michel Matera; Carine Clert (Focal point for the so- cial sectors), Lucy Bassett, Victoria Strokova, Anna Ocampo and Frieda Vandeninden; Andrew Sunil Rajkumar, Eleonora Cavagnero, Mirja Sjoblom, and Marion Cross; and Melissa Adelman, Tillmann Heydelk, Patrick Ramanantoanina, Axelle Latortue, and Marie Monique Manigat. The themes covered by the background papers produced by the World Bank include: the poverty profiles and evolution and poverty measurement, rural development, urban labor markets, the education sector, the health sector, shocks and vulnerability, social protection. The team at ONPES was Led by Shirley Augustin (Coordinator) and comprised Pierre Jorès Mérat (Assistant Coordinator), Jean Malherbe Fritz Berg Jeannot, Ilionor Louis, Lewis Am- pidu Clormeus, Josué Muscadin, Schmied St Fleur, Guy Alex Andre, Frantz Lamour, Hérard Jadotte, Dagobert Elisee, Lanier Sagesse, Emmanuel Michel David, Leonne Fatima Prophete (DPES/MPCE). The themes covered by the background papers produced by the ONPES include: the po- verty profiles and evolution and poverty measurement, labor makets and the working poor; vulnerability to natural disaster, households coping strategies in the face of poverty. The overall coordination and drafting of the report was led by Federica Marzo (Economist, GPVDR) and Shirley Augustin (Coordinator, ONPES). Written comments were received from external peer reviewers, including Jean-Yves Duclos, (Université Laval, Quebec), Tadashi Matzumotu (Organisation for Economic Co-operation and Development), Nathalie Brisson-Lameute (Consultant), Michael Clemens (Center for Global Development) and World Bank peer reviewers, including Ana Maria Oviedo, Gabriel Demombynes, Tom Bundervoet, and Ana Fruttero. The editorial work was conducted by Robert Zimmermann. The joint ONPES/World Bank Poverty Assessment team would like to thank Haitian institu- tions for the joint work done to produce the new official poverty line methodology used to base the analysis contained in this report, especially the Technical Inter-Institutional Com- mittee led by ONPES and including the Haïtian Institute of Statistics and Informatics (IHSI), the Direction of Economic and Social Planning (DPES) of the Ministry of Planning and Exter- nal Cooperation (MPCE), the Fund for Economic and Social Assistance (FAES), and the Natio- nal Food Security Coordination Unit (CNSA). The team would like to thank Michael Clemens (Center for Global Development) for his contribution to the study of remittances and migra- tion. Finally, the team would like to thank the Organization of International Migration in Haiti for facilitating data collection in the Internally Displaced Camps within the framework ofthe Enquête sur les Conditions de Vie des Ménages après le Séisme (ECVMAS 2012). | [page 16] WorldBank - ONPES | Abbreviations : Note: All dollar amounts are US. dollars ($) unless otherwise indicated | [page 17] [page 18] WorldBank - ONPES | Overview Despite a decline in both monetary and multidimensional pouerty rates since 2000, Haiti remains among the poorest and most unequal countries in Latin America. Two years after the 2010 earthquake, poverty was still high, particularly in rural areas. This report establishes that in 2012 more than one in two Haitians was poor, living on less than $ 2.41 a day, and one person in four was living below the national extreme poverty line of 51.23 a day. Progress is evident, but much remains to be done. Extreme pouerty declined from 31 to 24 percent between 2000 and 2072, and there have been some gains in ac- cess to education and sanitation, although access to basic services is generally low and is characterized by important inequalities. Urban areas have relatively fared better than rural areas, reflecting more nonagricultural employment oppor- tunities, larger private transfers, more access to critical goods and services and narrowing inequality compared to rural areas. Continued advances in reducing both extreme and moderate pouerty will require greater, more broad-based growth, but also a concerted focus on increasing the capacity of the poor and vuulnerable to accumulate assets, generate income, and better protect their livelihoods from shocks. Special attention should be given to vulnerable groups such as women and children and to rural areas, which are home to over half of the population and where extreme pouerty persists, and in- come inequality is increasing. Haiti is a country of contrasts, where the challenges are matched by the opportunities. With a population of 10.4 million people living in an area of 27,750 km2, Haïti is one of the most densely populated countries in Latin America.! While 22 percent of the total population lives in the Metropolitan area of Port-au-Prin- ce, the capital, slightly over half (52 percent) lives in rural areas; the rest reside in other urban areas outside the capital? Haitis strategic position in the middle of the Caribbean Sea, its potential as a tourist destination, its young Labor force, and its rich cultural heritage offer a wide range of economic and geopolitical oppor- tunities. Despite this, the wealth generated in the country is largely inadequate to meet the needs of the people: today, Haïitis per capita gross domestic product (GDP) and human development are among the lowest in Latin America and in the world (figure O1). 1 Based on available population projections of the Haitian Institute of Statistics and Informatics (IHSI 2012) and World Bank World Development Indicators (WDI). 2 Alldatain this briefing note are from the Enquête sur Les Conditions de Vie des Ménages après le Séisme (postearthquake household living conditions survey, ECVMAS 2072), unless otherwise indicated 3 Per capita GDP was $1,575 (purchasing power parity [(PPP] US. dollars) in 2013. Haïti ranks 161 among 186 countries in the Human Development Index of the United Nations Development Programme “Human Development Index (HDI) Value; United Nations Development Programme, New York, https.//data.undp.org/dataset/Human-Development-Index-HDl-value/8ruz-shxu EN [page 19] : Investing in People to Fight Poverty in Haïti Figure O1. GDP per capita in Haiti and in Latin America per capita GDP (2011 PPP Uss. dollars), 2012 35,000 30,000 C] 25,000 É 5 s @@ 20,000 + © # eo EÉ22ZY e Fi52s:°0e 15,000 57522: 00e 2E01852:-®e £ x "À 10,000 LESÈSs HEIN à < a = © D GS LE £ % 3 ge à 6 5 8 @ 5,000 e. F7 Ë Risssés RAT TS e sFÉSS SSI siEse o = 9 RSS sSssrs à ri ü S S 7 TZ Sources: WEO (World Economic Outlook Database), International Monetary Fund, Washington, DC, October 2013, http:/www.imforg/extemnal/pubs/ft/weo/2013/02/weodata/indexaspx; WDI (World Development Indicators) (database), World Bank, Washington, DC, http://data.worldbank. org/data-catalog/world-development-indicators. Poverty is widespread in Haiti; in 2012, the overall poverty headcount was 58.5 percent, and the extreme poverty rate was 23.8 percent. The new poverty mea- surement methodology developed by the technical agencies of the Haitian gover- nment reveals that almost 6.3 million Haïitians cannot meet their basic needs, and, among these people, 2.5 million are living below the extreme poverty Line, meaning that they cannot even cover their food needs (box 01).* The incidence of poverty is considerably greater in rural areas and in the North, in particular.5 More than 80 per- cent of the extreme poor live in rural areas, where 38 percent ofthe total population is not able to satisfy its nutritional needs, compared with 12 percent in urban areas and 5 percent in the Metropolitan Area (figure O.2). The poor are also geographically concentrated in the North, where the Nord-Est and Nord-Ouest Departments have an extreme poverty rate exceeding 40.0 percent (representing 20.0 percent of the overall extreme poor), compared with 4.6 percent in metropolitan Port-au-Prin- ce (representing only 5.0 percent of the extreme poor). The incidence of poverty among both man- and woman-headed households is about 59 percent; 43 percent of the population lives in woman-headed households.’ 4 These rates are based on per capita consumption and were calculated using the 2012 official moder- ate and extreme poverty lines of G 817 per capita per day (62.41 PPP of 2005) and G 41.6 per capita per day (51.23 PPP of 2005), respectively. 5 Forthe purpose of this study, Haïti is geographically divided into five regions: the North, the South, the Transversal (the Center), the Metropolitan Area, and the West 6 Based on a linear regression on poverty correlates, the sex of household heads is not correlated with poverty in any Location of residence. 7. This share is high for international standards, but is in line with other countries in the Caribbean region Antigua, Barbados, Dominica, Grenada, Saint Kitts and Nevis, and Saint Lucia present a share of wom- an-headed households above 40.0 percent (Ellis, 2003). [page 20] WorldBank - ONPES | Box O1. A new national poverty line for Haïti Using the new 2012 consumption data, for the first time the government of Haiti has produced a national poverty line, which thus becomes the new reference for the measurement, monitoring, and analysis of poverty in the country. Between October 2013 and February 2014 an interinstitutional technical commi- ttee led by the National Observatory of Poverty and Social Exclusion(ONPES) and including the Haitian Institute of Statistics and Informatics (IHSI), the Fund for Eco- nomic and Social Assistance (FAES), the National Food Security Coordination Unit (CNSA), and the Direction of Economic and Social Planning (DPES) of the Ministry of Planning and External Cooperation (MPCE) developed and certified the first official national poverty line for Haïti, with technical assistance from the World Bank, The poverty line is inspired by the cost-of-basic-needs approach and has values of G 817 (6241 PPP of 2005) for the moderate poverty line and G 41.6 ($1.23 PPP of 2005) for the extreme poverty line. The data used to produce the line are derived from the Enquête des Conditions de Vie des Ménages Après Le Séisme (post-earthquake household living conditions survey, ECVMAS 2012), the first Li- ving conditions survey conducted in Haiti since 2001. The poverty rates for 2012 and the associated profiles are therefore based on the new official national po- verty lines. The new methodology developed by the technical agencies of the Haitian government reflects international best practice. Consumption is considered a better measure of well-being because it captures living standards more accu- rately, unlike income, which generally underestimates well-being and overes- timates poverty®. Figure O.2. Incidence of poverty and number of poor in urban and rural areas a. Poverty incidence 6 80% 5 QT 5 70% 8 eo & 60% 5 ea ge 50% nn @ Metropolitan area 8 40% ÊT . @ Other urban Y 30% ns È . ue] © Rural > He @ Total 5 10% É £ où Extreme poverty Poverty 8 The poverty rates produced in 2001 by IHSI and FAFO (76% and 56%) were based on the international thresholds of 1 and 2 dollars a day (PPP) and on households income data EN [page 21] : Investing in People to Fight Poverty in Haïti b. Number of poor in rural and urban areas 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 Non poor Poor Extreme poor @ Rural @ Metropolitan Area @ Other Urban Source: Official poverty rates, based on ECVMAS 2012; World Bank and ONPES calculations. Vulnerability is extensive in Haiti. One million people live slightly above the po- verty line and could be pushed below the line by a shock; almost 70 percent of the population is either poor or vulnerable to falling into poverty (figure O.4).° Only 2 percent of the population consumes the equivalent of $10 or more a day, which is the region's income threshold for joining the middle class. À typical Haitian hou- sehold faces multiple shocks annually, and nearly 75 percent of households were economically impacted by at least one shock in 2012. The extreme poor are more vulnerable to shocks and the consequences of shocks: 95 percent experienced at least one economically damaging shock in 2012. Natural disasters, in particular, have a great disruptive potential partly because they so heavily affect agriculture, which is the main source of livelihood for a large share of the population, especially in rural areas. Indeed, the evidence shows that the most common covariate shocks are weather or climate related, while the most important idiosyncratic shocks are health related.'° 9 Inthe absence of panel or synthetic panel data, the vulnerable are defined as individuals living on a budget representing 120 percent of the poverty line: in other words, 20 percent higher than the poverty line. An alternative definition of vulnerability used by the World Bank for Latin America is tied to economic stability and a Low probability of falling into poverty. The threshold corresponding to this probability is $10 PPP a day, which is therefore used to identify the middle class in the region, while the vulnerable are defined as individuals living on between $4 and $10 PPP à day. 10 Covariate shocks affect large shares of the population of entire communities (such as natural disasters or epidemics), while idiosyncratic shocks affect individuals (such as sickness, death, or job loss) [page 22] WorldBank - ONPES | Figure O.3. Distribution of household per capita consumption (in Gourdes) — Extreme line — Moderate poverty line 200 . ility li E Vulnerability line 3 180 Æ 160 E 140 5 120 5 100 Ê 80 5 60 4O 20 [e] ©OQQQCQoQoQQCoQQoCoCCoooCoocoo LBSOIOOOODE0VOOOOOONNO ROROMORO0O10 M0 O 082080 SÉRIE GERESSOPERESES Annual per capita comsumption in gourdes Sources: ECVMAS 2012 and official poverty lines; World Bank and ONPES calculations. Significant economic, political, and natural shocks throughout the last decade had important impacts on people’s well-being'. The available data on poverty are cross-sectional, implying that they provide snapshots of welfare at the beginning of the 21st century and in 2072, but do not allow a disaggregated analysis of how each of these shocks affected households. However, a comparison of these two points in time suggests that welfare did improve despite repeated shocks. In particular, at the national level, the extreme poverty rate declined from 31 to 24 percent between 2000 and 2072 (figure O.5).? Improvements in urban areas drove this decline because the extreme poverty rate fell from 21 to 12 percent in urban areas and from 20 to 5 percent in the Metropolitan Area, but stagnated in rural areas, at 38 percent. While data from 2000 are not available to assess the relevant trends, moderate consumption poverty is also estimated to have modestly improved in the last decade.i 11 Among them the political crisis and floods of 2004, the hurricanes and increase in food prices of 2008, and the 2010 earthquake 12 The 2000 poverty rates are from the Fafo Institute for Applied International Studies (2001), a Norwe- gian research center, based on the IHSI Enquête Budget et Consommation des Ménages 1999/2000 (household income and expenditure survey, EBCM) (see (http://www.fafo.no/indexenglish. htm). The consumption poverty indicators for 2000 were calculated based on a national food poverty line estimated in a slightly different manner than the official 2012 methodology. The consumption aggregate in 2000 was developed using over 50 items in the food basket, while the 2012 aggregate was based on a food basket of 26 items that reflects 85 percent of the value of the food consumed among the reference population in all regions of Haiti (deciles 2-6). Furthermore, the aggregate for 2000 does not include imputed rents, while the aggregate for 2012 does. Simulations show that, even excluding imputed rents from the 2012 aggregate, the declining trend in extreme poverty holds. 13 Income-based measures suggest that moderate poverty declined from 77 percent in 2001 based on the Enquete des Conditions de Vie des Menages 2001 (survey on living conditions in Haiti 2001, ECVH 2001) to 72 percent in 2012 (ECVMAS 2012). Consumption-based poverty measures are considered the most accurate in capturing welfare levels, especially in countries with high rates of rural poverty and significant income volatility, the new, official Haitian poverty measure is consumption based. [page 23] : Investing in People to Fight Poverty in Haïti : . Figure O4. Evolution of extreme poverty in Haiti, 2000-2012 Despite a slight decrease in overall 40% extreme poverty in 35% 58 Haiti, the number 8 30% 31% of poor remains É 25% @2000 extremely high, E 20% 21% 20% © 207 especially in SE 15 rural areas. & 10% 5% 0% National Rural Urban Metropolitan area Source: EBC 1999/2000, en FAFO (201), seuils de pauvrete officels ECVMAS (2012) Nonmonetary welfare has also improved in Haiti since 2001 in both urban and rural areas (table O.1). The biggest gains have been in education, where participa- tion rates among school-age children have risen from 78 to 90 percent. However, the quality of service delivery is a concern: because of a combination of late starts, dropouts, and repetitions, only one-third of all children aged 14 years are in the appropriate grade for their age. Table O1. Access to basic services. Coverage rates, % National Urban Rural Indicator 2001 | 2072 2001 | 2012 2001 | 2012 Access to improved drinking water sources S S Habitat, nonhazardous building materials [us | 60 | n | æ | 3 | «| Sources: ECVH 2001; ECVMAS 2072; World Bank and ONPES calculations. Note: — = not available. WHO = World Health Organization. a. According to the international definition (WHO), access to improved drinking water is the proportion of people using improved drinking water sources: household connection, public standpipe, borehole, protected dug well, protected spring rainwater b. The expanded definition includes the international definition (WHO), plus treated water (purchased). c. includes electricity, solar, and generators. d. Rate of open defecation refers to the proportion of individuals who do not have access to improved or unimproved sanitation. This indicator is part of the Millennium Development Goals (MDG) and is a key element of discussion for the post-2015 agenda The open defecation rate declined from 63 to 33 percent nationwide between 2000 and 2012 reflecting gains in both urban and rural areas. e. improved sanitation is access to a flush toilet or an improved public or private latrine. [page 24] WorldBank - ONPES | The quality of sanitation access, remains low: only 31 percent of the population had access in 2012 to improved sanitation overall, and 16 percent had access in rural areas." Access to improved sources of drinking water is similar in urban and rural areas, at 55 and 52 percent, respectively. However, most of the remainder of the urban population (36 percent) purchases safe water directly from vendors; the rest (9 percent) use unimproved sources of drinking water. Meanwhile, most ofthe remainder of the rural population (44 percent) does not have this option and uses unimproved water sources (river water or unprotected wells) with a high probabi- lity of contamination. Access to energy (electricity, solar, or generators) expanded only slightly because of gains in urban areas, accompanied by stagnating levels in rural areas, which held at 11 percent. Over the same period, income inequality stagnated: the Gini coefficient was nu static at O.61 beginning in 2001. The richest 20 percent holds more than 64 Haiti is one of the percent of the total income of the country, against the barely 1 percent held by most unequal the poorest 20 percent. However, this hides opposing trends in urban and rural STI ITU p k D ’ 5 : world, in terms of areas, where inequality declined (from O.64 to O.59) and increased (from O.49 to both incomes and 0:56), respectively. These levels ofincome inequality place Haïti among the most outcomes. unequal countries in Latin America and in the world (figure O.6). Figure O.5. Income inequality in Haïti and in Latin America, circa 2012 a. Gini inequality coefficient, Latin America 07 O.6 O.5 O4 O3 O2 01 [e] E LISE FL LILOSESSX > EE > & L © a RS LE EE L 2 > à © & £ © S 5 2? = G Z ÿ; © 0 D S = PS8 & À ë 2 8 3 (e) 9 mi 14 Improved sanitation includes flush toilets and improved latrines. According to the United Nations Children's Fund and the World Health Organization, an improved sanitation latrine is one that hygien- ically separates human excreta from human contact. 15 The Gini has been calculated using the income aggregate for 2001 and 2012, comprising household per capita Labor income (including production for own consumption), nonlabor income, and imputed rent. The aggregate is built using the methodology of the Socio-Economic Database for Latin Ameri- ça and the Caribbean, as illustrated in CEDLAS and World Bank (2012) 16 Itis not possible to compare trends in consumption inequality because the 2000 estimate did not exclude outliers, which strongly affect inequality estimates. [page 25] : Investing in People to Fight Poverty in Haïti b. Income share, income quintiles, Haiti PS) 74.0 5 “Ÿ 64.6 ‘ « 54.8 D # 431 . .* 242, tt à 20.3 se” 7 418.6 RUCLS ee "160 Fit 62 09: 2 3 n 5 — Haiti —— LAC average ++. Haiti-urban ee. Haiti-rural Sources: ECVMAS 2012; PovStat 2014, data of the Center for Distributive, Labor, and Social Studies. Note: Average inequality in Latin America is based on income aggregates. The same methodology has been used to measure inequality in Haiti. However, comparability is not perfect because of differences in the questionnaires used to capture income. Despite improvements in basic services access, the poor face significantly Lar- ger barriers in accessing basic services. In 2012, 87 percent of 6- to 14-year-olds in poor households were in school, compared with 96 percent of children in nonpoor households. In the same year, child mortality in the highest welfare quintile was 62 per 1000’ live births, while it was 104 in the lowest income quintile. Similarly, the number of stunted children was four times greater in the lowest quintile relative to the highest. Fewer than 1 woman in 10 in the lowest quintile benefits from assisted delivery, versus 7 in 10 among the better off, which suggests that the poorest have limited access to maternal health services and are more likely to die during deli- very.° These facts show that poverty is an important barrier to both school enroll- ment and health service utilization: in 83 and 49 percent of cases, respectively, cost is the main reason for keeping children out of school or not consulting a doctor if they are sick.2° Households bear most of the burden of education costs (10 percent of their total budgets). In contrast, household health expenditures are relatively Li- mited (Less than 3 percent of total household budgets). These obstacles to invest- ment in human capital are greater in rural areas, where poverty is more extensive and the supply of services more limited. 17 Health related data presented in this study are from the survey DHS/EMMUS 2012 18 Welfare quintiles are based on a household asset index, not on household consumption 19 In 2012, the coverage of deliveries within institutions was 8.4 times greater among the highest welfare quintile (76 percent) than among the lowest welfare quintile (9 percent). Welfare quintiles are based on a household asset indicator, not household consumption: 20 According to the 2012 demographic and health survey (DHS), 7 in 10 women aged 15-49 years do not seek medical support for lack of money, while 43 percent do not do so for lack of transport (see chapter 3) EN [page 26] WorldBank - ONPES | Women and girls are particularly vulnerable because they face important obstacles to the accumulation and use of their assets, particular their human capital. Despite sizable progress in both education and health outcomes, adult women are still Less well educated than adult men and are more likely to beillite- rate, while their utilization level of health services is still very Low. Apart from initial differences in endowments, women in Haiti also face additional obstacles in parti- cipating in the Labor market where they are significantly Less likely to be employed and earn significantly Less than man (see below). Finally, gender-based violence and low participation in the public sphere are widespread in Haïti. Due to extreme levels of poverty and vulnerability, the social protection sys- tem in Haiti faces difficulties in adequately meeting the needs of the popu- lation. In the face of the high incidence of and vulnerability to idiosyncratic or covariate shocks, the poor and vulnerable have limited access to public support, given the low capacity of the State. Most assistance arrives to them in the form of remittances or support from churches, other nongovernmental institutions, and donors. In 2012, only 11 percent of the extreme poor received public social assistance through scholarships, food aid, or other transfers?! Despite recent efforts to expand social assistance provision under the EDE PEP framework, the majority of the poor continue to lack access to formal safety nets that could allow them to smooth their consumption over time, prevent irreversible loss of human capital, and avoid destitution. is needed on women, who face One of the key drivers behind the modest poverty gains in urban Haïti has disproportionate been greater access to nonagricultural income. The share of nonagricultural challenges in all income rose among all households in urban areas except for the poorest (figure aspects of life O7). The shift toward nonagricultural employment in urban areas likely reflects a DGERE transition toward better paid jobs in construction, transport, and telecommuni- cations, sectors that experienced positive value added growth during the period. The average hourly Labor income is two to four times higher in the informal and formal sectors than in the agricultural sector? In contrast, households in the first quintile saw their share of nonagricultural and agricultural income fall, while the contribution of private transfers (domestic and international remittances) in their income rose. 21 The coverage rate does not capture a number of larger programs such as school feeding and tuition fee waivers or new programs introduced under the government platform, EDE PEP (in Haitian Creole, help the people) 22 The informal sector is defined by the International Labour Organization as unincorporated enterpris- es (household businesses) that are not registered, do not keep formal accounts, and are not in the primary sector (agriculture) EN [page 27] : Investing in People to Fight Poverty in Haïti Figure O.6. Changes in per capita income composition in urban areas per income quintile, 2001-12 100% 80% 60% 40% 20% 0% 2001 | 2012 | 2001 | 2012 | 2001 | 2012 | 2001 | 2012 | 2001 | 2012 1 2 3 4 5 @ Production for home consumpion agriculture labor income @non-asricuiture labor income @rensions @ carital @ rrivate transfers @ Public transfers @imouted Rent Sources: ECVMAS 2072 and ECVH 2001; World Bank and ONPES calculations. Income generation opportunities in urban areas are limited by a two-sided pro- blem: the scarcity of jobs and the prevalence of low-paid employment. Unem- ployment affects 40 percent of the urban workforce, and almost 50 percent of the female workforce. Youth face unemployment rates above 60 percent, which trig- gers not only economic, but also social concerns?. The challenge of finding a job ends up producing high levels of discouragement. Haiti has a Low Labor force par- ticipation rate compared to the rest of the region: only 60 percent of working-age individuals (15-64) participate in the Labor market, compared, for example, with 70 percent in the neighboring Dominican Republic. Among those who find a job, 60 percent have earnings below the minimum wage and women earn, on average, 32% Less than men?*. Education plays a critical role in improving welfare in urban areas: Labor income is, on average, 28 percent higher among individuals who have completed primary edu- cation than among uneducated individuals. In this context, the urban poor resort to self-employment or informal microenterprises?® as a coping mechanism. Overall, almost 60 percent of the poor are in this type of occupation, and 75 percent of the poor are active in sectors such as trade, construction, and low-skill services. 23 Extended unemployment rate, which includes not only people in working age who do not have à job but are looking for one, but also those who are not Looking for a job because they are discouraged, waiting for a job answer, retired or sick, but would be immediately available if offered an opportunity. 24 Thisis so after one controls for age, education, experience, household size, number of young children in the household, urban Location, and sector of activity. 25 Composed of one or two persons (including the owner) [page 28] WorldBank - ONPES | The persistence of rural poverty reflects households growing reliance on a low-performing agricultural sector and production for home consumption. Over the decade, agricultural income (including production for own consumption and agricultural Labor income) rose in importance, representing between 48 and 59 percent of the incomes among the first three quintiles (figure O.8). Rural live- lihoods are highly dependent on agriculture: almost 80 percent of households engage in farming. Moreover, among half the households, farming is the sole eco- nomic activity. Unfortunately, the returns to agriculture are Low and unreliable, and the activity resembles a subsistence strategy rather than reliance on a productive economic sector. Lessons from better performing farmers suggest that impro- ving access to inputs, product markets and supporting crop diversification are the main channels to elevating productivity. Among the poor, only 20 percent use fertilizer and pesticides. Moreover, even though the area of cultivated land is only slightly smaller among the poor than among the nonpoor (1.2 hectares versus 1.6 hectares, respectively), the poor spend two to four times Less on fertilizer, pestici- des, seeds, and labor?’. Figure O.8. Changes in per capita income composition in rural areas per income quintile, 2001-12 100% 80% 60% 40% 20% 0% 2001 2012 | 2001 2012 | 2001 2012 | 2001 2012 | 2001 2012 1 2 3 4 5 @ Production for home consumpion @ agriculture labor income CU] Non-agriculture labor income C2] Pensions (2 Capital @ Private transfers @ruiic transfers @routed Rent Sources: ECVMAS 2072 and ECVH 2001; World Bank and ONPES calculations. 26 Since 2000, the sector has performed poorly, contracting by O.6 percent annually as à consequence of repeated adverse climatic shocks. In 2012, agricultural production narrowed by 13 percent fol- lowing a series of droughts, heavy rains, and hurricanes, which generated crop and seasonal income losses of 40 to 80 percent. The drop in production Led to a decline in the demand for Labor and a rise in the cost of Locally produced food. Poor households thus Lost income and faced higher con- sumption costs (prices). See “Haiti Food Security Outlook” (October 2012-March 2013), Famine Early Warning System Network, Washington, DC, http://www.fews.net/central-america-and-caribbean/ haiti/food-security-outlook/october-2012 27 Such a gap could arise from credit and liquidity constraints the poor face, as well as weak access to markets and knowledge about input use (Fritschel, 2002; Kydd et. al 2002; Jacoby, 1999) [page 29] : Investing in People to Fight Poverty in Haïti Participation in the nonfarm sector is key to emerging from poverty in rural Haïti. En- gaging in the nonfarm sector in rural areas reduces the probability of being poor by 10 percentage points. The typical nonfarm job in rural areas is a one- or two-person shop engaged in small retail. Still, the returns to this activity surpass those accruing to farming. About 40 percent of nonpoor households participate in the nonfarm sector, a participation rate that is 1.5 times higher than the participation rate among the poor. External financial flows, including remittances and international aid, have also contributed to the decline in poverty. The share of households receiving private transfers in Haiti rose from 42 to 69 percent between 2001 (ECVH 2001) and 2012 (ECVMAS 2012). Worker transfers from abroad have represented more than a fifth of Haiti's GDP in recent years; they originate mainly from the Dominican Republic and the United States. Furthermore, in the aftermath of the 2010 earthquake, the coun- try catalyzed international solidarity, resulting in unprecedented aid flows in money, goods, and services. These external flows contributed to poverty reduction over the period, especially in urban areas, which attracted most of the assistance. Migrating, both domestically and abroad seems to be a profitable income ge- neration solution for many households. An approximate cost-benefit comparison indicates that, on average, migration is profitable. À household with an out-migrant has forgone earnings of about G 5,000, but, in exchange, the migrant can expect to raise G 16,000 at destination (G 4,000 of which are sent in transfers). When con- trolling for individual and households characteristics, educated migrants earn on average between 20 and 30% more than their peer in rural areas. This report identifies three main areas for action in the fight against poverty and inequality in Haiti, to complement efforts for better governance and sus- tainable growth: i) Boosting income generation in rural and urban areas to pull households out of poverty; ii) Improving provision of basic services, such as health and education, to increase productivity potential and provide the poor and vulnera- ble with the means to improve their Lives in a durable manner; iii) Risk management and social protection policies to avoid livelihood losses. Policies to boost households’ income are essential to sustaining and accelera- ting welfare gains. In urban areas, achieving this objective will have to involve the creation of economic opportunities and better jobs, particularly among youth and women. À higher level of education, for example, is correlated with higher Labor in- come. In rural areas, the stagnation of both extreme poverty and income inequality observed between 2000 and 2012 reflects the increasing reliance on the Low-pro- ductivity agricultural sector. Because 80 percent of the extreme poor live in rural areas, it will be necessary to develop this sector by means of policies that support crop diversification and promote expanded access to inputs and to output markets. Furthermore, both in urban and rural areas it is necessary to improve the business EN [page 30] WorldBank - ONPES | environment in order to increase the profitability of employment. Policies aimed at improving the mobility of goods and people, such as investments in transport or financial inclusion, could contribute to this goal, while allowing households to harness the potential of migration (domestic and international) Enhancing access to education and health care is especially important in buil- ding individual and household human capital. In the context of limited economic opportunities, the public provision of services to increase the human capital accu- mulation capacity of the poor will be essential in breaking the vicious circle ofinter- generational poverty. Expanding access and the quality of services, while reducing costs among households will be critical to improving health and education outco- mes, particularly among children and women. Addressing early childhood develo- pment and gaining deeper knowledge about the determinants of school learning are essential in the education sector. Achieving universal primary enrollment will also require a short- to medium-term financing plan and an improved coordination with social protection programs. On the health care front, policies should aim at improving the accountability of service providers, increasing service utilization and quality, and expanding preventive health care services to reduce costs. In both sec- tors, furthermore, the establishment of an information system allowing for better identification and targeting of vulnerable populations, as well as for services quality control, will be critical in optimizing the use of available resources. In the face of recurring shocks and vulnerability, better targeting in social protection and better risk management strategies should be prioritized to protect households and individual livelihoods. One million people are vulnera- ble to shocks that could push them into poverty. Despite the significant expansion of social assistance provision within the EDE PEP framework, developing à targe- ting system is an essential step to enhancing human capital accumulation among the poor and vulnerable using, among others, a national poverty map allowing for the identification of pockets of poverty and therefore to expand coverage. Besi- des social protection measures, the ex-ante identification and understanding of disaster risks are crucial for the protection of the assets of the poor, as are risk reduction policies such as the retrofitting of critical buildings and the construction of protective infrastructures. It will also be necessary to improve the country's ca- pacity to manage disaster-related emergencies ex-post by strengthening institu- To combat poverty tional arrangements. and inequality in LL LL LL . a sustainable way, The regular monitoring of poverty and living conditions is a necessary step policies should focus to promoting evidence-based and effective policy making. One of the many in three key areas, obstacles to post earthquake reconstruction and emergency operations is repre- alongside strong sented by the lack of sound statistical information at the national Level. Strengthe- economic growth and ning the national statistical system through investments in this sector will allow better governance: the country to have reliable data from various sectors, through regular national Investing in People; censuses and surveys, such as ECVMAS, that will permit regular and systematic Boosting Incomes en ne ui : ne . and Opportunities; monitoring of poverty and households living conditions in Haïti, relying on the new and Protecting the reference rates for the country. At the same time, regular monitoring built on the Poor and Vulnerable solid baseline set out in this report will contribute to enhancing the design and from shocks. efficacy of antipoverty policy making. EN [page 31] : Investing in People to Fight Poverty in Haïti Background and introduction Haiti is one of the biggest, most densely populated nations in the Caribbean and one of the richest in challenges and opportunities. Haiti occupies the western half of Hispañola Island in the Caribbean Sea, sharing the island with the neighboring Domini- can Republic on its eastern border. With a population of10.4 million people (495 percent men and 50.5 percent women) according to the latest population projections of the IHSI (20M), Haïti is one of the most densely populated countries in Latin America (fifth, after four other Caribbean countries)#. While 22 percent of the total population lives in Port- au-Prince, a small majority of Haitians still live in rural areas (52 percent against 48 percent in urban areas). The population is highly concentrated in three departments: Ouest (35.6 percent, mainly urban), Artibonite (16.3 percent, mainly rural), and Nord (almost 98 per- cent). The fertility rate of 35 children per woman is reflected in a population growth rate of 1.6 percent according to the latest estimates, relatively low compared with other coun- tries at a similar Level of economic development? Haïti's strategic position in the middle of the Caribbean, its potential as touristic destination, its young Labor force, and its rich cultural heritage account for a wide range of economic and geopolitical opportunities. Despite this, the wealth generated in the country is largely inadequate to meet the needs of the people: today, Haïti has one of the lowest GDP per capita in Latin America and in the world ($1,575 in 2011 PPP dollars), while scoring161st among 186 countries according to the United Nations Development Programme’s Human Development Index (figure B1:1).5° Figure BI.1. GDP per capita in Haiti and in Latin America Per capita GDP, 2012 (in 2011 PPP U.S. dollars) 35,000 30,000 | @ 25,000 8 à F Ë ® 0e 20,000 RS . FL£Ées.- .e 15,000 & 5155600 £ 5 © d 2 8 2 = 10,000 RÉSESEÉ. 512000. a+ FFE SEec::c:00eee 5,000 e «A F7 & E3$2$55:-00ee A € SRÉSÉSÉSSÉSEESe o S DR LSozsws D ä $ Ê ZT Sources: WEO (World Economic Outlook Database), International Monetary Fund, Washington, DC, October 2013, http:/www.imforg/extemnal/pubs/ft/weo/2013/02/weodata/indexaspx; WDI (World Development Indicators) (database), World Bank, Washington, DC, http://data.worldbank. org/data-catalog/world-development-indicators. 28 World Bank World Development Indicators (WDI). 29 The demographic growth rate refers to extrapolations performed by IHSI on the basis of population projections for 2010—15. The previous growth rate was 2.5 percent and corresponds to the intercensus growth rate (1982-2003) 30 “Human Development Index (HDI) Value; United Nations Development Programme, New York, https:// data.undp.org/dataset/Human-Development-Index-HDl-value/8ruz-shxu [page 32] WorldBank - ONPES | The economic performance of the last 50 years has been disappointing; ave- rage growth has been among the lowest in the world. Between 1960 and 2010, Haiti recorded one of the lowest average growth performance in the world (below the averages in Latin America and Sub-Saharan Africa) (figure B1.2). In 1961-2000, average real per capita GDP contracted 1 percent a year, resulting in a cumulative reduction of 45 percent (World Bank 2006). In 2001-09, GDP growth was a mere O8 percent, and average real GDP growth per capita was -0.8 percent, while the earthquake of January 2010 caused a 5.5 percent contraction in the economy. The few periods of positive growth were short-lived. Historically, sustained economic growth only occurred in the 1970s, mainly because of favorable terms of trade and key public investments.5' Figure BI.2. GDP growth rate in Haiti and Latin America in 1980-2013 15 10 5 0 [e] + © œ © © æ@ © N\VWz vo CE œ ® © © à S $S © YO © Vs 06 s É SE S 8 $ FRS RSSSRS* -15 — Growth in Haiti —— Average growth in LAC (excludimg Haïti) Source: WEO (World Economic Outlook Database), International Monetary Fund, Washington, DC, April 2014, Attp:/wwwimforg/external/pubs/ft/weo/2014/01/weodata/indexaspx. Low economic growth, poor governance, and fragility are among the main cau- ses of substantial poverty and low human development outcomes. Poverty analysis based on household survey data has been extremely limited because of a lack of reliable data. The last Poverty Assessment produced by the World Bank dates from 1998.%? It describes poverty as widespread and the access to basic servi- ces as limited, particularly in rural areas. The report emphasized the huge economic gap between urban and rural areas, and attributed to it the migration from rural to urban areas, exacerbating the uneven distribution of public resources toward urban areas, especially Port-au-Prince. Despite this and as a result of demographic pres- sures, living conditions in the capital were characterized by relatively poor access to services and unhealthy housing conditions. Regardless of location of residence, the average poor household was Less well educated, had less access to wage inco- me or transfers (in absolute terms), and depended more on self-employment and 31 During this period, tourism grew considerably as well as the nascent export-oriented assembly-light manufacturing sector, which took advantage of the proximity of the U.S. market and tax incentives. The growing economy incentivized urbanization and boosted the construction sector in Port-au-Prince, fu- eling private consumption. Meanwhile, the government supported the momentum of growth by raising public investment in key infrastructure, such as telecommunications, energy, and ports. 32 The assessment was based on a series of surveys, including a rural livelihood survey and a micro survey of three urban areas (La Saline, St. Martin, and Tokio). No national poverty rate was provided because of the lack of a national survey, [page 33] : Investing in People to Fight Poverty in Haïti production for home consumption. The report listed a series of factors accounting for “the dire extent of poverty” (World Bank 1998), including poor governance and corrup- tion, inadequate growth caused by poor macroeconomic management and limited private investment, underinvestment in human capital, and the bad quality of public expenditure. It stated that “the interaction of these various factors, including high po- pulation growth, produces a ‘poverty trap' with one outcome: an increase in poverty and associated human, physical, social, and environmental degradation” (World Bank 1998). Based on the Enquete des Conditions de Vie des Menages 2001 (survey on living conditions in Haiti 2001, ECVH 2001) and a poverty line of $1.08 à day, a report of the Fafo Institute for Applied International Studies (Fafo) in 2004 provides similar stylized facts on income poverty in Haïti, suggesting that the situation had not signifi- cantly evolved at least since the late 19905 (Sletten and Egset 2004). A study of the World Bank (2006) confirmed that the causes of Haïiti’s weak economic performance since the beginning of the 1980s are to be sought in political instability, fragility, and poor economic governance, but also ack- nowledged the impact of external shocks. À poor business environment, decrea- sing investment in physical and human capital, the erosion of public expenditure efficiency, low growth, and, ultimately, the self-perpetuating persistence of poverty have been the result of the joint prevalence of political and economic factors of internal and external origin, as follows: * Political instability: Despite the glorious past that made Haiti the first indepen- dent black republic in 1804, the country's most recent history has been marked by several authoritarian regimes and popular uprisings, starting with the Duvalier era (father and son) that lasted 26 years, until 1986. Since then, political instability has worsened, and Haïti has seen a succession of 18 heads of state and few de- mocratic transitions. The last significant political crisis occurred in 2004, with the ousting of President Jean Bertrand Aristide by popular upheaval. + _ Economic mismanagement: Progressively, poor economic policy decisions from the late 1970s on have resulted in the creation of monopolistic public enterprises, the weakening of the private domestic sector and foreign investment, and the re- duction of productive public investment, such as in key infrastructure and human capital, leading to a deterioration in the country's potential for growth. + _ External shocks of economic origin: Haitis dependence on agricultural exports and imports as a source of revenue and for domestic consumption, respecti- vely, has made the country extremely vulnerable to external shocks, particularly shocks related to prices fluctuations for major exports (such as coffee and co- coa) or imported food (such as rice). Terms-of-trade shocks were experienced in 1981-92 (a fall in coffee prices), 2000-02 (a fall in coffee and cacao prices), and 2008 (a rise in imported food prices). + _ External shocks of political origin: In response to domestic political instability, Haï- ts main partners have repeatedly stopped or drastically cut official development assistance or trade relations. This was the case in the Duvalier era, when the country received almost no aid for development purposes, or the political crisis of the early 2000s. The embargo imposed by the United States between 1991 and 1994 had a EN [page 34] WorldBank - ONPES | devastating impact on the economy by significantly reducing productive capacity, thereby destroying the nascent export manufacturing—-assembly industries. + _ External shocks of climatic-natural origin: Its geographical position, compoun- ded by its dependence on agriculture, makes the country especially vulnerable to the impacts of climate-related shocks, such as hurricanes or droughts. Envi- ronmental degradation caused by deforestation and soil erosion has progressi- vely worsened the impact of these shocks, which greatly affect economic and agricultural activity.5 In 2004, floods aggravated the ongoing political crisis, cau- sing damage to the economy estimated at 55 percent of GDP. In 2008, Haïti was hit by four hurricanes, causing a contraction in agricultural production by more than 7 percentage points and a rise in domestic food prices. The 2010 earthquake was destructive and led to significant loss of human life and displacements, as well as damage to infrastructure, dwellings, and, to a Lesser extent, jobs. In 2012, the country was hit by two hurricanes (Isaac and Sandy) and one drought, leading to negative growth 0f1.3 percent in the national agricultural sector. In 2007, a new poverty reduction strategy was finalized, but its objectives have not been completely achieved. Subsequent to the World Bank Pover- ty Assessment (World Bank 1998) and the 2004 poverty profile of Fafo (Sletten and Egset 2004), a highly consultative Poverty Reduction Strategy Paper was developed in 2007 by the government and its partners within the framework of the Highly Indebted Poor Countries Initiative (MPCE 2008).5* However, setbac- ks related to the political environment, extreme weather events and the major 2010 earthquake impeded the achievement of the objectives set by Poverty Reduction Paper (MPCE, 2011). Despite this gloomy picture and the dramatic setback generated by the ear- thquake, positive signs have recently emerged. In 2005-09, Haïti experienced a period of continued economic growth (an annual average of 23 percent), with a peak in 2009 (31 percent), driven by agriculture and industry (figure B1.3).5° The re- turn to growth as well as other positive signals, sealed by the cancellation of most of the country's public debt through the Heavily Indebted Poor Countries Initiative, represented the difference with previous short-lived growth spurts and contributed to the generation of optimism in the country and among the country's partners. The democratic election of René Préval in 2004 and the onset of structural re- forms marked the return to macroeconomic and political stability. The earth- quake that hit the country on January 12, 2010 suddenly stopped the momentum. 33 The forested area shrank by 13 percent between 1990 and 2010 (United Nations Development Programme, Human Development Indicators, https://data.undp.org/dataset/Change-in-forest-area- 1990-2010-/77qj-63mn) 34 Based on the ECVH 2001 and a poverty line of $1.08 a day, the Fafo study (Sletten and Egset 2004) provides a picture of (income) poverty in Haïti similar to the World Bank 1998 study, describing it as a predominantly rural phenomenon, with 77 percent of the extreme poor Living outside the Metro- politan Area. At that time, poor households were more likely to be headed by women, especially in Port-au-Prince, have Less access to wage income or transfers (in absolute terms), and depend more on self-employment and self-production. 35 During this period, growth slowed only in 2008 because of food-price riots and the subsequent political crisis EN [page 35] : Investing in People to Fight Poverty in Haïti This tragedy caused over 200,000 deaths, considerable economic and infrastruc- ture damage, estimated at 120 percent of GDP, and the destruction of the state apparatus. Nonetheless, the country's progress toward political and economic sta- bilization resumed almost immediately after the earthquake, partly thanks to the solidarity of development partners. Post-disaster reconstruction and a strong in- flow of development assistance and remittances from the Haitian diaspora fueled economic recovery (a 5.5 percent growth rate in 2011), and the election of Michel Martelly in late 2010 was the first transition between two democratically elected presidents since 1996 and the first democratic political transition between opposing parties ever. Figure BI.3. Real and per capita GDP growth in 2001-2013 8% 6% 5.5% 4% 2 NA R$T 0% — _ ÈS — AN À — nIEG IN/S 8 5 E E\YE SE ml R RER 8 8 8 8 SR R& 6% 55% -5.5% —— Growth of per capita GDP (constant prices) ——— Growth of GDP (constant prices) Sources: IHSI 2014; World Bank and ONPES calculations. While the country aligns efforts to improve governance and set the eco- nomy on the path of sustained, broad-based growth, reducing dependence on international assistance, positive signals suggest there are new opportu- nities for poverty reduction. Government institutions were progressively rebuilt in the aftermath of the earthquake, since the new government has elaborated a Strategic document setting the goal of becoming an emerging economy by 2030 (Plan Stratégique de Développement d'Haïti (PSDH)). Associated with this vision, the government has resumed the path of reform (including à public financial ma- nagement reform plan adopted in June 2014.) and its midterm investment plan- ning activity. The government has made of poverty reduction a planning priority, and consistently expanded the budget devoted to social sectors. À new social protection strategy is currently under implementation, and aimes at reducing fragmentation, fostering coordination across government agencies, and increa- sing efficiency by enhancing monitoring and evaluation and the targeting of so- cial programs. However, much still needs to be done to see concrete results in improved governance in Haïti, in particular with respect to corruption, the gover- nment effectiveness and productive public investments. Indeed, data indicates that corruption and weak government effectiveness are still very much an issue in Haiïtisé, 36 The IMF (2011) ranks Haïti 56th out of 71 countries for which an index of efficiency of the public invest- ment management process is available. Furthermore, the 2013 World Bank Worldwide Governance Indicator for Haiti suggests that the country ranks in the lowest decile in measures such as control of corruption, government effectiveness and rule of law. EN [page 36] WorldBank - ONPES | The focus of this Poverty Assessment is to produce a thorough diagnosis of the levels, evolution, and drivers of poverty in Haïti, and to identify a set of priori- ty areas for policy action. The objective of this joint ONPES/World Bank study is to serve as platform to discuss and prioritize policies, and contribute with robust evidence to resource-allocation decisions. The joint work was built mostly, but not solely, on the newly available ECVMAS 2072, the first such survey in more than a de- cade, the progress made by the government in poverty measurement through the launch ofthe first official poverty numbers in May 2014 (Box BI1 provides a summary history of poverty measurement in Haiti.) and the deep knowledge ofthe country by the Haïtian institutions and the World Bank sectorial teams. À distinctive strength of this Poverty Assessment is that it identifies a set of priority areas of action in each of the sectors covered by the analysis. With the new evidence and identified priority areas, policy and resource allocation discussion and decisions of the Government and its partners will be enhanced and better informed. Box BI1. The history of poverty measurement in Haïti Historically, the analysis of monetary poverty in Haiti has been limited by the lack of credible, standard statistical information, as well as of an official measurement methodology. This Led to multiple attempts to measure poverty between 2001 and 2006, contributing to some confusion. The surveys available for this type of exercise include two household income and expenditure surveys (EBCM 1 and 11) conducted by IHSI, respectively, in 1986/87 and 1999/2000 and one living conditions survey collected in 2001 by IHSI in collaboration with Fafo (ECVH 2001). While the first two surveys include data on household consumption and expenditure, the third only co- vers income. Based on these data, two different types of analysis were con- ducted, as follows: + In 2001, a nonofficial national poverty line was defined on the basis of the cost-of-basic-needs approach and household consumption data from EBCM | and Il (Pedersen and Lakewood 2001). This exercise produ- ced relatively comparable poverty rates for two years, showing a decrease in consumption poverty from 59.6 percent in 1986/87 to 48.0 percent in 1999/2000. These results, however, were later contested on the basis of a methodological weakness in the definition of the line and of the nonfood component (Montas 2005). Alternative calculations suggest that the inci- dence of poverty did not change between the two periods. + Between 2003 and 2006, poverty was measured using the international poverty lines of the time ($1 and $2 PPP per head per day) applied to the income data of the ECVH 2001. Several agencies and researchers con- ducted the same analysis, producing different poverty rates based on a different use of PPP coefficients. Hence, the extreme poverty rate obtai- ned using the $1 PPP line and data from 2001 ranged from 48.9 percent (Verner 2005) to 539 percent (World Bank and SEDLAC 2005/06) and 5L.0 percent (UNDP 2003). EN [page 37] : Investing in People to Fight Poverty in Haïti The only poverty rate commissioned by the government and used re- gularly thereafter was produced in 2006 based on 2001 data. In 2006, the Ministry of Economy and Finance asked IHSI to produce a poverty pro- file for Haiti based on the ECVH 2001 to facilitate forthcoming discussions With the International Monetary Fund on a new program of assistance and to prepare the ground for the definition of a Poverty Reduction Strategy Pa- per (see MPCE 2008). The joint work of IHSI and Fafo, based on income data of the ECVH 2001, produced an extreme poverty rate of 56 percent and a poverty rate of 76 percent. The related report, published by Fafo in 2004, described poverty as a mostly rural phenomenon because 77 percent of the extreme poor were living outside the Metropolitan Area (see Sletten and Egset 2004). Between 2001 and 2012, no household survey was collected by IHSI, preventing any further attempt to update the poverty rate. The only national survey that took place regularly every five years was the demo- graphic and health surveys (DHS), which, however, do not facilitate the monitoring of monetary poverty. In 2011, the World Bank used data from two DHS (1995 and 2005) to study poverty based on household assets in 1990-2000, before the earthquake. The study showed an improvement by 5 percentage points between 1995 and 2005 and a deterioration of 3 percentage points between 2000 and 2004, reflecting the economic and political crisis characterizing 2001-04. In the aftermath of the earthquake, IHSI and its partners decided to collect a new survey on household living conditions, this time inclu- ding household consumption. Starting in 2010, the French research center DIAL (Développement, Institutions et Mondialisation), IHSI, and the World Bank collaborated to produce a survey that was representative at the national, urban-rural, and departmental levels and that had the goal of measuring living standards after the earthquake. The survey—ECVMAS 2012—to0k place in 2012. The availability of consumption data made the definition of a national poverty line possible, as well as the calculation of consumption-based poverty rates and the production of long-awaited poverty analyses. Between October 2013 and February 2014, an interinstitutional technical committee led by the National Observatory for Poverty and Social Exclu- sion (ONPES) and including IHSI, FAES, and the CNSA, developed the first official national poverty line for Haiti, with technical assistance from the World Bank. This threshold is inspired by the cost-of-basic-needs approach and has a value of G 817 per capita and per day ($1.98 in 2012 UsS. dollars) and G 41.6 per capita and per day ($1.00 in 2012 US. dollars). The poverty rates for 2012 and the associated profiles are therefore based on the new official national poverty lines (alternative measures of monetary poverty overtime are included in annex C). EN [page 38] WorldBank - ONPES | The analysis developed in the Poverty Assessment is framed around the im- portance of supporting the poor and vulnerable in building, using, and pro- tecting assets. Creating an environment that promotes greater growth and pros- perity is critical for the country, but, if growth is to be boosted and shared among the less favored, the assets of the poor and vulnerable must be built up, used, and protected. Al these three elements are necessary to achieve sustainable pover- ty reduction and shared prosperity. Improving access to assets such as human capital (education, health), and physical and financial capital is a key first step. Promoting utilization of those assets and fostering their returns is à second pi- Lar for genuine poverty reduction via income generation. . Finally, in a context of significant exposure to aggregate and idiosyncratic shocks, it will be essential to protect the assets of the poor through enhanced safety nets and social protection services for better risk management. Consistent with the conceptual framework presented above, the report is organi- zed in three parts: the first part offers a thorough diagnostic of poverty and inequa- Lity in the country; including levels and trends over time, and socio-economic and demographic profiles of the poor. The second part refers to the main drivers and obstacles to poverty reduction. It distinguishes three pillars: accumulation of key assets, namely education and health; urban and rural income generation; and risk management strategies to protect households livelihood, including disasters risk management and social protection. Finally, the concluding chapter summarizes the key messages and priority areas of action for policy. Within this framework, each chapter is organized in three main parts: an introduction, a diagnostics, and a key-messages concluding section. This approach implies that institutional and macro constraints to poverty reduction, such as issues of governance, fragility, and low economic grow- th, or issues of public resources availability, sustainability and allocation are not the focus of the analysis in this report. This choice has been motivated by the opportunity to use the newly available survey for a household-based analy- sis, as well as the parallel work being conducted by the World Bank, especially the Public Expenditure Review and the Systematic Country Diagnostic. The main objectives of these studies will be to draw a diagnosis of the major constraints to broad-based growth, with special attention to governance and public resource management. The Poverty Assessment, the Public Expenditure Review, and the Systematic Country Diagnostic will therefore provide a comprehensive picture of the constraints on poverty reduction in Haïti and the avenues for improvement. EN [page 39] [page 40] Part 1 Poverty and Inequality Diagnostic, 2012 [page 41] : Investing in People to Fight Poverty in Haïti Chapter 1: Poverty profile and trends Two years after the earthquake, monetary and multidimensional poverty is still Stark in Haïti, particularly in rural areas. In 2072, almost 60 percent of the popula- tion was poor, and one person in four was living below the extreme poverty line. Nearly half the households are considered chronically poor because they are living below the moderate poverty line and lack at least three of the seven basic dimen- sions of nonmonetary well-being. In rural areas, these numbers rise even higher: three-quarters of all households are monetarily poor, and two-thirds are considered to be living in chronic poverty. Compared with 2000, monetary and multidimensional poverty has improved sli- ghtly. Consumption-based extreme poverty declined from 31 to 24 percent between 2000 and 2072, and there have been some gains in access to education and basic infrastructure, although the levels and quality are Low. Income inequality is the hi- ghest in the region—at a Gini coefficient of O.61—and has been steady at that value since 2001. Urban areas have fared better than rural areas, reflecting larger private transfers, more nonagricultural employment opportunities, narrowing inequality, and more access to critical goods and services. Continued progress in reducing extreme and moderate poverty will require greater, more broadbased growth, but also a concerted focus on improving access to basic opportunities in rural areas, where more than half the population resides, extre- me poverty has stagnated, and income inequality is increasing. The regular moni- toring of social indicators will provide the evidence base necessary for informed decision making. This chapter presents the poverty profile in Haïti and trends in poverty since ear- ly 2000. This is the first time such a diagnostic has been possible in more than a decade. The analysis is based on the nationally representative post-earthquake living conditions survey conducted by IHSI in 2012 (ECVMAS 2012), except where otherwise indicated”. The poverty estimates are based on official national poverty lines developed by the government on the basis of fresh household consumption data. The new methodology has meant that comparisons across time are delicate. Comparisons have been conducted using two data sources produced by IHSI: the living conditions survey of 2001 (ECVH 2001), which provides information on the socioeconomic characteristics of the population, and the budget and expenditure survey of 1999/2000 (EBCM), which provides the only nonofficial poverty line and consumption-based poverty estimates. 37 The final sample of ECVMAS 2072 includes 23,555 individuals from 4,930 households. | [page 42] WorldBank - ONPES | The rest of the chapter is organized into three sections. The next sectionillustrates and explains trends in poverty and inequality since 2000. The subsequent section offers a description of the poverty profile in 2072. The final section concludes. 2.5 million people in national extreme Poverty is endemic in Haiti, with a poverty headcount at 58.5 percent and ex- poverty line, 80 treme poverty at 23.8 percent at the national Level in 2012 (table 111). These percent of whom live numbers indicate that almost 6.3 million Haitians cannot meet their basic con- in rural areas. sumption needs. Among these, around 2.5 million cannot feed themselves ade- quately. The poverty gap indicator is also high, at 24.4 percent at the national level. This indicator, the poverty deficit, represents the average distance from the pover- ty line.58 This means that, on average, the poor Live on Less than 60 percent of the value of the poverty line, hence, Less than G 48 per capita per day’. Table 11. Poverty and extreme poverty in Haïti, 2012 Moderate pouerty Estimate Standard error 95% confidence interuals Severity of poverty 134 0.0059 134 134 ECTS EE EE Share of the poor 23.8 0.0129 237 238 Sources: ECVMAS 2072; World Bank and ONPES calculations. Geographically, the poverty and extreme poverty rates are considerably hi- gher in rural areas. Rural residents are at significantly greater risk of poverty re- lative to urban residents. In 2012, the majority of the population was still Living in rural areas (52 percent compared with 59 percent in 2001), although the gap between the urban and rural populations was progressively disappearing because of constant migration from the countryside to the cities. Among the rural popula- tion, the poverty rate was as high as 74.9 percent, representing 67.0 percent of the total number of poor in the country. In contrast, the poverty headcount in urban areas was 40.6 percent. Port-au-Prince features the lowest poverty headcount in the country, at 29.2 percent, and hosts 11.0 percent of the total number of the poor. Extreme poverty follows a similar pattern (figure 11). 38 The equation of Foster, Creer, and Thorbecke (1984) to calculate poverty indicators is as follows Pa=l/nilyi<z[z-yija (1) where nis population size, jare individuals, yis the per capita measure of well-being (that is, consump- tion), zis the poverty line, and /is a function that takes the value of 1 if the statement is true and O if not. Ia = O, the resulting indicator is the headcount (the per capita poverty rate), if « = 1, the result is the indicator of the poverty gap; and, if « = 2, the result is the indicator of the depth of poverty. 39 Since the poverty gap can be written as the product of the poverty rate and the average distance of poor households to the poverty Line, a poverty gap of 244 and a headcount of 58.5 percent imply that the average poor household lives on 58 percent of the poverty line. | [page 43] : Investing in People to Fight Poverty in Haïti Figure 1.1. Incidence of moderate and extreme poverty in urban and rural areas, 2012. a. Poverty @ Urbain Qu: © Hi 749 58.5 + 40.6 35.5 24 20.6 123 = 56 13 ‘ ee Poverty Headcount Poverty gap Squared poverty gap b. Extreme poverty @ Urbain Qu: © Haiti 378 238 + 128 8.6 77 59 22 + 09 35 ES] Poverty Headcount Poverty gap Squared poverty gap Sources: ECVMAS 2072; World Bank and ONPES calculations. Not only is poverty greater and more widespread in rural areas, but it is also more entrenched. The poverty gap is 35.5 percent in rural areas, compared with 12.3 percent in urban areas. This means that the budget of the rural poor must rise by an average G 39 per capita per day if the poor are to step out of poverty, while G 25 would suffice in urban areas. The severity of poverty is also greater in rural areas, where the squared poverty gap indicator is almost four times higher than in urban areas (5.6 versus 20.6). This indicator takes into account the incidence and depth of poverty, as well as inequalities among the poor. The poorest departments are the farthest from the capital and the most isolated. They are geographically concentrated in the North (Nord-Est, at 79.3 percent, and Nord-Ouest, at 81.8 percent) and the South (Grand'Anse, at 79.6 percent). In the- se departments, poverty is also the deepest and the most severe (appendix A). 40 In the framework of this report, Haiti was divided in five geographical regions: the North, the South, the Transversale (the Center), the West, and the Metropolitan Area [page 44] WorldBank - ONPES | The same pattern characterizes extreme poverty (map 11). Nord-Est and Nord- Ouest have the highest poverty incidence rates. These two departments and Grand’Anse share two features: remoteness from the capital and isolation becau- se of the poor transportation infrastructure, which makes them almost inaccessi- ble during the rainy season. The departments hosting the three largest cities also have the lowest poverty rates: Ouest, at 391 percent, includes the capital, Port-au- Prince; Artibonite, at 60.5 percent, includes Gonaïves, the third largest city of Haïti and à dynamic trade center that also accounts for 70 percent of the domestic pro- duction of rice; and Nord, at 68.8 percent, includes Cap Haïtien, the second largest city and the second commercial and tourism port of the country.“ Map 1.1. Moderate and extreme poverty rates, by department, 2012 a. Poverty headcount b. Extreme poverty headcount nn, nn > es >, Sud , Sud - Est sud ne DT NN Sn s =) Below 40% From 10% to 25% Between 40% and 60% From 27% to 32% Between 60% and 78% From 33% to 37% Above 78% More than 37% Sources: ECVMAS 2072; World Bank and ONPES calculations. Since 2000, poverty has marginally decreased across the country, especially in urban areas (figure 1.2). Extreme poverty declined from 314 percent in 2000 to 24.0 percent in 2072, driven by progress in urban areas. While it moved from around 21 and 20 percent down to 12 and 5 percent in other urban areas and the Metropolitan Area, respectively, extreme poverty stagnated in rural areas. Data are not available to assess the relevant trends, but moderate consumption poverty is also estimated to have modestly improved in the last decade.+ 41° For the details on Gonaïves, see “Haiti-USAID Best Analysis; March 2013, Office of Food for Peace, US. Agency for International Development, Washington, DC, http.//www.usaidbestorg/docs/haitiReportpdf. 42 Income-based measures suggest that moderate poverty declined from 77 percent in 2001 (ECVH 2001) to 72 percent in 2012 (ECVMAS 2072). Consumption-based poverty measures are considered [page 45] : Investing in People to Fight Poverty in Haïti Figure 1.2. Trends in extreme poverty in urban and rural areas, 2000-2012 € £ ë à KR © 5 40% & D) à À à 35% D 2 5 30% a Es ë @2000 © 25% (NI rs à ©2072 £ 20% & L e S 15% F £ 10% ë E Ÿ ÿ 5% ©O O% £ = 3 2% Ë& Es National Rural Urban Metropolitan area Sources: EBCM 1999/2001; ECVH 2001; ECVMAS 2012; World Bank and ONPES calculations. More generally, living conditions measured according to indicators of access to basic services have improved in Haiti since 2001, but challenges remain (table 1.2). The biggest gains were in education, where participation rates among school- age children rose from 78 to 90 percent. However, the quality of education is a con- cern: because of a combination of late starts, dropouts, and repetitions, only one- third of all 14-vear-olds are in the appropriate grade for age.“ The open defecation rate declined from 63 to 33 percent nationwide between 2000 and 2072, reflecting gains in both urban and rural areas. However, the quality of sanitation access is Low. Only 31 percent of the population had access in 2012 to improved sanitation; in rural areas, the share was 16 percent.“ Access to an improved source of water is similar in urban and rural areas, at 55 and 52 percent, respectively. However, most of the remainder of the urban population (36 percent) can still access clean water by pur- chasing it, while the rest of the rural population (40 percent) use unimproved water sources, with a high probability of contamination. Access to energy (electricity, solar, or generators) expanded overall only slightly because of gains in urban areas, while it was constant in rural areas, at 11 percent. the most accurate in capturing welfare levels, especially in countries with high rates of rural poverty and significant income volatility; the new, official Haitian poverty measure is consumption based. Fur- ther poverty measures are presented in appendix C. 43 Education background paper (2014), Haïti Poverty Assessment, World Bank, Washington, DC. 44 improved sanitation includes flush toilets as well as improved latrines. According to the United Nations Children's Fund and the World Health Organization, an improved sanitation latrine is one that hygieni- cally separates human excreta from human contact. 45 Such as dwellings constructed with hazardous materials, untreated sources of water, and use of sur- face water (rivers, Lakes). [page 46] WorldBank - ONPES | Table 1.2. Access to basic services. Coverage rates, % Indicator 2001 2012 2001 2012 2001 2012 Access to improved drinking water sources de Habitat, nonhazardous building materials Sources: ECVH 2001; ECVMAS 2012; World Bank and ONPES calculations. Note: — = not available WHO = World Health Organization. a. According to the international definition (WHO), access to improved drinking water is the proportion of people using improved drinking water sources: household connection, public standpipe, borehole, protected dug well, protected spring rainwater b. The expanded definition includes the international definition (WHO), plus treated water (purchased) c. includes electricity, solar, and generators. d. Rate of open defecation refers to the proportion ofindividuals who do not have access to improved or unimproved sanitation. This indicator is part of the Millennium Development Goals (MDG) and is a key element of discussion for the post-2015 agenda The open defecation rate declined from 63 to 33 percent nationwide between 2000 and 2072 reflecting gains in both urban and rural areas. e. Improved sanitation is access to a flush toilet or an improved public or private latrine Haiti is a very unequal country, and income inequality in rural areas has in- creased. While income inequality—the Gini coefficient—has stagnated at O.61 over slightly more than 10 years, it deteriorated in rural areas from O.50 to 0.56. Furthermore, the top quintile of the distribution gathers more than 60 percent of the national wealth, and the top 1 percent of the population was living on more than 50 times the resources of the bottom 10 percent. These results suggest that Haiti is among the most unequal countries in Latin America (figure 1.3)*£. 46 The income aggregates in 2001 and 2012 were calculated using the same methodology (CEDLAS and World Bank 2012). To ensure comparability, the aggregate was not geographically deflated. Howev- er, the Gini coefficient would barely move in 2012 ifit were geographically deflated (from 0.608 to O.610). Part of the stagnation in inequality is driven by the imputed rent component of the income aggregate, which is of dubious quality in the case of the 2001 data. If this component is removed for both years, inequality decreases (from 0.67 to 0.63). Inequality would still increase in rural areas, although marginally, and decrease in urban areas. 47 The gap between the top 1 percent and the bottom 10 percent is most likely larger because the richest households tend not to be fully represented in the household data. The statistics are based on household income. 48 Appendix B presents Lorenz curves at the national, urban and rural levels. [page 47] : Investing in People to Fight Poverty in Haïti Figure 1.3. Income inequality in Haiti and in Latin America 20 percent of the . . . ; . . . richest households a. Gini coefficient (income based), selected Latin American countries, circa 2012 hold 64 percent of o7 total incomes 0.6 in Haiti O5 O4 O3 O2 o1 o > 5 œ 2 œ 5 œ œ [e] Lo} œ [®]} >» = œ Le} un 5 & © C 2 © ES Ô 2 À < S 8 5 HP LÉ + 3 % 8 6 à à À À À à © 2 à 9 Do ® £$ > & & GS © À 5 8 © mi 8 cs & CR S 5 < 8 8 b. Wealth accumulation, by income quintile ranked from highest to lowest À 74.0 EN 646 . Ü ‘ 9 54.8 . « e q : # 431 D * . Le 242, ut" 203 ... te TT. 418.6 UC Te 16.0 40 ot .° 38 > A TBrrrrrrrreege tt" 62 09: 2 3 à 5 — Haiti —— LAC average ++. Haiti-urban ++. Haïti-rural Sources: ECVMAS 2072; World Bank and ONPES calculations. [page 48] WorldBank - ONPES | In Haiti, 60 percent of households are considered multidimensionally poor in that they lack access to at least three of seven basic characteristics of we- ll-being (education of children and adults, improved sanitation, clean water, reliable energy, nonhazardous housing, and food security) (box 1.1). House- holds that are below the monetary consumption poverty line and lack access to basic goods and services are considered chronically poor because they face particularly difficult challenges in emerging from poverty compared with the transient poor, who may lack monetary resources, but who have access to basic services. Nearly half the households in Haiti are chronically poor, suggesting that their opportunities to emerge out of poverty and improve living conditions are highly constrained (figure 1.4). Almost 70 percent of rural households are conside- red chronically poor, compared with only slightly over 20 percent in urban areas, highlighting the particularly narrow opportunity to emerge from poverty in rural Haiti. Only 14 percent of households nationally are accounted for by the transient poor in that they lack monetary resources, but have access to basic services and are more likely to be able to rise above the poverty line, while another 12 percent are above the moderate poverty Line, but are made vulnerable by their deprivation in access to basic services. Box 11. The use of the multidimensional poverty index to identify the chronic poor The division of poverty into the chronic, deprived, and transient poor is based on a combination of the multidimensional poverty index (MP1) me- thodology and monetary poverty as measured by the poverty headcount indicator. The MPI, which includes nonmonetary poverty indicators, iden- tifies deprivation as the lack of access to basic services and infrastructu- re. We have included seven dimensions in the MPI indicator, namely, the educational attainment of children, the educational attainment of adults, health, water, sanitation, energy, and habitat, along with seven indicators, encompassing each dimension: education (the household head is litera- te; all school-age children are in school), health (the food security index), water (access to an improved source of drinking water), sanitation (access to improved sanitation), energy (access to a sustainable energy source), and habitat (access to a dwelling constructed of nonhazardous material- s)*. To be considered MPI poor, households must be deprived across at least three dimensions (Lépez-Calva 2013). In 2012, Haiti had a multidi- mensional poverty rate of 60 percent (figure B111). 49 For details of the methodology see appendix D. | [page 49] : Investing in People to Fight Poverty in Haïti Nearly half of the Figure B1... Poverty decomposition population (45 %) according to the MPI and monetary poverty is not only poor but also lack access to _—— : : er capita essential services consumption and infrastructure Deprived Poverty Line HTG 29,909.87 S ë Chronic poor Transient poor 8 7 6 5 4 3 2 1 0 Multi-dimensionally poor Numbers of deprivations Sources: ECVMAS 2072; World Bank and ONPES calculations. The chronic poor are those individuals living in households that are de- prived in terms of access to basic services, infrastructure or that are food insecure (deprived across at least three dimensions of the MPI) and poor in monetary terms. The deprived are those individuals living in households that do not consist of the monetary poor, but that are deprived across the MPI dimensions. The transient poor are those who are among the mone- tary poor, but who are not MPI deprived. Finally, to be defined as resilient, individuals must live in households that are neither deprived, nor among the monetary poor. Figure 1.4. Chronic and transitory poverty, service access deprivation and resilience in Haïti, 2012 @rationa rural Q@ urban 67% Li 45% Eh 29% 20% x 21% ee | He 12% 14% 0% L EPS 1% Chronic Deprived Transient Non-poor Sources: ECVMAS 2072; World Bank and ONPES calculations. [page 50] WorldBank - ONPES | This section provides answers to two main questions: Who are the poor? and Which factors are correlated with the escape from poverty? The answer to the first describes the demographic, social, and economic characteristics of the poor. The answer to the secondillustrates Levers and dimensions that reduce the risk of poverty. Each answer is presented in turn in the next subsections. Characteristics of the poor The poor live in larger households characterized by greater economic depen- dence and less educational attainment. While the average size of households among nonpoor households is 4.0 persons, the corresponding size among the pooris 5,3 (table 1.3). Among these poor households, 80 percent have five or more persons. The dependency ratio is 54 percent among nonpoor households, but 88 percent among poor households, pointing to greater pressure on the producti- ve population in these households.®° The poor live in households in which the heads have an average of three times fewer years of education, and as many as 61 percent of the heads of poor households are illiterate, compared with 34 per- cent among nonpoor households. These characteristics are exacerbated among the extreme poor and in rural areas, where poverty is more extensive and more entrenched, and are consistent with other studies on the issue”! (for example, see Sletten and Egset 2004; World Bank 2006; ONPES 2014). Table 1.3. Basic sociodemographic and socioeconomic characteristics of poor, extreme poor, and nonpoor households. Averages S S S Variable - à. - à. - à Household size, number® Literacy, heads w/>5 years of education, % 50 The dependency ratio is the proportion of household members aged 15-70 relative to the total number of members, regardless of age. Normally, the dependency ratio is defined on the basis of the 15-65 age-group, reflecting the formal retirement age. However, in the case of Haiti, where formality represents a small share of the Labor force, the age threshold of 65 is not realistic. 51 For the evolution of characteristics of households between 2001 and 2012 see appendix E. | [page 51] : Investing in People to Fight Poverty in Haïti Employed household members, number® Head employed in agriculture, % Head employed in the formal sector, % Head employed in the informal sector, % D CR private transfers (exclu- Households receiving remittances , % Average per capita consumption, HTG Average food share in total consumption, % ACCess to improved sanitation, % Access to a sustainable source of energy, % | 583 | 282 | 79 | m0 | 513 | 324 | 261 | 98 | 28 | Dwellings made w/nonhazardous materials, % Note: The estimates for the poor exclude the extreme poor. Variables reflect the percent share of individuals. a. Share of households relative to the average household. Poor urban and rural households have evolved in different environments that generate specific challenges. While average per capita consumption is similar among the rural poor and the urban poor, there are still important differences cha- racterizing rural and urban livelihoods. Thus, rural households devote a much hi- gher share of their consumption to food (63 percent), while urban households can afford a larger share of nonfood consumption (55 percent), including higher-value dwellings, more assets, and more access to services. These differences reflect a different composition of expenditure and better access to goods and services in urban areas. Most of the poor work, but their earnings are insufficient to lift them out of po- verty, particularly if they are working in the primary sector. Almost 70 percent of the heads of poor households have jobs (against 73 percent among the nonpoor). However, among the former, 61 percent work in agriculture, where average earnings are Less than 20 percent of the earnings in the formal sector. Among the remainder, 35 percent work in the informal sector, where earnings are Less than half those in the formal sector (4 percent) As a result, more than half of poor households under- take two or more income generating activities. Compared with the nonpoor, the poor rely significantly more on private transfers and production for home consumption and less on Labor income (figure 1.5). Al- though Labor income tends to be the main source of livelihoods among Haitian hou- seholds, this is not the case among the extreme poor, who depend more on private transfers (in urban areas) and production for home consumption (in rural areas). More broadly, livelihoods in rural areas are substantially Less well connected to markets and more dependent on a self-sufficiency economy. Of the means of livelihood in rural areas, 25 percent derive from production for home consumption. | [page 52] WorldBank - ONPES | Figure 15. Income composition in urban and rural areas and by poverty status a. Income composition, by area of residence ES Ëe eff = Metropolitan Area Other urban Rural National 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% © pensions @ Labor income © scholarships © capital revenue C] Imputed rent e Private Transfers @ Production for home consumption b. Income composition, by poverty status e LL Ÿ Ÿ @ mputed rent 100% 90% @ rrivate Transiers 80% 70% @ scholarships 60% n 50% @ canital revenue 40% n Pi 0% @ Pensions 20% @ Labor income 10% 0% @ Production for home consumption Non-poor Poor Extreme poor Sources: ECVMAS 2012; World Bank and ONPES calculations. The most abundant asset of the poor is human capital, but the poor face signi- ficantly higher barriers in access to health care and education.°? Children in poor households are Less likely to be in school: 87 percent of children aged 6-14 in poor households are in school, compared with 96 percent of children in nonpoor house- holds (see chapter 3). This suggests that poverty is an important barrier to school en- rollment, which is further supported by the fact that, in 83 percent of cases, cost is 52 Human capital is defined here broadly as a set of intangible assets, skills, and knowledge that can create economic value and generate more remunerative Labor outcomes. [page 53] : Investing in People to Fight Poverty in Haïti the main reason for keeping children out of school. Financial barriers are also the key obstacle to access to health care among the poorest, followed by lack of transporta- tion°* These barriers to investment in human capital are larger in rural areas, where poverty incidence is greater and service delivery more limited. Still, even though the education levels and health status of the poor are Low, human capital is their strongest asset because their access to physical or financial capital is constrained. The poor suffer from poor nutrition early in life and from food insecurity, which also affects their investments in human capital. Food insecurity is significant in Haïti, at 28 percent nationwide and 34 percent in rural areas.“ Poor household members are much more likely to report frequent hunger or lack of food at bed time relative to the members of nonpoor households (figure 1.6). Households with children under the age of 5 are much more likely to experience repeated food shortages.°° As a result, one-fifth of under-5-year-olds are chronically malnourished (DHS 2012). This is a particular cause of concern because proper nutrition in early life is crucial for brain development and subsequent life outcomes (Alderman and King 2006). Figure 1.6. Food insecurity in Haïti, 2012. Food availability among the poor and nonpoor and among households with and without young children 100 80 60 40 20 0 Non Poor | Non Poor | Non Poor e Non Poor | Non Poor | Non Poor poor poor poor poor poor poor No food |Went to bed| Entire day No food |Went to bed| Entire day hungry |with no food (child<s5) hungry |with no food (child<5) (child<5) @never (O days) [2 Rarely (3-10 times/month) @ Often (more than 10 times / month) Sources: ECVMAS 2072; World Bank and ONPES calculations. Note: The survey questionnaire asked how often, over the past four weeks, a household had experienced ‘no food at all” or that at least one household member ‘went to bed hungry” or “spent all day without eating” 53 Because of a lack of money, 7 in 10 women aged 15-49 do not seek medical support, while 43 percent do not seek the support for Lack of transportation, according to the DHS 2012 (see chapter 3) 54 According to the National Food Security Coordination Unit, the food insecurity rate was 28 percent nationwide and 48 percent in rural areas in 20%. To measure food insecurity, the unit uses à compos- ite indicator composed of both quantitative and qualitative measures. The numbers contained in this chapter, on the other hand, refer exclusively to the food security indicator of the Food and Agriculture Organization of the United Nations, which is based on food intake. 55 Shared prosperity background paper (2014), Haïti Poverty Assessment, World Bank, Washington, DC. [page 54] WorldBank - ONPES | The poor in Haiti are particularly vulnerable to shocks and are more likely to re- sort to strategies that are harmful to human and physical capital accumulation (figure 1.7). À typical Haitian household faces multiple shocks annually: nearly 75 per- cent of households suffer economic consequences following a shock°° Households in poverty are more vulnerable, particularly those in extreme poverty. AmMong poor households, 95 percent experience at least one economically damaging shock per year. In most cases, households cope through monetary support provided by others (27 percent) or by changing nutritional inputs (16 percent)? However, the extreme poor receive relatively Less financial support (17 percent versus 37 percent in resilient households) and change their food consumption habits more frequently (22 versus 10 percent). In particular, if the shock hits the entire community, a staggering 56 percent of households in extreme poverty change their nutritional behavior, as opposed to 37 percent ofresilient households. The extreme poor are also more likely to remove their children from school because of shocks, particularly if the household is experiencing a change in composition (such as the birth or death of a household member) or a decline in the monetary support from outside the household, which is often used to pay school fees. (Box1.2 examines the issue of gender inequality, another determinant of poverty.) Figure 1.7. Share of the population affected by a climatic shock and poverty Level, by department @ai Poor @Extreme poor 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 3 ë Ë OS £ 7 à ? 8 È = 2 3 5 * Source: ECVMAS 2012; World Bank and ONPES calculations.Note: The poverty line i5G2990987 The extreme poverty line is G15,240.03. Climatic shocks include hurricanes, floods, droughts, and excessive rainfall The survey questionnaire asked ‘during the last 12 months, was your household affected by one ofthe following?" 56 Shocks background paper (2014), Haiti Poverty Assessment, World Bank, Washington, DC. 57 The latter strategy includes decreasing the quantity of food, the number of meals consumed, or food qua consuming food harvested before maturity, gathering food in the wild; and reliance on seeds | [page 55] : Investing in People to Fight Poverty in Haïti Box 12. Gender inequalities generate great vulnerabilities in Haïti Women and girls in Haiti face significant obstacles when accumula- ting assets, including human capital, and register Lower education and health outcomes. Despite sizable progress in school enrollment among younger cohorts, adult women are still Less well educated than adult men and are more likely to be illiterate. Adult men have, on average, two more years of education than women and are over 10 percentage points more likely to be literate. Early withdrawal from school can have lifelong conse- quences. Underage marriage and teen pregnancy, for instance, represent additional threats for girls who are not in school: 17 percent of Haitian wo- men are married in adolescence, compared with 2 percent of men, while this number drops among girls with higher education (Cicmil 2013). Maternal mortality, at 380 deaths per 100,000 live births, is still five times higher than the regional average (WHO 2014a)%. Fertility rates are also way above regional figures, particularly among less well educated women hou- sehold heads: those with no formal education have twice the number of children relative to women with at least upper-secondary schooling. Poor nutrition is also a threat for both children and mothers: according to the DHS 2072, 22 percent of children are stunted or too short for their age, while near- ly half of women aged 15-49 have anemia. The prevalence of HIV/AIDS is higher among women (27 percent) than men (17 percent), reflecting, among other things, knowledge differentials: only 15 percent of young women have correct information on how to prevent sexual HIV transmission, versus 28 of young man. (Boesten and Poku 2009). Furthermore, poor education and gender norms interact with health outcomes: there is anecdotal evidence that cultural reasons play a major role in the high percentage of birth delive- ries in Haïti that take place outside a health care facility (65 percent), genera- ting more risk of maternal mortality. Women are significantly disadvantaged in using their assets and obtaining the relevant returns, particularly in the Labor market. Apart from initial diffe- rences in endowments, women in Haiti seem to face additional obstacles in participating in the Labor market. Holding constant several social and demo- graphic characteristics, one finds that women are 20 percentage points more likely than men to be unemployed and, if working, 6 percentage points more likely to be in the informal sector. Wages among women are also 32 percent lower than wages among men. Statistical tests show that over two-thirds of this difference is unexplained by observable characteristics, suggesting that discrimination could play a role in accounting for the result. Gender-based violence and low participation in the public sphere are wi- despread in Haïti. Gender-based violence is a chronic problem: according to the DHS 2012, 13 percent of Haitian women have experienced sexual violence, and 29 percent of women who have ever been married have ex- 58 This number is not accepted by the Ministry of Public Health and Population (MSPP) | [page 56] WorldBank - ONPES | perienced spousal violence, whether emotional, physical, or sexual. Vul- nerability is particularly high among internally displaced people in camps and areas affected by the 2010 earthquake: a survey in 2011 indicated that 64 percent of 981 pregnant adolescent girls who were interviewed had become pregnant after being raped (PotoFi 2012). Raising awareness, improving security and legislation, and creating economic opportunities for women are important measures to address the immediate and long- term needs of women and girls against gender-based violence. Only 4 percent of all parliamentary seats are occupied by women, placing Haïti 136th of 142 countries and well below the regional average of 26 per- cent. At the national level, as of April 2014, 8 of 23 ministers and 3 of 20 secretaries of state were women: At the local level, women account for only 12 percent of all mayors. The government has taken steps to expand women's representation, including by creating the Gender Equality Office in Parliament and amending the Constitution to stipulate a quota of at least 30 percent women in all public offices, but there is no enforcement mechanism, and implementation remains low at all levels of formal po- litical life. a. CEPALSTAT (database), Statistics Division, United Nations Economic Commission for Latin America and the Caribbean, Santiago, Chile, http://estadisticas.cepal.org/cepalstat/WEB _ CEPALSTAT/Portada.asp?idiomazi. Risk factors associated with povertys Larger households and children are more likely to be poor. Poverty is three ti- mes more widespread among households with more than six members relative to households with fewer than three members (73.6 versus 24.6 percent) (table 1.4). In particular, the presence of young children more often translates into higher poverty rates. Poverty is more extensive among children and relatively Less exten- sive among adults. Almost 70 percent of preschool-age children (under 5 years of age) live in poor households, highlighting the vulnerability of this age-group. The poverty rate among school-age children (ages 5-14) is the second highest, at 66 percent, representing 27 percent of all the poor. 59 For the results of linear regressions to identify poverty correlates, see appendix F. The regressions take into account demographic and socioeconomic characteristics such as educational attainment among heads of household, household composition, and Labor market participation to predict per capita consumption (Log and normalized by poverty line). | [page 57] : Investing in People to Fight Poverty in Haïti Table 1.4. Poverty incidence, by category of household Characteristic Poverty Extreme pouerty | Population | Poor Extreme poor Area of residence [um Joe | w Ju |[wm) m D D Household size, persons Do Dm D Le Le Household compositiona Gender, head Status ofthe head Dm Te | w au fol Educational attainment, head | [page 58] WorldBank - ONPES | Labor force status, head Sector of activity, head Socioeconomic position, head Total 585 2237 100 100 100 Sources: ECVMAS 20172; World Bank and ONPES calculations. a. Poverty measured at the individual level, by age-group. b. See the text for an explanation of plaçage. Poverty incidence does not differ by gender, but it does differ by marital status. Unlike in 2001, the poverty rate among individuals living in woman- or man-headed households is not statistically different, at 58.3 and 59.0 percent, res- pectively. Meanwhile, 72 percent of the poor Live in households in which the head is in a formal relationship, either married or placé. Plaçage is a form of customary union. Plaçage is common in Haiti, particularly in rural areas, where it involves 36.2 percent of all household heads. Poverty incidence is more than 10 percentage points higher among households in which the heads are placés than among hou- seholds in which the heads are married. The poverty rate is higher among households in which the heads are relatively uneducated. Poverty incidence is more than four times greater among households headed by a person with no education relative to households with heads who have completed secondary or higher education (776 versus 178 percent). Households with uneducated heads represent more than 50.0 percent of the poor, while a stag- gering 605 percent have heads who have not completed primary education. [page 59] : Investing in People to Fight Poverty in Haïti The poverty rate is higher among the unemployed, but only in urban areas. Par- ticipation in work is associated with somewhat Less poverty incidence only in urban areas, where unemployment increases the poverty rate by more than 10 percentage points (figure 1.8). Almost 40.0 percent of those in urban areas who work do not earn enough to stay out of poverty. The corresponding share is 75.5 percent in rural areas, and, nationwide, there is no statistically significant difference in poverty rates among those who work and those who are unemployed, though those not in the Labor force show a slightly higher poverty rate. The poverty rate is especially high among hou- seholds in which the heads work in the primary sector, at 76 percent (for example, in agriculture, forestry, or fishing), or in the informal sector, at 45.2 percent, which emplo- ys 73.0 and 32.6 percent of the total urban and rural Labor force, respectively. Figure 1.8. Poverty rate by region, economic situation and household head's sector of activity. a. By area of residence and economic status 755 784 8Oh 778 598 608 639 615 50.6 39.8 457 448 218$ 852$ 8 2 28) 5)8e a L es C e [= £ — £ — £ — [0] [0] [0] 5 5 5 Ea aff = Urbain Rural National b. By sector of activity @ Poverty headcount @ % population @ % poor 71 54 51 39 43 42 21 10 4 4 © © Primary Formal Informal Sources: ECVMAS 2072; World Bank and ONPES calculations. [page 60] WorldBank - ONPES | of the poorest More than 10 years after the last household living conditions survey, the avai- households in Haiti lability of new data has made a fresh diagnosis possible. The use of the recent have access to postearthquake living conditions survey (ECVMAS 2012) and the official poverty improved sanitation; lines developed by the government has served as a basis for the identification of compared with 65 the poor, a description of the main characteristics ofthe poor, and a determination percent of the richest . . . households. ofthe principal risks associated with poverty. Poverty is widespread in Haïti, and it is deeper and more severe in rural areas. In 2072, the overall poverty headcount was 58.5 percent, and the extreme pover- ty rate was 23.8 percent. The incidence of poverty is considerably higher in rural areas and in the North. More than 80 percent of the extreme poor live in rural areas, where 38 percent live in extreme poverty, compared with 12 percent in ur- ban areas and 5 percent in the Metropolitan Area. The progress in reducing poverty has been modest in urban areas, but the stagnation in rural areas generates concern. At the national Level, the extreme poverty rate declined from 31 to 24 percent between 2000 and 2012. However, advances in urban areas were behind this decline, while poverty stagnated in ru- ral areas. Almost 70 percent of rural households are considered chronically poor, compared with 20 percent in urban areas, highlighting a double deprivation—in monetary terms and in access to basic services and infrastructure —and the parti- cularly narrow opportunities to emerge from poverty in rural Haiti. Inequality is still wide in terms of both income and access to basic services, preventing the poor from accumulating and effectively using their human ca- pital and improving their well-being. Income inequality is the highest in Latin America; the Gini coefficient was O.61 in 2012 and the richest 20 percent of the distribution gathers more than 60 percent of the national wealth. Although ac- cess to basic services has improved since 2001, levels are still Low, particularly in rural areas, and the quality of services is limited. Furthermore, access increases with wealth, and the poor have markedly Less access to services, including educa- tion and health care, because the cost represents a considerable burden on the budgets of the poor and an important barrier to human capital accumulation. In particular, educational attainment, which correlates strongly with welfare, is Low among the poor, affecting the capacity of the poor to generate income. Women and girls are particularly vulnerable because they face important obstacles to the accumulation and use of their assets, particularly their hu- man capital. Despite sizable progress in both education and health outcomes, adult women are still less well educated than adult men and are more likely to be illiterate, while maternal mortality is still dramatically high. Apart from initial diffe- rences in endowments, women in Haiti also face additional obstacles in participa- ting in the Labor market because they are significantly Less likely to be employed and earn more than 30 percent Less than men. Finally, gender-based violence and low participation in the public sphere are widespread in Haiti. | [page 61] : Investing in People to Fight Poverty in Haïti In light of this diagnostic, the following messages emerge as key for additional, sus- tainable poverty reduction: The regular monitoring of poverty and living conditions is a necessary step to promoting evidence-based, effective policy making. One of the many obstacles to postearthquake reconstruction and emergency operations was the lack of sound statistical information at the national level. Strengthening the national statistical system through investments in this sector will allow the country to have reliable data from various sectors, through regular national censuses and surveys, such as ECVMAS, that will permit regular and systematic monitoring of poverty and house- holds living conditions in Haïti, relying on the new reference rates for the country. At the same time, regular monitoring built on the solid baseline set out in this report will contribute to enhancing the design and efficacy of antipoverty policy making. Policies should encompass ways to boost the income generation capacity of the poor and to protecting their assets from shocks more effectively, while overall economic growth remains a prerequisite for any poverty reduction. This diagnostic highlights that the poor in Haiti face significant obstacles to accumula- ting, using, obtaining the returns to, and protecting their assets. In urban areas, the poor struggle to find a (decent) job and heavily rely on private transfers; in rural areas, the poor are highly dependent on subsistence agriculture, the productivity of which is severely affected by frequent natural disasters and which is associated with significant food insecurity. Three-quarters of Haitians and 95 percent of the poor suffer from at least one economically damaging shock per year. Human capi- tal accumulation to seize the best opportunities, protection from shocks to reduce losses and damage, and ex ante and ex post coping strategies are the priority areas of the actions needed to reduce chronic poverty and promote shared prosperity. | [page 62] Part Il Drivers and Constraints for Poverty Reduction [page 63] : Investing in People to Fight Poverty in Haïti Chapter 2: Income generation in rural and urban areas Sustainable poverty and inequality reduction builds on strengthening the capacity of rural and urban populations to generate income in a reliable form. Haïiti’s popula- tion is equally split: half the people live in rural areas, and half live in urban areas. While there is a trend toward greater urbanization, half the country still depends on income sources that are subject to rural realities and end up with a poverty inciden- ce of 75 percent. The other half strives to find job opportunities that may propel them above the poverty Line today, but render them highly vulnerable to recurrent adverse social and economic shocks tomorrow. This chapter outlines the challenges and opportunities for income generation in Haiti. It is organized as follows.°° The introduction discusses the role of income in the poverty trends observed in the past decade. The next section delves into the rural reality of income generation and the constraints faced in productive farming. The following section addresses Labor opportunities in urban areas and the case of self-employment, one ofthe most salient aspects of the urban job market. The sub- sequent section presents migration and foreign and internal transfers as a strategy to complement Labor income and enhance well-being. The final section concludes. The key driver of the poverty gains in Haiti was increased access to nonagri- cultural income in urban areas. In a context of limited economic growth (see the Background and introduction), the share of nonagricultural income rose among all households in urban areas except for the first quintile, the extreme poor (figure 21). The shift toward nonagricultural employment in urban areas likely reflects a tran- sition toward better paid jobs in construction, transport, and telecommunications, sectors that experienced positive GDP growth during the period. The average hourly labor income is two to four times higher in the informal and formal sectors than in the agricultural sector.f! In contrast, households in the first quintile saw their share of nonagricultural income fall, while the contribution of private transfers (domestic and international remittances) in their income rose. The movement out of the agri- cultural sector has been accompanied by à deepening in migration from rural to urban areas, where access to economic opportunities and services is greater. 60 This chapter draws on Atuesta, Cuevas, and Rodella (2014), Coello et al. (2014), ONPES (2014) and Cuevas, Marzo and Scot (2014) background papers prepared for the study by the World Bank and Ob- servatoire National de La Pauvreté et de l'Exclusion Sociale (ONPES). 2014. Investing in People to Fight Poverty in Haiti, Reflections for Evidence-based Policy Making. Washington, DC: World Bank. 61 The informal sector is defined by the International Labour Organization as unincorporated enterprises (household businesses) that are not registered, do not keep formal accounts, and are not in the prima- ry sector (agriculture). | [page 64] WorldBank - ONPES | Figure 21. Change in per capita income in urban areas, by income quintile, 2001-2012 2001 | 2012 | 2001| 2012 | 2001| 2012 | 2001| 2012 | 2001| 2012 1 2 3 4 5 @ Production for home consumpion @ agriculture labor income @ Non-agriculture labor income @ Pensions @ capital @ Private transfers Public transfers @imouted Rent Sources: ECVMAS 2072 and ECVH 2001; World Bank and ONPES calculations. Income generation opportunities in urban areas are limited by a two-sided problem: the low growth and scarcity of jobs and the prevalence of Low-qua- More than half . of the working lity employment. Unemployment affects 40 percent of the urban workforce and poor operate in almost 50 percent of the female workforce. Youth face unemployment rates that agriculture and more are above 60 percent, which triggers not only economic, but also social concer- than 40% work in ns®2. The steep challenge of finding a job ends up producing high levels of discou- the informal sector, ragement.£ Haiti has a Low Labor force participation rate compared to the rest of mainly as self- the region: only 60 percent of working-age individuals (15-64) participate in the employed. labor market, compared, for example, with 70 percent in the neighboring Domi- nican Republic. Among the people who find a job, 60 percent earn Less than the minimum wage, and women earn, on average, 32 percent Less than men.£* Education plays a critical role in improving welfare in urban areas: Labor income is, on average, 28 percent higher among individuals who have completed primary education than among uneducated individuals. In this context, the urban poor re- sort to self-employment or two-person businesses as a coping mechanism. Ove- rall, almost 60 percent of the poor are in this type of occupation, and 75 percent of the poor are active in sectors such as trade, construction, and low-skilled services. 62 Extended unemployment rate, which includes not only people in working age who do not have a job but are looking for one, but also those who are not Looking for a job because they are discouraged, waiting for a job answer, retired or sick, but would be immediately available if offered an opportunity. 63 Finding a job is made difficult by the limited opportunities as well as lack of information on job opportunities, as formal channels to access job offers are generally unavailable: two wage workers in three use personal connections to Look for and find jobs (ECVMAS 2012) 64 This is so after one controls for age, education, experience, household size, number of young chil- dren in the household, urban location, and sector of activity. [page 65] : Investing in People to Fight Poverty in Haïti The stagnation in rural poverty reflects an increasing reliance on the low-performing agricultural sector and production for home consumption. Over the decade, agricul- tural income (including production for own consumption and agricultural Labor in- come) grew in importance, representing between 48 and 59 percent of the incomes among the first three quintiles (figure 2.2). Rural livelihoods are highly dependent on agriculture: almost 80 percent of households engage in farming. Moreover, among halfthe households, farming is the sole economic activity. Returns to agriculture are low and unreliable, and the activity resembles a subsistence strategy rather than reliance on a productive economic sector‘ The experiences of more successful farmers suggest that improving access to inputs and supporting crop diversification are the main channels to elevating productivity (see below). Among the poor, only 20 percent use fertilizer and pesticides. Moreover, even though the average area of cultivated land is only slightly smaller among the poor than among the nonpoor (1.2 hectares versus 1.6 hectares, respectively), the poor spend two to four times Less on fertilizer, pesticides, seeds, and Labor. Figure 2.2. Change in per capita income in rural areas, by income quintile, 2001-2012 100% === — x RER" 60% » 40% 20% 0% 2001 2012 | 2001 2012 | 2001 2012 | 2001 2012 | 2001 2012 1 2 3 4 5 © Production for home consumpion [2 Agriculture Labor income © Non-agriculture Labor income @ Pensions © capital @ Private transfers Public transfers @ imputed rent Sources: ECVMAS 2072 and ECVH 2001; World Bank and ONPES calculations. 65 Since 2000, the sector has performed poorly, contracting by 0.6 percent annually as a consequence ofrepeated adverse climatic shocks. In 2072, agricultural production contracted by 13 percent fol- lowing a series of droughts, heavy rains, and hurricanes, which generated crop and seasonal income losses of 40 to 80 percent. The drop in production led to a decline in the demand for Labor and a rise in the cost of locally produced food. As a result, poor households lost income and faced higher con- sumption costs (prices) (Haiti Food Security Outlook, Famine Early Warning System Network, October 2012-March 2013) [page 66] WorldBank - ONPES | Participation in the nonfarm sector is key to emerging from poverty in rural Hai- ti. Engaging in the nonfarm sector in rural areas reduces the probability of being poor by 10 percentage points. The typical nonfarm job in rural areas is a one- or two-person shop engaged in small retail. Still, the returns to this activity surpass those accruing to farming. About 40 percent of nonpoor households participate in the nonfarm sector, a participation rate that is 1,5 times higher than the participation rate among the poor. External financial flows, including remittances and international aid, have also contributed to the decline in poverty. The share of households receiving private transfers in Haïti rose from 42 to 69 percent between 2001 and 2072, including both domestic and international transfers. Per capita remittances increased by 26 percent between 2001 and 2012 (in real terms) Worker transfers from abroad have repre- sented more than a fifth of Haïti's GDP in recent years; they originate mainly from the Dominican Republic and the United States. While transfers from the former are more li- kely to reduce poverty because they tend to benefit poorer households located in rural areas, the remittance flows from the United States are larger. Furthermore, in the after- math of the 2010 earthquake, the country catalyzed international solidarity, resulting in unprecedented aid flows in the form of money, goods, and services. These external flows also contributed to poverty reduction over the period, especially in urban areas, which attracted most of the assistance. Although the main economic activity in rural areas involves farming, there are also opportunities for diversification into the nonfarm economy. Agriculture is the do- minant economic activity in rural Haiti, about 78 percent of households are enga- ged in the sector, but almost a third of agricultural households also manage to diversify and perform nonfarm activities (figure 2.3). Overall, about half the house- holds in rural Haïti undertake farm activities exclusively; a quarter of households work only in the nonfarm sector; and a quarter work in a mix of activities.5’ 66 Based on remittance inflow data (balance of payments, government of Haïti, 2014) 67 The farm-only category is defined as households in which all economically active members are engaged in a farm activity. This includes households in which all members are only engaged in agri- cultural wage activities. The nonfarm-only category refers to households in which all economically active members are engaged in nonfarm activities, whether a household enterprise or nonfarm wage or salary work. The both farm and nonfarm category refers to households in which economically active members are engaged in a combination of farm and nonfarm activities. Some examples of nonfarm activity include selling shoes, soap, and packaged foods such as rice or candy. | [page 67] : Investing in People to Fight Poverty in Haïti Figure 2.3. Farm and nonfarm labor force participation, rural households [] Farm only PER î @ Nonfarm only PS 54% CT © Both farm and 72% Nonfarm Source: ECVMAS 20172; World Bank and ONPES calculations The West has the highest participation in nonfarm activities (32.4 percent). It has the highest education levels and literacy rates, which are important factors in the participation in nonfarm activities. It is also the closest to Port-au-Prince and therefore has better access to infrastructure such as electricity and safe sources of water, which are especially relevant for nonfarm activities. The vast majority of the working population in rural Haïti is involved in house- hold economic activities (90 percent) either as self-employed or unpaid family labor. This means that most individuals are working on household farms or in hou- sehold-run nonfarm enterprises to which they contribute as unpaid workers or ow- ners. Wage work is especially limited in rural Haiti; only a small share of individuals {O percent) are employed as wage workers there (figure 2.4). Figure 2.4. Labor force participation, by type of employment (] f À = @ Wage earner Dre e [2] Self-employed @ Unpaid Labor 53% Source: ECVMAS 20172; World Bank and ONPES calculations [page 68] WorldBank - ONPES | Among both agricultural and nonagricultural households, self-employment is the most common type of work (figure 2.5). Within farm households, there is an almost equal share of self-employed individuals and unpaid household Labor. However, in the nonfarm sector, the self-employed and wage earners are much more common than unpaid household Labor. Self-employment is the most com- mon type of work in all four rural regions. Figure 2.5. Employment, by farm and nonfarm participation 100 80 60 40 " . se O Farm only Nonfarm only Both farm and Nonfarm @wage earner C] Unpaid Labor C] self-employed Source: ECVMAS 20172; World Bank and ONPES calculations Among rural households, not being poor is strongly related to engaging in the nonfarm sector. More than 80 percent of farm-only households are poor. If agricul- tural households are able to diversify, they are significantly Less likely to be poor. Among diversifying households, poverty reaches 75 percent. The importance of the nonfarm sector in reducing poverty is most clearly seen among nonfarm-only hou- seholds, among which poverty incidence is below 55 percent (figure 2.6). Figure 2.6. Economic activity, by poverty Level 100 [] NonPoor 80 C2) Poor 60 40 20 À 2 À = 0 Farm only Nonfarm only Both Farm and Nonfarm Source: ECVMAS 20172; World Bank and ONPES calculations 68 Agricultural households are defined as households that have crop, livestock, or agricultural wage activity. Some of these households also perform nonfarm activities. [page 69] : Investing in People to Fight Poverty in Haïti If one holds household sociodemographic characteristics constant, a multiva- riate analysis of the correlates of rural poverty shows that (box 2.1): + _Accessing income from nonfarm activities is associated with a reduction by 10— 12 percentage points in the probability of being poor. + _ Receiving remittances from abroad is associated with a reduction by 9 percenta- ge points in the likelihood of falling into poverty. *_Inagriculture, the number of crops matters, rather than the type: every additional crop reduces poverty by 1.25 percent; there is no significant association between poverty and cultivating cash crops. *_ For every additional year of education of the household head, the probability of poverty falls by 1 percentage point. + _ The gender ofthe household headis not a predictor of poverty status in rural areas. Box 21. The correlates of poverty and food security We estimate the correlates of poverty and food security using à Logit mo- del of the following form: Pop=Bot+B1Px re + B2Pn rw DZ+0Z+QX+A+Ee (B2.1.1) P=Bo+B1Px re BP rw t+pZ+QZ+QX+À+E (B2.1.1) where Pr = if household consumption expenditure is above the natio- nal poverty line of $198 a day (nonpoor); P, = 1 if households are defi- ned as food secure based on the household dietary diversity score of the Food and Agriculture Organization of the United Nations; P,,. = 1if at least one household member is participating in a nonfarm enterprise activity; Pxw = 1 if at least one household member is participating in a nonfarm wage activity; Z is a vector of farm household characteristics; X is a vector of household characteristics; department-fixed effects are captured by À; and & is the idiosyncratic error term. We estimate the models (B2111) and (B21.2) on the whole rural sample and on the farm household subsample to discover if any correlates are more likely to affect farm households. For details on the estimated model please see appendix G. Agriculture In Haïti, agriculture is an economic activity mostly conducted to produce for home consumption, with limited market connectivity. The average household in rural areas consumes most ofits output. The ratio between the value of the output sold to the value of the output produced, a proxy measure of the connection to markets, is below 40 percent. The poor are Less well connected to markets than the nonpoor; the ratio is 37 for the poor to 43 for the nonpoor. [page 70] WorldBank - ONPES | Factors of production Agricultural households in Haiti tend to cultivate relatively small plots of land of approximately 1.3 hectares, similar to the size of the plots in Sub-Saharan African countries such as Ethiopia, Lesotho, and Malawi, where over 80 per- cent of landholdings also tend to be smaller than 1.5 hectares. Landownership rates in rural Haïti are high, nearing 90 percent. Poor and nonpoor households are equally likely to own the land they work. However, the average amount of land cultivated by nonpoor households is more than 30 percent larger than the corres- ponding amount among poor households. The size of the plots leased in or leased out by households is small relative to the size of owned plots: the average leased in plot is about O3 hectares (table 21). Most likely to increase soil fertility, many farmers practice self-fertility as evidenced by the substantial share of households that leave some land fallow. It may also be that the cost of cultivating infertile land is high relative to the expected gains, making it more practical to leave land fallow. Table 2.1. Land acquisition. Percent, unless otherwise indicated Indicator Allrural | Women | Men T-test Poor Nonpoor T-test *#* p <O.O1 ** p <O05 * p <O1 Nonpoor households enjoy better access to productive factors, including both Labor and nonlabor inputs. Given the intensity of the planting and harvest periods, households rely on hired Labor to supplement their own Labor (table 2.2). Nonpoor households are not only more likely to use both household and non- household labor; they also use higher numbers of workers relative to the poor.f° A similar trend also holds for fertilizer, seeds, and pesticides: the nonpoor are more likely to use these inputs and also spend more on them./° However, as a share of total production value, poor and nonpoor households spend equally. 69 Nonhousehold Labor cannot be separated into paid and unpaid Labor (for example, exchange Labor) because this information is not available in the survey. 70 The survey does not offer adequate information to distinguish between farmers who purchase im- proved seeds and farmers who purchase regular seeds. [page 71] : Investing in People to Fight Poverty in Haïti Table 2.2. Agricultural inputs. Percent, unless otherwise indicated Indicator All rural Women Men T-test Poor Nonpoor |T-test Labor inputs Nonlabor inputs [restes mdence | 201 | me | w | sm | we | x | 2 | ##* p <O.O1 ** p <O05 * p <O1/ *** p <O.01 ** p <O.O5 * p <O1 Types of agricultural activities Virtually all farm households grow food crops, while nearly half also grow at least one cash crop. Among households growing food crops, 84.3 percent sell part of what they grow." In addition to growing crops, 75 percent of households raise cattle and other livestock, and 304 are engaged in forestry activities (figure 27). There are no salient contrasts among the types of agricultural activities that poor and nonpoor households undertake, with the exception of cash crops. Relative to households below the poverty line, households above the poverty line are more likely to cultivate cash crops, thereby improving their income generation prospects (table 2.3). Figure 2.7. Share of households, by farm activity EN Food crop ;] Livestock R Cash crop 9 Forestry NZ) Fisheries o 20 40 60 80 100 Source: ECVMAS 2072: World Bank and ONPES calculations 7 The agricultural module does not provide information on the quantities produced, sold, or consumed, but does provide the relevant values. This limits the ability to analyze the share of production sold, consumed, or otherwise [page 72] WorldBank - ONPES | Table 2.3. Activities of agricultural households Percent Indicator Cash crop Food crop Livestock Fisheries Forestry Gender of head Dom | ms | ms | 0 | 0 | nm Poverty status Food security status a. Cash crops are defined as the sale of mangos or coffee. Crop diversification is common in rural Haiti. Poor and nonpoor households are equally likely to diversify (figure 2.8). The top three crops are maize, bana- nas, and Cassava or yams. Among cash crops, mangoes are more common: over 40 percent of households grow them, compared with about 17 percent growing coffee. On average, farm households cultivate about five crops each, and 70 per- cent of households grow at least four different crops on their plots (table 2.4). Figure 2.8. Farm crops grown % of households producing Maize Ü Bananas …—%— Cassava, Yams etc D 4 Green beans LES Mango [ Millet ÿ Coffee 1 Ÿ Rice € Vegetables Peanuts o 20 40 60 80 100 Source: ECVMAS 20172; World Bank and ONPES calculations | [page 73] : Investing in People to Fight Poverty in Haïti Table 2.4. Diversity among the crops grown Indicator Crops grown, average number Farms that grow four or more crops, % Region Gender of head Poverty status Food security status The livestock sector is characterized by small animals such as chickens and goats, and there are no remarkable differences across poor and nonpoor hou- seholds except for the use of nonlabor inputs. Poultry are the most common li- vestock raised in rural Haïti (figure 2.9). Although the poor and nonpoor do not differ in the use of Labor inputs to raise livestock, nonpoor households have better access to nonlabor inputs (for example, veterinarians) in their livestock activities (table 2.5). Figure 2.9. Percentage of households, by livestock raised Chicken nd Goats æR cattle F. d Pigs ee Equine Ca d Sheep Lil Other poultry ÿ Rabbits o 10 20 30 40 50 60 70 Source: ECVMAS 2072; World Bank and ONPES calculations Table 2.5. Livestock inputs. Percent Indicator All rural Women |Men T-test Poor |Nonpoor |T-test Labor input ienceoftbor "À 555 ui | 552 | oo | 56 | 7 | Nonlabor input, vet and other Ecdenceofnonabor | mn | @u | m7 | os" | é6 | wi |" # p <001 ** p <O05 * p <O1 | [page 74] WorldBank - ONPES | Agricultural productivity increasing agricultural productivity is still considered the key engine to reducing poverty and improving food security in developing countries (World Bank 2007). About 80 percent of rural households are engaged in the sector, and improving agricultural productivity is one of the main levers for pro-poor growth, but also to alleviate food insecurity. This highlights the importance of examining the sector to identify factors that are crucial to raising productivity (box 2.2). Box 2.2. Estimating correlates of agricultural productivity Although the data used for the analysis are cross-sectional, they help in understanding the main factors of production and the contextual charac- teristics that correlate with greater productivity in the agricultural sector. The analysis, therefore, does not claim causality, but rather aims to esta- blish robust correlations. Other studies have examined the determinants of the agricultural sector in Haïti (see Verner 2008). However, the availa- bility of new data allows us to update the information. The measure of agricultural productivity used is the value of the total harvest per hectare. In line with the literature, we include the following variables as covariates: household characteristics such as gender of head, education of head, age of head, and household size; land size; physical inputs such as fertilizer, seeds, and pesticides; Labor inputs; and other plot characteristics. We explore the correlates of agricultural productivity using a simple household ordinary least squares specification in the form: In Y=B + L+BL + @.nP +2 y.InD +QX+1+e, (B2.21) id À j J where Y'is the total value of harvest per hectare; L is the total Land size cul- tivated by the household in hectares; P: and D. represent the physical and Labor inputs, respectively, used by the househott: Xis a vector of other hou- sehold and plot characteristics; department-fixed effects are captured by À, and E is the idiosyncratic error term In addition to estimating the regression for the entire rural sample, we also estimate the model for poor and nonpoor agricultural households separately to find if there are noticeable differences in significant factors of production between these households. a. All physical input variables —log fertilizer use, log pesticide use, log seeds use—refer to input costs that have been divided by the hectare size cultivated and normalized by log transfor- mation. The household growth of a cash crop is a dummy variable for whether the household produces either mango or coffee. The assistance postearthquake variable is a dummy for whether the household received help in the form of agricultural physical inputs such as fertili- zer since the earthquake. There is no information on the amounts received in the module. The number of working-age men and women refers to household members aged 15-64. | [page 75] : Investing in People to Fight Poverty in Haïti There is an inverse relationship between farm size and agricultural productivity, a common finding in developing countries with limited access to input markets. Controlling for relevant farm and household characteristics, one finds that larger plots are Less productive than smaller plots (table 2.6). In particular, a 1 percent increase in farm size is correlated with a drop of O.6 percent in agricultural productivity. This inverse relations- hip arises because of a lack of access to credit markets, irrigation, and Labor and agricultural input markets that impedes the exploitation of larger plots with the same intensity as smaller ones./? Table 2.6. Correlates of agricultural productivity Independent variable All rural Poor Nonpoor Land size TV NT NT D TT Physical inputs NT NT ON NT D) 0% | © | vw | TT Labor inputs D) om | © | en | oo | con | vx Other agricultural/plot characteristics D A NT D A NT D TNT NT 72 ACcording to Barrett et al. (2010), the inverse relationship (IR) between farm size and productivity likely arises for one of three main reasons: (a) imperfect factor markets, (b) omitted variables, or (0) statistical issues related to the measurement of plot size. As Carletto (2013) describes, imperfect factor markets (Land, Labor, insurance) are linked to differences in the shadow price of production factors that, in turn, lead to differences in the application of inputs per unit of Land in ways that are correlated with farm size. Carletto (2013) assesses the concerns about measurement issues and finds that, with better Land measurements, the IR finding is strengthened, not weakened. This supports studies by Unal (2008), who shows that the IR exists in Turkey, where it is driven by failures in the Labor market. Masterson (2007) and Vadivelu et al. (2001) also find empirical evidence of an IR in India and Paraguay. For other examples, see Eastwood et al. (2010); Lipton (2009). | [page 76] WorldBank - ONPES | TNT TT Household characteristics: BE PE PS ES D om | om | om | D | cm | vw | vw | D | vo | ve | vw | RE NT D NT NT Household composition 7 NT © SN NT co | con | cw Other economic activities NT NT TT D 0m | om | om | D | co | vom | vw | NT D 7 Note: The dependent variable is the log ofthe total crop value per hectare. Ordinary least square point estimates with robust standard errors are shown in parentheses. The results for state-fixed effects are not shown. * p <O10 ** p <005 *** p <OO1 Greater access and use of inputs are correlated with increased productivity in both poor and nonpoor agricultural households. There is a positive correlation among physical inputs (fertilizer, pesticides, and seeds), Labor inputs (household and nonhousehold Labor), and agricultural productivity. À 10 percent rise in non- household labor use per hectare is correlated with about a 2 percent expansion in agricultural productivity. However, household Labor is more important in poor | [page 77] : Investing in People to Fight Poverty in Haïti households, where a 10 percent increase in household labor use per hectare is co- rrelated with a 2.6 percent rise in productivity. Nonpoor households can hire non- household Labor more easily. Crop diversification is correlated with greater agricultural productivity in both poor and nonpoor households. While causality cannot be implied, diversification seems beneficial as a risk management strategy. This finding may also hint at the benefits of intercropping practices. Growing cash crops (mangos and coffee) does not appear to be significantly correlated with agricultural productivity. Agricultural productivity does not vary by household demographic characteris- tics. Whether the household head is a man or a woman, young or old, Less or better educated, the crop value per hectare is not affected (if all else is equal). Population pressures and environmental degradation are additional important factors contributing to the declining productivity in agriculture. In a country that is already densely populated, steady population growth continues to put pres- sure on the natural resource base; average farm size has declined over time; and farms have become Less productive (Dilley et al. 2005). Moreover, Haïiti's exposure to frequent hurricanes and tropical storms, combined with high rates of soil erosion that have reduced soil fertility and adversely affected crop output, cause annual productivity Losses in agriculture ranging from OS to 12 percent. Extensive defores- tation in many parts of the country has worsened the erosion problem and led to the loss of enormous quantities of fertile topsoil (Dilley et al. 2005; Verner 2008). The nonfarm sector The nonfarm sector in rural Haiti is predominantly characterized by trade and commerce, and, as the more reliable source of income in rural areas, it is a main source of livelihoods among nonpoor households. About 40 percent of nonpoor households participate in the nonfarm sector (table 2.7). Nonpoor households have 50 percent more access to the nonfarm sector than the poor; the difference is sta- tistically significant. Within nonfarm households, the nonpoor participate relatively more in higher-skilled industries or sectors such as education and health care, while the poor concentrate more around low-skilled services (table 2.8). 73 The forest cover is now Less than 2 percent of the country (Library of Congress 2006) | [page 78] WorldBank - ONPES | Table 2.7. Nonfarm activity, by type of household. Percent Indicator Household enterprises Salaried/wage nonfarm Other nonfarme D EE RE RTE Gender of head Lame | me | ns | » | Poverty status RE EE EX Note: * Indicates statistically significant differences within each category. a. Other nonfarm activity includes unpaid apprenticeship and household labor. ##% p <OO1 * p <O05 * p <O1 Table 2.8. Household participation in non farm activities, by industry. Percent Indicator Industry and construction ee cr Transport ECRECONR Other services commerce and health TU AE TE NT EU EC ET Gender of head CC ES NT EXT MRC TEE Poverty status EC RE TS RTS EE EE EE Most of the nonfarm enterprises in rural Haiti operate on a small scale and are in the informal sector, mostly selling prefabricated products. Nonfarm bu- sinesses are micro in nature and have an average of 1.5 workers, including the owner (table 2.9). A limited share of businesses hire laborers, only 7 percent among nonpoor households and 5 percent among poor households. The most common reasons provided by households for starting nonfarm enterprises are to increase income and because of the unavailability of wage employment; the biggest mar- ket they serve are other households. [page 79] : Investing in People to Fight Poverty in Haïti Table 2.9. Household enterprise profile Workforce Type of business Indicator Number Hired® Household workers? Informal, % [uen }s Jrun Jin | Region CE RE EEE EE EE EEE su [5 uw [ul vw | 0 | Gender of head Poverty status a. Conditional on using hired labor. b. Conditional on having other household members working for the enterprise besides the owner. Box 23. The government strategy for rural development Rebuilding the nation's agricultural production base ranks among the top priorities of the governmenta Promoting the development of the rural nonfarm sector is also considered important because an expanding non- farm sector could absorb surplus unproductive labor as it exits from the agricultural sector, slowing rural-to-urban migration, while creating oppor- tunities to boost household income (Lewis 1954; Verner 2008). The Ministry of Agricultural Resources and Rural Development has imple- mented key agriculture policy reforms. In 2010, the government launched a short- to medium-term strategy and investment plan for 2013-16. The plan identifies four main objectives for the agricultural sector: (1) modernize the ministry to enable better governance; (2) raise agricultural productivity to improve food security and increase revenue; (3) develop agricultural value chains, with particular emphasis on increasing exports; and (4) adopt and promote ecological agriculture to preserve natural resources. Other major agricultural policy reforms have dramatically changed the way direct su- pport to farmers is handled. For the first time, subsidies for agricultural inputs are being provided through voucher schemes, which are Less distortionary than traditional subsidies applied across the board according to input prices. [page 80] WorldBank - ONPES | The use of vouchers has encouraged greater participation by the private sector in the provision of inputs, allowing for a general positive spillover effect on non-beneficiaries. Progress has also been made in strengthening the capacity of the key institutions charged with the provision of agricul- tural public goods and services, especially in animal and plant health, but also in research and development and extension services. a. The objectives of the National Plan of Agricultural Investments (2011-16) includes (1) raising the productivity and competitiveness of the agricultural sector, (2) increasing the contribution of agricultural productivity to national food availability by 25 percent, G) reducing the number of individuals experiencing food insecurity by 50 percent, (3) boosting agricultural income among at least 500,000 households, and (4) enhan- cing the resilience of the population in the face of natural hazards (Arias et al. 2013). Labor force participation in Haiti is Low compared to Latin America’ and comparable with levels in Sub-Saharan Africa. Less than two-thirds of the wor- king-age population participates in the Labor market. Labor force participation is slightly higher in urban areas than in rural areas (table 210). Table 210. Labor market indicators geographically disaggregated. Percent except where otherwise indicated » Unemploy- | Informal | Invisible unde- | Urban/ru- Participa- | Employ- Location ment rate, | employ- | remployment, | ral popula- tionrate | mentrate ; : extended ment minimum wage | tion ratio Munume | w | us | se | | we | w | Qu uote ef » [nu Regions Dom [sm | sm | um | w | w un fu fes jm w | De jeu ls) « | x. Source: ECVMAS 2012. Note: See appendix G for the definition of concepts. n.a. = not applicable. 74° When compared to the rest of the region, the participation rate is calculated for the population aged 15-64, whereas in table 210 the rate is calculated for the population over 15 years, which explains the difference in rates (60% vs 647%) [page 81] : Investing in People to Fight Poverty in Haïti Differences between urban and rural settings arise if one looks at unemploy- Compared to ment rates, which tend to be higher in urban settings.” The unemployment rate : formal workers, in urban areas is almost twice the rate in rural areas (39.6 and 22.3 percent, respecti- agricultural workers DR : earn on average vely) (see table 210). Because the levels of participation are similar, this means that 75% less, and overall employment rates are Lower in urban areas. These facts are perfectly reflec- , informal workers ted in the employment and unemployment rates of the regions, where the regions earn more than With the highest percentage of urban population, such as the Metropolitan Area and 50% less. the North, have the lowest employment rates and the highest unemployment rates, while the opposite occurs in Less urban regions, such as the South. Because of the importance of Labor income in all urban Haitian household budgets, a rate of unem- ployment of almost 40 percent in urban areas is a matter of concern. Labor market earnings are particularly Low among the vast majority of the wor- kers in both urban and rural areas. Around 60 percent of workers in urban areas earn less than the minimum wage; this goes up to 80 percent in rural areas, where most workers are employed in agriculture (see table 210). Moreover, slightly Less than 70 percent of the workers in urban areas are in the informal sector. In urban areas, poor individuals present higher average unemployment and unde- remployment rates than the nonpoor. The poor have a harder time finding a job, and, if they find a job, it is most often associated with low-quality status; thus, two-thirds of poor workers hold jobs with earnings below the minimum wage (table 211). 75 More than one definition of unemployment, underemployment, and informality is available, but, for ease of exposition and considering the definitions most well adapted to the Haitian context, this chapter presents only the results based on the definitions of extended unemployment, invisible underemploy- ment, and informal employment. See appendix H for these definitions as well as those also considered, but not presented in the main text. Results based on other definitions are available upon request 76 Informal employment is defined as all contributing family workers, all independent workers in the informal sector, and all employees without written contracts and not benefiting from social protec- tion. This definition does not include people working in the primary sector (agriculture). The informal sector is defined as all unincorporated enterprises (household businesses) that are not registered or do not keep formal accounts. This definition also does not include people working in the primary sector (agriculture). The definition of underemployment here corresponds to invisible underemploy- ment, which includes all employed individuals who earn Less than a minimum amount of money an employee should earn by law (in this case, G 250 per day = G 7,500 monthly, which was the minimum wage before October 2012). Admitting that the concept of underemployment is used repeatedly throughout this chapter (in part with the intention of international comparison), the relevant definition, that is, the proportion of people earning Less than the minimum wage, might not be the most appro- priate indicator of job quality and competitive wages in the Haitian context. Indeed, Labor earnings vary widely across industries and types of occupations, and the minimum wage is not enforced equally in all industries. For these reasons and following Herrera and Merceron (2013), who write on underem- ployment and job mismatches in Sub-Saharan Africa, the next section presents rates of people earning less than the average Labor income within industries and occupations as à proxy for job quality and competitive wages in the Labor market. [page 82] WorldBank - ONPES | Table 2.1. Labor market indicators in urban settings, by poverty level. Percent Indicator Nonpoor Poor Non-—extreme poor Extreme poor Invisible underemployment [OO sa | 6er | 555 | eo | Note: Invisible underemployment captures the proportion of people earning less than the minimum wage. An analysis of individual characteristics and Labor market outcomes shows that women, youth, and less well educated individuals are at a significant di- sadvantage. A first analysis looks into the issue of unemployment. Holding cons- tant social and demographic characteristics, one finds that women are almost 20 percentage points more likely than men to be unemployed./? Young inexperienced workers are disfavored: for every year of additional experience, the probability of unemployment is reduced by about 1.5 percentage points. Education plays a subs- tantive role, and the role is more sizable, the higher the Level of education. Complete unemployment is 7 points Less likely among those with lower-secondary education than among those without education, while, in upper-secondary education, the di- fference is 15 points. Gender and age are important correlates of the probability of earning less than the minimum wage, that is, being among the invisible underemployed. All else being equal, women are 6 percentage points more likely than men to earn less than the minimum wage (see appendix 1, table 11). This difference holds even after we control for the type of industry that women and men select themselves into. Invisible underemployment is also a more severe issue among younger wor- kers (aged 15-24): the likelihood of earning Less than the minimum wage among them is 13 percentage points higher than among workers aged 25-54./8 Education appears to be a strong mitigating factor in invisible underemployment. Higher levels of education are correlated with a decrease in the probability of earning less than the minimum wage (appendix |, table 11). The Labor market recognizes the accumulation of and investments in skills. There are also returns to experience; thus, higher levels of experience reduce the chances of invisible underemployment. Gender, age, and education are closely associated with the likelihood of in- formal employment. Women are more likely than men to have an informal job. 71 The analysis makes use of ordinary least squares and Probit regressions to study factors associated with the likelihood of unemployment. The regression results are included in appendix | 78 The government defines youth as people 15-24 years of age. | [page 83] : Investing in People to Fight Poverty in Haïti Other factors being equal, women workers are 6 percentage points more likely to be Women are 20 employed informally. Similarly, youth are more affected by informality; workers aged p.p. more likely 15-24 are 5 points more likely than workers aged 25-54 to be informally employed. than men to be But the most sizable difference is associated with education, and the difference wi- unemployed and dens, the higher the level of education. Compared with workers without education, canne less workers with lower-secondary educational attainment are 20 points Less likely to be an men: informal, while workers with upper-secondary educational attainment or more are more than 40 points Less likely to be in the informal sector (appendix I, table 11). Hourly earnings, a measure of labor market productivity, confirm that edu- cation, experience, and gender matter substantially. The Labor market rewards formal education. Even completing basic education yields almost 30 percent more earnings per hour than not attending or not completing the primary level. Moreover, returns to education increase steeply with the level of attainment. Hourly earnings among workers with lower-secondary education are almost 50 percent higher, while, among workers with upper-secondary education or more, earnings are 125 percent higher compared with workers without education (appendix I, table 11). Experience is also rewarded in the market. Five additional years of experience are associated with a 15 percent increase in earnings per hour. Women make 32 percent less per hour than men. The difference holds even after comparing workers of similar education and working in the same sector. Box 2.4 analyzes whether there are signs of discrimination affecting women in the Labor market. Box 24. Zooming in on the gender earnings gap using the Oaxaca-Blinder decomposition If earnings gaps between men and women appear in the Labor market, it is conceivable that the difference, to some extent, may be explained by diffe- rences in individual characteristics between men and women, for instance, if men are, on average, better educated than women. After one controls for such characteristics, the Labor earnings of women and men should be the same if no gender discrimination is present. However, the results presented in appendix J, table J1 show that the hourly labor income among women is around 32 percent lower than among men after one holds education, experience, and even industry of employment constant. Is this a sign of discrimination? To refine the drivers of the differences between the hourly Labor incomes of men and women in urban areas in Haiti, Oaxaca-Blinder decompositions were used (Jann 2008). The Oaxaca-Blinder decomposition therefore pro- vides additional elements to help us understand the extent to which the gender earnings gap can be accounted for by observable and unobserva- ble characteristics. For this purpose, we define three specifications. The first includes age and level of education as the individual characteristics that may explain the gen- der earnings gap. The second specification includes the same observable | [page 84] WorldBank - ONPES | characteristics as the first, plus the number of children in the household. The third specification includes those included in the second, plus dum- mies for the industry of activity. The results are summarized in figure B2.41. Figure B2.41. Oaxaca-Blinder decomposition results for different specifications, urban Haiti G) = Q@) + Industry 35.71% of Activity ; (2) = (1) + Number 167 @ © of children in the HH Du ) = Age and Level o LA of Education IAE [| | 0% 20% 40% 60% 80% 100% @Exptained @Unexplained Sources: ECVMAS 2072; World Bank calculations. Based on the third specification, observable characteristics such as age, the Level of education, the number of children in the household, and the industry of activity explain almost 36 percent of the gender earnings gap, but the other 64 percent remains unexplained. The existence of the gen- der gap unexplained by observable characteristics suggests some gender discrimination takes place in the Labor market.a The fraction of the gender wage gap unexplained by observable charac- teristics in urban Haïti is higher than in African and other Latin American countries, alerting to the urgency of addressing this particular dimension. According to Nopo (2012), the part of the gender wage gap attributable to differences between men and women that cannot be explained by ob- servable characteristics in Latin American and Caribbean countries avera- ges around 18 percent (in circa 2007). There is, however, a large variation across countries. For instance, the widest reported gap occurs in Nicara- gua, at 28 percent, and the smallest occurs in Colombia, at 7.3 percent, but none is wider than the one in urban Haïti. However, Nordman, Robilliard, and Roubaud (2013) show that, in the main cities in seven French-spea- king countries in Africa, the corresponding results ranged from 40 to 67 percent in 2001/02, which is a little closer to the urban situation in Haiti in 2072. For example, the unexplained part of the gender gap in Lomé (Togo) is around 45 percent after one controls for sector. a. À caveat of these results is that they might include some selectivity bias in the sense that the gender gap is calculated only for people who are working and who are thus selected into the Labor market. There is also a high probability of self-selection into particular industries of activity. [page 85] : Investing in People to Fight Poverty in Haïti Despite the high rates of unemployment, informality, and underemployment, urban settings are much better connected to markets and services and there- fore present undeniable opportunities for poverty reduction in Haiti. While a large share ofthe employed population continues to earn Low wages and is not protec- ted by safety nets, overall urban areas offer comparatively the best income generation prospects because of their links to domestic and international markets, their dynamic tertiary sector, and better access to services. Understanding the sectoral structure of the Labor market Although certain industries provide opportunities for better earnings, most jobs are in the low-earnings trade sector. Figure 210 shows that earnings are not only higher in industries such as education and health care, transportation, cons- truction, and other services than in trade and agriculture, but also more equally distributed. The case of trade is particularly important since it employs about 40 percent of urban workers. Trade workers receive earnings that are both lower in level, but also higher in variability. Figure 2.10. The distribution of hourly Labor income in urban areas, by industry 6 4 > 5 [+ oo Le} 2. o -2 o 2 4 6 Log of Hour Labor Income @asricutture @ Construction @ Trade @ Transportation @ Other services @ Education and/Health kernel=epanechnikov, bandwith=0.4597 Sources: ECVMAS 20172; World Bank and ONPES calculations. Note: Outliers have been eliminated from the calculation. An outlier is defined as an observation with a value larger than the median, plus three times the standard deviation. In urban areas, O.91 percent ofall observations were discarded. The sectorial structure of the Labor market has women at a disadvantage. In the trade sector, the low-earnings and high-variability sector of the Labor market, the vast majority of workers are women. About 70 percent of the trade jobs are held by women, whereas, in the better-pay sectors of education and health care, fewer than half the workers are women (table 212). [page 86] WorldBank - ONPES | Table 2.12. Gender, poverty and labor income in urban areas, by industry à , » , 4 | Womenineach | Hourly labor income, Sector Observations | Weighted observations | Workers, % Sector HTG, 2005 prices Other services 951 364,896 26.0 128 618 Sources: ECVMAS 2012; World Bank calculations. Note: Outliers have been eliminated from the calculation. An outlier is defined as an observation with a value larger than the median, plus three times the standard deviation. In urban areas, 0.91 percent ofall observations were discarded: The majority of trade workers are involved in self-employment. Figure 211 re- veals that industries with Low earnings potential and high variability, such as trade, tend to hold a larger proportion of the self-employed than other industries. Overall, almost 37 percent of all workers are self-employed. On the other hand, industries with better pay prospects, such as education and health care, transportation, and construction, are more likely to have executives, qualified and semiqualified wor- kers, and laborers. Figure 2.11. Composition of occupations in urban areas, by industry 100% _. 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2 5 8 5 £ 8 2 5 œ 5 rs 2 = © © Lo] Fa 3 5 F b I Le] E - Lo] É Fe] e] S Le] ün 2 Q [= re < C] e 8 £ (e] S ® S F 9 © TD LU © Family aide @semi-auatified worker @ patron @ Executive @setfEmployed @ Qualified Worker @iborer Sources: ECVMAS 2072; World Bank calculations. [page 87] : Investing in People to Fight Poverty in Haïti Overall, trade and self-employment are, respectively, the industry and the oc- cupation with the largest numbers and percentages of women, the poor, the least well paid, and the least well educated workers in urban Haïti. Is this nearly 40 percent of urban workers destined to be poor and remain in poverty? Or do they have a chance to mobilize out of poverty? Can public policies be implemented to improve the Labor market and economic conditions among this large share of the urban population? Self-employment: scope for improvement? In the short run, improving the labor situation of the self-employed in urban areas could significantly enhance the well-being of at least 40 percent of wor- kers. Self-employment covers a wide range of situations. While it is a relatively low-earnings occupation, there are workers among the self-employed who manage to receive earnings comparable with those in other occupations. Moreover, becau- se it is the most common form of occupation, notably among women and poorer individuals, one should try to learn some lessons about what might help improve self-employment. Looking at the positive deviation among the self-employed, that is, those wor- kers deviating from the norm and obtaining better outcomes than the rest, one finds that a small investment in skills can have big payoffs. Within self-employ- ment, table 213 compares those people receiving more than the average in hourly earnings—around one-fourth of the self-employed—with those people earning Less than the average. Perhaps the most salient result is that, with an average of only 1.3 more years of education, the self-employed who earn more than the average ofthe occupation have an hourly Labor income of HTG 105, while those earning Less than the average obtain only around G 12 per hour. Table 213. Differences between the self-employed who earn more or less than the average wage of the occupation, urban areas Indicator More than average Less than average Difference | Significance Sources: ECVMAS 2012. World Bank calculations. ** p <O.01 ** p <005 * p <O | [page 88] WorldBank - ONPES | An encouraging lesson is that the income prospects of the self-employed could be substantially enhanced by relatively modest improvements in skills. Among the self-employed earning more than the average, two-thirds have attai- ned a level of education equal to or higher than primary school, while 50 percent of the self-employed earning less than the average are in this group (figure 212). Passing from five to six years of education (thus completing the primary Level is associated with an increase in salary of almost G 95 per hour. This striking result implies that a little investment in years of education or remedial training for skills acquisition among people who have been out ofthe school system for a long time could substantially increase the Labor income of the urban poor. Figure 2.12. Education among the self-employed earning Less or more than the average hourly Labor income, urban areas Less than average @ @ More than Average { { 0% 20% 40% 60% 80% 100% C] None @ Primary non-completed @ Primary completed & Sec1 non-compl_ @3 Sec1 completed & Sec.2 non-com. Sec2 completed & university Sources: ECVMAS 2072; World Bank and ONPES calculations. International migration is an important complement to household income in Haiti and, despite the relatively limited share of the population migrating, the returns to migration are significant. For political and economic reasons, large numbers of Haitians emigrated throughout the 20th century, mostly to Canada, the Dominican Republic, France, and the United States (see Jadotte 2008; Oroz- co 2006). As of 2010, over a million Haitians (10 percent of the population) were estimated to be living abroad, half in the United States.”° An important economic dimension associated with migration is remittances, which, in Haiti, account for almost 20.0 percent of GDP. Among all the countries on which data are available for 2012, only El Salvador and Guyana (16.4 percent), Honduras (15.7 percent), and Jamaica (14.5 percent) registered remittances as a share of GDP larger than 10 percent. The fact that several of these countries have larger diasporas relative to their population suggests that the economic ties of Haïiti’s migrants are rather 79 See ‘Bilateral Migration Matrix 2010 Bilateral Migration and Remittances (database), World Bank, Washington, DC, http://g0:worldbank.org/JITCINYTTO. [page 89] : Investing in People to Fight Poverty in Haïti strong. and that migrants have a higher income potential (their income as migrants is disproportionately higher than what they could have earned in Haïti). Internal migration and transfers are also relevant, particularly among the rural population. The decision of a household to send one ofits members abroad can be seen as an investment: families incur upfront costs (airplane tickets, visa fees, and so on) to reap future income gains from better labor opportunities.89 If the initial costs are too high for poorer households, however, moving within the country can be a second-best option.f! In Haiti, over one-fifth of the population was not born in the department of residence, and the majority of the internal migrants are now lLi- ving in the Ouest Department (65 percent). In 2012, more than half the population in the Metropolitan Area had migrated from other departments (ECVH 2001; ECVMAS 2012). According to the available data, the share of internal migrants has marginally risen, from 20.4 percent in 2001 to 23.9 percent in 2012, probably attracted by the new opportunities generated in and around Port-au-Prince during the postearth- quake reconstruction or seeking escape from the continued deterioration in agri- cultural productivity. Migrants are generally better off, and migration overseas bears significantly different results from migration to the Dominican Republic or within Haïti. Mi- grants are generally better educated than nonmigrants. However, while migrants to the countries of the Organisation for Economic Co-operation and Development are far more likely to have secondary and tertiary education, domestic migrants are relatively Less well educated. In all, the Dominican Republic option for international migration is more similar to domestic migration than to migration overseas. Even if less well-off relative to international migrants, internal migrants are, on average, better off than nonmigrants in terms of education, quality of employment (they are more likely to be wage earners and to be formal workers), and welfare in general. Relative to men migrants, women migrants are Less well educated and more likely to be self-employed and to work in the informal sector: these differen- ces are even more pronounced than among nonmigrants. Compared with men migrants, women migrating to the Metropolitan Area are significantly Less well edu- cated and more likely to be unemployed (60 percent, against 41 percent among men), inactive, or to be working in the informal sector. These characteristics are even more pronounced among migrants than among man and women in general. Despite the difficulties encountered by women in the Labor market, migrating wo- men are generally better off than their nonmigrant peers. Migrant women are also more likely to be single or separated. Because the returns on both domestic and international migration are high, labor income in Haiti is supplemented significantly by private transfers. 80 For details on this approach to migration and remittances, see Clemens and Ogden (2013). Clemens (207) estimates that unskilled Haitian farmers migrating to the United States could increase their annual incomes by a factor of 20 81 Clemens (2014) gathers evidence that migration tends to increase with income until a certain thresh- old, suggesting that poorer households would like to migrate, but do not have the means to do so | [page 90] WorldBank - ONPES | An approximate cost-benefit comparison indicates that, on average, migration is profitable. A household with a migrant has forgone earnings of about G 5,000 be- cause the migrant is no longer working at the origin, but, in exchange, the migrant can expect to raise G 16,000 at destination (G 4,000 of which are sent back home in transfers). Although these numbers may Look similar, both the migrant and the migrant's household at origin are better off because the migrant receives greater labor income, and the household shares resources among fewer people and ob- tains the transfer besides. When controlling for individual and households charac- teristics, educated migrants earn on average between 20 and 30% more than their peer in rural areas. In rural areas, half of all income derives from labor, a quarter from production for home consumption, and 13 percent from private transfers. In urban areas, private transfers account for about 20 percent of household income, while Labor represents two-thirds.52 Monetary transfers, especially remittances, are predominantly an urban phe- nomenon and contribute more to income, while nonmonetary transfers are more widespread, but represent less value. For the country as a whole, over 35 percent of urban households receive remittances, while only 20 percent of households in rural areas do so. Domestic monetary transfers are more equitably distributed (26.7 percent in urban areas versus 26.4 percent in rural areas), while nonmonetary transfers are slightly higher in rural areas (521 percent versus 501 percent in urban areas). Monetary transfers are often larger in value, meaning that their contribution to total income (an average of 24.5 percent in the country) is larger than that of in-kind gifts (12.2 percent). Households headed by unemployed or inactive individuals or by women are much more likely to receive private transfers. Remittances from migrant re- latives can be an important source of insurance against Labor market and other shocks. Conditional on a set of observable characteristics, a household with a head who is unemployed or inactive is 10 percent more likely to receive remit- tances and 11-18 percent more likely to receive domestic private transfers. Wo- man-headed households are also 8-9 percent more likely to receive private monetary transfers. While both the poor and the nonpoor have equal access to transfers origi- nating in Haiti, the nonpoor have more than twice as much access to foreign remittances. A little more than a quarter of both poor and nonpoor households receive remittances originating in the country. However, more than a third of nonpoor households receive remittances from abroad, while fewer than a third of the poor have access to such remittances. Nonpoor household remittances are also more likely to be regular, and to become available more than once à year. Not only are remittances more frequent, but they are also larger among the nonpoor, more than double the average received by poor households (table 214). 82 The remainder of the income is derived from imputed rents, which reach about 13 percent of total household income (see chapter 1) | [page 91] : Investing in People to Fight Poverty in Haïti Table 2.14. Remittances and other income percent unless otherwise indicated Indicator All rural Poor Nonpoor T-test Transfers Private transfers, Local and foreign are regular 66386" Other income sources # p <001 ** p <O05 * p <O1 Transfers are most commonly used for food and then to pay for education (ta- bles 215 and 216). Among both poor and nonpoor households, the main allocation of transfers is for food. Among almost two-thirds of the recipients, private transfers support food purchases. While the share is greater among poor households, remit- tances among the nonpoor aid in covering food expenses in more than 60 percent of cases. There are no substantial differences between the poor and the nonpoor in the shares of transfers allocated to education expenses. Table 2.15. Uses of transfers in rural areas Percent Use All rural Poor Nonpoor T-test Construction or repairs to housing Family events (deaths, weddings, and so on) Economic activity (buying tools, raw materials, and so on) # p <O01 ** p <005 Table 2.16. Uses of transfers in urban areas Percent Use All urban Poor Nonpoor Construction or repairs to housing Family events (deaths, weddings, and so on) Economic activity (buying tools, raw materials, and so on) [page 92] WorldBank - ONPES | Private transfers reduce poverty and inequality. Because over 60 percent of poor and extreme poor households rely on some sort of transfer, private transfers have a sizable effect on poverty headcounts.55 Without transfers, extreme poverty would increase from 23.8 percent to 28.9 percent, whereas moderate poverty would rise to 63.0 percent from 58.5 percent. Poor households have less access to re- mittances, and, thus, by excluding these, international transfers would raise extre- me poverty to 255 percent and moderate poverty to 607 percent5* In line with evidence on the region, without remittances, the Gini coefficient measuring inco- me inequality would rise to 0.614, and it would rise to O.618 if all private transfers were excluded.5° Treating migration and remittances as another income generation strategy could contribute to a more productive debate and improve income oppor- tunities among households. Because money is fungible, it makes more sense to focus on how to expand the opportunities for income generation, rather on what households can do with their remittances. Thus, regardless of the source of remit- tances, the more productive focus is on how to improve the capacity of households to invest their scarce resources. At the same time, this helps clarify the reverse ques- tion: how can households obtain more resources. Analysts point out that temporary migration agreements offer a smart opportunity. Box 2.5 expands on the issue. Box 2.5. Remittances as a return on investment It is difficult to overstate the importance of migration and remittances to the income of the poor in developing countries. Remittances going to the developing world totaled $401 billion in 2012 and are projected to reach $515 billion by 2015. Likewise, a 20 percent increase in the stock of remittance-sending migrants would lead to an increase of $20 billion in the new resources flowing to developing countries, more than the entire G-7 gave in bilateral aid in 2011. However, there are differing perspectives on how migration should be addressed in development economics and policy. Clemens and Ogden (2013) argue that, rather than windfall inco- me, such as lottery winnings, development economics should focus on migration and remittances as part of a productive investment portfolio 83 Official poverty rates are based on consumption, not income. The exercise above consists in subtracting transfers from total consumption and recalculating poverty rates, thus relying on the assumption that households consume all of the income they receive, but only that income and no Savings. 84 Acosta et al. (2006) use the ECVH 2001 to estimate the effect of remittances on poverty. Using an in- come-based welfare measure and the international poverty lines of $1 and $2 a day for extreme and moderate poverty (at the time), respectively, they find that excluding remittances increased extreme poverty from 53 to 60 percent and moderate poverty from 71 to 76 percent. 85 Acosta et al. (2006) show that, for most countries surveyed in the region, nonremittance income is more unequally distributed than total income. Using the ECVH 2001, they estimate that the Gini coefficient would increase from 0.669 to 0.670, the smallest increase in their sample (apart from Nicaragua and Peru, where inequality actually decreases). Our result of a 12 percent rise in the Gini is in Line with what is recorded in countries such as the Dominican Republic (2004), Ecuador (2004), Guatemala (2000), and Paraguay (2003). | [page 93] : Investing in People to Fight Poverty in Haïti for poor families. Moving to a city or à foreign country is one of the few investments households can make that has a potential return in the hun- dreds of percent and that can boost income far more than less-productive economic activities that would be available were migrants to stay home. Ifmigration is treated as a return on investment, more productive questions can then be asked, and more productive public policy be pursued. In ge- neral, there is not a significant difference between the investment by poor families of their income from remittances and their income in general, in- dicating that they tend to view the remittances (and migration) as another part of their investment portfolio, rather than as exogenous income. Thus, instead of examining the barriers to the investment of income from remit- tances, policy makers might reformulate the question and address the sig- nificant barriers to investment in migration, potentially the most profitable part of a household financial portfolio. In the context of Haïti, a massive barrier to migration is the Lack of U.S. tem- porary worker visas among Haitians. Clemens (2011) calculates that, were such visas available, each worker admitted through the program would, on average, raise their average income by $10,000 a year. Of this amount, 30-40 percent would be sent back to Haiti, and the multiplier effects of investment would mean that each dollar sent back would expand the Hai- tian economy by $3 or more. Currently, there is virtually no legal path for Haitians to enter the United States for employment, thus representing a significant barrier to such investment. Even unskilled agricultural work is à massively profitable return on investment for Haïitian households, but ac- cess to the labor market in the United States is generally unavailable for those households that do not already have family in the United States or that cannot claim asylum. Haitis population is equally split: half lives in rural areas, and half Lives in urban areas. À sustainable reduction of poverty and inequality needs to be built on stren- gthening the capacity of rural and urban populations to generate income in a re- liable form. In this regard, priority zero in terms of the implications for policy to boost income generation is to reach a path of consistent economic growth. While important, this is common knowledge without examining a living conditions survey. This chapter shows that, given the macroeconomic situation, certain microecono- mic determinants are critical in fostering inclusive income generation able to propel poverty reduction. Four priorities can thus be distilled for the attention of policy makers, as follows: 86 The size of remittance multiplier effects is still Little understood in the research literature and this topic would deserve further study. | [page 94] WorldBank - ONPES | Priority 1: Boost agricultural productivity. Because 75 percent of the rural popu- lation is living in poverty and because the vast majority relies heavily on agricultu- 5 focus an re, it is imperative find ways to raise productivity in the agricultural sector. A LUE and the self- a. Access to basic inputs (fertilizer, pesticides, seeds, knowledge) is at the top of the Este ne be F n list. The evidence presented in this chapter shows that households in Haïti are res- ne in tricted in their access to productive inputs and that this is a particular constraint on Haiti going forward poorer households. Past experience suggests that distribution systems inefficien- cies are among the major constraints to inputs availability. Addressing potential market failures in the provision of these inputs, for instance by engaging more with the private sector, represent a key first step to a more reliable and food-secure farming sector. Increasing the knowledge of farmers through trainings adapted to their context is also critical. b. Improuing the links to output markets is crucial. Because Less than 40 percent of total production now goes to the market, a next phase in agricultural development after the consolidation of production through quality and reliability enhancements is to integrate the sector with markets, improve value chains, explore export oppor- tunities, and exploit geographical location advantages. As Haitis food system transitions from its current subsistence orientation to become more market-orien- ted, food quality and safety will become increasingly important, as well as infras- tructure investments, mostly in roads, to facilitate access to markets and decrea- se losses during transport. This report can motivate future research that uses the agricultural census to inject more granularity and depth into responses to issues of productivity, inputs, and market integration.2? c. Promoting diversification of agricultural production into cash crops can contribute to raise incomes and food security. This chapter shows that, relative to poor hou- seholds, non-poor, food secure households are more likely to cultivate cash crops. Given the benefits of diversification, households that rely on agriculture as a major livelihood source should be encouraged to diversify out of food crops. d. Fostering the sustainable use of natural resources is essential. Over the longer term, the welfare of rural households in Haïti will be linked to the quality ofthe natural resource base on which agriculture depends. Stark population pressure, combined with the unregulated exploitation of natural resources and unsus- tainable farming practices, has exacted a heavy toll, leaving vast areas of the country with little or no forest cover, heavily eroded landscapes, and severely depleted soils. Great effort is needed to repair the damage of decades of mis- management by reversing land degradation, restoring soil fertility, reestabli- shing the vegetative cover, and conserving and protecting increasingly scarce water resources. The obvious place to begin would be through the promotion of more environmentally friendly agricultural production practices, combined with regulations (and enforcement) to control the exploitation of common-pool re- sources, especially trees. 87 More detailed and concrete policy actions are also suggested in "Rural Development in Haiti: Chal- lenges and Opportunities” (2014), background paper, Haiti Poverty Assessment, World Bank, Washing- ton, DC. | [page 95] : Investing in People to Fight Poverty in Haïti e. Priority 2: Facilitate the off-farm jobs option for rural workers. The availability of nonfarm income sources has made à difference among rural households. Such jobs can be related to agriculture on the upstream (input suppliers) or on the downstream (value-adding and processing) or be separate to the sector (such as small retail). Productive investments, training and other actions to promote Labor and physical mobility and diversify rural incomes are needed. Priority 3: Invest in skills. In urban areas, Labor markets, even in the constrained environment of Haïti, reward skills and education significantly. Workers with greater educational attainment can attain substantially better results than others. Among new cohorts of students, there is gender parity. Among older cohorts, however, wo- men are at a stark disadvantage. a. Ensuring coverage and enhancing the quality of education are key (see chapter 3). Education is a key asset for better performance in the Labor market. Disseminate an entrepreneurial culture among young people could also help them navigate a difficult job market. b. Consider improvements in technical and vocational training. For the adult population, the avenues for the accumulation of human capital pass through job training (rather than going back to school). The supply of job training centers has increased exponen- tially in recent years. The Institut National de Formation Professionnelle can play a role in more effectively regulating and monitoring the surge in informal and uncertified trai- ning options. Moreover, better coordination with the private sector on the type of skills that are in short supply for current and future demand will aid in job creation. c. Harness international migration. While the more highly skilled are more likely to mi- grate overseas, the investment in their skills is not Lost because they remit transfers that play an important role in the capacity of households to stay out of poverty. A be- tter Local business environment will enable remittances to be turned into productive uses (which leads to the next priority). Priority 4: Invest in basic infrastructure and work toward a more enabling busi- ness environment. For both employers and the self-employed, having better ac- cess to basic inputs, such as electricity, is important in promoting growth, elevating productivity, and creating jobs. While one- or two-person businesses in trade and commerce are typical in Haitis market, a share of firms are currently providing wage employment to à minority segment of the workforce, thereby helping wage workers achieve the better Labor outcomes to which many other workers aspire. Self-em- ployment is an entry point to the Labor market used mainly by youth and women, two groups facing relatively higher barriers to wage-paying jobs. A large share of the self-employed are self-employed by necessity rather than because of entre- preneurial ability. There is scope for boosting the performance of both employers and the self-employed by undertaking complementary investments in basic infras- tructure, for example, electricity, and removing the constraints on access to inputs, including credit and skills. Future research could explore the extent to which the self-employed are able to grow into small businesses or exit self-employment by obtaining wage jobs in larger firms. Analysis of this dynamic could help in reaching an understanding of the process of job generation within the context and given the constraints in Haïti. | [page 96] WorldBank - ONPES | Chapter 3: Challenges to human capital accumulation Health and education outcomes and service utilization have improved in Haïti®8. However, they have been relatively inadequate, especially among the poor. There are clear signs of the intergenerational transmission of poverty in Haiti, a trend that could be broken through improvements in educational attainment. Indeed, education positively influences health outcomes and is a strong determinant of labor earnings. It should therefore be prioritized in the effort to reduce chronic poverty and vulnerability. Cutting the costs and raising service supply in education and health care will be key to enhancing service utilization and outcomes, parti- cularly in rural areas. More sustainable sources of financing are needed to avoid overburdening households, particularly in education and health care. The diagnostic provided in chapter 1 highlights that human capital accumu- lation in Haiti is key to improving well-being in monetary and nonmonetary terms, but still presents important challenges that need to be addressed to reduce poverty. Low educations levels, food insecurity, and poor access to basic services are associated with chronic poverty in Haiti, particularly in rural areas. This chapter aims to provide an in-depth description of the current state of human ca- pital accumulation in Haiti and how it has evolved in terms of access to and, to the extent the data allow, the quality of health care and education services. Education and health care are critical to building Labor productivity and ad- vancing the welfare of individuals. On average, an additional year of education generates a 10 percent increase in earnings, and this effect tends to be stronger in developing countries.5°? In addition to boosting the earnings of individuals, educa- tion can contribute to economic development. One of the most robust correlates of GDP growth across countries is average scores on international standardized tests taken by secondary-school students (Hanushek and Woessmann 2009). Li- kewise, an improvement in life expectancy and child health can create tremen- dous returns in economic development and poverty reduction. Undernutrition, 88 This chapter is based on Adelman et. al (2014) and Cross et al (2014), two background papers of the study by the World Bank and Observatoire National de la Pauvreté et de l'Exclusion Sociale (ONPES). 2014. Investing in People to Fight Poverty in Haïti, Reflections for Evidence-based Policy Making, Washington, DC: World Bank. 89 Barro and Lee (2012) and Montenegro and Patrinos (2012) highlight the correlations. However, several studies have estimated the causal effect of greater educational attainment on earnings and find effects that are of the same order of magnitude as the correlations (see Card 1999; Duflo 2001; Psacharopoulos and Patrinos 2010). 90 Each 10 percent improvement in Life expectancy at birth is associated with a rise in economic growth of at least O3 to O4 percentage points per year, holding other growth factors constant {Sachs 2001). Another study using a panel of countries observed from 1960 to 1990 found that a one-year improvement in à population's life expectancy contributed to an increase by 4 percent in EN [page 97] : Investing in People to Fight Poverty in Haïti which affects mainly poor households, also pushes up the incidence and severity of disease and is an associated factor in over 50 percent of all child mortality (OECD and WHO 2003). Sickness and disease generate economic losses estimated at be- tween 17.4 and 35.0 percent of GDP.° In Haiti, the better educated are better off (figure 31). Among households in which the heads have never attended school, 78 percent are living in poverty; this is L.5 times the poverty rate among households in which the heads have completed upper-secondary school or above (see chapter 1.22? In urban areas, Labor income is, on average, 28 percent higher among individuals who have completed primary edu- cation than among uneducated individuals. Adults who have completed primary school are about 30 percent more likely than adults with no education to be living outside their department of birth, thereby accessing better economic opportuni- ties:* Among all internal migrants, 65 percent moved to the department of Ouest, the center of economic activity and education in the country. Haitians who migrated Improved access to to the United States are substantially better educated than those remaining in Haïti. and quality of basic Education is linked with Lower fertility rates: the fertility rate is high in Haiti, at 3.2 services, can have children per woman, against 21 in the region; however, better educated adults are a huge impact, not more likely to be married and to have fewer children. Among household heads, wo- only the current generation, but the men who have completed at least upper-secondary school have, on average, half , i j ing 24 Ferconelmeltell as many children as women with no formal schooling. economic output after controlling for other structural factors and other human capital factors, such as education and work experience (Bloom 2003). Sachs (2001) also reports that poor countries with an infant mortality rate between 50 and 100 deaths per 1,000 Live births enjoyed an annual growth of 37 percent a year, whereas similar poor countries with rates above 150 showed average growth of only O1 percent a year. Such results are confirmed by the “Global Health 2035" report, which finds that a decrease in mortality accounted for about 11 percent of recent economic growth in Low-income and middle-income countries as measured by national income (Jamison et al. 2013). However, although various studies have shown an association between health and economic development (see Barro 1996; Bhargava 2001; Bloom 2003, Bloom and Sachs 1998), the positive effect of health on economic growth is not yet conclusive. After controlling for exogenous factors such as new chemicals and drugs, international health campaigns, and main diseases, Acemoglu (2007) demonstrates that there is no evidence that an increase in life expectancy leads to more rapid growth in income per capita 91 The economic losses associated with disease are calculated by converting disease-induced Loss- es into dollar terms. Using disability-adjusted life years, economists have estimated that the loss in income because of malaria in Sub-Saharan Africa represented 174 percent of gross national product in 1999, while the economic loss because of AIDS was 351 percent of gross national product. 92 Throughout this chapter, primary is used to refer to the first two basic cycles of the Haitian education system, lower-secondary school is the third basic cycle; and upper-secondary school is the equivalent of secondary school. 93 individuals or their families may migrate to take advantage of educational or economic opportunities. Thus, education may be a cause as well as a consequence of migration 94 Global evidence shows that women's education and fertility are significantly negatively correlated after one has controlled for other relevant factors, such as wealth and urban status, suggesting that educa- tion is a cause of lower fertility (Bongaarts 2003) [page 98] WorldBank - ONPES | Figure 31. Welfare and educational Level in Haïti, 2012 a. Household poverty status by educational attainment of head, % 100% 90% 80% 70% 60% 10% @ Not Poor 30% 20% @rcc 10% 0% No schooling Incomplete Complete Complete lower Complete upper primary primary secondary secondary and above b. Adults (15+) living outside the department of birth, by educational attainment, % 35% 301% 30% 27% 25% 2% 23.2% 20% 19.3% 15% 10% 5% 0% (a F4 % © © M Om & ee À No schooling Incomplete Complete Complete Complete upper primary primary lower secondary and secondary above c. O- to 18-year-olds, by educational attainment of woman household head, number Complete upper secondary and above Ê | 0.77 Complete Lower secondary LA à 1.26 : | 2 oschooine | ss DS 7 [e] 1 2 3 Source: ECVMAS 2072; World Bank and ONPES calculations. Note: The incomplete primary category includes individuals who attended preschool. [page 99] : Investing in People to Fight Poverty in Haïti Health outcomes in Haiti are below the regional average. In Haïti, life expectancy at birth is 62 years, which is aligned with other low-income countries, but well below both the regional and the world average.” The adult mortality rate is also high relative to the rest of Latin America, particularly among women (227 in Haiti against 89 in Latin America) (table 31). Food insecurity is substantial, particularly in rural areas (34 percent), among the poor, and among households with children, which explains the high rates of malnutrition and stunting among children. This may impair the development of chil- dren and may help perpetuate poverty. Sickness-related shocks have been identified among the most common and most severe shocks in economic terms (see below). Table 3.1. Basic health indicators Indicator Gender Haïti Regional average | Global average Total fertility rate per woman BR ES RE EE Life expectancy at birth, years Life expectancy at age 60, years Health life expectancy at birth, years Under-5 mortality rate per 1,000 Live births Adult mortality rate Euomn | a | w | w | MMR per 100,000 live births | mo | es | x | HIV prevalence per 100,000 population D Jus | 5 | sn | Tuberculosis prevalence per 100,000 population lo | 26 | so | ww | a. The adult mortality rate represents the probability of dying between 15 and 60 years of age per 1000 population This chapter examines human capital accumulation and related trends in Haiti in terms of access to health care and education services. The analysis builds on the poverty pro- files presented in chapter 1 and, therefore, on the data from the recent postearthquake living conditions survey (ECVMAS 2012). Additional data sources include the govern- ment and the DHS series, which allow more meaningful comparisons over time. Haiti has made substantial progress in expanding educational attainment over the last two decades. Younger Haitians have more education, on average, relative to older Haitians, suggesting that attainment has been increasing (box 3.1). However, the relationship between age and educational attainment may derive in part from selective international migration (see above). Figure 3.2 shows that, among young adults aged 15-19, educational attainment and literacy have been improving stea- dily. In 1994, 13-14 percent of men and women had never attended school; by 2072, the share had dropped to 3 percent. Among these same cohorts, a growing portion are reaching lower-secondary school or above. 95 See ‘Haiti: Country Profiles, GHO (Global Health Observatory) (database), World Health Organization, Geneva, http.//wwwwhoint/gho/en/. EN [page 100] WorldBank - ONPES | Box 31. The intergenerational persistence of education: educational gap analysis How persistent is educational attainment across generations? The answer is important to understanding the extent to which education offers all Haitian children an opportunity to build their human capital and improve their welfare. While data are not available on the educational attainment of adults and their parents, ECVMAS includes data on adult educational attainment and the current grades attended by their children (age 10 and older). This allows one to calculate the educational gap, that is, the diffe- rence between a child’s potential and actual educational attainment. For example, because formal schooling Legally begins at age 6, the potential educational attainment of a 10-year-old would be four years. If that child had never attended school, the gap would equal 4, and, if the child was in grade 2, the gap would equal 22 Results show that the average educational gap among children aged 10-14 is largest among children in the poorest households, at over 2.5 years, reflecting the lower rates of school enrollment and higher rates of overage for grade among this population segment (figure B311). The ave- rage across quintiles, at nearly 2.0 years, is substantially above the gaps elsewhere in the region, where the average gap among 15-year-olds was about 15 years in 2009 (Ferreira et al. 2013). Figure B3.1.1. Educational gap among children 10-14 by per capita consumption quintile Average Q5 Q4 Q3 Q2 Q1 o 05 1 15 2 25 3 Figure B3.1.2 shows that parental education has a significant effect: even wi- thin consumption quintiles, children (10- to 14-year-olds) of better educa- ted parents show lower educational gaps. À one standard deviation increa- se in parental attainment is associated with a decrease in the educational gap of O.84 years in the poorest quintile and O.68 years in the top quintile. On average across quintiles, the effect is O.86 years, substantially higher than the regional averages, which are O3 years among 10-year-olds and O.6 years among 15-year-olds (Ferreira et al. 2013). This suggests that the persistence of educational attainment is particularly strong in Haïti. EN [page 101] : Investing in People to Fight Poverty in Haïti Figure B3.1.2. Average reduction in education gap given a standard-deviation increase in parent educational Level, by per capita consumption quintile _ [e] F] 28 S 8 -02 RS 22 04 LS 85 Se œ æ “O6 Se $e cg -08 pe: £ © 250 DE 1 5 È 8 L @c O0 Os O2 @ Qc Note: Each bar represents the average reduction in the educational gap associated with a one standard deviation increase in parental educational attainment by household per capita consumption quintile. Other covariates included in the regression are the child's gender, department-fixed effects, and an indicator for living in an urban area. Only children living in a household where the head is one oftheir parents are included in this analysis. a. This measure approximates years of educational attainment by grade because the actual years spent in school are not known. Therefore a year repeating a grade is treated as O years ofeducational attainment a. This measure approximates years of educational attainment by grade because the actual years spent in school are not known. Therefore, a year repeating a grade is treated as O years ofeducational attainment. Figure 3.2. Educational Level of adults and young adults a. Adult educational attainment 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 15 18 21 24 27 3O 33 36 39 42 45 48 51 54 57 60 63 Age (years) @ Complete upper secondary and above @ Complete lower secondary @ Complete primary @ Incomplete primary e No schooling [page 102] WorldBank - ONPES | b. Attainment among 15- to 19-year-olds 60% 50% 40% © 30% @ 2000 20% © 2005 2012 10% © 0% Women Men Sources: Chart 3: ECVMAS 2012; World Bank and ONPES calculations, chart b: DHS final reports. Despite this progress, educational attainment among adults is still only rela- tively modest, which affects earning capacity. Compared with its Latin American and Caribbean neighbors, Haïti has the highest share of adults with no education. Literacy rates in all departments, including Ouest, are Lower than the regional ave- rage and, in several departments, are close to the global average among Low-inco- me countries (map 31). Nationally, the adult literacy rate is about 77 percent, mid- way between the average in low-income countries and the average in the region. While the average number of years of education among young men and women is the same, adult men (24-64) have an average of almost two years more education than adult women. In the region, the trend is opposite: adult women are, on ave- rage, better educated than men. One of the determinants of unemployment and underemployment is the Low level of education, particularly in urban areas, where unemployment is associated with poverty and vulnerability. The completion of primary education by an adult living in an urban area results in an increase of 25 percent in Labor earnings. Therefore, investing in adult education, including basic literacy and numeracy skills as well as the technical skills demanded in the Labor market, seems to be key to reducing poverty in Haiti. [page 103] : Investing in People to Fight Poverty in Haïti Map 3.1. Literacy rate in Haïti, 2012 Adult (15+) literacy rate by department D 80-85 75-80 . 70-75 : 65-70 60-65 % 55-60 eo 915%: LAC average [== F on 612%: Low-income country average è 715%: Haiti average = Sources: ECVMAS 20172; World Bank and ONPES calculations; WDI (World Development Indicators) (database) Worid Bank, Washington, DC fttp://dataworldbankorg/data-catalog/world-development- indicators. Note: Data on literacy in countries other than Haïti are taken from various sources. Methodological differences may therefore affect the comparisons between Haiti and other countries. Youth face additional challenges in the Labor market, despite their higher litera- cy levels, which suggests that higher-quality education and some professional training might help. Although the level of educational attainment is higher among better educated young adults relative to older cohorts, it still lags behind the corres- ponding levels in the rest of the region (see map 311). This penalizes young adults on the Labor market, particularly in urban areas (see chapter 2). Indeed, people between the ages of 15 and 24 in urban areas have not only the lowest rates of Labor partici- pation and employment, but also the highest rates of unemployment and informal employment. This suggests that the average quality of education they have received is Low (see below). Youth who have completed primary school may therefore still lack basic skills, in addition to needing more job-relevant training. À wide range of market failures likely contributes to the situation (failures in the Labor market, the credit mar- ket, and the market for education, including information scarcity). Investing in youth training alone may thus be insufficient to improve youth employability. School participation among children and progress in school and in learning Despite advances in recent decades, about 10 percent of 6- to 14-year-olds are not in school. Figure 3.3 shows that the majority of preschool-age children and 90 percent of children of official primary-school age (6—11) are in school. (Box 32 offers a picture of the structure of the education system.) This represents progress: in 2001, participation rates among the same age cohort were around 78 percent. Within the Latin America and Caribbean region, however, enrollment rates are below 95 percent among this age-group only in Nicaragua (88 percent), Guatemala (92 percent), and Hon- duras (94 percent). School enrollment begins to drop off around age 15 in Haïti, but 73 96 Data in SEDLAC (Socio-Economic Database for Latin America and the Caribbean), Center for Distrib- utive, Labor, and Social Studies, Facultad de Ciencias Econémicas, Universidad Nacional de La Plata, La Plata, Argentina; Equity Lab, Team for Statistical Development, World Bank, Washington, DC, http:// sedlac.econo.unlp.edu.ar/eng/statistics php. [page 104] WorldBank - ONPES | percent of 18-year-olds report they are still in school (ECVMAS 2072). The numbers indicate that about 200,000 children aged 6-14 are currently out of school?’ Figure 3.3. School enrollment for children in Haiti, 2012 100% 80% 60% 40% 20% 0% 3 45 6 7 8 9 10111213141516171819202122232425 Age (years) Source: ECVMAS 20172; World Bank and ONPES calculations. Note: Children in preschool are considered enrolled in school. Enrollment is based on answers to the survey question asking if children are currently in school, rather than on administrative enroliment records. Box 32. The education system in Haïti Formal education in Haïti is structured across four levels: preschool, basic education, secondary school, and higher education (figure B3.21). Preschool is meant to serve children from age 2 to 5, and is considered to have four levels based on these ages: poupons, petits, moyens, and grands. However, this structure is not formally mandated by public policy. The first two cy- cles—grades 1-6 for children aged 6—-11—are considered primary education. Thereafter, children may enter into vocational programs or continue to the third basic cycle (lower-secondary school), which consists of three grades for children aged 12-14. Similarly, vocational programs are available after lower secondary, or children may continue on to secondary (upper-secon- dary) education, which consists of three or four grades depending on the model followed by the school. Higher (tertiary) education includes a range of university, technical, and vocational programs. Figure B3.2.1. The formal education system DEEE mandated by Primary Secondary Government CE CE OS Sources: Data ofthe Ministry of Education and Vocational Training; World Bank estimates. 97 Estimates based on ECVMAS (2072), enrollment rates, and population projections in IHSI (2007) [page 105] : Investing in People to Fight Poverty in Haïti Most children are overage for grade because of a late start and slow progres- sion. Figure 34 shows attendance rates in primary, secondary, and tertiary educa- tion. In 2001, the national net enrollment ratio for primary school stood at about 60 percent and, by 2072, had risen to 72 percent. Similarly, the overall secondary (Lower and upper) net enrollment ratio rose from 22 to 47 percent. These increases reflect progress in raising the share of children in school and improving school progression according to the appropriate age for grade. However, substantial distortions between age and grade remain, leading to large differences between the net enrollment rate and the gross enrollment ratio at every level, until participation drops off steeply at the tertiary level. These distortions are driven by the widespread practice of starting primary school late and by the high rates of grade repetition and dropout. Figure 3.4. Enrollment rates in primary, secondary, and tertiary education 130% 120% 10% 100% 90% 80% 70% 60% 50% @"xr 40% 30% O©car 20% 10% 0% Primary Lower Secondary Upper Secondary Tertiary © PS & LE Source: ECVMAS 20172; World Bank and ONPES calculations. Note: Net enrollment rate (NER) = enrollments at a given level of education among the age-group that officially corresponds to that level expressed as a share ofthe same age-group in the population. Gross enrollment ratio (GER) = the number of children who are attending school at that level regardless of age, divided by the number of children in the age-group that officially corresponds to that level Children start primary school an average of two years late and progress slowly, such that fewer than 60 percent will reach the last grade of primary. While the official age for beginning primary school is 6, the average child enters first grade at 7.8 years, after spending two or more years in some form of preschool. This distortion grows over time because about 10 percent of children repeat, and 2-6 percent drop out of each grade of primary, such that there is a three- to four-year gap between the average age of students and the prescribed age from second grade onwards (table 32). Using a simulated cohort approach, these rates imply that only about 58 percent of children in first grade will arrive at sixth grade, and only 29 percent will reach the final year of upper secondary. Therefore, identifying and addressing the drivers behind late primary-school starts as well as the high repetition and drop-out rates are critical to boosting educational attainment. The available data allow some analysis of the in- dividual and household characteristics correlated with overage-for-grade status, but additional research is needed into the systemic causes. [page 106] WorldBank - ONPES | Table 3.2. The average students completes primary school at nearly 16 years of age Grade AUerage age Prescribed age |% expected to repeat % expected to drop out BEA PE RE PS RE A RE BE EP PE AS EE > SC RER ET RE VS PE ES EE BC PE PS SP A Se ARE PRET AE RE ES d > 5 5 à 28 2 [po [208 À 18 | 9 | 50 | Source: World Bank estimates based on data in DHS 2012 Note: Rheto = grade six Philo = grade seven. Children in poor rural households are much less likely to be in school or to be in the appropriate grade for age (figure 3.5). Among all poor households, 88 percent of children aged 6-14 are in school, compared with 96 percent of children in nonpoor households. Similarly, among poor households, 62 percent of children aged 10-14 are overage for their grade (70 percent in rural areas), against 38 percent among the nonpoor. These results suggest that poverty is an important barrier to school enroll- ment. If other characteristics are held equal, evidence shows that, for a G 1,000 increa- se in annual household per capita consumption (worth about 4 percent ofthe national poverty line), the probability of school enrollment rises by O2 percentage points, while the probability of being overage for grade declines by 3 percentage points. Many fac- tors drive the correlation between poverty, enrollment, and overage status, including the cost of schooling, which may delay school enrollment or cause children to drop out temporarily. Indeed, about one-third of children aged 10—14 who are not in school are working, and only 60 percent of the children in the lowest welfare quintile are in school, but not working. Many children continue to serve as restavecs, which can affect their enrollment and progression in school? The costs associated with education are the primary reason children are not in school in 83 percent of cases. Other factors asso- ciated with poverty, such as malnutrition, poor health (see below), and lack of stimula- 98 To examine the effects of several household characteristics within the same framework, we have car- ried out a probit regression of school enrollment and overage status characteristics among individuals (Xi), households (Hi), and area of residence (Zi), as follows: in school i = a+B1Xi + B2Hi + B3Zi + ei (3.1) The results are presented in appendix K 99 Restavecs (reste avec in French means stays with) are children in poor, usually rural households who are sent at an early age to live with wealthier families, usually relatives, in urban areas, in the hope of a better life. Frequently, these children are used as servants by the host families, who typically disrespect the children's most basic human rights. Restavec children are difficult to identify in house- hold survey data. In ECVMAS (2012), only 91 observations include household members identified as “domestique = restavèk: Yet, some studies have found that the problem is significant. For instance, a 2009 study by the Pan American Development Foundation found that there may be as many as 225,000 restavecs in Haïti (Pierre et al 2009). [page 107] : Investing in People to Fight Poverty in Haïti tion, can have lasting effects on children's cognitive development. If they do not receive Girls start dropping adequate stimulation in early childhood, children may enter school ill prepared and be out of school at the more likely to perform poorly, to repeat grades, and to drop out of school relative to chil- age of 14, earlier dren whose cognitive skills and overall school readiness are more suited for entry into than boys, exposing primary school.19° themselves to longer terms . k consequences such Figure 3,5. School enrollment by area of residence, as early marriage poverty status, and gender, % andilliteracy a. Enrollment rates overage for grade 100% 90% 80% 70% 60% 50% Female 40% d 30% © male 20% 10% 0% Poor Not Poor Poor Not Poor Rural Urban b. Share of 10- to 14-year-olds 100% 90% 80% 70% 60% 50% Female 40% d 30% @ Male 20% 4 © e ÆA 1 e “ Kéd EXD Kâd EX. 0% Poor Not Poor Poor Not Poor Rural Urban Source: ECVMAS 2072; World Bank and ONPES calculations Note: Children aged 6—14 are included in chart a (gender differences are not statistically significant, but the differences poor/nonpoor are). Data limitations prevent grade-spedific overage analysis on children under age 10. Children are classified as overage if they are at least two years older than the prescribed age for their grade. Gender is not statistically significant for enrollment, but is so for students who are overage for grade. The presence of parents in the household and their education level, the area ofresidence, and disability are correlated with enrollment and regular progres- sion in school. Compared with their peers in school, children who are not in school are much less likely to be the son or daughter of the household head and much more likely to be disabled. They are also more likely to be in households in rural areas and 100 For example, see Currie and Thomas (1999); Feinstein (2003), Heckman and Masterov (2007); Pianta and McCoy (1997), Reynolds et al. (2001). [page 108] WorldBank - ONPES | in households in which the heads have little education. Among children in school, those who are overage are much more likely to be boys and much less likely to be the children of the household head. As with out-of-school children, they are also more likely to be living in rural areas and in households in which the heads have little education. While boys are more likely to be overage, girls start dropping out of school sooner than boys, at around age 14. Many students learn little, particularly in poor communities. According to as- sessments administered in early grades in selected schools, basic skills are acqui- red slowly or not at all, particularly in schools in poor communities. For example, assessments conducted in schools in Artibonite and Nippes found that the average third grader could only read 23 words per minute, well below the estimated speed of 35-60 words per minute required for comprehension of a basic text (RTI Inter- national 2010; USAID 2012). Weak learning outcomes are not surprising because instructional quality and the provision of learning materials are generally believed to be limited (MENFP 2013). For example, in French-language and mathematics as- sessments of primary-school teachers in the Central Plateau, where the questions were drawn from teacher training institute examinations, only 10 percent (French) and 22 percent (mathematics) of teachers were able to answer at least half of the questions correctly (Gallié and Marcellus 2013). National examinations are first administered at the completion of grade six and have been criticized for archaic content and a reliance on memorization. Students sitting for the examinations are a relatively select group, considering that many children do not reach beyond sixth grade and that sitting for the first two exa- minations requires the payment of a fee (G 250 in sixth grade; G 350 in ninth grade). Passing rates were about 75 percent in grades six and nine in 2013; they were 29 percent in Rheto (grade six) and 38 percent in Philo (grade seven). These rates vary by department, as does the share of students actually sitting for the examinations. Given the weaknesses in basic skills suggested by small-scale studies, nationally re- presentative learning assessments are needed to understand the challenges faced by the majority of Haitian students. Household expenditures and the supply of education The supply of public schools in Haiti is limited. According to data ofthe 2010/11 school census, only 12 percent of the 17076 schools in Haiti are public, and they host 22 percent of primary pupils and 27 percent of secondary students. While the majority of children are in nonpublic schools, 61 percent of children living in poor households attend nonpublic schools, compared with 78 percent of nonpoor chil- dren. Among poor children attending nonpublic schools, over 70 percent attend either community schools or private schools that are not religiously or community affiliated. The offer of nonpublic primary schools has increased exponentially in re- 101 In both cases, the communities involved were targeted by the government and its international part- ners for assistance because of their poverty and vulnerability. Therefore, conclusions about Learning in Haitian schools more broadly cannot be drawn from these selected examples. EN [page 109] : Investing in People to Fight Poverty in Haïti cent years (figure 3.6). Slightly fewer than half of nonpublic primary schools are re- ligiously affiliated; Protestant-affiliated schools make up the majority of these. Few data on nonpublic schools are systematically collected beyond the basic information voluntarily provided in the annual school census.i°? Over half of all primary schools are not yet officially recognized by the government, which is currently developing a decentralized Licensing system with multiple levels of official recognition. Figure 3.6. Number of public and non-public schools, by year 16,000 14,000 2 8 12,000 = ä 5 10,000 5 8,000 É © El Z 6,000 ' SCHOOL 4,000 2,000 o rer M © OO O En © + None tr nn 0 OO « in DE BED RSR SD SÉÉSRERESSSRS ROSE a PRÉPEPISrrrrÉreee8r000 Year —— Public —— Non public ———Public (C2003) —— Non public (C2003) Source: School census 2002/03, 2010/11. Despite the weak progression through grades and the poor learning outcomes, households spend a substantial amount to send children to school. Among all households with children aged 6-14 in school, 93 percent report positive education expenditures. These expenditures are substantial on average, and households repor- ted spending 10 percent of total annual household consumption on education (for all children) during the 2011/12 school year. This share is uniform across poor and nonpoor households. The cost is about 50 percent higher in nonpublic schools re- lative to public schools. The higher cost is driven by higher tuition fees (figure 3.7). According to the 2002/03 school census, fees are positively correlated with school infrastructure (latrines, electricity), smaller class sizes, and more teaching materials (Demombynes, Holland, and Leon 2010). 102 In some schools, such as those participating in the government's Education pour Tous (Education for AU tuition waiver program, additional data on enrollments, school materials, and other characteristics are periodically collected. [page 110] WorldBank - ONPES | Figure 3.7. Educational expenditures by category, children aged 6 to 14 years 100% 80% 70% Other 60% © Transport 50% @Uniforms 40% Books 30% Ld @ruition 20% 10% 0% Poor Not Poor Source: ECVMAS 2072; World Bank and ONPES calculations. Regardless of the type of school their children attend, households also spend substantial amounts on uniforms, books, and transportation. These expenditu- res represent a particular burden for poor households, and cost is cited as the primary reason if households are asked why children are out of school. Because poor households have more school-age children and lower total consumption, they spend Less than half as much per child compared with nonpoor households, G 3,600 compared with G 11,400 per child per year. Overall, estimates based on 2012 data show that households spend more than G 21 billion ($500,000) per year on education. Households bear most of the cost of education; they are sometimes helped by private transfers because public expenditure on education is Low. House- holds cover 64 percent of the total cost of education, while, according to the Mi- nistry of Economy and Finance, the government only covers 30 percent, equal to 3.5 percent of GDP. Donors only cover 6 percent of the total cost, and their contri- bution is declining (figure 3.8). There is evidence that private transfers help cover education costs. Only 4 percent of poor households and 3 percent of nonpoor households with children aged 6-14 report they receive transfers specified for schooling. However, among all households that receive private transfers, the ave- rage amount received is over G 45,000, substantially more than total average edu- cational expenditures across households. These households also report spending much more on education relative to other households in the same consumption quintiles. Because money is fungible, transfers not specified for schooling may still provide resources that go toward educational expenditures. [page 111] : Investing in People to Fight Poverty in Haïti Figure 3.8. Financing sources for education a. Source of annual education financing Donors 6% q Public resources 32% ft @ Households 62% b. Donor contributions for education, in billion HTG 9 8 ; À 6 e @ conmmitments 5 . (©) — 2 Disb: n ©) [] isbursements e fA [eo] FY10 FYN FY12 FY13 Source: ECVMAS 2072; World Bank and ONPES calculations. Recognizing that education costs represent a barrier to access and a substan- tial burden on households, the government has taken on greater financing responsibility for primary education. Since 2007, it has provided tuition waivers to nonpublic schools with the assistance of development partners. These waivers have allowed hundreds of thousands of children to attend school without paying tuition. More recently, the Martelly-Lamothe administration initiated the Program- me de Scolarisation Universelle Gratuite et Obligatoire (free and compulsory uni- versal education program, PSUGO), which is intended to finance primary education for hundreds of thousands of additional students. These efforts have provided relief to households burdened by education expenses and may also draw in children who have been kept out of school because of costs. However, given that 50 percent of household education expenditure does not go for tuition, some children are li- kely to continue to be kept out of school if these costs are not reduced as well. [page 112] WorldBank - ONPES | Complementary social protection initiatives such as conditional cash transfers may help families meet these nontuition costs. If they are well targeted and well designed, evidence points to the positive impact of such transfers on school at- tendance and the reduction of child Labor in a wide range of countries (Ribe, Ro- balino, and Walker 2010). The Ministry of Education and Vocational Training’s current strategic plan recognizes that, in addition to poverty, there are many obstacles to school enrollment. The strong correlation of school enrollment with individual and hou- sehold factors, particularly disability and living in a household not headed by one's parents, points to important barriers besides costs. While vulnerable groups are currently served primarily by nongovernmental organizations (NGOSs), the gover- nment is expected to carry out studies to understand the needs of these groups and to provide public support for their education (MENFP 2013). The government also intends to build new classrooms and schools in areas lacking capacity. The ministry’s plan includesinitiatives to improve progression through school and to increase learning, but little is planned for early childhood develop- ment. The majority of children start primary school late, and the gap between the appropriate and the actual age for grade is growing. To address this problem, the government is developing accelerated programs for overage students and studying ways to encourage parents to send children to primary school at age 6. Investments in teacher training, learning materials, and other steps are also an- ticipated to address the poor learning outcomes (MENFP 2013). The ministry is leading in the establishment of an early childhood development policy, but the initiative has been delayed, and a timeline for completion and implementation has not been fixed. Health outcomes and service utilization Health outcomes have improved during the past decade. Despite the devas- tating earthquake in 2010, key maternal and child health outcomes have shown progress. The infant mortality rate decreased by 9 percent, from 70 deaths per 1,000 live births in 2005-06 to 64 deaths per 1,000 in 2072, and the under-5 mortality rate dropped by 10 percent (figure 3.9). The number of underweight and stunted children fell by 35 and 24 percent, respectively (table 3.3). Despite impro- vement in many child health outcomes, the prevalence rate of acute respiratory infections (ARIS) went up by 56 percent between 2005-06 and 2012. The 2010 earthquake could explain this sudden rise because the incidence of ARIs usually goes up during crises (Bellos et al. 2010). EN [page 113] : Investing in People to Fight Poverty in Haïti Figure 3.9. Infant and under-5 mortality rates, by wealth quintile index. Number of deaths per 1,000 live births a. Infant mortality rates 78 77 80 70 73 64 62 67 58 61 51 45 ® @ 2005-06 DHS 2012 DHS @rxa @iovest @second @niddle @rourth OHighest b. Under-5 mortality rates 125 114 110 où 102 96 98 92 88 83 62 55 ® @ 2005-06 DHS 2012 DHS @rxa @iovest @second @middle @rourth @ Highest Source: Data in STATcompiler (DHS Program STATcompiler) (database), ICF International, Rockville, MD, Attp:/wwwstatcompilercom/. [page 114] WorldBank - ONPES | Table 3.3. Health outcomes among children, by wealth quintile index, 2005-06 and 2012 Indicator Q1 Q2 Q3 Q4 Q5 Total DHS 2005-06 CNRS EEREEE Cream 0 on or À DHS 2012 CTMRSERRENSREEE Sources: DHS 2005-06, 2012 from STATcompiler (DHS Program STATcompilen) (database), ICF International, Rockville MD, http:/www.statcompilercom/. Note: Data for stunting and underweight rates for 2005-06 were taken from the STATcompiler database, where the data take into account the new World Health Organization methodology for calculating these rates. Similarly, health care service utilization improved between 2005-06 and 2012. The coverage of cost-efficient health interventions such as oral rehydration therapy, which is used to treat diarrhea (the most important cause of mortality among children), improved by 32 percent between 2005-06 and 2012, and vac- cination coverage increased by 10 percent. Although still Low, there was also an improvement by 9 percent in the number of children treated against ARIs. Despite some improvements, maternal outcomes and maternal health service utilization rates in Haïti are among the worst in the region. The MMR fell by 43 per- cent between 1990 and 2013, from 670 deaths per 100,000 live births in 1990 to 380 deaths per 100,000" in 2013 (figure 310). Although national estimations provide a much lower rate (157 per 100,000 according to MSPP), the MMR in Haiti remains much higher than the regional average of 68 per 100,000 (WHO 2014a). Progress occurred in maternal health service utilization. There was a 64 percent increase in assisted births in institutions between 2005-06 and 2012. The number of deliveries assisted by staff skilled in obstetrics, such as doctors, midwives, and nurses, rose by 42 percent, and the share of women who received at least four antenatal visits rose by 24 percent (table 3.4). However, the prevalence of health facility delivery, deliveries attended by skilled staff, and skilled antenatal care visits is much lower in Haïti relative to all Lower-midd- le-income Central and South American countries (figure 311).10* 103However, these MMR figures are not as reliable as figures that are based on household survey data {such as the figures on infant mortality and other indicators) because the MMR figures are based on estimates from the World Health Organization and others. Survey data—which are much more reliable—cannot be used to capture the trend in the MMR because the MMR has not been measured in recent household surveys in Haiti 104Trained matrones, a type of traditional birth attendant in Haïti, are not considered skilled in obstetrics (DHS 2012), which may not be the case in lower-middle-income countries in the region [page 115] : Investing in People to Fight Poverty in Haïti Figure 3.10. The maternal mortality ratio, 1990-2013 Number of deaths per 100,000 live births 800 300 670 580 600 510 470 500 380 400 300 C} 200 100 [e) 1990 1995 2000 2005 2013 Source: WHO 2014b. Table 3.4. Maternal and child health service utilization, by wealth quintile index, 2005-06 and 2012 Indicator (ojl Q2 Q3 Q4 Q5 Total DHS 2005-06 ÉTÉ RE RE RE EE [sttegatendent [| --l- | - | x) Heath fai delivery | 5 | 5 | 7 | 5 | «= | 7 DHS 2012 [stteabn ll - | 01 [stegamengant | --- | - |») [eatncitydeien | 9 [2 | 5 [sn | w | x | Sources: DHS 2005-06, 2012 from STATcompiler (DHS Program STATcompiler) (database), ICF International, Rockville, MD, http:/wwwstatcompilercom/. Note: — = not available EN [page 116] WorldBank - ONPES | Figure 3.11. Health care service use, Haiti and selected lower-middle-income Latin American countries. Percent Guyana Honduras Nicaragua Dominican Republic, 2013 Bolivia @ Haïti - 50 100 @ Health Facility Delivery @ Skilled Birth Attendance @ Skilled ANC (4 visits) C] Treatment of ARI @ Treatment of Diarrhea ] Full Immunization Sources: Data on Bolivia (DHS 2008), Guyana (DHS 2009), Haiti (DHS 2012), Honduras (DHS 2011-12), and Nicaragua (DHS 2001) taken from STATcompiler (DHS Program STATcompiler) (database), ICF International, Rockville, MD, http://www.statcompilercom, Child health outcomes are also a concern. The under-5 mortality rate is nearly six times the regional average of 16 (WHO 2010). While the prevalence rate of children with ARIS in Haiti (14 percent) is Lower than the corresponding rates in most Latin American countries, the latter have greater immunization coverage and better treat- ment of children with ARIs. Child health service coverage indicators are much lower in Haïti than in other lower-middle-income Latin American countries. The poorest exhibit systematically worse health outcomes and health service utilization rates. Despite improvements among the lower wealth quintiles since 2005-06, large inequalities persist: the poorest quintiles do worse in health out- comes and service utilization."® For instance, child mortality in the highest income 105 Wealth quintiles here refer to the wealth indicator computed using DHS data, not to consumption quintiles. [page 117] : Investing in People to Fight Poverty in Haïti quintile was 62 deaths per 1,000 live births, while it was 104 among the lowest inco- me quintile (see figure 3.9). Compared with the highest quintile, the number of stunted children was four times greater in the lowest quintile in 2012 (see table 3.3). Among children who had ARIs, 52 percent in the highest wealth quintile were given treatment versus 23 percent in the Lowest wealth quintile (see table 3.4). The share of assisted births in institutions was eight times greater among the highest wealth quintile (76 percent) than among the lowest wealth quintile (9 percent) in 2012, which highlights that the poorest have limited access to maternal health services. Health outcomes among children are influenced by the level of education of their mothers. In 2012, 34 percent of children whose mothers had no education were stunted versus 12 percent of children whose mothers had attained secondary or hi- gher education (table 3.5). Women with no education have three times more children who are underweight compared with women with a secondary or higher education. Similarly, 33 percent of children whose mothers have no education are vaccinated versus 51 percent of children whose mothers have a secondary or higher education, and 59 percent of women with secondary or higher education deliver at health faci- lities compared with only 13 percent of women who have no education, a difference of 354 percent. Table 3.5. Children's health outcomes and service utilization, by educational attainment of the mothers Education of the mother | Stunted Underweight | Vaccination ARltreatment | Diarrhea treatment | Health delivery Secondary or higher 12 6 51 51 60 59 Source: DHS 2012 Sickness is one of the most important shocks experienced by the Haïitian popu- lation, affecting their earning capacity. Over a calendar year, 37 percent of house- holds suffer from health-related problems (excluding cholera), and, for 28 percent, these are the most severe shocks experienced during the year (figures 312 and 313). Overall, health shocks are the second most common shock experienced by hou- seholds, after hurricanes and floods, but the most severe. The cholera epidemic, which has been devastating parts of the country since 2011, is among the first 10 shocks in terms of incidence and the fourth in terms of severity (box 3.3). The high level of job informality and the Low access to social security suggest that such shoc- ks may have a direct impact on the ability to generate income in the household (see chapter 5). [page 118] WorldBank - ONPES | Figure 3.12. Share of households encountering problems over the previous 12 months, 2012 Interruption of financial transfer from government } 1 Only 53% and 31 % of the population Equipement, tools down 3 have access to Non-farm wage/income loss 5 safe water and | improved sanitation Death of one family member 6 respectively: Increase on seed price 7 improvement a . in access will Support of additional family member 7 contribute to better Interruption of financial transfer from parents 10 health outcomes, ; including eradicating Bankruptcy of non-farm business 10 a : pecy the cholera epidemic Diseases of crops/plants 15 Animal disease 15 Cholera 15 Theft of money, good or harvest 17 Irregular rains 23 Drought 30 Scarcity of basic food on the market 30 Sickness (other than cholera and accident) 37 Cyclones, flood 44 O 10 20 30 40 50 Sources: ECVMAS 20172; World Bank and ONPES calculations Figure 3.13. The five most severe shocks among Haitian households, 2012. Percent CT 30 25 20 Ru” Y mn 15 | D (TRS 5 O Sickness (other Cyclones, flood Scarcity of basic Cholera epidemic Drougth than cholera and food on the accident) market Source: ECVMAS 20172; World Bank and ONPES calculations [page 119] : Investing in People to Fight Poverty in Haïti Box 3.3. Cholera epidemiological evolution and current policy actions Despite a reduction in the incidence rate of cholera since 2010, cholera still poses a significant challenge. The current cholera outbreak in Haïti be- gan 10 months after the devastating earthquake of January 12, 2010. Over 705,207 cases and 8,559 deaths were recorded over the next three and a half years (table B3.3:1). Since the outbreak, concerted national and inter- national efforts have cut the number of new cases and deaths considerably each year. The number of cases declined from a monthly average of 29,336 in the first full year of the epidemic (2011), to 1,240 through 2013. The num- ber of deaths decreased correspondingly, from 4,101 in 2010 to an expec- ted 64 in 2014. The incidence rate is the lowest since the beginning of the epidemic and below the 1 percent target rate set by the World Health Orga- nization® Yet, Haïti is still dealing with cholera. Eliminating the disease will require sustained action from the government and development partners. Table B3.31. Epidemiological evolution of cholera in Haiti, 2010-14 Vear Oct-Dec 2010 | 2011 2012 2013 June 2014 | Total Source: Data of the Ministry of Public Health and Population. An enduring solution will require substantial investments to boost ac- cess to water and sanitation and improve hygiene. Access to water and sanitation is Low in Haiti: only 53.2 percent of the population has access to an improved water source, while 31.3 percent have access to improved sanitation facilities.® However, these figures hide the divide between urban and rural areas. Thus, improved water sources are available to 55.0 and 517 percent of the population, respectively, in urban and rural areas, while ac- cess to improved sanitation is at 47.9 and 15.9 percent, respectively. Cholera cannot be sustainably eliminated without addressing the primary vectors in the spread of the disease, such as the lack of a safe water supply and inadequate waste management and sanitation. Gains in health care and in water and sanitation will also help prevent other illnesses and increase general preparedness and resilience in the face of other diseases and disasters. Boosting capacity in addres- sing these issues establishes a stronger platform for disaster preparedness (including epidemics) and ultimately contributes to reducing poverty and enhancing the lives of the poor. a. ECVMAS 2072 for the data. For context, see WHO (20140). b. “At a Glance: Haïti” United Nations Children's Fund, New York, http:/www.unicef.org/infobycountry/haiti _ statistics.htmlL. EN [page 120] WorldBank - ONPES | The most vulnerable to health shocks are the elderly and children because of their more vulnerable health status and their reliance on support from their families. The main risks among the elderly are associated with the limited covera- ge pensions (contributory or noncontributory schemes), the lack of access to health care, and the need to depend on family or charity for survival. Indeed, the elderly do not usually live alone in Haïti. Overall, more than 85 percent of the elderly live in hou- seholds with nonelderly people, and this share is more than 10 percentage points hi- gher among the poor elderly (92 compared with 81 percent among the nonpoor). This is an indication that the elderly must rely on the support of younger generations to a much larger extent in Haïti than in other countries, which may constitute a source of vulnerability. The share of people aged 60 and above living with nonelderly people in Haiti is one of the highest in Latin America and the Caribbean, 88.6 percent compared With the regional average of about 71.0 percent.i°6 While health shocks affect similarly the poor and nonpoor, cholera dispro- portionally affects the poor in rural areas. Among households that experience health problems, 55 percent are nonpoor and 53 percent are in urban areas. Meanwhi- le, cholera mostly affects the extreme poor, but also households in rural areas (table 3.6). The latter are almost twice as affected by cholera relative to households in urban areas, which is not surprising given that cholera is a waterborne disease arising becau- se of poor sewage and poor sanitation and mainly caught by vulnerable populations that do not have regular access to a protected source of drinkable water: access to improved sanitation in rural areas is at 15.9 percent, and over 46.0 percent of the rural population drink water from unsafe sources (see chapter 1). Table 3.6. Proportion of households that consider sickness and cholera the most severe problems, by poverty line, residence, and gender. Percent Indicator Sichkness Cholera epidemic Poverty line Area of residence Gender Total 28 7 Source: ECVMAS 2072; World Bank and ONPES calculations. Note: Sickness does not include cholera or accidental injury or death. 106This is a simple average based on ASPIRE environmental indicators, see ‘ASPIRE: The Atlas of Social Protection, Indicators of Resilience and Equity; World Bank, Washington, DC, http://datatopicsworld- bankorg/aspire/ EN [page 121] : Investing in People to Fight Poverty in Haïti Limited supply and lack of financial resources are the two most common reasons the poorest do not use health services. In 2013, at the national level, the top reason for not seeking care among the entire population suffering from a health problem was lack of money (49 percent). The poorest suffer even more from the financial constra- int: 65 percent did not consult a health provider because of lack of money, against 39 percent among the top quintile (figure 314). The problem of financial barriers to access was equally prevalent across all departments (between 78 and 84 percent). Between 2005-6 and 2013, the situation did not change, and cost and distance remain the main reasons people do not seek medical treatment (figure 315). Figure 3.4. Causes of non-access to health services, by per capita consumption quintile, 2013 100% 90% 80% & [re] 70% Ÿ & & 60% = $ & Re $ 50% 5 S S & 40% a &æ ES À Be &e & BC 30% æe À RME Shgzx CSS. OMS RE Te Es 20% | CR & æ . ë 10% Ÿ : © 0% Lowest 2nd Middle Ath Highest Total @rc: necessary @ Too expensive/Lack of Money @ Automedication Other Source: World Bank and ONPES calculations based on ECVMAS 2 2013 Figure 3.15. Obstacles in access to health care services, by wealth quintile index a. 2005-06 Highest TI 60 4th 1 16 79 Middle 83 2nd 21 8 Lowest 5 92 Total T7 78 - 20% 40% 60% 80% 100% [page 122] WorldBank - ONPES | b. 2012 Highest 57 4th - 77 Middle 83 2nd 86 Lowest Lu 1 30 Total 76 - 20% 40% 60% 80% 100% @ Not willing to go alone @ Distance to health provider @ Not having money for treatment Not having permission to go for treatment Source: World Bank and ONPES calculations based on DHS 2005-06, 2012 The supply of health services and household expenditure The number of medical staff has increased recently and the density of both medical staff and hospital beds is high relative to Low-income countries in Africa.!°? Currently, 17,736 medical and paramedical personnel as well as commu- nity health workers work in Haïti; that is 16.75 medical staff for every 10,000 peo- ple. (There are 9.5 doctors [generalists and specialists], nurses, aid-nurses, and midwives per 10,000 inhabitants [figure 316, chart a].) The number of medical personnel rose in absolute terms between 2071 and 2013./% The coverage of me- dical staff is higher in Haiti than in most low-income countries in Africa. The den- sity of medical personnel in Benin, Burkina-Faso, and Mali is, respectively, 8.3, 62, and 51 medical personnel per 10,000 inhabitants (WHO 2013a). (Box 3.4 provides an overview of the health care system in Haiti.) Additionally, Haïti has 7 beds per 10,000 inhabitants, which is also higher than many low-income African countries: Benin, Burkina-Faso, and Mali have respectively 5, 4, and 1 beds per 10,000 inha- bitants (WHO 2013a). Haiti has the same bed density level as Honduras, but Less than that of other lower-middle-income Latin American countries (WHO 2013a). 107 World Bank estimates based on IHE and ICF International (2014). 108World Bank estimates based on IHE and ICF International (2014) and MSPP (2011). There may be methodology differences between the World Bank and MSPP estimates because the latter do not provide a definition of staff categories. Physicians include both generalists and specialists in the World Bank estimates, but this may not be the case in the MSPP data. The results should therefore be interpreted carefully. [page 123] : Investing in People to Fight Poverty in Haïti Figure 3.16. Coverage of health services a. Density of medical staff per 10,000 inhabitants 16 E & 14 12 . 10 — 8 off 6 4 2 Urban Rural total @rhysician @Nurse Q@Aid-Nurse (Auxiliaire) @ Midwife Source: World Bank estimates based on IHE and ICF International 2014, b. Density of beds per 10,000 inhabitants 12 on 10 ei À 10 | 8 + 6 7 4 “à 4 2 Urban Rural Total Box 34. The health care system in Haiti Health system governance includes units and directorates at the central level of the Ministry of Public Health and Population, 10 departmental heal- th directorates, and 42 arrondissement health units. Services are provided at various levels of the health care system, which includes 907 facilities. The formal health service delivery system includes (1) a first Level of 784 health centers and dispensaries—129 health centers with beds, 298 heal- th centers without beds, and 359 dispensaries providing primary care in the communes and municipalities—and 105 community referral hospitals in the arrondissements; (2) a second level consisting of 8 departmental [page 124] WorldBank - ONPES | hospitals providing secondary health care; and (3) a third level made up of 8 national referral or teaching hospitals providing tertiary health care (IHE and ICF International 2014). Primary care is organized into two tiers linked in a referral system be- tween primary health service providers and the community referral hos- pitals (figure B3.4-1). At the community level, the first tier includes basic health care institutions delivering a basic package of services, including health promotion, disease prevention, and curative care. The package covers child, adolescent, and women's health, emergency medical and surgical care, communicable disease control, health education, and the provision of essential drugs. The secondary tier in the health service pyra- mid network includes the community referral hospitals, which offer four basic services, namely, medicine, pediatrics, obstetrics, and surgery. At the secondary level are the departmental referral hospitals, which offer additional specialized services, including ophthalmology, orthopedics, urology, and dermatology. Since the cholera outbreak, some facilities at the primary and secondary levels have put in place cholera treatment centers or related units (depending on the number of beds), which are generally located in tents. However, because the funding for cholera pre- vention and the funding for treatment are separate, parallel emergency response systems have been established in an unstructured manner. The Ministry of Public Health and Population is now seeking to integrate these emergency responses to treat all acute diarrheal diseases. To this end, it has launched the Cholera Elimination Plan, with the support of the Regio- nal Coalition for Water and Sanitation to Eliminate Cholera Transmission in the Island of Hispaniola. At the top of the health service delivery pyra- mid is the most specialized national referral hospital, the Hospital of the State University of Haïti. Figure B3.4.. The health service delivery pyramid Sp al PRIMARY - 1st tier Primary Health Service Providers - Health Centers - 784 EN [page 125] : Investing in People to Fight Poverty in Haïti At the community Level, rally posts, mobile clinics, community agents, and local birth attendants provide health services. Although not all communi- ties have such services, physical access to health care is improved conside- rably where they exist. For example, a dispensary and a health center are, on average, two hours away, while a rally post is 20 minutes away; a health agent, 40 minutes away; and a mobile clinic, an hour away. However, the services provided do not include all the basic health services necessary for the community. Services also include oral rehydration points established in remote areas to address mild cases of cholera and refer more complicated cases to the cholera treatment centers or units. Sources: IHE and ICF International 2014; World Bank 2013b The density of medical staff and beds per 10,000 inhabitants across depart- ments and areas of residence is unequal, thereby compromising access and quality of health care service delivery in some areas, especially to the poo- rest.\9° The density coverage of medical staff in five departments is narrower than the national average. These departments include the second-most populated and least poor (Artibonite), the least populated and the third-least poor (Nippes), and two sparsely populated departments that are also the poorest (Grand'Anse and Nord-Ouest). Consistently, the highest density coverage is in the department of Ouest, which has the highest population density and the highest number of poor in the country. Additionally, the density of medical staff has limited correlation with the density of the poor per 10,000 inhabitants (0.47) at the departmental Level, hi- ghlighting the inadequate health service coverage among the poor (figure 317). The coverage of medical and paramedical staff is 2.5 and 2.0 times lower in rural areas than in urban areas. Although the density coverage of community agents is higher in rural areas (4.7 agents per 10,000 inhabitants versus 3.8 in urban areas) because they typically work in inaccessible areas, the number of agents seems inadequa- te and illustrates the access problems experienced in rural areas. While medical infrastructure is more or Less well distributed across departments based on popu- lation shares, medical infrastructures, particularly secondary health care facilities, Still fail to reach the rural population, especially the poorest, who are often living in the remotest areas. Indeed, the density of beds is 4 per 10,000 inhabitants in rural areas; two times lower than in urban areas (see figure 316, chart b). 109The World Bank estimates the density of medical personnel and number of beds per 10,000 inhabi- tants based on IHE and ICF International (2014) and the latest population figures from the IHSI (2014) The density of medical personnel and bedbs is estimated for all 907 health facilities in Haïti EN [page 126] WorldBank - ONPES | Figure 3.17.. The density of medical staff: ratio medical staff/poor population o nr 49 8 140 :- 9 10 10 2 £ 9 ° 0 À 52 100 : 7 6 7 8 £ 2e 80 : 5 = cs Q 3£ 40 - # gd 5 0. 2 5 8 CA 5 5 2 # 8 gp 2 8 ÿ % à 8 3 5 2 Ô 5 © z £ an Z ë & DT 9 < O 2 2 = Poor population /10,000 inhabitants —@—Medical staff for 10,000 inhabitants Source: World Bank and ONPES calculations based on IHE and ICF International 2014; IHSI 2014. The government manages one-third of the health facilities, but donors and households bear much of the financial burden of health services. The Minis- try of Public Health and Population manages 38 percent of health facilities, while the nonprofit sector manages 18 percent, the mixed sector 20 percent (both the ministry and nonprofits), and the private sector 24 percent. Yet, donors and house- holds bear much of the financial burden of health services. In 2011-12, 64 percent of total health expenditures were funded by donors, 29 percent by households, and 7 percent by the government. Additionally, contributions from donors decrea- sed by 161 percent between 2012/13 and 2013/14, while the state budget increa- sed slightly, thereby imposing sustainability risks for financing in the health sector. The majority of Haitians, including the poor, consult public health care pro- viders, and only a minority, concentrated in rural areas, turn to traditional medicine (table 3.7). In case ofillness, 46 percent of the poor and 41 percent of the nonpoor consult public health care providers. Similarly, the nonpoor are three times more likely than the extreme poor and twice as likely as the poor to consult private health care providers. Only 5 percent of the population consult with tra- ditional healers; however, the prevalence is higher in rural areas (8 percent) and among the moderate poor (7 percent) and extreme poor (6 percent). Poorer po- pulation segments tend to use self-treatment more often than do the nonpoor: 8 percent of the moderate poor and 10 percent of the extreme poor buy medicines from street medicine sellers, while 5 percent of the nonpoor do so. EN [page 127] : Investing in People to Fight Poverty in Haïti Table 3.7. Health care providers, by the location and poverty Level of the population served. Percent Location Poverty level Provider Total Metropolitan Urban Rural Nonpoor Moderate poor Extreme poor Community health workers For Ji fée 2 | 3 | 8 |. Traditional birth attendant fon do nl o | oo | 1 | o. Other (e] 5 6 n nm 8 5 Source: World Bank and ONPES calculations based on ECVMAS 2 2013. The burden of health expenditure is higher among the extreme poor. On ave- rage, individuals spend 17 percent of their budgets on health (table 3.8). In terms of shares of their total budget, the extreme poor spend 5.5 and 117 percent more than the moderate poor and the nonpoor, respectively. In terms of absolute value, though, the extreme poor spend slightly Less than a fifth of the amounts spent by the nonpoor. Table 3.8. Per capita annual out-of-pocket health expenditures, by poverty line. Percent of the per capita consumption aggregate Measure Overall, N = 23,555 Nonpoor, n = 10,000 Extreme poor, n = 5,646 Moderate poor, n = 7909 Source: ECVMAS 2012; World Bank and ONPES calculations Note: Out-of-pocket health payments cover consultation, examination, medicines, treatment material hospitalization, and expenses for spectacles and prostheses. Expenditure is estimated based on the total number ofindividuals (N = 23,555). The per capita annual average expenditure is estimated based on the total number of individuals, whether or not they have all made health care expenditures. Health expenditures are significantly higher in urban areas, reflecting the be- tter supply of health care services there (table 3.9). Urban residents spend two times more than rural residents on health care. Most health care facilities in rural areas are dispensaries, which do not offer laboratory or x-ray equipment. Health service supply is greater in urban areas (see figure 316, chart b and table 37), with a higher density of medical staff and beds per 10,000 inhabitants, thereby offering more incentive for urban residents to spend on health care services. EN [page 128] WorldBank - ONPES | Table 3.9. Per capita out-of-pocket health expenditures, by gender and location Characteristic Auerage, HTG Auerage, S Gender Location Source: ECVMAS 2012; World Bank and ONPES calculations Note: Out-of-pocket payments for health cover consultation, examination, medicines, treatment material, hospitalization, and expenses for spectacles and prostheses. Medicine is the main driver of out-of-pocket expenditures. Households spend, on average, G 3,175 ($75) per year on health care services, of which 60 percent (G 1,891 or $45) is spent on medicines (table 310). Consultations represent the second driver of health expenditure (G 484 or $12), followed by hospitalization (G 386 or $9). In other countries in the region, medicine is one of the highest health expenditure items. Between 30 and 60 percent of health expenditures in Latin America go for medicines (UNECLAC 2009). Table 310. Household out-of-pocket health expenditures, by service type. (N = 4,929) Item Amount, G Amount, $ Share, % Prosthesis and glasses Total 3175 75 100 Source: ECVMAS 20172; World Bank and ONPES calculations The incidence of catastrophic health expenditure is greater among the ex- treme poor. Catastrophic health expenditure represents one way to assess the financial hardship caused by sickness. Health spending is considered catastrophic when households spend a certain threshold of their incomes or nonfood con- sumption on health. There are various methodologies to measure the threshold. There seems to be some agreement on the use of the 25 percent threshold of nonfood consumption to measure the necessary level of financial protection (WHO and World Bank 2013). In Haiti, 3.4 percent of households encounter catas- trophic health expenditures (figure 318). Poorer and rural households incur such expenditures more often. The incidence is 3.7 percent among the moderate poor, EN [page 129] : Investing in People to Fight Poverty in Haïti L.O percent among the extreme poor, and 17 percent among the nonpoor. The inci- dence is three times higher in rural areas (5.0 percent) than in urban areas (1.6 per- cent), suggesting that the poor and households in rural areas are more vulnerable to health shocks relative to the nonpoor and households in urban areas. Figure 3.18. Incidence of catastrophic health expenditure in Haïti, 2012 Men 34 Women 3.5 Urban 1.6 Rural 5.0 Lower quintile 41 2nd quintile 61 3rd quintile 32 Ath quintile 21 Highest quintile 15 Extreme Poor 5 Moderate Poor 37 Non-Poor 17 Total 3.4 o 2 4 6 8 Health expenditures of households are risks. There are different methods to considered catastrophic when households measure the level of catastrophic health spend a certain threshold of their income or expenditures. There is a growing consensus of their non-food consumption on health. to use 25% as threshold of non-food The level of catastrophic health expenditure consumption, hence, the methodology allows health policy makers to measure the chosen by the WB to estimate the Level of level of financial protection against health catastrophic health expenditures in Haiti. Source: ECVMAS 2072; World Bank and ONPES calculations Haiti has a low incidence of catastrophic health expenditure compared with other African and Latin American countries (figure 3.19). Haiti has one of the lowest incidences of such expenditure, at 34 percent. The incidence is above 30 percent among low-income countries of Africa. Half the population incurs the ex- penditure in Burkina Faso (511 percent), and two-thirds in Mali (74.2 percent). The expenditures are Lower in Ghana (34.5 percent), Kenya (23.4 percent), and Senegal {75 percent), but are still high relative to Haïti. Other Latin American upper-midd- le-income countries (the Dominican Republic, Ecuador, and Paraguay) also have a high incidence (above 30 percent). However, it is difficult to compare the incidence in Haïti and in these countries. Indeed, the data used to estimate the expenditure in countries presented in the Poverty Assessment are from the World Health Survey 2002-04, but Haitis analysis is based on ECVMAS 2012.10 Researchers have noti- ced that the World Health Survey provides higher estimates of health payments, but lower estimates of total consumption, leading to overestimates of catastrophic health expenditure compared with other surveys (Moreno-Serra 2013). 110 WHO World Health Survey (database), World Health Organization, Geneva, http: //wwwwhoint/health- info/survey/en/ EN [page 130] WorldBank - ONPES | Figure 3.19. The incidence of catastrophic health expenditures in Africa and Latin America Senegal, 2003 18% Mali, 2003 74% Kenya, 2004 23% Ghana, 2003 35% Burkina Faso, 2003 52% Uruguay, 2003 1% Paraguay, 2003 33% Ecuador, 2003 50% 2minic Republic, 2003 40% Haïti, 2012 3% 0% 20% 40% 60% 80% Source: World Bank 2012 Further research should be conducted to clarify the causes of the lower in- cidence of catastrophic health expenditure in Haiti relative to low- and lower-middle-income countries. One hypothesis is that the Low levels of catas- trophic health expenditure arise because of the limited utilization of many types of health services in Haiti compared with Low- and lower-middle-income coun- tries in Africa and Latin America (see figure 319). Indeed, the incidence of health expenditures is mainly driven by the cost of medicine in Haïti, which may imply a high degree of self-treatment. The significant levels of external funding may also contribute to the lower incidence of such expenditures in Haiti. However, it is not clear whether donor funding is efficiently and equitably distributed among the 10 departments, and research should be conducted on this issue. Furthermore, the low quality of Local health services because of factors such as shortages of drugs and medical supplies may deter patients from consulting health facilities, as demonstrated by a recent study conducted in three departments (IHE and ICF International 2014). In some cases, patients must buy drugs and medical supplies on their own to receive care, which may impede more frequent reliance on health facilities. The extremely high poverty headcount is certainly another key explana- tion for the Low incidence of catastrophic health expenditure. Households may be too poor and therefore unwilling to face the financial hardships caused by reliance on health services. Indeed, the DHS shows that lack of money is the key reason people do not visit health facilities. The use of savings and borrowing money from friends or relatives for health care are not accounted for in estimating catastrophic health expenditure; neither are expenditures for traditional healers, thereby un- derestimating the incidence of such expenditure. Indeed, most households facing health and cholera shocks use savings or borrow money from friends or relatives (see appendix K). [page 131] : Investing in People to Fight Poverty in Haïti Asset accumulation in health care and education are essential to building human capital and are instrumental in increasing economic opportunity and improving welfare in Haiti as elsewhere in the world. Education and health service utilization and health and education outcomes have improved in Haiti; however, both are relatively limited, especially among the poor. Adult Literacy and enrollment among school-age children are significant- ly Lower among poor households. A number of factors could explain this. À large number of poor children have to work while attending school, raising the probability of dropping out or becoming overage for grade. Similarly, poor households spend substantially Less on school fees, which are associated with the quality of the servi- ce and the infrastructure provided by the school. Child and maternal mortality indi- cators show a similar pattern: child mortality and malnutrition as well as maternal mortality are more prevalent among the poorest, suggesting the Lower utilization of health services and the higher impact of health shocks on poor households. In particular, the Low levels of both outcomes and service utilization among women are à serious concerm. There are strong signs of the intergenerational transmission of poverty in Haïti, which could be curtailed by boosting educational attainment. The average edu- cational gap among 10- to 14-year-olds is largest among the children in the poo- rest households, at over 2.5 years. The better educated the parents, the narrower the gap, and the more likely the children are in school and at the appropriate age for grade. Moreover, the children of better educated parents face Less risk of being undernourished or stunted, both of which affect cognitive and physical develop- ment and future prospects. The children of better educated parents thus have more chances to perform well in school, thereby increasing their future earning capacity and their chance to escape poverty. Based on this diagnostic, this study offers a set of suggestions for policy priorities and suggested actions in education and health care are listed below. Education Priority 1: Sustain and expand the access to primary education. While primary-school enrollment rates have increased substantially in recent decades, enrollment is still not close to universal, particularly among the most disadvantaged children, including the poorest, those living without their parents, and those with disabilities. At the same time, declining donor financing and a recent decision!" by the Ministry of Education and Voca- tional Training to stop funding tuition waivers for new cohorts of first graders in nonpu- blic schools are threatening the gains in access made in recent years. Achieving universal 111 The Ministry of Education and Training announced on August 8, 2012 a number of measures (12), including the number 7, on the interruption of tuition waivers funding for new cohorts the first grade of primary school http://Lenouvelliste com/lenouvelliste/article/134312/Les-12-mesures-de-Manigat. html EN [page 132] WorldBank - ONPES | primary enrollment will therefore require several critical actions by the government and its development partners (taking into account the different needs and service access in rural and urban areas), including the following: a. Produce and implement a short- to medium-term financing plan for primary education, increasing the resources available for the sector. Through the do- nor-financed Tuition Waiver Program, PSUGO, and the national school feeding program (Program National de Cantines Scolaires), the financial burden of tuition and nutrition at school has shifted substantially from households to the gover- nment and contributed to increases in enrollment and attainment. These gains are now threatened due to a lack of financing. The Tuition Waiver Program has stopped taking on new cohorts in the first grade because donor financing throu- gh the sixth grade cannot be guaranteed. At the same time, ongoing funding for school meals has not been secured from donors. The creation of the Fonds Na- tional d'Education (national education fund), which is financed through interna- tional phone communications and transfer taxes, provides a new funding stream for education and has been used to support PSUGO. However, the receipts ofthe fund do not appear to be sufficient to back tuition waivers, school meals, and PSUGO. Additional resources are therefore required so that the government can eventually assume full financing responsibility over primary education. National policies and medium-term financing plans focusing on tuition waivers and school feeding are therefore urgently needed. b. In coordination with social protection programs, determine medium- to long- term strategic plans for service delivery by type of provider at all levels of edu- cation, starting with primary education. The majority of schools at all levels in Haiti are nonpublic and operate with little oversight or accountability. The g0- vernment has built several new public primary schools in recent years and has decided to strengthen public service provision in primary education by no lon- ger funding tuition waivers in nonpublic schools. Starting in the 2014/15 school year, PSUGO will only fund new first grade cohorts in public schools. While a medium- to long-term shift in financing from nonpublic to public schools is feasible, this shift threatens the access of hundreds of thousands of students who live too far from a public school or who may be unable to enroll because of limited capacity. In addition, preprimary, secondary, and postsecondary edu- cation are also largely nonpublic, and strategies for increasing access to these levels within the government's fiscal constraints are also needed. Because fi- nances are cited as the main reason children are out of school, targeted cash transfer programs can serve as an incentive to send poor children to school and can help poor families meet the associated expenses (see chapter 5). c. Establish a robust information system of beneficiaries, including a targeting me- chanism. Although instruments exist at the program level to identify benefi- ciary schools for the various programs offered by the Ministry of Education and Vocational Training (including the Tuition Waiver Program or PSUGO), no system integrating all programs is currently in place, nor one identifying beneficiary students. À robust information system is needed to avoid the duplication of EN [page 133] : Investing in People to Fight Poverty in Haïti efforts and to strengthen the supervision capacity of the ministry. Such a sys- tem would also contribute to monitoring the new measures adopted by the g0- vernment in issuing teaching licenses and school certifications. An information system that facilitates the identification of geographical areas and schools in need of resources and that utilizes poverty data and data from social protection programs would allow the government to allocate more effectively its limited resources where they are most necessary. Priority 2: Improve learning and the quality of service delivery in education to avoid school abandonment. As presented in chapter 3, early assessments suggest that learning is limited in primary schools, particularly in poor and rural communities. Other indicators of the quality of education, including teacher knowledge and learning materials available in schools, suggest that many children, but particularly poor children, are receiving low-quality primary education. This contributes to high repetition and drop- out rates, and ultimately to low educational attainment because children with weak basic skills are unable to complete primary education and continue to secondary education or otherwise gain little from school. Raising quality will require several key steps, including the following: a. Build up the educational information system and collect better data on learning, school progression, and other outcomes in education. Haiti Llacks a national Lear- ning assessment system, limiting the government's ability to identify and address the barriers to basic skill acquisition. Assessments of learning based on representa- tive samples beginning in the early grades would provide a foundation for planning interventions and measuring their success. Such information would also facilitate the tracking ofinequalities across areas of the country, which existing data suggest are substantial. Furthermore, it would allow elucidating some of the outstanding critical questions, such as the importance of the language used for teaching in primary school for student learning (Creole versus French). Plans to pilot early gra- de reading and mathematics assessments on nationally representative samples, as well as recently announced plans by the Ministry of Education and Vocational Training to develop national examinations prior to the first one now taken at grade six, represent productive steps toward accomplishing this goal. Among other me- asures, these data would help the government to design efficient policies against school abandonment. b. Increase public oversight through targeted and well-implemented measures and systematic data collection to hold schools accountable. Several reform measu- res announced by the ministry in August 2014 hold the promise of increasing public oversight in primary schools. These include plans to phase in a manda- tory teaching license based on demonstrated competencies; an in-service tra- ining program for teachers; a mandatory school identity card, leading to even- tual certification; and ministry supervision of schools with low passing rates in national examinations. Data from learning assessments, as well as from other sources such as the school census, could also be used to inform parents about the quality of schools, as a basis for creating contract incentive systems between the government and schools, and to hold schools accountable for outcomes EN [page 134] WorldBank - ONPES | (Leveraging PSUGO, which is effectively unconditional financing, as a starting point).? These measures, if implemented effectively, would contribute to in- creased quality, learning, and, ultimately, educational attainment. Given the breadth of these activities and the ministry's limited capacity, careful prioriti- zation and planning, followed by vigorous implementation, will be critical to making these measures effective. c. Address preprimary education to give children a solid foundation for skill buil- ding. Investing in children, particularly poor children, before they reach primary school is critical because malnutrition, lack of stimulation, and other depriva- tions are common (see chapter 5). As a result, children enter primary school two years late, on average, putting them at a substantial disadvantage in Lear- ning and educational attainment. In Haiti, preprimary education is provided mainly by the nonpublic sector and, Like other levels of education, is Largely unregulated. Yet, the majority of children attend at least one year of preprimary education prior to entering first grade, creating an opportunity for the govern- ment to help lay the foundation for human capital accumulation. The efforts in preprimary education should ideally be carried out in coordination with a broader government strategy for early childhood development that includes health care, social protection, and other sectors. Health care Priority 1: Expand the coverage, utilization, and quality of health care services. Maternal mortality and child mortality have decreased significantly since 2000. This is significant given the devastation of the 2010 earthquake. This progress is mirrored by the better coverage of key interventions (for example, diarrhea treat- ment and antenatal care). Yet, mortality indicators remain unacceptably high, which can be attributed to persistently limited service utilization and inadequacies in the coverage of basic interventions such as assisted births in health care faci- lities and treatment of ARIs. The shortcomings in coverage and service utilization are still accompanied by important inequalities linked to poverty, area of residen- ce, and gender. Improvements in both areas will therefore require several critical actions by the government and its development partners, including the following: a. As in education, establish an information system for a unified beneficiary targe- ting mechanism. Exempting particular population groups from the payment of health care fees and eliminating the fees for certain services are likely to boost access, especially among the poor. Because external financing is expected to decrease substantially in coming years, effective targeting is even more criti- cal. The development of appropriate targeting tools, including a deprivation and vulnerability index, is critical. Several actors in the social protection sector (FAES, the Ministry of Social Affairs and Labor, and others) are involved in the de- velopment of these tools, which will be used to reach vulnerable populations. 112 The power of school-level data was recently demonstrated when school year 2013/14 national examination results were released for the first time among schools. The low performance in some schools created a public outcry and helped spur the ministry to announce several reform measures in August 2014. EN [page 135] : Investing in People to Fight Poverty in Haïti b. Focus on programs with a prouen record in enhancing the utilization of health care services, especially in primary health care and within communities. The re- levant interventions can focus on a number of fronts. Thus, by paying providers according to the quantity and quality of maternal and child health services they deliver, results-based financing has the potential to improve efficiency in service delivery and the quality of care, which may encourage patients to use health care facilities. Building on existing experience, the Ministry of Public Health and Population is currently working with the U.S. Agency for International Develo- pment and the World Bank to develop a national results-based financing mo- del for Haiti.!5 This model will contract public and nonpublic providers through a results-based financing mechanism to provide to the population a minimum package of services with a strong emphasis on preventative services. Focusing on communities is likely to expand the utilization of primary health care services among the poor (including preventive health services) and, hence, reduce the risk faced by the poor of incurring catastrophic health expenditures linked to hospitalization and expensive medicines. The World Bank—supported Kore Fan- mi Program addresses demand- and supply-side barriers to service utilization to help improve maternal and child health outcomes, particularly among the poor. To address social determinants, a network of community agents (Kore Fanmi) Will deliver certain basic preventative services, promote behavior change, and link households to services and opportunities. c. Fill knowledge gaps to understand the low-usage, low-spending conundrum. Two remarkable features of Haïtis health care system are the limited utilization and out-of-pocket spending. In facing a health problem, 55 percent of the popu- lation does not rely on public services, and households spend only 17 percent of their budgets on health. Catastrophic health expenditure is experienced by only 34 percent of households in Haiti, a 10th of the levels in comparable countries in Africa and Latin America. More research is needed to clarify these findings. The low-usage, low-spending patterns raise the key question of whether the cost of the services is too high relative to the perceived quality, but the extent to which this is true needs to be understood. Among the possible determinants of Low service utilization are the influence of culture on health service usage and the low quality of the services provided. Both issues are worthy of further study."* Priority 2: Develop innovative donor coordination mechanisms. Budget allo- cations from external sources declined by 62 percent between 2012/2013 and 2013/2014, and this trend is likely to continue in the near future. It is thus impe- rative to develop much better mechanisms to coordinate the plethora of external donors in the sector and to find meaningful ways to enhance efficiency and reduce 113 The US. Agency for International Development Result-Santé pour Le Développement et la Stabilité d'Haïti Program, which has certain features of results-based financing and covers selected health facilities in all the departments, has shown some dramatic improvements in child and maternal health utilization through the payment of incentives to nongovernmental and public facilities (Zeng et al. 2012) 114 There is anecdotal evidence that cultural factors play a major role in the high share of birth deliveries in Haiti 65 percent—that take place outside health facilities. EN [page 136] WorldBank - ONPES | overlaps, at the same time ensuring that the government's priorities for interven- tion are systematically taken into account ‘”. Without this, there is a severe risk that quality in health service delivery and health utilization service levels may fall even further. Possible mechanisms to enhance donor coordination include esta- blishing a well-staffed subunit devoted to donor coordination and harmonization of the relevant initiatives, adopting a sector-wide approach, and gradually shifting to pooled funding mechanisms. 115 A framework for donor coordination already exists within the MPCE: the Coordination of External Development Assistance (CAED). This mechanism coordinates the activities of donors through the joint program of aid effectiveness (PCEA). EN [page 137] : Investing in People to Fight Poverty in Haïti . Le Chapter 4: Shocks and vulnerability Haïti is prone to shocks of various kinds, from covariate weather-related and eco- nomic shocks to idiosyncratic economic and health shocks'é. The country's vul- nerability to these shocks is increased by institutional weaknesses and resource shortages that hamper efforts to prepare for, mitigate, or cope with the shocks at the macro and micro levels. Poor households are more likely to experience shocks: 95 percent of households in extreme poverty experience at least one economically damaging shock each year. Rural households are more likely to be impacted by climatic shocks, which are often compounded by agricultural setbacks, while urban households are more likely to be affected by nonagricultural economic shocks. The poor are Less able to cope with shocks, and their coping strategies are more likely to impede future economic activities or human capital accumulation because the poor generally cope by selling assets, taking on more debt, or reducing food in- take. In the case of cholera or weather-related shocks, which are far more prevalent among the poor, the most common coping strategy involves doing nothing, sug- gesting that the poor possess few means to protect their livelihoods from shocks. Risk is inevitable and has important consequences in the lives and decision processes of people exposed to it, particularly in poor countries, which have neither the financial nor institutional means to respond to shocks. individuals, households, communities, and countries are all exposed to risks that depend on factors such as geographical location and geological environment, but, Like indivi- duals, they have different capabilities to prepare for and deal with shocks. The shoc- ks may involve covariate or systemic risks, such as a financial or political crisis, a na- tural disaster, crime, an epidemic episode, or idiosyncratic risks such as the loss of employment among individuals. According to the World Development Report 2014 (World Bank 2013a), the majority of households in developing countries are con- fronted by at least one shock each year, and some are exposed to more than one. Although some individuals may be able to protect themselves from the potentially catastrophic effects of shocks, the majority of the world's poor have limited access to formal insurance.i? This is because of a lack of collateral and high information 116 This chapter is based on ONPES (2014) and Raeza-Sanchez, Fuchs and Matera (2014), background papers for the study by the World Bank and Observatoire National de La Pauvreté et de l'Exclusion So- ciale (ONPES). 2014. Investing in People to Fight Poverty in Haiti, Reflections for Evidence-based Policy Making. Washington, DC: World Bank. 117 Formal mechanisms work through the market, such as purchasing insurance. Informal mechanisms are arrangements within and between households, including using savings, selling assets, receiving mon- etary or other aid from family and friends, altering the consumption pattern by purchasing Less expen- sive items, or taking on additional employment. Both mechanisms can be adopted ex ante to protect the household, that is, purchasing insurance or diversifying employment, or ex post as a response to the shock, such as obtaining credit or selling assets. If the totality of formal and informal mechanisms is not sufficient to maintain the household at the same level of consumption as before the shock, the household will have to reduce its consumption temporarily, or, if the shock is sever enough, the effects may be persistent (Dercon 2004). [page 138] WorldBank - ONPES | and administrative costs, often resulting in sudden drops in consumption in the face of shocks.\8 This is exacerbated in rural settings, where livelihoods depend on rainfall and good temperature and humidity, as well as fertilizer quality, control of crop diseases and personalillnesses, a healthy political situation, favorable tra- de policy, and many other factors. Small islands and extremely poor countries such as Haiti face a combination of extensive and intensive risks, inadequate resources, and low institutional capa- city to prepare for and cope with shocks and are thus especially vulnerable.i? Preparedness, in particular, plays a key role in mitigating the impact of shocks, particularly if systemic. In the event of such crises, responses often need to be managed through formal public instruments, because the systemic impact gene- rates important market failures and the disruption of informal mechanisms of risk sharing, resulting in widespread inadequacy of self-insurance, particularly among the poor and extreme poor. While NGOs and partner countries can provide finan- cial and logistical support, the role of national governments in crisis management remains preeminent in ensuring preparedness and mitigation (Marzo and Hidecki 2012). While Haiti's 2010 7-point Richter-scale earthquake killed 230,000 people, a much larger earthquake in Chile (8.8 on the Richter scale) was destructive, but there were far fewer fatalities, only 525. Haitians are subject to frequent covariate and idiosyncratic shocks. At the ma- cro level, covariate shocks are often related to natural disasters, which are common because of the geographical position of the country (earthquakes, hurricanes, and floods) and the effects of which are exacerbated by widespread deforestation and land degradation. These shocks have à large impact on economic and agricultu- ral activity: in 2012 alone, the country was hit by two hurricanes (Isaac and Sandy) and one drought, leading to negative growth of 1.3 percent in national agricultural production. Economic shocks are also common in Haiti because the country has an open economy and suffers from international fluctuations, arising mainly from increases in import prices, declines in export prices, and the volatility of remittances (for instance, because of a shock in a destination country, such as the Dominican Republic or the United States). Political instability, linked to the institutional fragi- lity that characterizes the country, can also influence the welfare of households if it results in an interruption or slowdown in economic activity or official develop- ment assistance, such as in the early 19905 or early in the first decade of the 2000s. Haitians must also address considerable idiosyncratic shocks such as death, 118 The cost of acquiring the information needed to assess risk, monitor borrower performance, and enforce contractual obligations is high 119 intensive risks stem from low-probability, higher-impact events, whereas extensive risks are associ- ated with high-probability, lower-impact events. Examples of the former are major earthquakes, hur- ricanes, or epidemics, while examples of the latter are Localized flooding, disease among individuals, or unemployment. Extensive risks also include idiosyncratic risks. Another useful distinction is the fact that intensive risks are typically associated with large metropolitan areas, where highly concen- trated economic activities are exposed and vulnerable to catastrophic hazards. In contrast, extensive risks can be associated with rural areas and periurban areas and with the poor living in these areas. 120 These are only the last of yearly natural disasters: among the recent ones preceding the earthquake of 2010, one may List the floods of Fonds-Verrettes and Mapou and the Jeanne Cyclone in 2004, and the hurricanes Fay, Gustav, Hanna, and Ike in 2008 (ONPES, forthcoming). | [page 139] : Investing in People to Fight Poverty in Haïti illness, job loss, and lower wages. Because market and institutional mechanisms are not available to them, the consequences of shocks can be considerable in ter- ms of income losses, despite the existence of informal mechanisms such as support from family and friends. Indeed, studies suggest that, in determining vulnerability to poverty in Haïti, idiosyncratic shocks and shocks at the Local Level are more important than covariate shocks affecting larger regions (Échevin 2013; Jadotte 2010). Poverty is significant in Haiti, but so is the vulnerability to poverty, and shocks can drive an additional million Haitians into poverty. This is illustrated in figure 41, representing the distribution of wealth across the population. The high histograms around the poverty lines indicate that most of the population lives on budgets that are close to the poverty threshold.?1 Thus, the figure shows the substantial vulne- rability of the population to poverty, given that the households close to the line are most likely to transit in and out of poverty as à consequence of a shock. Such a shock could push 1 million people into poverty and 2.5 million into extreme poverty.?? The consumption level of only 2 percent ofthe population exceeds $10 a day, which is the regions income threshold for joining the middle class. Figure 4.1. Vulnerability to poverty in Haïti, 2012 —————— Extreme line — Moderate poverty line æ 200 Ce E ——————— Vulnerability line E E 140 5 120 5 100 & 80 5 60 40 20 O OoQQQQOoQCQoQCCCCoQocoeoocoo SD0000O0VOOOLOOO000000 Om OROMOMOMm OMR OM AO RO SHRSRESRGGSÉSSROLESRÉSS Annual per capita comsumption in gourdes Sources: ECVMAS 2072; World Bank and ONPES calculations. 121 In the absence of panel or synthetic panel data, the vulnerable are defined as individuals living on budgets representing 120 percent of the poverty line. According to this definition, almost 10 percent of the population would be vulnerable, and, together, the poor and the vulnerable would represent two- thirds of the population. An alternative definition of vulnerability used by the World Bank in the case of Latin America is tied to economic stability and the probability of falling into poverty. The threshold corresponding to this probability is $10 a day (in PPP US. dollars), a sum that is therefore used to identi- fy the middle class in the region, while the vulnerable are defined as individuals living on between $4 and $10 PPP a day (Lépez-Calva 2013). If we were to use this definition, the share of the poor and vulnerable would be 98 percent because only 2 percent of the population is living on budgets of more than $10 a day, 122 These numbers are obtained by measuring the effect of a 20 percent reduction in household con- sumption, thereby simulating the impact of a shock such as a natural disaster. [page 140] WorldBank - ONPES | The purpose of this chapter is to describe and shed Light on the relationship be- tween shocks and poverty in Haiti. In particular, the correlation between poverty incidence and shocks are analyzed. The coping mechanisms of Haitian households faced with shocks (such as using savings, receiving aid from friends, changing nu- tritional inputs, or taking children out of school) and the Links with human capital accumulation and economic opportunities (the next section) are also examined. In consideration of the prominence and severity of natural disasters, a section is dedicated to a discussion of vulnerability to this kind of shock, with a focus on the impact of the earthquake of 2010.12: © 2. Shocks, impacts, and household coping mechanisms | Doerenes ch experiences at least one economically damaging shock Prevalence of shocks per year; nearly Shock incidence is high in Haiti and is similar across departments.?* À typical 1 mon LEE Haitian household faces multiple shocks annually; 78 percent of households in ELOMT Lrel 2x0 Port-au-Prince, 89 t of households in other urb d 94 t falling into poverty ort-au-Prince, percent Or nousenolds In other urban areas, an percen following such a of rural households experienced at least one shock. In 2012, between two-thirds shock and three-quarters of the population of 7 of the 10 departments were affected by a climatic shock. There was some geographical variation: 43 percent of the population in the department of Ouest, and 78 percent of the population in the department of Sud-Est were affected (figure 4.2). The impact of diseases seems to be more evenly distributed across the country. From 64 to 67 percent of the popu- lation were affected in half the departments. Disease-affected population shares above 70 percent were found in three departments: Centre, Grand’Anse, and Nord. Economic shocks were more generalized: near or above half of the populations in 123 Worldwide between 1975 and 2008, only 23 mega-events led to almost 1.8 million Lost lives, and O26 percent of all the events during the period accounted for nearly 80 percent of disaster-linked mortality (United Nations 2009). The events were concentrated in time and space: at least half of the deadliest disasters took place between 2003 and 2008, and 84 percent of the deaths and 75 percent of the destroyed housing were associated with only O7 percent of the loss reports (United Nations 2009). These types of events represent an intensive risk because of their Low-probability of occurrence, but their high impact when they do occur, in contrast to the more common high-prob- ability, lower-impact events that represent a more extensive risk. However, the magnitude of the impacts that intensive-risk events entail masks the extensive risks to which millions of people around the world are exposed each year. in a sample of 12 countries between 1970 and 2007, the United Nations (2009) finds that over 99 percent of local governments reported that 16 percent of the deaths and 51 percent of the damaged housing were associated with such events. 124 We rely on data of the first round of ECVMAS 2012 collected in the fourth quarter of 2012. The survey included a module on the shocks experienced by households and on the coping mechanisms, if any, employed in response to the shocks. Information on 18 separate shocks was collected. For the purpose of the analysis, shocks were grouped into three broad categories: idiosyncratic economic shocks, covariate community-wide economic shocks, and covariate weather-climatic shocks. The id- iosyncratic economic shocks were disaggregated into six categories: health, household composition, agricultural setbacks, loss in nonagricultural economic activity, decrease in outside help, and crime (See appendix L for a list of the specific shocks in each category) The respondents were also asked to identify the three shocks that most affected their households economically and the principal cop- ing mechanism they used when faced with the problem. À total of 35 strategies were accounted for, including no strategy. We have grouped these into 12 groups of coping mechanisms. (See appendix M for a List of the groups and the component strategies.) [page 141] : Investing in People to Fight Poverty in Haïti almost every department was affected by economic downturns. Economic shocks are the most prevalent in the department of Ouest. Crime has become à serious concern: between 16 and 20 percent of the population has been affected by inse- curity across the country. Compared with other low-income economies, these per- centages are high. Heltberg, Oviedo and Talukdar (2013) report a lower prevalence of shocks in Afghanistan (16.4 percent among urban households and 49.0 percent among rural households), Bangladesh (14.0 percent among urban households and 15.9 percent among rural households), Malawi (40.0 percent among urban house- holds and 66.8 percent among rural households), Tanzania (83.4 percent among urban households and 83.3 percent among rural households), and Uganda (297 per- cent among urban households and 56.2 percent among rural households)."° Figure 4.2. Population shares affected by shocks, by department @ Climatic © Disease Economic security 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% a a TD à v v TG v Lrd a ê ê 5 ê È ë à Ê d & el L Z L © [= < = Q [e) TZ D v = Q] 2 5 5 2 OU oO] T Z a 2 E Ê 8 < s o Ô Z Source: ECVMAS 2072; World Bank and ONPES calculations. Note: The survey question was ‘during the last 12 months, has your household been affected by one of the following problems?" Climatic shock = hurricanes, floods, droughts, or irregular rainfall Disease shock = disease other than cholera, cholera epidemic diseases, animal diseases, or crop and plant diseases. Economic shock = death of a household member hosting new members supported by the household: food scarcity, market price increase for seeds, fertilizer, or equipment; broken tools; bankruptcy of a nonfarm household: household lost wages or other income; loss of transfers from parents; loss oftransfers from the government. Security shock = theft of money, goodS, or harvest. 125 By construction, the percentages for Tanzania are higher. The survey in Tanzania reports shocks experi- enced in the past five years, instead of the past 12 months as in ECVMAS and the other country surveys. [page 142] WorldBank - ONPES | While all households in Haiti face multiple economic shocks annually, the poor are more likely to be hit. Households in extreme poverty experienced an average of nearly three shocks per year, whereas resilient households only ex- perienced 2.54 shocks (figure 4.3).7’ Only 4 percent of extreme poor households are unaffected by shocks, against 16 percent among resilient households. Several factors may explain this result. Households in extreme poverty may be more likely to perceive events as shocks because they have fewer means to cope. Alternati- vely, households in extreme poverty could be more prone to shocks given their locational and occupational decisions. Figure 4.3. Number of shocks by welfare levels a. % of households facing shocks, by household poverty status 30% 25% @ one 20% ©: ©? 15% ©: 10% ©: 5 5% 6+ O% Extreme Poor In poverty, but Vulnerable, but Resilient not extreme not poor 126 The poor are those households in which per capita expenditures are above the extreme poverty line, but below the poverty line. Vulnerable households are those that are above the poverty line, but in which per capita expenditures are Less than 20 percent above the line. Resilient households are those in which per capita expenditures are more than 20 percent above the poverty line (that is, the nonpoor). Under these definitions, 29 percent of a sample of 4,930 households are among the extreme poverty; 19 percent are among the poor, 10 percent are vulnerable; and 42 percent are resilient. 127 Total shocks may be slightly underestimated because the survey did not collect information on the number of times a particular type of shock was experienced [page 143] : Investing in People to Fight Poverty in Haïti b. % of households facing shocks, by household location 90% 80% 70% @noshocks 60% @ Idiosyncratic household shock 50% @ covariate Shock 40% 30% 20% 10% 0% Port-au-Prince Other Urban Rural Source: ECVMAS 2012; World Bank and ONPES calculations. Rural households are significantly more likely to be hit by a shock relative to ur- ban households. Rural households experience almost twice the number of shocks experienced by households in Port-au-Prince, 3.29 and 1.85 shocks, respectively. In general, households in the metropolitan area of Port-au-Prince are two times Less likely to experience a shock of any kind compared with households in other urban areas and three times Less likely than rural households. Health shocks are the most common shocks hitting the Haïitian population, followed by weather-climate and economic covariate shocks. Two-thirds of the Haïitian population is regularly hit by idiosyncratic shocks, poor and nonpoor alike. Health-related shocks are the most common. However, 50 percent of the poor and vulnerable face health-related shocks, against 43 percent among the resilient. The most common covariate shocks are weather-climate related. While households in Port-au-Prince report a similar prevalence of climatic shocks as in other low-inco- me countries, shares are much higher in other urban areas and in rural areas, at 44 percent and 73 percent, respectively."?8 Economic or agricultural setbacks affecting the community are the third-most common shocks. The poor in rural areas are more likely to be hit by agricultural and climatic shocks, while, in urban areas, economic shocks affecting labor incomes and private transfers are more common. The poor are much more likely to be hit by agricultural setbacks (33 percent among the extreme poor, against 18 percent among resilient households) and climatic shocks (73 percent among the extreme poor, against 46 percent among the resilient) (table 41). Climatic-weather shocks are likely to be associated with agricultural setbacks (a correlation coefficient of 128 Although the pattern is similar to what is found in other Low-income countries, the actual prevalence of weather shocks is generally higher in Haiti than in the other five Low-income economies (Heltberg, Oviedo, and Talukdar 2013). The comparison economies report climatic shocks among between 32 and 39 percent of the total population [page 144] WorldBank - ONPES | O3) and to be more prevalent among the rural population: 73 percent of rural households have been economically affected by such shocks.?° The relatively low incidence and impact of climatic shocks among the resilient is most probably linked to the fact that most of these households are in urban areas (68 percent), against only 18 percent of the extreme poor. The shocks associated with an inci- dence that increases as welfare is enhanced are those shocks affecting economic activity or involving crime, which is more common in urban areas. Idiosyncratic economic shocks (19 percent)—such as failure of a nonagricultural business or wage Losses not caused by illness—and economic shocks caused by a decrease in the transfers received from family, friends, or the government (15 percent) are more common in urban areas, reflecting the higher reliance of urban households on Labor income and private transfers. Table 41. The prevalence of types of shocks faced by households, by poverty status Type of shock In extreme pouerty In pouerty, but not extreme | Vulnerable, but not poor Resilient Idiosyncratic household shocks Economic shock affecting community Source: ECVMAS 2072; World Bank and ONPES calculations Man-headed households with children are more likely to be hit by shocks. Households with children are more likely to be affected by a health shock, a household composition shock, an agricultural setback, a loss in econo- mic activity, or a covariate economic shock than households without children (ta- ble 4.2). Woman-headed households are Less likely to experience a shock than man-headed households. Because man-headed households are prevalent in ru- ral areas (61 percent of rural households are headed by men), only 16 percent of woman-headed households overall experience agricultural setbacks, whereas 31 percent of man-headed households experience such shocks; likewise, 62 percent of man-headed households are affected by a weather-climatic shock, compared with only 49 percent of woman-headed households. This evidence may reflect the fact that most of the women are employed in nonfarm activities (see chapter 2) even in rural areas; 50, they are Less vulnerable to climatic or agricultural shocks. 129 These differences do not imply that a weather-climatic event is necessarily more likely in rural areas, but that such events are more likely to be felt economically by rural households versus urban ones | [page 145] : Investing in People to Fight Poverty in Haïti Table 4.2. The prevalence of types of shocks, by household type Type of shock With children Without children Man-headed Woman-headed Idiosyncratic household shocks Economic shock affecting community Source: ECVMAS 2012; World Bank and ONPES calculations ** p <O.01 * p <005 The impact of shocks Households, especially resilient households, perceive idiosyncratic shocks as more severe than covariate shocks."° For more than 60 percent of the households, idiosyncratic health shocks are the most severe shocks experienced in economic ter- ms. The second- and third-most severe shocks are covariate: weather- or climate-re- lated shocks and economic shocks or agricultural setbacks affecting the community, respectively. Among resilient households, 60 percent perceive idiosyncratic shocks as the most severe, against only 25 percent for covariate shocks because of the relatively infrequent incidence of weather-climate shocks among this category of households. Contractions in income or assets or in food consumption are the main econo- mic consequences of shocks.!‘! Health, weather-climatic, and economic shocks all lead to a reduction in incomes, which is perceived as their biggest consequence by all households, but particularly by the vulnerable. Income losses are followed in im- portance by reductions in assets and in food purchases. For the main shock (which is, most often, health related), 53 percent of households in extreme poverty suffered from reductions in food production, against 34 percent among resilient households, reflecting the greater significance of production for home consumption among the poor (table 4.3). Among the extreme poor, reductions in food production are the se- cond-most severe impact after income losses in the case of the second and third shocks in order of severity (weather-climate and economic shocks).5? 130 Idiosyncratic shocks are the most important ones for 60 percent of the population. If a second shock occurs, it is as likely to be an idiosyncratic as a covariate shock, and, if a third shock occurs, it is more likely to be covariate. The pattern holds even if the sample is limited to those households that have experienced both kinds of shocks at least once. 131 The survey explored the types of economic impacts that the three main shocks had on households. The potential self-reported economic impacts of shocks are categorized as a reduction in (a) income, (b) assets, (c) food production, (d) food stocks, and (e) food purchases. 132 The shocks are organized according to their importance to the household, not chronological order. | [page 146] WorldBank - ONPES | Table 4.3. The economic impact of shocks, by household poverty status. Percent reductions among households in each category unless indicated otherwise Indicator AL In extreme pouerty In pouerty, but not extreme Vulnerable, but not poor Resilient Source: ECVMAS 2072; World Bank and ONPES calculations Coping mechanisms The solidarity of friends and family and reductions in food intake are the main coping strategies adopted by households faced with shocks. The most com- mon mechanisms for coping with the most important shocks are monetary help from others (27 percent), changing nutritional inputs (16 percent), and doing no- thing (5 percent). Changes in nutritional inputs are especially important for coping with covariate economic (48 percent) and weather-related shocks (24 percent), which most probably affect agricultural production or Labor incomes. Thus, for the most part, households are able to cope with idiosyncratic shocks without resor- ting to changing their nutritional inputs; however, nutritional inputs are Less well protected if a household experiences a covariate economic or weather shock. The most common strategy for coping with idiosyncratic shocks is monetary help from outside the household or not using any coping strategy at all (box 41). Among households that experience health shocks, 41 percent resort to asking for mone- tary help from others. | [page 147] : Investing in People to Fight Poverty in Haïti Box 41. Formal and informal mechanisms for risk management: financial inclusion Formal financial services can be important instruments for poor peo- ple to cope with shocks; however, the overall access to and usage of them is relatively limited in Haiti. According to Global Findex data, based on a survey conducted in 2011 on 504 individuals, only 27 percent of Haïitian adults (age 18 and older) have an account in a formal financial institution, compared with 45 percent in Latin America and the Caribbean, and 29 per- cent in other Low-income countries. Among the entire population, only 11 percent have medical or health insurance. Only 24 percent of Haitians (and 8 percent of the poor) reported having saved formally in the previous year. The population in the bottom 40, those with Low levels of formal education, and youth reported the lowest levels of usage of formal financial services. Lack of resources and access are among the main reasons households do not use formal financial services. The scarcity of resources (to make use of bank accounts or to open and maintain them) is the main reason reported by the Haitian population for not using formal financial institu- tions (figure B411). The second-most important reason is the lack of a carte d'identification nationale (national identity card, CIN) or necessary docu- mentation, reflecting partly the high level of informality in the economy and the weakness of institutions. Despite the limited coverage of banks and cooperatives in the territory (currently only 273 branches exist in the country, most of which are located in Port-au-Prince and a few other major urban areas), physical access does not seem to be among the main cons- traints facing Haitians in opening formal accounts. Figure B41.1. Reasons for not having an account at a financial institution 35% 29% 31% 20% 18% 15% 16% 15% 12% 5% 4% 4% ax 4% They aretoo Theyaretoo Idon'thavethe Idon'ttrust ldon'thave Because of Because far away expensive necessary them enough money religious someone in my documentation to use them reasons family already (ID, wage slip) has an account fn C] Haïitian adults (% age 18+) e Income, bottom 40% (age 18+) [page 148] WorldBank - ONPES | Despite the limited access to formal financial institutions, Haitians do need money to invest in the future and cope with risk and often use informal institutions to access it. Among the Haitian adult popula- tion, 67 percent (and 55 percent of the poor) reported they had taken out a Loan in the year preceding the survey, reflecting a significantly higher usage of Loans than in other Low-income countries. Only 10 percent ofthe respondents cited institutional lenders as their source of credit because most loans were provided by family or friends (at a significantly higher rate than in other Latin American or low-income countries), followed by private money lenders, whose services are reportedly expensive. In Haïti, lending is particularly important for coping with health issues and emer- gencies and to pay school fees: 27 percent of Haitians over 15 years of age reported they had taken out a Loan to cope with health issues or emer- gencies, and 28 percent said they had done so to pay school fees (compa- red with 16 and 7 percent, respectively, among all low-income countries). To facilitate the access of the poor to financial services and to improve the quality of these services, the government of Haiti has engaged with the private sector in launching several initiatives. These initiatives aim at facilitating access to financial services through digital wallets and cell phones, at increasing the number of points of service through nonbank agents, and at piloting innovative schemes, such as the delivery of condi- tional cash transfers (for example, the payments in the Ti Manman Cheri scheme) through mobile phones and remittance agents. Efforts are also under way to define a more comprehensive strategy for financial inclu- sion to would help address, in an integral manner, a wide range of issues that hamper the supply and use of financial services by the poor and by micro, small, and medium enterprises, such as the absence of a proper consumer protection framework; deficiencies in the legal and supervisory framework that governs financial cooperatives, microfinance institutions, and insurance companies, or the difficulties that many poor people face in gaining access to financial services because they do not have proper identification. Shocks are more likely to impede the future economic activities of house- holds in extreme poverty, while resilient households largely rely on private transfers. Households in extreme poverty are two times more likely than resilient households to sell assets to cope with shocks, at 10 and 4 percent, respectively. They are also marginally more likely to take on debt: 16 percent of households in extreme poverty use debt as their main coping strategy, against 12 percent of resi- lient households. Meanwhile, resilient households are two times more likely than households in extreme poverty to rely on (nonloan) monetary help from outsi- ders, at 38 and 16 percent, respectively. In particular, 54 percent of resilient house- holds tap into monetary help from others to counter the effects of health shocks, whereas only 26 percent of households in extreme poverty are able to do so. | [page 149] : Investing in People to Fight Poverty in Haïti Shocks can generate important losses in human capital, especially among the poor. Changes in household composition (the death or birth of a household member) or a drop- off in monetary help from outside the household are the two events that are more likely to lead to the removal of a child from school. The use of this mechanism of removing a child from school is prevalent among households in extreme poverty. Households in ex- treme poverty are also two times more likely than economically resilient households to change their nutritional inputs, at 23 and 10 percent, respectively. If a covariate economic shock strikes a community, 58 percent of households in extreme poverty change their nutritional profile as opposed to 36 percent of resilient households. Not only are house- holds in extreme poverty more likely to change their nutritional intake, but households in extreme poverty report a higher incidence of covariate shocks (see table 41). Haitians are less able to cope with intensive-risk disasters than with exten- sive-risk events. Extensive-risk shocks such as idiosyncratic health or economic shocks are usually high-probability, lower-impact events that Haitians have learned to address primarily by selling assets or by relying on their extended social network for loans or on transfers from family, friends, or NGOSs (figure 4.4). Over two-thirds of households affected by idiosyncratic unemployment shocks and over 70 percent of households that reported they were affected by idiosyncratic diseases were able to cope by selling assets or by relying on their friend and family networks. However, if an intensive shock occurs, such as a climatic event (for example, hurricanes, floods, or droughts) or an epidemic (such as cholera), the ability to sell assets plummets to around 10 percent of these households, and not much more is gained through the networks of these households. It is plausible that, because assets are damaged in some climatic events, they lose their market value, and, because an entire region can be affected by a climatic or health shock, households are Less able to rely on ne- tworks. Government aid plays almost no role in helping Haitians cope with a shock; a strategy to prepare for, mitigate, and respond to disasters is therefore needed. | [page 150] WorldBank - ONPES | Figure 4.4. Coping strategies, by type of shock 100 In the face of shocks, » 5 households often 6% 80% make costly trade- É ë 60% offs that sacrifice $ E 40% long-term welfare 5 for immediate 5 à 20 benefit: 56 percent 8 o of households in Disease Unemployment Social transfers Cholera Cyclones and Drought extreme poverty stopped floods change their food @rore @ other consumption, @ migration @ child's school withdrawal which can lead @ Government @:ors to malnutrition, , @ New work activities @ Natural resources stunting or anemia. @ reduced consumption/expenses © relied on social network @ sold assets Source: ECVMAS 2012; World Bank and ONPES calculations. Note: Cholera, hurricanes, floods, droughts, disease, unemployment, and a termination of social transfers are shocks that can be considered intensive-risk events. Multivariate analysis Covariate weather-climate and economic shocks have a negative impact on wel- fare. The cross-sectional analysis determines the extent to which income groups resort to various coping strategies depending on the type of shock and after one controls for household characteristics. The results confirm that covariate shoc- ks are associated with lower per capita expenditures across the population (see appendix N). Covariate economic shocks are associated with about 12 percent Less per capita expenditures, and covariate weather shocks are associated with about 15 percent less per capita expenditures. Furthermore, households that resort to changing nutritional inputs to counter the effects of a shock spend significantly less per capita than households that have not experienced a shock: 24 percent less in the case of covariate economic shocks, 30 percent Less in the case of cova- riate weather shocks, and 25 percent Less in the case ofidiosyncratic shocks. Hou- seholds that take on debt or use another strategy besides the five main strategies to cope with a weather-related shock spend Less per capita than households that have not experienced a shock. | [page 151] : Investing in People to Fight Poverty in Haïti The poverty-disaster vulnerability relationship In most regions in Haiti, the poor are more likely to be affected by a climatic shock, which is reflection of the incidence of poverty. The share of people affected by a natural shock varies considerably from one department to another. Nonetheless, in all departments, the poor are disproportionately affected. Indeed, in the poorest departments (Grand’Anse, Nord-Est, and Nord-Ouest), between 78 and 82 percent of the affected population is poor. In contrast, Ouest is the least vulnerable: only 43 percent of the population is affected by shocks; of these people, only 19 percent are poor, while 23 percent are nonpoor (figure 4.5). Figure 4.5. Climatic shocks and poverty, by department, 2009 90% Fe R2 = O.61745 Nord-Est - F S 80% ‘6 st Grand ASe O\cid Ouest [°3 = © centre po 70% = rd Nippes Le) Sud S 60% © $ [®@) | Sud-Est 2 50% Artibonite Q [°: 5 £ 40% uest = un 30% 40% 45% 50% 55% 60% 65% 70% 75% 80% Percentage of population affected by a climatic shock (Vulnerability) Source: ECVMAS 2012; World Bank and ONPES calculations. Note: Climatic shock = hurricanes, floods, droughts, and irregular rainfall The poverty line is set at G 2990987 Departments are classified by level of vulnerability. Vulnerability levels are set according to the share of people affected by a climatic shock The size ofthe bubbles in the figure refers to the relative size ofthe relevant population in 2009 There is a direct relationship between a department’s vulnerability to natural disasters and the level of poverty among the population. The poorer an indivi- dualis in Haïti, the more vulnerable is the individual to natural disasters (figure 4.6). People may be vulnerable to disaster because they Live in places susceptible to one or more natural hazards or because their behaviors and local and national regula- tions are inadequate to reduce risk. A proxy for vulnerability to natural disasters is the number of people in a department who are affected by a given event. The use of ECVMAS to calculate a poverty headcount for every department and to determine the degree to which the department is vulnerable to natural disasters according to the share of the population affected by natural shocks makes it possible to show that there is a direct relationship between vulnerability and poverty. [page 152] WorldBank - ONPES | Figure 4.6. Poverty and vulnerability in Haïti. Population in poverty by vulnerability zones & ë 8 à ë S 70% e S & 5 5 60% & ä È e 3 £ ES 2 50% È S È [= Del S 40% & e à 30% 2 £& 20% © 2 10% £ mn O% Vulnerability Very low Low Medium High Very high @rovety headcount extreme poverty headcount Source: ECVMAS 2072; World Bank and ONPES calculations Note: The poverty line is set at G 2990987 The extreme poverty line is set at G 1524003. Departments are classified according to their vulnerability. Vulnerability levels are based on the share ofthe total population affected by a climatic shock. The categories were (1) very low Ouest), 2) low (Nord and Nord-Est), (5) medium (Artibonite Grand Anse and Nippes) () high (Centre Sud and Nord-Ouest) and (5) very high (Sud-Est). Haïiti's hazards Haiti is one of the countries most exposed to hazards in the world, making it particularly vulnerable to the associated economic losses. Over 93 percent of Haiïti's surface and more than 96 percent of its population are at risk of exposure to two or more hazards. According to these measures, Haiti ranks fifth in the world in exposure to risk to two or more hazards (World Bank 2005). With every event, whether hurricane, flood, earthquake, landslide, or drought, there is an economic consequence: 56 percent of the GDP of Haiti is linked to areas exposed to risk stemming from two or more hazards. While Haitis vulnerability derives in part from its geographical location, part also derives from internal or institutional factors. The comparison between the Dominican Republic and Haiti, which share the island of Hispaniola, highlights three key differences (table 4.4). First, the number of weather events from 1980 to 2010 was 63 percent higher in Haiti than in the Dominican Republic, suggesting that the greater vulnerability of the former causes particular hazards to become disaster events more easily. Second, although both countries experienced similar numbers of storms, Haiti had more than twice as many floods as a consequence of the stor- ms (figure 47). Floods are one of the most common weather-related events that affect Haiti and partly arise because of the severe deforestation that has weakened and impoverished the land, unlike the Dominican Republic. Third, Haiti's greater 133 In 2009, forest coverage was 3 percent in Haïti, against 47 percent in the Dominican Republic (see ONPES forthcoming, based on Collier [2009)). [page 153] : Investing in People to Fight Poverty in Haïti vulnerability is reflected in the consequences of these events in terms of human and economic losses, which reflect also chaotic migration from rural to urban areas, the consequent inadequacy of buildings and building codes, and lack of diversification in sources of income.** While the events occurring in Haiti since 1980 resulted in over 230,000 deaths and nearly $9 billion in damage, the Dominican Republic had fewer than 1,500 deaths and $2.6 billion in damage. At an annual average of over 5284 million, Haïtis costs are more than three times higher than the costs ofits neighbor. Table 4.4. Disasters in the Dominican Republic and Haiti compared, 1980-2010 Disaster Haiti Dominican Republic Source: EM-DAT (OFDA/CRED International Disaster Database), Centre for Research on the Epidemiology of Disasters, Université Catholique de Louvain, Brussels (data version: 1108) http:/ wwwemdatbe/database Note: The 2010 earthquake in Haiti was responsible for 95 percent of the deaths and over 90 percent ofthe economic damage during the period. Excluding the 2010 earthquake from the table would yield a different picture in terms of economic consequences: the Dominican Republic would become more susceptible than Haiti to economic loss. One possible interpretation of this is the greater asset exposure ofthe Dominican Republic relative to Haïti However, the number of deaths is still 26 times higher in Haïti than in the Dominican Republic. Figure 4.7. Number of disaster events, by type, Dominican Republic and Haïti, 1980-2010 45 40 35 30 25 20 15 10 5 u EN US > e a TT D EE = 4 À Drought Earthquake Flood Storm Other Epidemic @ Haiti @ Dominican Republic Source: EM-DAT (OFDA/CRED International Disaster Database), Centre for Research on the Epidemiology of Disasters, Université Catholique de Louvain, Brussels (data version: v11.08), http:/www.emdatbe/database 134 The population of Port-au-Prince rose from 400,000 to 3 million over the last 40 years (ONPES forthcoming) [page 154] WorldBank - ONPES | Coastal areas and Port-au-Prince are the most vulnerable to weather and other natural events. The most common extreme weather events in Haiti are storms, floods, draughts, earthquakes, and landslides. Al such events are com- mon across the country, but particularly in coastal areas and Port-au-Prince. Stor- ms, floods, and droughts are all caused by a lack of watershed protection and deficiencies in irrigation (World Bank 2013b). Urban areas and rural populations in coastal areas are particularly vulnerable because the forces favoring urbanization and market pressures to seize land for agriculture usually remove vegetation and thereby destroy buffer zones, which increases the vulnerability of these areas. Both urban and rural areas suffer the consequences of these shocks. Urbani- zation patterns account for part of the economic losses caused by these events because of the damage to housing and infrastructure, the disruption in Logistics and transportation chains, and loss of life. Rural areas bear a larger share of the costs in terms of losses in agricultural produce, which have an impact on food security and livelihoods. Haiti’s hazards have larger consequences not only because of the country's geological, geographical, and developmental challenges, but also because of institutional weaknesses, including inadequate planning and lack of regu- latory enforcement. Hazards can be divided in Haiti into natural and anthropic hazards. The former includes earthquakes, floods, and torrential floods, and the latter includes fires and accidents during the transport of dangerous materials. Historically, the deadliest hazard in Haïti is seismic activity, of which the triggers are still being researched. However, the consequences of seismic activity are sig- nificantly related to human decisions on how and where to build. Building codes, deficient urban planning, and other institutional weaknesses amplify these conse- quences (table 4.5; box 4.2). Similarly, urban planning and building codes are also amplifying factors in the consequences of floods. Given that laws in Haïti impose restrictions on building in natural drainage areas, it is possible that, as with anthro- pic hazards, a lack of regulatory enforcement may aggravate the consequences of disasters (CIAT 2013). Natural hazards can slow or stop growth and development, gen rating destruc- tion and diverting public investment to emergency reconstruction operations. | [page 155] : Investing in People to Fight Poverty in Haïti Table 4.5. Triggers and consequences of hazards in Haïti Hazards Predisposing factors Triggering or aggravating factors Socioeconomic amplification factors Natural hazards Proximity to large structures, Inadequate building codes, deficient Floods Surface flatness and organiza- Storms, hurricanes, exceptional rain | Inadequate building codes, deficient tion ofthe hydrographic network | in terms ofintensity and duration urban planning, institutional weaknesses Torrential floods Settlements in alluvial fans or Storms, hurricanes, exceptional rain | Inadequate building codes, deficient glens in terms of intensity and duration urban planning, institutional weaknesses Anthropic hazards Transport of dan- | Dangerous products, inadequate . . . . Fire REmnENE CERENUERE Human accidents, traffic congestion | Lack of regulation or enforcement materials Source: Adapted from Mathieu et al 2003. Box 4-2. The disaster risk management strategy in Haïti Haiïtis Système National de Gestion des Risques et des Désastres (National Disaster Risk Management System, NDRMS) was established in 2001 by 10 key line ministers and the president of the Haïitian Red Cross. The NDRMS has achieved significant results in disaster preparedness and response since its inception: while the 2004 hurricane season resulted in 5,000 casualties and over 300,000 affected people, the Fay, Gustav, Hannah, and Ike Hurri- canes resulted in a combined total of fewer than 800 casualties and over 865,000 affected people. Strong collaboration between the key members of the NDRMS and its technical and financial partners was critical to improving the speed and efficiency of the response capacity. However, the 2010 crisis following the earthquake was beyond the capacity of the NDRMS. A focus on the impact of the 2010 earthquake In January 2010, a magnitude 7.0 earthquake struck Haiti, the most powerful in over 200 years, causing hundreds of thousands of deaths and injuries, sending thou- sands more into homelessness or displacement, and inflicting tremendous infras- tructural damage on water and electrical infrastructure, roads, and port systems in the capital, Port-au-Prince, and surrounding areas. The force of the earthquake was felt miles from the epicenter. Shaking intensity can be measured according to the Modified Mercalli Intensity scale, which is a measure of the severity of an earth- quake. Gauged according to the scale, there was a major area of intensity close to the epicenter, the departments of Ouest and Sud-Est (map 41). Haiti was then struck by a cholera epidemic in October; four months later, some 4,500 deaths had been reported. Following the disaster, the human toll was extremely severe: 2.8 million people were affected by the earthquake, which caused over 200,000 deaths and more injuries. Over 97,000 houses were destroyed, and some 188,000 were dama- ged. Over 600,000 people fled to nonaffected regions (Échevin 2011). [page 156] WorldBank - ONPES | Map 4.1. The shaking intensity of the 2010 earthquake Legend nn. CL] Departments Modified Mercalli Intensity-damage 1 Damaging 1. Imperceptible : 8. Heavely damaging 2. Scarcely felt . 9. Destructive 3. Weak D. 10. Very destructive 4. Largely observed . 11. Devastating 5. Strong k 12. Completely devastating 6. Slightly damaging n Source: Based on data of"Shakemap us2010a6; Earthquake Hazards Program, United States Geological Survey, Reston, VA http:/earthquakeusgsgov/earthquakes/shakemap/global/shake/20105a6/. Although the earthquake affected all the communal sections (sections commu- nales) of the country, the impact was mainly concentrated in two departments. Among these communal sections, 30 were destroyed; 38 were heavily damaged; and 54 were damaged (figure 4.8). These communal sections are mainly in the departments of Ouest and Sud-Est, where almost 40 percent of the national po- pulation was concentrated in 2009. Additionally, 56 communal sections where slightly damaged, mainly in the departments of Artibonite, Centre, and Nippes. The remaining communal sections were largely or considerably damaged. Figure 4.8. Damage among communes as a result of the 2010 earthquake. Communal sections and population affected, by shaking intensity @ communal sections @% National population y 300 35% 6 29% 5 5 250 27% 30% 2 245 ë _ 25% g 200 20% ce È 20% © £ 150 S 8 147 15% à 5 100 10% ê = 8% 10% = rs] 6% [°] 2 Be E 50 56 54 5% 3 38 [e] 0% Largely Strong Slightly Damaging Heavily Destructive observed (mmi 5) damaging (mmi7) damaging (mmi 9) (mmi 4) (mmi 6) (mmi 8) Source: Based on data of ‘Shakemap us2010rja6,' Earthquake Hazards Program, United States Geological Survey, Reston, VA, http:/earthquakeusgs.gov/earthquakes/shakemap/global/ shake/2010rja6/; population estimates: IHSI. [page 157] : Investing in People to Fight Poverty in Haïti The 2010 earthquake destroyed large numbers of dwellings and resulted in the loss of many jobs, though to a lesser extent. Nationwide, 41 percent of all dwellings were damaged. The share was much larger in the departments of Ouest (61 percent) and Sud-Est (54 percent). The share of dwellings damaged and the remuneration lost were smaller in Ouest (13 percent) and Sud-Est (12 percent); national figures stood at 8 percent. In 7 percent of the cases, only dwellings were damaged, but no remu- neration was lost; the share was slishtly larger in the departments that bore most of the damage. À large share of the families that saw their dwellings damaged had no employment prior to the earthquake. The dwellings were damaged of more than one in four families that had no jobs before the earthquake. The corresponding share was higher in Sud-Est (33 percent) and Ouest (37 percent). Among households, 75 percent believe their living standards have deteriorated since the earthquake (figure 4.9). Figure 4.9. Perceptions of Living standards after the earthquake @oesraded @aintained @ improved 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% £ & ÿ a £ £ g 3 £ a EH Ë + ä ? È 3 < $ Ü Source: ECVMAS 20172; World Bank and ONPES calculations. Vulnerability is extensive in Haiti. One million people live slightly above the po- verty line and could be pushed below the line by a shock; almost 70 percent of the population is either poor or vulnerable to falling into poverty. The consumption level of only 2 percent of the population exceeds $10 à day, which is the region's income threshold for joining the middle class. [page 158] WorldBank - ONPES | Haïitians face frequent covariate and idiosyncratic shocks. The most common covariate shocks are weather-climate related. Economic shocks are also common in Haïti because of international fluctuations in import or export prices and the volatility of remittances. Political instability has haunted the country for several decades and can affect welfare if it results in an interruption or slowdown in eco- nomic activity or official development assistance. Haitians also face considerable idiosyncratic shocks such as death, illness, job loss, and declining wages. While a typical Haitian household faces multiple shocks each year, poor and rural areas are even more vulnerable. Nearly 75 percent of households are eco- nomically impacted by at least one shock each year. The extreme poor are more vulnerable to shocks and the consequences of shocks: 95 percent experience at least one economically damaging shock each year. Rural households experience almost twice the number of shocks affecting households in Port-au-Prince. Health shocks and covariate weather-related shocks are the most common and most severe and affect the poor more heavily. The poor in rural areas are more likely to be hit by agricultural and climatic shocks, while, in urban areas, economic shocks affecting labor incomes and private transfers are more common. The poor are much more likely to be hit by agricultural setbacks (33 percent among the extreme poor, against 18 percent among resilient households) and climatic shocks (73 percent among the extreme poor, against 46 percent among the resilient). Both types of shocks are more pre- __ valent in rural areas. Idiosyncratic economic shocks and economic shocks caused IE criticalto : . . É : have effective risk by a decline in the transfers received from family, friends, or the government are management and more common in urban areas, where most of the resilient live, reflecting the hi- social protection gher reliance of urban households on labor income and private transfers. strategies in place to Lessen the The poor are less successful in coping with shocks, and, if they have a stra- impact of shocks on tegy, it is more likely to impede future economic activities or human capital Haïti's poorest and accumulation. Haitians, especially the poor, lack formal instruments to manage mort Vunerables risk effectively, such as social protection programs and formal financial products, ability to cope with and rely on informal mechanisms, such as private transfers or debt, to mitigate their effects the chocks ex post. As a result, most households do nothing (as in the case of cholera and weather-related shocks), suggesting that the poorest are unable to cope with shocks and adopt coping strategies that damage human capital. Ove- rall, 23 percent of households in extreme poverty changed their nutritional profile in response to a major shock, and 58 percent did so in response to a covariate economic shock. Natural disasters have a great disruptive potential because of the geographi- cal position of Haïti, institutional weaknesses, and shortages of the resources needed to prepare for, mitigate, or cope with shocks at the macro and micro | [page 159] : Investing in People to Fight Poverty in Haïti levels. Haitis hard-earned development gains are often jeopardized by adverse na- tural events that generate destruction of key human and infrastructural resources and divert development funds to emergency and relief operations. In light of the high incidence of shocks, several priorities for policy actions emerge, as follows: Priority 1: Assess the needs in social protection and possibly expand coverage among the poor and vulnerable to protect their assets and livelihoods. in the face of the high incidence and vulnerability to idiosyncratic or covariate shocks, the poor and vulnerable have limited access to public support. Most assistance is provided through remittances or support from churches, other nongovernmental institutions, and donors. Access to formal safety nets could allow the poor and vul- nerable to smooth their consumption over time and prevent irreversible losses of human capital, as well as avoid destitution. In order to define the most appropriate support measures, however, a thorough understanding of the hazards and house- hold coping strategies is needed: this study represents a first step in that direction. Priority 2: Mainstream disaster risk management activities into all growth, de- velopment, and poverty reduction strategies to facilitate the transition from an approach based on living at risk to an approach based on living with risk. To ensure that the transition from the emergency response phase and the recons- truction phase after the earthquake is effectively brought to an end, it is important to continue strengthening and mainstreaming disaster risk management activities and make sure disaster risk management becomes à core component of a sus- tainable poverty reduction and economic growth strategy. Disaster risk manage- ment has already been included as a key cross-cutting priority in the government's Poverty Reduction Strategy Paper (2008-11) and as à principle pillar of the Uni- ted Nations Development Assistance Framework (2009-11), as well as the World Bank's Country Assistance Strategy (2009-11), the Joint Aid Effectiveness Program (Programme conjoint d'efficacité de l'aide - PCEA) for 2013-2016 of the of Coordina- tion framework of the External Aid for Development (Coordination de l’aide externe au développement - CAED) and other disaster risk management projects (see box 41. More recently, the Post-Earthquake Disaster Needs Assessment 2010 and the Action Plan for National Recovery and Development of Haiti present disaster risk management as a cross-cutting priority in both the public and private sectors and as an opportunity to promote (1) decentralization, (2) a stronger civil society, and (3) an innovative private sector. Overall, this demonstrates a growing consensus within the government and among technical and financial partners of the importance of integrating disaster risk management as a critical component of a successful pover- ty reduction and economic growth strategy. Priority 3: Strengthen the capacity of the NDRMS; this capacity is Low and has no legislative underpinning. To achieve and sustain growth, Haïti requires the robust | [page 160] WorldBank - ONPES | institutional and operational capacity to manage multiple risks and respond to di- sasters. This calls for à broad review of the NDRMS and a discussion of institutional and policy options for each of the following actions: a. À key initial step toward upgrading the management of disaster risks inuol- ues improuing the identification and understanding of disaster risks in Haiti by quantifying and anticipating the potential impacts of natural hazards on Haïi- tian society and the economy. The Civil Protection Directorate and the Ministry of Economy and Finance could start by enhancing their disaster data management system and damage and loss assessment procedures and by keeping track of the historical disaster data. This information is essential to the assessment of disaster risks and the design of any disaster risk financing instrument. Beyond assessing the damage and losses from actual events, the development of hazard maps, the building of an exposure database, and the spatial analysis of risks are key elements in fostering appropriate investment and territorial planning, a fact that is recognized in article 149 of the decree of October 12, 2005 (CIAT 2013). b. Reducing existing risk and avoiding the creation of new risks by integrating risk awareness in public policies and investments. Disaster-risk-related information can guide investments in addressing existing risks. The retrofitting of critical buil- dings, the construction of emergency safety infrastructure, and the rebuilding of natural ecosystems are examples of the disaster mitigation investments needed in Haïti. However, these engineered structural measures must be accompanied by, for example, adequate policies and programs to promote improved territorial planning and building regulations and practices to avoid the creation ofnew risks. c. Improuing the capacity to manage disaster-related emergencies. Strengthening institutional arrangements for emergency and preparedness, including à fully functioning National Emergency Operations Center, remains à top priority. Esta- blishing a fully operational chain of command supported by emergency respon- se plans, simulation exercises, and adequate warning and communication-sensi- tization systems requires strong national leadership and corresponding political traction. d. increasing the resilience of government and households. Financial protection strategies, particularly if they are designed to meet the needs of the population in extreme poverty, can help protect the government and households from the economic burden of shocks and disasters. Haitis government is à member of the Caribbean Catastrophe Risk Insurance Facility, which allows the country to purchase insurance coverage to finance immediate postdisaster recovery needs, based on parametric triggers (the occurrence of a predefined event rather than an assessment of actual losses). Other parametric insurance instruments, such as index-based insurance or weather-based agricultural insurance products, could also be explored, but often face technical difficulties in modeling the risk. | [page 161] : Investing in People to Fight Poverty in Haïti Coverage of correctly identified risks is also a key condition of successful risk transfer. If the probability of the insured event is too large (for all high-probability, low-scale events), then the cost can become prohibitive in the absence of a sub- sidy. Finally, the use of subsidized microinsurance services could also be piloted as an alternative to social safety nets for vulnerable populations. e. Clearly define the institutional and budgetary framework of the NDRMS, in- cluding roles and responsibilities for the series of institutions that are invol- ved in civil protection and risk management. Providing the NDRMS with a new legal and institutional framework thus becomes paramount. In addition, the NDRMS requires long-term budget planning, particularly planning for staff and recurrent expenditures. | [page 162] WorldBank - ONPES | Chapter 5: Poverty and social protection This chapter describes the access to social protection in Haïti. The findings show that, in the face of significant poverty and numerous vulnerabilities throughout the life cycle, few of the poor have access to social protection or to safety nets. First, access to social security is out of the reach of most Haitians, but particularly the poor. Second, only a small share of the Haïitian population benefits from social protection. Third, because of Low coverage and limited generosity, social protec- tion benefits are inadequate and play only a marginal role in reducing poverty and inequality and in improving opportunities for the population. The poorest groups—rural residents and children, especially young children—receive a dispro- portionately small share of the benefits. The costs generated by this lack of effec- tive protection for the poorest households are thus high. Meanwhile, public policy has recently begun focusing on strengthening social protection so as to accelera- te poverty reduction. The government's umbrella initiative, EDE PEP, represents a positive effort to create new programs to address important social risks. However, administrative data and the authors’ confirm that the coverage of social protec- tion programs and the coordination and coherence across programs, especially among the poorest and in rural areas, must still be significantly improved. The previous chapters highlight the potential role that effective and well-targe- ted safety nets could play in mitigating the significant poverty and vulnerability in Haiti (chapter 2), enhancing the access and use of health care and education and promoting human capital (chapter 3), and helping the poor manage shocks and risks (chapter 4), while connecting them to skills-building and income generating opportunities. Looking at evidence from the ECVMAS and complementary sources, this chapter explores the access of Haitian households, especially the poorest, to the public provision of safety nets. This issue is important because the gover- nment has traditionally been able to play only a limited role in providing à safety net to the poor and vulnerable. Devereux's Catch 22 (2000) of social protection, “the greater the need for social protection, the Lower the capacity of the state to provide it” is particularly true in fragile states such as Haiti (Harvey et al. 2007, 4). In the face of economic shocks or natural disasters, the poor have limited access to public support, and most assistance continues to be provided through remit- tances or support from churches, other nongovernmental actors, and donors. The literature points to the Low coverage and often ad hoc or limited nature of pro- grams, which frequently cover only small geographical areas or narrowly defined sets of beneficiaries (Lamauthe-Brisson 2013; Lombardo 2012). The massive 2010 earthquake that ravaged the country both exposed and exacerbated Haiti’s lack of 135 This chapter is based on Strokova, et al. (2014), a background paper for the study by the World Bank and Observatoire National de la Pauvreté et de l'Exclusion Sociale (ONPES). 2014. Investing in People to Fight Poverty in Haiti, Reflections for Evidence-based Policy Making. Washington, DC: World Bank | [page 163] : Investing in People to Fight Poverty in Haïti a coherent safety net system. However, quantitative evidence has been lacking on the actual access of households to social protection; hence, the value added of this chapter at a time when the government has set out to develop a social protection strategy as part of its antipoverty program. This chapter provides evidence on access to social protection instruments in Hai- ti, mainly based on ECVMAS 2012. Given the fragmentation and low coverage of programs, it is nearly impossible to obtain comprehensive data on these interven- tions through nationally representative household survey data. Nonetheless, the results are indicative and could serve as a basis for a more detailed study based on alternative data sources, such as administrative data. The chapter does not aim to be comprehensive. For example, the analysis of programs is mainly limited to recent government platforms or initiatives and does not reflect the wide scope of donor initiatives. The chapter is organized as follows. The next section outlines the conceptual and policy framework. Based on the ECVMAS findings and complementary sources of information. The following section summarizes the needs of Haitians for social pro- tection interventions based on the diagnostic provided in previous chapters and using the life-cycle approach to social protection needs. The subsequent section assesses the extent to which these needs are addressed today in Haïti. The final section concludes. Policy framework Social protection includes a variety of interventions that can be tailored de- pending on purpose, target group, and context. Typical interventions include conditional or unconditional cash or part cash transfers, food transfers through food distribution, nutrition programs, school feeding programs, sales of subsidized food, universal subsidies to cover food and energy expenditures, labor-intensive public works programs (cash-for-work programs), and fee exemptions for basic services in health care or education (such as PSUGO in Haiti). Depending on the needs of the target population and the objectives, the va- rious interventions may be short term or medium to long term. In Haïti, for example, the lessons learned from the 2010 earthquake suggest that both short and medium term responses are needed. À social safety net should be able to provide support over the medium term basis if it is designed to address chronic vulnerability or promote access to health care and education through transfers linked to school attendance or health visits. If a crisis or disaster occurs, the system should be able to respond rapidly to needs either by scaling up existing schemes or 136 The goals of social protection are understood here as resilience, equity, and opportunity. The scope of social protection encompasses programs that range from noncontributory social assistance, including humanitarian aid, to contributory social insurance or social security, and instruments that can link households to skills-building and income generating opportunities through access to the Labor market or self-employment (World Bank 2012). | [page 164] WorldBank - ONPES | diversifying interventions and implementing temporary short-term programs that are properly targeted. The evidence suggests that social safety net programs must be part of a broader social protection and promotion system if they are to contribute effectively to poverty reduction and enhance resilience and equity. Promo- tion here refers broadly to interventions that favor increasing the human capi- tal and the livelihood opportunities of the poor, including bridging coverage gaps in conditional cash transfers programs to boost the investments in health care and education and enhance the employability of beneficiaries and augment the access of beneficiaries to self-employment or small entrepreneurship programs. This broader social protection and promotion system encompasses three main types of interventions: (1) social insurance instruments or contributory schemes usually linked to formal occupation (contributory pensions, health, or unemplo- yment insurance), (2) active Labor market programs that foster employability and facilitate Labor market insertion, and (3) noncontributory programs (social assis- tance) that support productive activities, for example, among poor farmers (agri- cultural inputs) or self-employment promotion among the extreme poor (such as microcredit schemes). The fragility that Haiti experiences poses additional challenges in the design and implementation of effective and sustainable social protection interven- tions. First, households in fragile states face a mixture of acute and chronic needs that require a combination of flexible, short-term responses, as well as long-term interventions”. Humanitarian aid and emergency responses are challenged to become part of a system of long-term, yet responsive safety nets. Second, most analysts and observers recognize that the lack of conceptual clarity about what constitutes social protection adds another layer of difficulty, but argue that the objectives and types of instruments of social protection should be the same in fragile states as in other development contexts. Third, the extent to which currently available instruments, financing mechanisms, delivery arrangements, and actors (the government, NGOs, donors) are prepared to cope with the fragility context needs to be assessed. À summary of the key risks faced by various age-groups throughout the life cycle and the general implications for policy is presented in figure 51. Each age-group may be described as follows. Young children (under 5 years old) in Haïti are at high risk of malnutrition and mortality. The main risks among this age-group are low birthweight, i ade- quate nutrition, debilitating disease, and lack of early stimulation, all of which 137 While few conclusions are supported by robust evidence, the literature offers insights into social pro- tection in fragile states that are relevant for the Haitian case (Barrientos 2008; Carpenter et al. 2012; Harvey et al. 2007, IEG 2013; World Bank 2011). | [page 165] : Investing in People to Fight Poverty in Haïti may impair development and may contribute to perpetuating poverty#. Acute malnutrition and chronic malnutrition are still a concern among poor children under 5 years of age because they are primary indicators of the long-term, cumulative effects of undernutrition among young children. Households with children are more likely to suffer from food shortages. The related inequalities in health outcomes and access to health care are large, and the poorest quintiles do Less well (chapter 3). Figure 51 Key risks, the life cycle, and social protection in Haiti: a summary Key facts on risks + Malnutrition: 22 percent suffer from chronic malnutrition (DHS 2012) + Mortality: the mortality rate is 92 deaths per 1,000 live births, nearly 6 times higher _.... than the regional average of 16 (World Health Organization) el ou key implications ESC + Programs aimed at malnutrition address supply and demand barriers - Amore comprehensive early childhood policy Key facts on risks + Nonenrollment or school drop-out rates: approximately 200,000 children (ages 6-14) are estimated to have dropped out of school (ECVMAS 2012) + Child Labor: 20 percent of children are involved in a work activity; 6 percent of chil- . dren (ages 10-15) in the poorest quintile work and do not attend school Schoolage + Restavecs: up to 225,000 children are restavecs (PAHO 2009) children (ages 6-17) : Key implications + Social promotion programs to cover the direct and indirect costs (opportunity costs of child Labor) of education for poor children Key facts on risks + Unemployment: 28.3 percent; up to 38.0 percent in urban areas and Port-au-Prince + Not earning adequate income: 80 percent of the rural poor are in households hea- ded by a working person (61 percent in urban areas and 56 percent in Port-au-Prince) + Gender: Women are at a disadvantage in the Labor market Key implications + Urban areas: temporary income generation programs that also incorporate improve- ment of training and skills-building + Rural areas: holistic programs for the extreme poor combining consumption smoo- thing and food security, productive projects, and access to financial capital - Address specific constraints for women (care for children, the elderly or disabled) Key facts on risks + Lack of stable income: more than half of people over 65 are poor + Lack of access to health care or other care Key implications - Targeted social pensions, solutions for access to health care or other care 138 Because of data constraints, the report focuses on nutritional and health status during childhood as well as inferences on early childhood development. [page 166] WorldBank - ONPES | More generally, children under 5 years of age are particularly vulnerable to poor developmental outcomes because of multiple and complex risk factors related to pouerty; this can have lasting effects throughout the life cycle. Lack of stimu- lation, Low levels of parental education, and other risk factors such as maternal stress and depression can have lasting impacts on children's cognitive develo- pment.*? Without adequate stimulation in early childhood, children may enter school ill prepared and are more likely to have poor academic performance, to repeat grades, and to drop out of school relative to children whose cognitive skills and overall school readiness are higher upon primary-school entry (Currie and Thomas 1999; Feinstein 2003; Heckman and Masterov 2007; Pianta and McCoy 1997; Reynolds et al. 2001). The data on child development in Haiti are limited, but, according to the 2012 DHS, 81 percent of children (2- to 14-year-olds) experienced physical punishment. À growing body of research indicates that children who have experienced physical punishment tend to exhibit more aggressive and antisocial behavior (Durrant and Ensom 2012). These findings call for a pro-poor early child- hood development approach whereby social protection instruments help link fa- milies and parents to adequate services (for example, food security, health care, education, prevention of violence in the home). School-age children (6-17 years old) from poor backgrounds are at a significant disadvantage in school attendance (chapter 3). For this age-group, the major risks are nonattendance or dropping out of school for several reasons, particularly mo- ney-related reasons or early pregnancy. A nonnegligible share of children are involued in child Labor, and many continue to serve as restavecs. À nonnegligible share of school-age children work and do not attend school. Restavec children who work as domestic servants outside their own households are difficult to identify in household survey data; some studies indicate the problem is significant. For instance, a 2009 study by the Pan American Development Foundation found that there may be as many as 225,000 restavecs in Haïti (Pierre et al. 2009). Adults also face important risks in Haïti (chapter 2). Many adults (18-64 years of age) in Haiti face a risk of unemployment or lack of sufficient income, which reinforces the need to frame social protection more broadly as promotion be- cause poor households need to be connected to skill enhancement and better opportunities to earn income. The principal risks facing adults are unemployment, underemployment, low and variable income, informality, working but earning in- sufficient income to cover basic needs (the working poor), unstable livelihoods, and lack of access to physical and financial capital. Women are at a disadvantage in many respects (chapter 2). The unemployment rate is twice as high among women relative to men, but the disparity is greater in rural areas, where women are almost 3 times more likely to be unemployed than men. The relationship between unemployment and poverty varies by area. 139 Elevated levels of maternal stress during pregnancy have been found to be associated with poorer cognitive functioning among offspring at 1 year of age (Davis and Sandman 2010). | [page 167] : Investing in People to Fight Poverty in Haïti Overall and in other urban areas, the unemployed are split almost equally between the poor and the nonpoor, the majority of the unemployed in Port-au-Prince are nonpoor, and, in rural areas, the unemployed are primarily poor, particularly among women. In rural areas, woman-headed households have less access to agricultural inputs (such as seeds) which could lead to lower productivity, thereby creating a gender gap. Youth face additional challenges in becoming active on the labor market. In urban areas, young people between the ages of15 and 24 exhibit not only the lowest rates of Labor market participation and employment, but also the highest rates of unem- ployment and informal employment (chapter 2). The elderly (65 years and older) in Haïti are vulnerable to pouerty and have to rely on support from their families. The main risk among the elderly is the lack of any pension (contributory or noncontributory scheme) or access to health care and the reliance on family and charity for survival (chapter 3). Given the dynamics of demo- graphy in Haïti, the elderly tend to be somewhat neglected in antipoverty programs. The elderly represent Less than 5 percent of the poor; however, poverty is still pre- valent among this group because more than half of people above age 65 are poor (see below). Persons with disabilities are likely to suffer specific disadvantages. Although data limitations in the ECVMAS 2072 have led to an underreporting of disability, the EC- VMAS analysis of education outcomes shows significant differences in terms of enrollment among children with and without disabilities.*° This likely reflects the limited resources available for special education as well as the physical and social barriers to access (Beeston 2010). However, a better understanding is needed ofthe types of disabilities children have and the nature of the related barriers. This section examines the extent to which the social protection needs presented above are addressed in Haiti today. It assesses the extent to which the current mix of programs fit with the poverty and vulnerability profiles of Haitians. What are the recent trends in social protection? Are they moving in the right direction? Is the performance of social protection policies in terms of coverage, equity, and ade- quacy appropriate? The section first presents key findings from the ECVMAS data and then brings together the available evidence based on recent assessments of social protection sectors in Haiti, interviews and discussions with stakeholders, and administrative data.1*1 140 While only 2 percent of children aged 6—14 are identified as physically or mentally disabled in the ECVMAS 2012 data, these children are 50 percentage points Less likely to be in school, meaning only 41 percent are in school (Adelmann 2014). 141 Field interviews were conducted with a representative sample of donors, government agencies, and international and local NGOSs in October 2013. The findings and analysis presented in this report also benefited from a consultation workshop, ‘Strengthening Social Protection and Promotion in Haïti; in May 2014 | [page 168] WorldBank - ONPES | Key findings based on ECVMAS data Fact 1: Access to social security (contributory programs) is out of reach for most Haitians, especially the poor, leading to a lack of protection in old age or in case of sickness or disability. Only wage employees working in the formal sector have access to the limited social insurance schemes existing in Haiti. Social security in Haïti covers formal private sector wage-earning employees (administered by the National Security Office for Old Age and the Office of Workers Compensation Insurance, Sickness, and Maternity) and public civil servants (administered by the Direction of the Civil Pension and Self-Insurance Program). Among the active population, that is, those in the Labor force, employees in wage employment constitute only one-fifth ofthe total, which corresponds to Less than 10 percent of the population. Because of high levels of informality, only 11 percent of wage workers have access to social security, primarily concentrated in the upper quintiles of the popula- tion. Among wage workers, only a small share (11 percent) have access to social security, while the overwhelming majority do not (figure 5.2). Access to social se- curity is greatest among individuals in the richest quintile of per capita consump- tion. Two-thirds of employees with social security are in the top quintile, while only 5 percent are in the second poorest quintile, and virtually no one in the bottom quintile has access. Given the prevalence of informality in rural areas, access is concentrated in urban areas, particularly in Port-au-Prince. Figure 5.2. Access to social security by quintile of per capita consumption a. Access by quintile, % 25 FA Lo 8 $ 20 = = 8 " ë 5 ee E ® 10 9 è @ 5 E 5 Fe] © Les o Qi Q2 Q3 Qu Q5 Total 142 individuals employed as wage workers who contribute to social security or receive social security benefits, such as paid sick Leave or maternity or paternity leave, are considered here to have access to social security. The survey questions used for this analysis refer to individual employees. [page 169] : Investing in People to Fight Poverty in Haïti b. Extent of access by quintile @c 02 ©0: Q4 @c: Source: ECVMAS 20172; World Bank and ONPES calculations Access to health insurance through employment in a firm registered with the Office of Workers Compensation is also low. Only a small percentage (Less than 4 percent) of the Haitian population has access to health insurance administered by this agen- cy. Most households with the insurance are in the highest consumption quintile and live in the Metropolitan Area. The insurance is only available to employees of formal firms and their families, and the social contributions by both employers and employees and the coverage are voluntary (Cross et al. forthcoming). Because they lack access to contributory programs, poor Haitians have limited pro- tection against poverty in old age or in case of disability or sickness. Because the access to social security is limited, few people are eligible for contributory pensions when they retire and those who are eligible tend to be much better off. ECVMAS 2072 data show that only 2.6 percent of the elderly (aged 65 years and older) receive pensions (old age, disability), and the majority are nonpoor. Pension beneficiaries overwhelmingly reside in urban areas (92.0 percent), and almost half (43.2) percent live in the Metropolitan Area. These results are consistent with the fact that access to social security is limited in rural areas. Fact 2: Social assistance coverage is alarmingly Low and well below the level of identified needs, particularly among young children. Only about 8 percent of the Haitian population received noncontributory social assis- tance benefits in 2012. According to ECVMAS 2072 data, the benefits included scho- larships, food aid, and other transfers (figure 5.3). (But see box 51 for the limitations of ECVMAS 2072 data.) Overall coverage, defined as the share of the population receiving benefits, is slishtly higher in rural areas, primarily because of the larger share covered 143 Preliminary findings from ECVMAS 2013 also confirm that overall coverage is on the order of 16 percent for social protection and about 13 percent for social assistance, not including assistance from NGOS and religious organizations, the coverage of which is estimated at about 55 and O8 percent of the population, respectively. 144 Both direct and indirect beneficiaries are taken into account, ie. if one member of the household receives social protection benefits, all members of the household are considered beneficiaries. [page 170] WorldBank - ONPES | by food aid (8.8 percent compared with 53 percent in urban areas).#° More than 60 percent of food aid beneficiaries reside in rural areas, while the benerficiaries of scholar- ships and other transfers are slightly more likely to be in urban areas.'*6 Figure 5.3. Coverage of social assistance programs and distribution of beneficiaries. Population covered, percent a. Total and by poverty status 12.0 113 101 10.0 95 83 79 8.0 71 56 @rtu 60 47 @ Extreme poor 4.0 @ Moderate poor 20 12 9908 07 03 04 O4 02 Non-poor 0.0 ALL SA Scholarship Food aid Other public transfers b. Beneficiaries, urban vs rural 100 90 80 10 @vian 60 @rura 50 40 30 20 10 oO & M © ; | Ÿ RS ALLSA Scholarship Food aid Other public Poor transfers Note: Direct and indirect beneficiaries. Source: ECVMAS 2072; World Bank and ONPES calculations 145 In contrast, more than half the population benefits from remittances, which arguably play the role of an informal safety net in Haiti. (See the Shared prosperity background paper [2014], Haiti Poverty Assessment, World Bank, Washington, DC) 146 The types of assistance discussed here are more permanent and do not cover the emergency humani- tarian assistance that was provided after the 2010 earthquake. A retrospective module in ECVMAS 2012 shows that a large share of the population (about 70 percent) received some humanitarian assistance. [page 171] : Investing in People to Fight Poverty in Haïti Box 51. Methodology and limitations of ECVMAS data on social protection Data on the coverage and performance of social protection programs are limited in Haïti, and ECVMAS provides an important baseline; however, it is not without limitations. Social protection programs are highly fragmented and often small in scale and coverage. Thus, household survey data do not capture many beneficiaries of such programs. Ifthe coverage is low among the general population, there would be few observations in a nationally represen- tative survey, thereby limiting the analysis possible with the data. The main cash transfers or benefits identified in the ECVMAS 2012 sur- vey are pensions (old age, disability, and so on), scholarships, and other transfers (food aid, survivor benefits, and so on). Only 114 individual-level observations report receipt of at least one of these social protection bene- fits. There are 309 households that report receiving food aid from the gover- nment, NGOSs, or associations (table B511). The small number of observations represents a limit on the possible conclusions, especially for specific bene- fits, and, overall, the analysis should be taken as indicative and reflective only of the programs discussed. Most EDE PEP programs are not likely to be reflected in the statistics, with the possible exception of food aid. Similarly, other (nonfood) assistance received from NGOs is also not included.a Table B5.11. Sample and population sizes for social protection variables in ECVMAS 2012 Sample size Population 8 2 a 8 2 a e] œ € e] S = Indicator oo 3 o] 5 = © a = e a = eo a £ ® 8 E 5 TZ £ œ I £ œ Source: ECVMAS 20172; World Bank and ONPES calculations Note: The sample size columns show the number of households, individuals, and recipients of social protection programs in the survey. The population columns show the number of households, individuals, and recipients of social protection programs, expanded to the population through the use of expansion factors. a. Many other programs have been shown to have limited coverage, with a few exceptions, such as the school meal program (the National School Canteens Program) or PSUGO (Lamauthe-Brisson 2013; Lombardo 2072). In 2013-14, the school meals program and its partners covered almost O9 million students (according to the Ministry of Education and Vocational Training), and PSUGO had a coverage of around one million students (Lamauthe-Brisson 2013). [page 172] WorldBank - ONPES | While social assistance coverage appears to be progressive, it varies somewhat by program type. About 11 percent of the extreme poor receive some social as- sistance benefits, compared with 10.5 percent among the moderate poor and 5.6 percent among the nonpoor. While the coverage of food aid is lower among the nonpoor, it is Less so for scholarships and other transfers. The social assistance coverage of various population groups is not even; young children are underrepresented among social assistance beneficiaries, which is a concern given the vulnerability of this group. Children under 5 have the lowest coverage: only 74 percent of all children under the age of 6 benefit (indirectly) from social assistance benefits (figure 5.4). This is a particular concern given that this group suffers from the highest poverty rates (see above). While the coverage of school-age children is also quite Low, these children are much more likely to benefit from programs targeted at schools, such as school feeding programs or PSUGO, which are not captured by the survey. Figure 5.4. Coverage of social assistance programs, by age-group 65+ yo 18-64 yo 6-17 yo O-5 yo 0% 2% 4% 6% 8% 10% 12% Percent population in each group covered @ausA @ Food aid @ Scholarship € other public transfers Source: ECVMAS 20172; World Bank and ONPES calculations Note: The figure shows direct and indirect beneñciaries. Limited access to a national identification document (CIN) can be an obstacle in gaining access to social protection and other services. Analysis of ECVMAS data also points to the fact that access to CIN is more limited in rural areas and among the poor, especially among female heads of households, who are most likely be the ones seeking social assistance or services (Box 5.2). [page 173] : Investing in People to Fight Poverty in Haïti Box 52. Limited access to a national identification document (CIN) can be an obstacle in gaining access to social protection and other services Access to a national identification document (CIN) is more limited in rural areas and among the poor, especially the extreme poor in the Centre and Nord departments.® Among adults, 72 percent have a valid CIN, while almost 165 percent of adults have never had a CIN (the remai- ning 10 percent have had a CIN, but either have Lost it or it was not renewed after expiration) (figure B5.21). This share is higher in rural areas, where al- most 1 in 5 adults has never had a CIN, and for the poor: while 77 percent of the nonpoor have CIN, the share is 67.5 and 62.5 percent among the mo- derate and extreme poor, respectively. The poor in the Centre and Nord departments have the least access to CIN, since only 55 and 57,5 percent of the extreme poor have a valid CIN, respectively. Figure B5.2.1. Availability of national ID among adults 18 years and older a. by residence area @tHaveD @Hadid but lostornotrenewed @NeverhadiD @NA/Missing Rural Urban Total [page 174] WorldBank - ONPES | b. By poverty status 90 80 77 70 67.5 62.5 - 60 S S 50 3 Q & 2 40 Le} * 30 20 10 0 | Non-poor Poor Extreme Poor Source: ECVMAS 2072; World Bank and ONPES calculations. Heads of poor households are Less likely to have a CIN, particularly if they are women, with important consequences for household access to social protection programs. While only about 8 percent of all heads of households have never had a CIN, 67 and 73 percent of heads of extre- me poor and poor households have a CIN, against 83.6 percent among nonpoor households. Furthermore, while woman heads of households are Less likely to have a CIN than man heads of households, the gap is much larger for poor women. Among the extreme poor households, for example, only 62 percent of woman heads have a CIN, while 707 percent of man heads have a CIN. In comparison, among the nonpoor, 78.3 per- cent of woman heads and 83.2 percent of man heads of household have a CIN. Because heads of households are more likely to be the ones applying for services, including social aid or accessing other forms of assistance, having a CIN is especially important for the heads of poor households. a. ECVMAS 2012 has a question on whether those above 10 years old have a CIN, but only those 18 years and above are eligible for a CIN; hence, the analysis is limited to adults 18 years of age and older. Fact 3: The targeting of social assistance benefits could be improved because a large share accrues to the nonpoor. As much as half of social assistance be- nefits accrue to the nonpoor. Among social assistance benefits, as much as half go to the nonpoor (figure 5.5). While this may be somewhat puzzling given that the 147 This share is reflective of only the programs captured in ECVMAS. It is not currently possible to esti- mate what share of other assistance provided by the government or NGOSs goes to the poor. [page 175] : Investing in People to Fight Poverty in Haïti share of nonpoor beneficiaries is Less than half, the size of the transfers tends to be bigger among better off quintiles, leading to a much more regressive distribution of benefits.*£ This is especially the case in other transfers, but also holds for food aid and, to some extent, for scholarships.*? Another issue is that some government subsidies, such as gasoline subsidies, are hi- ghly regressive; as much as 95 percent ofthe subsidy accrues to the richest quintile.">° Figure 5.5. Incidence of social protection benefits, by quintile of per capita consumption and poverty status FA ; M (4 100% S 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% All social Other public Scholarship Food aid assistance transfers @Extreme poor @moderate poor @ Non-poor Source: ECVMAS 2072; World Bank and ONPES calculations Note: The figure shows direct and indirect beneñciaries. Fact 4: Adequacy of social assistance benefits is Low. The value of most social assistance benefits (cash or food) is small and, hence, contribute relatively little to the consumption of beneficiaries. With the exception of other transfers, which are larger in absolute terms, social assistance benefits are much less generous (figure 5.6, chart a). Scholarships, albeit small in absolute va- lue, contribute a large share (almost 33 percent) to the consumption of the extre- me poor (figure 5.6, chart b). Overall, however, social assistance benefits contribute only 11 percent to this consumption. The contribution to consumption tends to fall among the moderate poor and the nonpoor because their consumption is larger compared with the value of benefits; so, the benefits are relatively more important among the poorest. 148 Households are ranked according to consumption, net of social assistance transfers, so, this is not simply a result of households moving up the quintiles merely because of the transfers. 149 The number of observations for other transfers is small; so, these estimates are Less reliable 150 Based on an analysis of fuel subsidies currently being carried out by the World Bank. [page 176] WorldBank - ONPES | Figure 5.6. Benefit amounts and the contribution : : Le The groups with to the consumption of beneficiaries the highest a. Average annual per capita transfer poverty receive a disproportionately 40,000 small share of the 35,000 benefits. 30,000 25,000 20,000 15,000 10,000 5,000 | A © & % À Scholarship Food Aid ALLSA Other public Consumption transfers b. Average share of benefits in consumption by benefit type 4.8 Food Aid 9 5.9 ee 91 e 3.3 Scholarship 2.3 32.4 38.7 Other public transfers 422 202 e 5.8 Al Social Assistance &y 7.2 109 O 10 20 30 40 50 [ Non-Poor @ Moderate Poor Extreme Poor Source: ECVMAS 20172; World Bank and ONPES calculations Note: The figure shows beneñciary households only. Households are ranked into quintiles on the basis of per capita consumption, net of social assistance transfers. Fact 5: Social protection programs have limited impact on poverty and in- equality because of the low coverage and limited value. For instance, without social protection transfers, including pensions, the poverty headcount would be less than half a percentage point higher relative to the current poverty rate. In contrast, without remittances, the poverty rate would be almost 4.5 percentage points higher: 63 percent instead of 58.5 percent. Fact 6: Some types of programs perform better than others in reducing the poverty gap. Despite the overall Low impact, some programs are able to redu- ce the poverty gap to a greater extent than others. The cost-benefit ratio, or the [page 177] : Investing in People to Fight Poverty in Haïti reduction in the poverty gap obtained for each G 1 spent on the program, varies greatly by transfer type and the degree of poverty. Among the moderate poor, for instance, food aid and scholarships reduce the poverty gap by, respectively, about G 0.56 and G O44 for each G 1 transferred to households (figure 5.7). Among the extreme poor, scholarships are more effective than food aid. Because of the charac- teristics of the beneficiaries, pensions are not effective at reducing the poverty gap. Figure 5.7. The cost-benefit ratios of various social protection transfers a. Moderate poor Food aid 0.56 L e © Scholarship 046 F1 Other transfers 0.36 $ Remittances 0.28 à Pension Xe ii HTG O O2 O4 0.6 b. Extreme poor Scholarship O.31 9 e © Food aid 020 M e Remittances Cu | Other transfers 010 à Pension 0.01 fi HTG O O1 02 O3 O4 O5 0.6 Source: ECVMAS 2012; World Bank and ONPES calculations Note: The figure shows the reduction in the poverty gap obtained for each G 1 spent on the programs. [page 178] WorldBank - ONPES | Insights from other complementary sources: the underlying factors of the inadequate social protection in Haïti Given the ECVMAS data limitations and the need for a more comprehensive analysis of social protection program provision, the examination of other relevant sources of information is useful. These include previous analyses of the social protection system in Haiti (UNECLAC 2013; UNICEF 2012); a preliminary analysis of public expenditures (as part of the ongoing World Bank Public Expen- diture Review); stakeholder interviews, and a review of the recent strategy deve- loped by the government to accelerate poverty reduction, the Plan d'Action pour l'Accélération de la Réduction de la Pauvreté (PAARP), which encompasses the EDE PEP platform, the umbrella for social protection programs. The key conclu- sions are summarized as follows. First, complementary sources of information corroborate the ECVMAS fin- dings regarding the lack of adequate social protection coverage given the needs of the population. Recent studies confirm that there is an underprovision of social protection in Haiti (Lamauthe-Brisson 2013; Lombardo 2012). Considering the high level of poverty and the poor social indicators, which are exacerbated by the high risk of economic shocks or natural disasters, the poor have limited access to public support. Most assistance continues to be supplied through remittan- ces or support from churches, other nongovernmental actors, and donor projects. Existing programs are characterized by limited coverage, are often ad hoc, cover small geographical areas or narrowly defined sets of beneficiaries, and are scatte- red across numerous institutions. Second, the complementary sources of evidence shed light on the underl- ying interrelated factors behind the inadequate social protection provision in Haiti. The weak implementation capacity characteristic of a fragile country such as Haiti is exacerbated by the multiplicity of actors operating in social protection, including numerous donors and NGOs. Following the earthquake, the number of humanitarian organizations on the ground increased dramatically, and, despite coordination efforts (the United Nations-led topical clusters), there was a mul- titude of simultaneous interventions in the same geographical areas, sometimes benefiting the same households. The postearthquake period has also made the problems more evident: the multi- plicity of actors, the lack of coordination mechanisms, and the lack of a common targeting approach. The development of an overarching social protection strategy has also been more difficult during the postdisaster period. Such a strategy would have allowed, at a minimum, the identification of priorities and greater clarity in appropriate institutional roles, thereby reducing fragmentation and duplication within the government and across donors and NGOs. The focus has primarily been on emergency response rather than on building the foundations of a long-term social protection system, such as a solid targeting | [page 179] : Investing in People to Fight Poverty in Haïti mechanism, an integrated information system, and a vision of the types of social protection interventions necessary to meet the needs of the people. The majority of recent policies and programs has focused more on the supply side of public provision in education (PSUGO), health care, infrastructure, or microcre- dit, and has neglected the development of the abilities of the poor and vulnerable to access these facilities through social protection interventions (Lombardo 2012). Some observers have regretted the lack of social protection instruments such as conditional cash transfers that could effectively promote investments in health care and education by poor households. Government spending on social protection continues to be low." Data on expenditu- re related to poverty reduction activities suggest that social protection spending is a relatively small share of overall spending, Social protection spending peaked at about O9 percent of GDP in 2009-10, but has fallen since (figure 5.8). However, since 2009, the spending on promoting employment has been around O7 percent of GDP and jumped to 12 percent of GDP. The expenditure on food security increased from about O3 percent of GDP between 2009 and 2011 to O4 percent in 2011-12. Still, combined, these three areas continue to be dwarfed by spending on infrastructure (power su- pply, transport) and access to basic services (sanitation, drinking water). Figure 5.8. Poverty-related spending as a share of GDP 9.0% 8.0% 70% 6.0% 70% LS 15% 5.0% 19% 16% 21% 09% 05% 40% ” » O4% O7 04% 05% 3.0% 22% XXE 2.5% 19% 20% 21% 21% Ds O0.5% 10% O.7% 07% OT7% O.6% O.7% 0.0% 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2012-2013 Employment @ Health @ Food security @Education @social protection @ other Source: World Bank and ONPES calculations based on data ofthe Direction des Etudes et de la Programmation Budgétaire. Note: Social protection expenditure includes public pensions, health insurance and social assistance activities from MAST and MCFDF. It excludes EDE-PEP spending, PSUGO and PNCS, as well as non-public social insurance schemes. Other includes power supply, transport, sanitation, equipment, housing, and access to drinking water. It does not include extrabudgetary expenditure. 151 The Public Expenditure Review under way will help capture the spending of donors and NGOs that is not captured here. [page 180] WorldBank - ONPES | Recent trends in social protection: encouraging developments Despite the challenges, there have been some encouraging recent develop- ments in social protection. This includes efforts to establish a national social protection strategy, which has started by laying the foundation of the key blocks needed for a social protection system. In recent years, the government of Haiti has taken several steps to develop a national social protection strategy. The government'’s Action Plan for National Re- covery and Development of Haïti of March 2010 views the establishment of a social protection system as a critical factor in the recovery and growth of the country. Sin- ce then, several initiatives have been launched, such as the fight against hunger, the extreme poverty initiative Aba Grangou and EDE PEP. In May 2014, the Prime Minis- ter's Office launched PAARP, which is organized around EDE PEP and identifies the elements underpinning the implementation of a social protection system, such as a national targeting system, a unique beneficiary registry that can be used in various s0- cial programs, and an integrated service delivery model aimed at communes through a network of multisectoral agents and local coordination in social protection. EDE PEP has recently emerged as the framework for several government flagship programs. The goal of EDE PEP is to protect the vulnerable living in ex- treme poverty throughout the life cycle to ensure long-term investment in hu- man capital and to provide opportunities to overcome the condition of extreme poverty. The program is implemented primarily by FAES, with some programs un- der the Ministry of Social Affairs and Labor and the Ministry of Public Health and Population. It is based on four complementary pillars: (1) social inclusion, (2) the development of human capital, (3) economic inclusion, and (4) the development of a decent environment (figure 5.9). Figure 5.9. Main programs under EDE PEP Kore Moun Panye Family - Ranje Kay Andikape Solidatité Planning PSUGO LOCLENEEr Katie Kore Ti Fight against School : “ Gran Moun Kantin Mobill cholera feeding TIGEEIE HO Community _ Community Literacy Restaurants Ten SR ENEE health centers Programs Ti Manman : Cheri Carte Rose Kore Etidyan Source: FAES 2014, [page 181] : Investing in People to Fight Poverty in Haïti Because of the expansion of EDE PEP programs, spending on social safety nets has recently grown in comparison with the corresponding spending in other low-income countries, though it is still Low.'°? Government spending on social safety nets in 2012-13 is estimated at O.84 percent of GDP (figure 510).°* Budgetary expenditure on social safety nets by the Ministry of Social Affairs and Labor and the Ministry of Women's Affairs and Women's Rights represents only O.4 percent of GDP. However, the recent expansion of EDE PEP programs, estimated at O.5 percent of GDP in 2012-13, has nearly doubled the spending on safety nets, though it is still low relative to the corresponding spending in other low-income countries. Finan- ced through extrabudgetary sources (Petrocaribe), this increase in spending sug- gests that one should take a closer Look at the effectiveness ofthese new programs in realizing their objectives and improving their targeting to ensure they reach the most vulnerable. Figure 510. Social safety net spending as a share of GDP, low-income countries 4% 35% 3% 25% 2% 15% 1% 05% O% = ÉRRSZ Frs se és ses + à s à > € SO à @spending on EDE PEP (est) Os Spending Sources: Haiti: World Bank calculations based on Ministry of the Economy and Finance and FAES data; other countries: World Bank 2014. Note: The figure shows social safety net spending in various years (2009-11). For Haïti, the data correspond to poverty-related social protection expenditure in 2012-13. The expenditure on EDE PEP is estimated. The government is engaged in the development of a national targeting sys- tem that will be coupled with a social registry of beneficiaries to improve the efficiency and effectiveness of social protection programs. This is necessary 152 Because of the fragmentation of social protection programs across various institutions and agencies within and outside the government, collecting comprehensive data on spending is challenging. Bor- garello (2009) finds that total spending on social safety nets was only O7 percent of GDP in 2009, ex- cluding fuel and electricity subsidies, and 24 percent, including them. This includes spending through the state budget and multilateral and bilateral agencies. 153 includes budget expenditures of the Ministry of Social Affairs and Labor and the Ministry of Women's Affairs and Women's Rights and extrabudgetary expenditure on EDE PEP. | [page 182] WorldBank - ONPES | because no common targeting approach exists, and the targeting of government programs varies by program and may not be systematic. À technical committee to develop a national targeting tool has been created under the leadership of the Ministry of Social Affairs and Labor, and a proposal for a national targeting tool was approved by the committee in early 2014.°* This tool needs to be tested in the field, and it will undergo a broader consultation and validation process, including with actors in other sectors such as health care and education. In the short term, the tool will be used primarily in donor-supported programs that fall outside the scope of EDE PEP, but may eventually be used in EDE PEP programs. The government also plans to start consolidating public social programs that have similar objectives and more effectively coordinating donor programs. The PAARP envisions better coordination and consolidation among programs that are currently duplicated across various institutions. For example, Kore Moun Andi- kape is a cash benefit for the disabled and the elderly that is currently administe- red separately by FAES and the Social Assistance Fund under the Ministry of Social Affairs and Labor, with different benefit levels and eligibility criteria. Furthermo- re, the government is also seeking to coordinate donor programs to fill the gaps identified in the plan more effectively (for instance, coverage gaps). PAARP envisions the use of a network of agents to accompany vulnerable families and coordinate at the communal level; an example is the Kore Fanmi project. This model aims to improve the efficiency of social service delivery in Haiti. Kore Fanmi is an initiative of the government and is supported by the United Nations Chil- dren's Fund and the World Bank. It aims to lay the foundation for a cost-effective and sustainable strategy for integrated social service delivery by providing à com- mon platform for the coordination of social interventions by all service providers at the local level. It is being implemented by FAES in partnership with United Na- tions agencies, especially the United Nations Children's Fund and the World Food Programme. The initiative serves as the Link between demand and supply and is helping to lay the groundwork of a social protection system (box 5.3). In addition, Kore Fanmi has also been able to play a role in emergency response. In the case of flooding in one commune and acute food insecurity resulting from drought in another, Kore Fanmi was able to use information from community agents to iden- tify affected families, request an immediate response, and coordinate the delivery of assistance to the appropriate beneficiaries 154 This committee includes representatives of FAES and major donors, such as the United Nations Chil- dren's Fund, the United Nations Development Programme, the U.S. Agency for International Develop- ment, the World Bank, the World Food Programme, and international NGOs such as CARE and Action Contre la Faim. The committee sought to develop a targeting tool that would respond to the specific needs of two major programs, Kore Fanmi and Kore Lavi (a nutrition and food voucher program), as well as serve the country more broadly with a national tool. | [page 183] : Investing in People to Fight Poverty in Haïti Box 5.3. Kore Fanmi Kore Fanmi seeks to improve the access and efficiency of social service delivery in the rural areas of Haiti. The approach involves direct family ac- companiment and support for the basic human rights of families. Kore Fanmi relies on a network of multisectoral community agents who work directly with and are accountable to a specific set of families. These agents deliver direct life-saving services and essential commodities (for example, nutrition supplements, vaccinations, mosquito nets, and soap), promote po- sitive behavioral change, and refer families to the available social services. Before initiating family support activities, the program undertakes a map- ping exercise, which is an inventory of the services available to the popu- lation in the target area through various service providers. This inventory, called the opportunity map, is used as a basis for referral. A tailored family development plan for each family that outlines a set of Life objectives is created based on a socioeconomic survey of each family’s vulnerabilities. The type and intensity of family coaching vary depending on the needs and vulnerabilities of each family. Kore Fanmi uses a dynamic and integrated management information sys- tem to analyze each family's conditions and vulnerabilities, propose key actions, and track progress. Thus, Kore Fanmi creates a mechanism to reach poor and vulnerable fami- lies, generates an objective way of identifying the most vulnerable families and analyzing their needs, coordinates the provision of services in munici- palities, and strengthens the capacity of local governments to oversee the provision of services within their jurisdictions. Recent trends in social protection: persisting challenges Despite the recent progress, significant challenges remain, especially with re- gard to closing the coverage gaps affecting certain population groups, such as young children. Given the limited coverage of social protection, EDE PEP seeks to reduce covera- ge gaps. Coverage is still narrow in the regions with the highest poverty rates. While it is not possible to gauge the coverage of EDE PEP programs using ECVMAS, admi- nistrative data available from FAES shed light on the extent of coverage over the pre- vious two years. The coverage of in-kind programs, such as mobile canteens or the distribution of food kits, is much wider than that of cash transfers (figure 5.11). Cash transfers cover about 3 percent of the population, while in-kind programs, excluding food distribution, cover as much as 8 percent."°° But even in-kind programs have Limi- ted coverage in the departments with the highest poverty rates (Centre, Grand'Anse, 155 Information on unique benefciaries is not available; so there may be some double counting. [page 184] WorldBank - ONPES | Nord-Ouest). Estimating the coverage of food distribution is problematic, but data show that, by far, the majority of meals are distributed in the Ouest department (in Port-au-Prince, in particular), where poverty rates are the lowest. Figure 5.11. Coverage of EDE PEP programs, by type and by poverty rate and departmen, 2012-13 20% 90% 18% . e e 80% e 16% 70% . L] e 14% e 60% e e 12% 50% 10% 8% 40% 8% ° 30% 6% 4% 3% 20% : | (] [] | [ | h 0% 0% E 5 £ 5 3 2 5 2 g 5 5 Ê à Zz 9 2 2 D = £ E < ) 2 @cast (6 of pop) @in-kind (% of population) @rc C6 of pop), RHS Source: World Bank calculations based on data of FAES and ECVMAS 2012. Note: Coverage is calculated using administrative data and capture only direct benefciaries of the following programs cash and in-kind transfers: Ti Manman Cheri, Kore Etidyan, Kore Moun Andikape, Bon Solidarite, Bon Dijans, Panye Solidarite, Kore Paysan (seed), Kore Paysan (fish). It does not include PNCS, PSUGO and Kantine Mobile. EDE PEP proposes a life-cycle approach, but seems to lack sufficient focus on early childhood. The plan for the reduction of extreme poverty includes a few programs in the early childhood window; however, most of the interventions (community restaurants, disaster response programs, and health interventions) are insufficient and not tailored to the needs of this age-group (table 51). For example, health insurance is available only in urban areas and is contributory; so, it is unlikely to reach the most vulnerable. Meanwhile, community pharmacies do not focus on preventive health care and malnutrition, which is a critical priority among young children. [page 185] Eu LA Investing in People to Fight Poverty in Haïti Table 51. Alignment of EDE PEP programs With risks and vulnerabilities across the life cycle .e Planned number of Life-cycle stage Risk Projects under EDE PEP beneficiaries (2016) hood PSUGO 1,500,000 2 Gdtiere ED DS ae School feeding 1200,000 chidhood Ti Manman Cheri 100,000 3. Youth Poor educational outcomes Kore Etidyan 30,000 Unemployment, lack of access TiKredi 6500 to credit : “os _ . 4. Adulthood Unemployment, low income HIMO (public works) EE Poor living conditions, poor Ranje Kay Kartier/ Banm BCE sanitation Lumie- Banm Lavi Illiteracy Alphabétisation 150,000 5. Old age Kore Ti Gran Moun 30,000 Disability Kore Moun Andikape 30,000 REMRREMERSIEAIRES Resto Communautaire | 150,000 curity Panye Solidarité 600,000 Natural disaster/emergency Kantin Mobil 1,000,000 RE ES Campaign for prevention of 6. All cycles Cholera Disease/lack of access to health care Community health centers ES Inadequate living conditions Ranje Kay Kartier/ Banm (lack of access to sanitation, : . nn Lumie- Banm Lavi but insu- drinking water or waste mana- ; fficient gement) EC CS ER Source: World Bank, based on data of FAES Note: — = not available. [page 186] WorldBank - ONPES | Additionally, considering the main risks affecting each stage of the life cycle, there are discrepancies between programs and needs not only in terms of the risks addressed, but also in terms of the scope of the programs. For example, malnutrition is a major risk affecting children under 5; however, under the new stra- tegy to reduce extreme poverty (based on the EDE PEP framework), there is little effort directed at preventing malnutrition or proactively improving children's deve- lopment potential early in life. While child Labor and restavecs are considerable phe- nomena, the risks are not addressed by EDE PEP programs. Half the people above age 65 are poor, and the coverage of contributive pensions is limited. Yet, the non- contributory cash transfer program will only be able to cover 30,000 people. Risks related to poor living conditions, that is, related to health or disaster vulnerability, are not addressed under the current strategy. The programs un- der the fourth pillar of EDE PEP (the development of a healthy environment and fostering access to decent lodging) have limited coverage. For example, Ranje Kay Kartier and Banm Lumie-Banm Lavi, which seek to improve urban neighborhoods, are only expected to cover 25 neighborhoods. Public works are also included un- der the pillar, but it is not clear what they will cover. The gap between current programs and needs may still be bridged by re- thinking the design of some of the flagship programs. For example, the target groups could be expanded to include young children, and the benefit mix could be modified to support human capital formation more effectively. In the face of large and entrenched poverty rates and numerous vulnerabi- lities, few of the poor have access to social protection or formal safety nets. First, access to social security is out of reach for most Haitians, particularly the poor. Second, only a small share of the population benefits from social protection. Because of narrow coverage and limited generosity, social protection benefits are inadequate and play only a marginal role in reducing poverty and inequality and in improving opportunities among the population. The groups with the highest poverty—rural residents and children, especially young children—receive a disproportionately small share of the benefits. The costs generated by this lack of effective protection for the poorest hou- seholds are high, especially for future generations, and lead to missed oppor- tunities in the formation and accumulation of human capital. These costs are borne principally by children; this is of great concern because the consequences can become irreversible if children do not obtain proper support in their first thou- sand days, if they do not receive early childhood stimulation, or if they are kept out of school for too long. ECVMAS data show that children had the lowest social protection coverage, despite the high poverty rates and risks they face. On the positive side, the recent development of public policy is focusing on stren- gthening social protection to accelerate poverty reduction. The government's | [page 187] : Investing in People to Fight Poverty in Haïti umbrella initiative, EDE PEP, represents a positive effort to create new programs to Only 11 percent address important constraints and risks such as the high cost of school tuition (the ofthe cure PSUGO program) and disabilities (Kore Moun Andikape). It also applies a life-cycle Public social approach that responds to some of the needs identified through the ECVMAS and assistance through complementary sources such as the DHS. food ali cr ou Overall, the findings presented here confirm the urgent need for social protection transfers. and promotion interventions that would enable the poorest households (especially those in rural areas and with young children) to overcome the hurdles to building and preserving human capital in the face of repeated shocks. This could include instruments such as a cash transfer targeted on families with pregnant women and children under 5 years of age, interventions to effectively reduce the costs of schooling, programs to provide productive opportunities, and programs to improve living conditions. The challenge and opportunity now involve deciding how these key findings can translate into elements of a strategic agenda and how to establish priorities in a fiscally constrained institutional environment. Setting the following four priorities may be useful. Priority 1: Build the foundational blocks of a social protection and promo- tion system, starting with a targeting system. This priority would include the following actions: a. Implement the new national targeting tool and establish a system of monitoring and evaluation. The targeting tool was developed by the government and donor partners to improve the equity and efficiency of social protection spending and to reduce the gaps in coverage. À system of monitoring and evaluation, including impact evaluations of existing programs, would allow identifying obstacles or implementation problems and evaluate the effectiveness of interventions (com- paring impacts of cash vs. in-kind transfers, for example). b. Build on existing government efforts to formulate a strategy based on the po- verty and vulnerability profile emerging from ECVMAS and focusing on a mini- mum package of social protection and promotion interventions. The interven- tions should have clear objectives and target the poorest populations (especially young children) and the geographical areas most in need (especially rural areas). According to most stakeholders, both EDE PEP and, more recently, PAARP repre- sent initial steps and can benefit from feedback and improvements. c. Define and reinforce institutional arrangements and sustainable coordination me- chanisms within the government and with interested donors to reduce fragmen- tation and enhance efficiency. Within the government, more clarity could be esta- blished by defining and reinforcing the roles and responsibilities of ministries (the Ministry of Social Affairs and Labor) and agencies (FAES), especially in terms of plan- ning and coordinating. Among interested donors, the revival of a social protection donor roundtable (within the CAED) is encouraging. These efforts are shifting away from an emphasis on emergency response and short-term actions and toward an emphasis on the introduction of medium-term social protection interventions. The harmonization of social protection approaches and indicators is advancing thanks | [page 188] WorldBank - ONPES | to multiagency initiatives such as the Social Protection Inter-agency Cooperation Board.°* In the case of Haïti, a common approach to communal agents linking users and beneficiaries to services and opportunities would be highly relevant. Va- rious programs and donors are supporting the government in this approach (the United Nations Children's Fund and the World Bank in Kore Fanmi, the U.S. Agency for International Development in Kore Lavi, and so on). d. Moue forward with the development of a unique registry of social protection beneficiaries in priority areas. Given the difficulties of implementation, this effort could be restricted to priority areas of focus such as all social safety nets aimed at children or at the geographical areas with the highest levels of pover- ty. It can also support efforts to ensure that national identification is available to all poor and vulnerable to allow access to social assistance and services. A phased approach could also be envisaged. Priority 2: Increase the coverage of social safety nets, especially among hou- seholds with children, while insuring sound targeting and improving the qua- lity of relevant programs, particularly those able to enhance human capital promotion. This priority would entail some of the following actions: a. Take advantage of existing potential. Extend the coverage of relevant programs that promote human capital accumulation among the poor, while improving pro- gram design and effectiveness. For example, the Ti Manman Cheri conditional cash transfer currently targets school-age children already in school, but would be more effective at supporting human capital formation if it covered younger children and encouraged children who are out of school to enroll in school. In addition, initiati- ves aiming at improving the efficiency of social service delivery, such as Kore Fanmi, could be helpful in linking poor households to services and opportunities. b. Given the close links between pouerty and education outcomes, take the fo- llowing steps: (1) continue to exploit the synergies between initiatives removing supply-side constraints (the removal of school tuition fees through PSUGO using funds channeled to schools) and demand-side barriers (Ti Manman Cheri to ad- dress nontuition costs), (2) intensify efforts to identify and include children current- ly out of school, and (3) target the poorest areas identified in ECVMAS; these are mainly in the north. c. For the minimum set of social protection and promotion interventions suggested above, ensure the gradual improvement in quality standards in service delivery through financial incentives. A possible avenue would be to link additional donors or budget funding to the use of targeting, user feedback, and monitoring and eva- luation mechanisms. 156 This initiative includes the Department for International Development Cooperation of Finland, the Deutsche Gesellschaft für Internationale Zusammenarbeit, the European Commission, the Food and Agriculture Organization of the United Nations, the International Labour Organization, the Interna- tional Policy Centre for Inclusive Growth, the United Nations Children's Fund, the United Nations Development Programme, and the World Bank. | [page 189] : Investing in People to Fight Poverty in Haïti Priority 3: Pursue articulation efforts and watch for agile implementation on the ground. a. Make social protection productive in addressing the risk of volatile and insuffi- cient income among poor adults. Promote the articulation of well-targeted social protection programs that promote human capital and productive initiatives, with some adaptation for differences in rural and urban areas. In rural areas, the govern- ment and interested donors could consider scaling up promising pilot initiatives with good track records such as the Fonkoze multipronged initiative Chemen Lavi Miyo (pathway to a better Life) for extremely poor women in the Plateau Central.’ b. Address regional disparities by building on and improuing the geographical Plans Spéciaux (territorial action plans for poverty reduction) and mainstream in- clusion in targeted social protection interventions. c. c. Continue to assess the comparative advantage of various actors (the govern- ment, NGOS, foundations) in the implementation of social protection and promotion initiatives, with a view to achieve flexible, agile, and swift implementation even in the most far-flung areas. This is based on the recognition that the government is not currently able to ensure the delivery of social protection interventions at scale. d. Complement demand-side interventions through sectorial policies to improve the access, affordability, and quality of services, especially in health care and education. Common targets in priority regions identified by ECVMAS and other sour- ces could be determined. e. Strengthen the links between structured programs designed to address chronic po- verty or human capital promotion and emergency disaster response mechanisms. Priority 4: Address the issue of predictable, efficient, and sustainable financing for social protection. This action entails seizing the opportunity of the current Pu- blic Expenditure Review co-led by the government and the World Bank. In this con- text, a few issues are emerging. While the data call for greater spending on social protection to ensure better coverage, there might also be ineffective and regressive expenditures that could be reallocated, such as fuel subsidies. The fuel subsidy reform that is currently under way could provide an opportunity to reallocate some of the savings to support interventions with the highest potential to reduce poverty and promote investments in human capital. Candid and constructive discussions should also be encouraged on the sustainability of investments. Sustainability re- lates to institutional sustainability; hence, the need to achieve progress in establi- shing the institutional framework highlighted in priority 1, to ensure efficiency and equity, to speed up the targeting reforms, and to focus on results. 157 Following the graduation approach promoted by the Ford Foundation and other partners, Fonkoze, a Haitian microfinance NGO, rigorously selected extremely poor women in the Plateau Central and provided them with social protection and productive opportunities: consumption support through small cash benefits, support for free health services and housing improvements, access to savings, asset transfers, technical training, and coaching. Three years after the start of the program, 962 percent of the women participants had lowered their poverty Level, and 70 percent were sending children to school, versus 10 percent at the beginning of the program | [page 190] Part Il: Reflections to Promote Evidence-based Policy Making [page 191] : Investing in People to Fight Poverty in Haïti Chapter 6: The way forward: key messages and priority areas of policy actions For the first time in a decade, it is possible to study the extent, evolution, and drivers of poverty in Haiti based on household characteristics and behaviors throughout the country and across rural and urban settings. The collaboration between ONPES and the World Bank, the efforts to collect the living standards me- asurement survey ECVMAS 2012 and the official poverty lines recently developed by the government have made this possible. TWo years after the 2010 earthquake, monetary and multidimensional poverty was still severe in Haiti, particularly in rural areas. In 2012, almost 60 percent of the population was poor, and one person in four was living below the extreme po- verty line. Nearly half of all households were considered chronically poor because they were living below the moderate poverty line and lacked at least three of the seven basic dimensions of nonmonetary well-being. In rural areas, these numbers were even higher: three-quarters of all households were monetarily poor, and two- thirds were living in chronic poverty. Compared with 2000, monetary and multidimensional poverty has improved slightly, but inequality in both income and access to basic services remains the highest in the region. Extreme poverty declined from 31 to 24 percent between 2000 and 2072, and there were gains in access to education and basic infrastruc- ture, although the levels and quality were low. Income inequality is the highest in the region—at a Gini coefficient of O.61—and has been steady at that value since 2001. At the same time, access to basic services such as water and sanitation and to economic opportunities are characterized by huge inequalities dictated by poverty, location of residence, and gender. Women and girls are particularly vulnerable because they face important obs- tacles in the accumulation and use of assets, particularly human capital. Despite sizable progress in education and health outcomes, adult women are still Less well educated than adult men and are more likely to beilliterate, while maternal mortality is still five times higher than the regional average. Apart from initial differences in en- dowments, women in Haiti also face additional obstacles in participating in the Labor market: they are significantly Less likely to be employed, have Less access to inputs, and earn more than 30 percent Less than man. Gender-based violence and low parti- cipation in the public sphere are widespread in Haïti, reflecting weak agency. The analysis in this report is framed around the importance of asset-building and protecting the poor and vulnerable. This report builds on new evidence to provide stylized facts and analysis to contribute to an informed debate on the challenges and opportunities in poverty reduction. Creating an environment that promotes greater growth and prosperity is critical for the country, but, if growth is be increased and sha- red with those less favored, the vulnerable must be supported in building and protec- ting their assets. Better access to education and health care as well as physical and EN [page 192] WorldBank - ONPES | financial assets improves income generation opportunities across the board. But in a context of significant exposure to aggregate and idiosyncratic shocks, it is also essential to protect asset accumulation among the poor through access to safety nets and social protection services for improved risk management. The regular monitoring of poverty and living conditions is a necessary step to promoting effective, evidence-based policy making and policy implemen- tation. One of the many obstacles to post earthquake reconstruction and emer- gency operations is the lack of sound statistical information at the national Level. Making sure that the next household survey is implemented within a reasonable time frame will help prevent the recurrence of the shortage of information. Regu- lar monitoring built on the solid baseline described in this report will contribute to enhancing the design and efficacy of antipoverty policies. While overall economic growth remains a prerequisite for poverty reduction, policies should seek to raise the capacity of the poor and vulnerable to accu- mulate assets, generate income, and protect their livelihoods from shocks. Special attention should be paid to vulnerable groups, such as women and chil- dren, and to rural areas. In the following paragraphs, the main priority areas for policy actions emerging from the diagnostics produced in the previous chapters are listed. These areas of actions will provide a new platform of dialogue for Government and its partners. This evidence-based dialogue will allow the various players to define and prioriti- ze actions, and allocate resources accordingly. Challenges Incomes have stagnated in rural areas, where 80 percent of the extreme poor are concentrated. The stagnation reflects the problems with reliance on the low-perfor- ming agricultural sector and production for home consumption. Rural livelihoods are highly dependent on agriculture. Almost 80 percent of rural households engage in far- ming, and, in 50 percent of households, farming is the sole economic activity. However, the returns to agriculture are Low and unreliable, and the activity resembles a subsis- tence strategy rather than reliance on à productive economic sector. This situation has led to constant migration from rural areas to urban areas. Participation in the nonfarm sector helps rural households emerge from pover- ty. Engaging in the nonfarm sector in rural areas reduces the probability of being poor by 10 percentage points. The typical nonfarm job in rural areas is a one- or two-person shop engaged in small retail. Still, the returns to this activity surpass those accruing to farming. About 40 percent of nonpoor households participate in the nonfarm sector, a participation rate that is 1.5 times higher than the participa- tion rate among the poor. EN [page 193] : Investing in People to Fight Poverty in Haïti Urban areas have fared better than rural areas, reflecting larger private transfers, more nonagricultural employment opportunities, narrowing inequality, and more access to critical goods and services. While urban areas offer comparatively bet- ter opportunities to escape poverty, access to services is affected by overpopula- tion; unemployment affects 40 percent of the urban workforce; and 60 percent of workers earn less than the minimum wage. The urban poor must therefore resort to self-employment or two-person businesses as a coping mechanism. Overall, al- most 60 percent of the poor are in this type of occupation. Internal and international migration is an important livelihood strategy among ru- ral and urban households. Physical mobility from rural to urban areas and from Haïti to other countries is a strategy that households commonly adopt to improve labor market incomes and obtain higher returns to human capital. About 20 percent of Haitians have migrated internally; 10 percent live abroad; and private transfers (do- mestic and foreign) account for 13 percent and 20 percent of income in rural and urban areas, respectively. Policy guidance While consistent economic growth is a prerequisite for poverty reduction, policies should focus on increasing the income generating capacity of the poor. Microeco- nomic determinants are equally critical for fostering economic opportunities that are inclusive and contribute to poverty reduction. The in-depth analysis of living conditions described in this report allows three priorities to be distilled for policy makers, as follows: + In rural areas: *_ Boost agricultural productivity through improved access to basic inputs (fertili- zer, pesticides, seeds, Labor and distribution chains) and output markets; encou- rage the diversification of crops, the acquisition of skills and knowledge specific to the Haïtian rural context, as well as the sustainable use of natural resources. *_ Facilitate off-farm jobs opportunities as a way of generating additional revenue and managing risk by undertaking interventions designed to improve the quality of the rural Labor force (eg, basic education, vocational training) and to generate increased rural non-farm employment opportunities (eg, programs to encoura- ge expansion of rural enterprises, support to rural financial institutions). * In urban areas: + Invest in skills because workers with better educational attainment are able to obtain substantially better results than workers without education. Focus on entrepreneurial knowledge to improve the profitability of self-employment. Pay special attention to women and youth, who are particularly disadvantaged in Labor markets. [page 194] WorldBank - ONPES | * In both rural and urban areas: + Invest in basic infrastructure (including electricity, water, and roads) and seek a more enabling business environment to boost the performance of farmers and the self-employed. + Harness migration: private transfers play an important role in the capacity of households to stay out of poverty. Challenges The utilization of education and health care services and health and educa- tion outcomes have improved; however, the related indicators are still re- latively low, and inequality is still substantial. Adult literacy and enrollments among school-age children are significantly lower among poor households. Seve- ral factors may explain this result. A large number of poor children have to work while attending school, raising the probability of dropping out or being overage for grade. Similarly, poor households spend substantially Less on school fees, which are associated with the quality of the service and the infrastructure provided by schools. Child and maternal mortality indicators show a similar pattern: child mor- tality and malnutrition and maternal mortality are higher among the poorest, sug- gesting Less reliance on health services and a greater impact of health shocks on poor households. The health outcomes and service utilization among women are particularly worrisome. The financial burden and inadequate service supply constrain health and education service utilization and outcomes, particularly in rural areas. Hou- seholds spend, on average, 10 percent of their budgets on education and 3 per- cent on health care.l® Sickness is considered the most severe shock in economic terms. The low levels of household health expenditure suggest that households cannot afford to pay more or do not have access to health services. Cost is the main reason children are kept out of school or do not benefit from medical care. Distance from a service provider is the second-most important reason. As donor support declines, the incidence of these expenditures is likely to rise, and service utilization is likely to narrow, impacting outcomes. Policy guidance Policy makers should seek to raise the human capital accumulation capacity of the poor and vulnerable, considering the importance of this capacity in impro- ving welfare. 158 Conditional on registering positive expenditures for education and health care. [page 195] : Investing in People to Fight Poverty in Haïti * Ineducation: + _Sustain and expand access to primary education. Achieving universal primary enrollment will require several critical actions by the government and among development partners, including (a) the production and implementation of a short- to medium-term financing plan in primary education to increase resour- ces available to the sector; (b) in coordination with social protection programs, the determination of medium- to long-term strategic plans for service delivery by type of provider at all levels of education, starting with primary education. +_Improve learning and the quality of service delivery in education. increasing quality will require several key measures, including (a) increasing public over- sight through targeted and well-implemented measures and systematic data collection to hold schools accountable, and (b) addressing problems in pre- primary education to give children a solid basis for skill building. + Inhealth care: + Expand coverage, utilization, and the quality of services by building on pro- mising service delivery models. The government and development partners should focus on programs with a proven record of enhancing the utilization of health services, especially primary health care and in communities, inclu- ding results-based financing and community service provision. + _Develop innovative donor coordination mechanisms in the health sector, ta- king into account national priorities. + In both education and health care: +_Establish an information system with a unified beneficiary and targeting me- chanism. + _Narrow the knowledge gap, particularly the determinants of Low school pro- gression, learning and abandonment, as well as the low usage, Low spending conundrum in health care services. Challenges Vulnerability is extensive in Haiti. One million people live slightly above the po- verty line and could be pushed below the line by a shock; almost 70 percent of the population is either poor or vulnerable to falling into poverty. A typical Haitian hou- sehold faces multiple shocks annually; 78 percent of households in Port-au-Prince, 89 percent of households in other urban areas, and 94 percent of rural households experience at least one economically damaging shock each year. Haiti’s hazards have larger consequences compared with other countries not only because of the country’s geological, geographical, and developmental challenges, but also because of institutional weaknesses, including inadequa- te planning and lack of regulatory enforcement. Haitis hard-earned develop- ment gains are often jeopardized by adverse natural events. [page 196] WorldBank - ONPES | In the face of the high incidence and vulnerability to idiosyncratic or cova- riate shocks, the poor and vulnerable have limited access to public support. Recently, the government undertook significant efforts to expand social assistance provision through the EDE PEP framework. However, substantial challenges remain, especially in closing coverage gaps affecting certain population groups, such as young children, or regions with highest poverty rates, such as Centre, Grand'Anse, and Nord- Ouest. EDE PEP proposes a life-cycle approach, but Lacks sufficient focus on the early childhood window. Policy guidance In light of the high incidence of shocks, two types of interventions are needed to increase resilience: assess social protection needs and expand coverage among the poor and vulnerable where possible and mainstream disaster risk manage- ment activities in all poverty reduction strategies. + In social protection: + Build the foundational blocks of a social protection and promotion system, starting with a targeting and monitoring system. + Increase the coverage of social safety nets, especially among house- holds with children, while insuring optimal targeting and improving the quality of relevant programs, particularly those able to enhance human capital promotion. *_Pursue the capacity building and coordination of efforts across ministries and agencies and ensure effective implementation on the ground. -_ Address the problems in the provision of predictable, efficient, and sustaina- ble financing for the overhauled social protection and promotion package. * In disaster risk management: + _Improve the identification and understanding of disaster risks in Haïti by quantifying and anticipating the potential impacts of natural hazards, and deepen the knowledge about households’ coping strategies. + Reduce risks and avoid the creation of new risks by integrating risk mana- gement in public policies and investments. Information on disaster risks can be used to guide investments so as to address risks. The retrofitting of critical buildings, the construction of protective infrastructures, and the rebuilding of natural ecosystems are examples of disaster mitigation in- vestments needed in Haiti. + _Improve the country's capacity to manage disaster-related emergencies by strengthening the institutional arrangements for emergency and prepared- ness, including à fully functional capacity for the National Emergency Ope- rations Center, and focusing on the importance of public sensitization and communication campaigns. *_Increase the resilience of the government and households by adopting fi- nancial protection strategies (for example by promoting financial inclusion that allows the mobilization of savings or access to insurance systems). [page 197] : Investing in People to Fight Poverty in Haïti Appendix À. Poverty indicators, disaggregated by department and area of residence, 2012 Table A1. Poverty indicator, disaggregated by department and area of residence, 2012 jcaton | ea | 0 | Poney | VUREPE | roapoo | Per | She count gap Artibonite Total "| 60 À 24 | 15 | 1755602 | 106205 | 100 | 00 | Centre Lun [#7 | 5 | 6 | mo | ses | © | 0 | Grand'Anse Total" |" 20 À 36 | 7 | 508% | sous | 100 | 00 | Nippes fra | 6 | 50 | 1 | ns | men | 5 | & | Total "À" 66 À 29 | 15 | 8:57 | 730306 | 100 | 00 | orel [uen | 5 | mn | 8 | am | nos | w | w feu | me [| 7 | su | mou | sm | « | [ro | 6 | 5 | 6 our | 2 | 0 | 0 | Nord-Est Nord-Ouest feu e [ ou À 2 | sum [sos | 5 | © | Ouest [urban | 5 | o | 4 | sono |1009%0| » | à | Cor | 5 | 5 | © | sun iso | 10 | 00 | Sud Lutter | w | n | 8 | um | su | © | 5 | feu & | 5 | 0 | se | vom | mn | = | Sud-Est Lun | 55 | ù | s | se | ss | 5 | s fre [| © [| » | 6 | um | 5600 | & | » | EN [page 198] WorldBank - ONPES | Appendix B. Income Inequality — Lorenz Curves Figure B.1. Lorenz Curves at National, Urban and Rural levels, 2012 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% O 5 10 15 20 25 30 35 4O 45 50 55 60 65 70 75 80 85 90 95 100 @) — Perfect Equality Curve — Total — Urban — Rural Sources: ECVMAS 2072; World Bank and ONPES calculations. EN [page 199] : Investing in People to Fight Poverty in Haïti Appendix C. Poverty rate comparisons Table C1. Poverty rates based on different poverty lines and welfare measures, 2000-12 À. Consumption Type of line PPP convertor Line 2000 2001 2012 Modeste | we00s | vw | 25 | veas | na | m | 60 | B. SEDLAC official income aggregate | | | | | Type of line PPP convertor Periodicity of line Line Currency 2001 2012 loueme | exo | y | 10 | mraur | 5 | u | | [oueme [res | | 15 | mraus | 5 | m | | Mmacemte | eos | | 2 | mraur | » [u | | Dnademe fees [| 25 | emeus | on [um — Mmaceme | eos | | 4 | rar | & [u | Yes: 2005 day 10 PPP dollars 97 NA C. SEDLAC unofficial income aggregate Type of line PPP convertor Line 2001 2012 Joueme | ess | dy | 1œ | mu | 5 | 5 | [oueme | eos | | 1 | mes | & |» | | Dnademe fees [| 2 | mes | 7 [| — Dnodemme [sos | un | 25 | wvœus | m | | | Mmacemte | 200 | | 4 | mraur | w | & | | Yes: 2005 day 10 PPP dollars 98 97 D. SEDLAC unofficial without imputed rent Type of line PPP convertor Line 2001 2012 Ioueme | ess | y | 1œ | mauws | & | 5 | auene ] es | | 5 | rar | & | s | Modems [sos | un | 2 | wwœus | » | 5 | | Mode | weoos | | 25 | mas | & | | | Duademe | venoes [ u [ « | mmœus | n | es | Vulnerable Yes: 2005 day 10 PPP dollars 98 97 E. Fafo Type of line PPP convertor Line 2001 2012 Jouene | es | mat | 7 |taumguce | 5 | | | Note: Panels À and B do not include the income referent to members of secondary households. All aggregates refer to per capita values and include production for home consumption. For the estimation of the household size, ALL members—both main and secondary —where included as long as they all fulfill our definition of household member * 2072 line. ** Source Fafo. Lines for 2012: 23,912044 and 17995.531, for poverty and extreme poverty, respectively. Link to the methodology for the estimation of poverty line in 2000: http://www.fafo no/ais/other/haiti/ poverty/PovertyLineForHaitipaf EN [page 200] WorldBank - ONPES | Appendix D. The methodology for determining the MPl and identifying the categories of the poor, 2012 The dimensions of the MPI 1. Food security score (Food and Agriculture Organization of the United Nations): This indicator is based on the Dietary Diversity Index, defined on a scale from O to 12. À household is considered as food secure if its score is above 8 (see Swindale and Ohri-Vachaspati, 2005, Crush et al. 2012) 2. Kids in school age are all enrolled in school : The household is not deprived if all school age children are enrolled in school 3. The household head has at least 5 years of education 4. Access to a protected water source (drinking water) : The household is not deprived if it has access to one of the following sources a. Private tap / DINEPA (Direction Nationale de l'Eau Potable et de L'Assai- nissement) b. Public Fountain c. Artesian / Drilled Well d. Protected well e. Protected water source f. Rainwater g. Kiosk (seller of treated water) h. Treated water (truck, bottle, bag, dock, gallon) 5. Hazardous material : The household is not deprived if its dwelling is construc- ted with the following non-hazardous materials: a. Walls: wood/boards, cement/blocks, bricks/stone b. Ceiling: cement/concrete, metal sheet c. Floor: cement, wood/boards, tiles, ceramic, marble 6. Source of Sustainable Energy: The household is not deprived if it has access to one of the following sources of energy: a. Electricity (individual meter EDH, collective meter EDH or without meter) b. Generator (Delco) c. Solar panel. 7. Improved Sanitation: The household is not deprived if it has access to one of the following improved sanitation facilities: a. Flushing toilet (WC) b. Individual/private improved latrine c. Public/collective improved latrine EN [page 201] : Investing in People to Fight Poverty in Haïti The two dimensions of poverty: 1. Income poverty: The household is poor in the monetary dimension if its annual consumption per capita is Less than the poverty Line (29,909.87 gourdes). 2. Non-income poverty: The household is poor in the non-monetary dimension if it is deprived in 3 or more dimensions of the multidimensional index. The categories of pouerty: 1. Chronic: monetary and non-monetary poor. 2. Deprived: non-monetary poor. 3. Transient: monetary poor. 4. Resilient: not poor in all dimensions. [page 202] WorldBank - ONPES | Appendix E. The evolution of the characteristics of households (poor and nonpoor) Table E1. Characteristics of poor households, 2001 and 2012 | Variable National Urban Rural National Urban Rural Puis | | ml | | Dengenmberoïcéenegedo-apemosene | 05 | 06 [os | 0 | ous | os | Sources: ECVH 2001; ECVMAS 2072; World Bank and ONPES calculations. EN [page 203] Investing in People to Fight Po n Haïti Appendix F. Poverty correlates Table F1. Linear regressions to identify poverty correlates, by area of residence Model 1 Model 2 aperdue/pouerg tre MS RE CE EE SE expenditure/povuerty line Nbr children 00-04 (0.010) (O.0147) (0.0165) (0.010) (O.0147) (0.0166) Nbr children 5-14 (0.00716) (0.00973) (0.0106) (0.00718) (0.00977) (0.0107) -0.0787*** -0.0823*** -0.0776*** -0.0783*** -0.0815*** -0.0779*** Nbr adults 15-64 (0.00546) (0.00658) (0.00910) (0.00546) (0.00660) (0.00910) -0.0538** -0.0285 -0.0610* -0.0529** -0.0303 -0.0580* Nbr adult 65+ (0.0233) (0.0322) (0.0343) (0.0233) (0.0322) (O.0344) -0.0125 O.00814 -0.0627* 0.0469 -0.00278 O.0653 Woman head (0.0221) (0.0271) (0.0362) (0.0525) (0.0835) (O.0715) Age head 0.0143*** O.00499 O.0224*** 0.0135*** O.00466 O.0214*** ge hea (0.00363) (0.004871) (0.00569) (0.00368) (0.00489) (0.00578) -0.0001726*** -2.22e-05 -0.000205*** -0.000120*** -197e-05 -0.000200*** Age head squared (G.6le-05) (4.98e-05) (547e-05) (.63e-05) (5.03e-05) (5.50e-05) =1if household receives -0.0230 -0.0364* 0.007289 -0.0235 -0.0359* 0.00139 private transfers (excluding remittances) (0.0165) (0.0210) (0.0260) (0.0165) (0.0210) (0.0260) 21 if Household receives 0133" 0.268*** 0189°** 0133" 0.269°** remittances (0.0187) (0.0218) (0.0325) (0.0187) (0.0218) (0.0326) Primary not completed (0.0226) (0.0321) (0.0332) (0.0226) (0.0321) (0.0332) REVONLEES 0.292“* 0.259** O.291* 0.292** 0.258" not completed (0.0263) (0.0331) (0.0431) (0.0263) (0.0332) (0.0432) non-com (0.0295) (0.0356) (0.0533) (0.0295) (0.0357) (0.0533) 0.619*** O.584*** 0.626*** O.615*** O.587*** O.604*** Sec2 completed & university (O.0441) (O.0484) (0.0978) (0.044) (0.0486) (0.0982) -0260°** oz" -0263°** -0315** -0162 Unemployed (0.0887) (0102) (0.221) (0.0887) (0102) (0.221) -0.337% -0.259 0.339" -0.353"* -0258 Inactive (0.0894) (0106) (0.219) (0.0894) (0106) (0.219) -00958 | 00563 “00650 | -o1 Skilled worker (0.0740) (0.0726) (0192) (O.0741) (0.0727) (0192) [page 204] -0191** -0164** -0165 -0192* -0165** -0179 Unskilled worker (0.0790) (0.0784) (0199) (0.0790) (0.0785) (0.200) -0.255*** -0.228** -0.251 -0.257** -0.233** -0.245 Laborer (0.0824) (0.0837) (0.201) (0.0824) (0.0838) (0.202) & O.0505 0137 0.0161 O.0484 0132 O.0171 ner (0.0863) (0.0914) (0.200) (0.0863) (0.0914) (0.200) -0.0326 -0.0469 -0.00576 -0.0354 -0.0541 -0.00533 Self employed (0.0867) (0.092) (0.201) (0.0867) (0.0913) (0.201) -O.0171 -OT4 0.00750 -0.0254 -0109 0.017129 Family aide (0124) (0.230) (0.228) (0124) (0.230) (0.229) O174*** 0163* 0165* O173*** 0164** O161* Industry/construction (0.0468) (0.0707) (0.0824) (0.0469) (0.0709) (0.0824) TE O155*** 0102* O.246*** 0155*** 0103* O.255** rade (0.0320) (0.0613) (0.0494) (0.0321) (0.0617) (0.0494) O.272** O.272** O.314** O.267*** O.272** 0.316** Transportation (0.0609) (0.0797) (0136) (0.0609) (0.0801) (0136) -O145** -0194** -0.0837 -O145** -0190* -0.0813 Education/health (0.061) (0.0802) (0136) (0.0610) (0.0804) (0136) 0143*** 0153* 0.0209 0142*** 0153** 0.0264 Other services (0.0382) (0.0637) (0.0673) (0.0382) (0.0640) (0.0673) 0.0672 0.0538 0.0576 0.0665 0.0516 O.0547 Large private enterprise (0.0570) (0.0567) (0139) (0.0570) (0.0567) (0139) -0.0519 -0159 0139 -0.0440 -0150 O135 Small formal (0.0970) (0103) (0.202) (0.0970) (0103) (0.202) -0.256*** -0.297*** -0154 -0.257*** -0.298*** -0159 Small informal (0.0672) (0.0706) (0147) (0.0672) (0.0706) (0147) -0.0524 -0.0487 O.OT4 -0.0532 -0.0456 0.00703 Association, NGO (0.0722) (0.0754) (0160) (0.0722) (O.0754) (0160) -0186*** -0.233** -0.0893 -O197** -0.236*** -0.0966 Household (0.0713) (0.0726) (0163) (0.0713) (0.0727) (0163) placé -0.0574*** -0.00602 -0.0828** -0.0643** -0.00832 -0.0797** acé (0.0219) (0.0298) (0.0329) (0.0260) (0.0360) (0.0380) -0.000993 0.0195 -0.00587 0.0959 0102 0121 Cohabiting (0.0417) (0.0498) (0.0713) (0.0587) (0.0697) (0101) gi -0.0493 -0.0382 -0.0419 -0.0791 -0.0430 -0.0585 ingle (0.0406) (0.0497) (0.0669) (0.0528) (0.0668) (0.0836) O456** O430*** 0.302 O.822*** 0.783*** O153 Divorced (0155) (0140) (0.984) (0.223) (0198) (0.986) __. -0142* -0160** -0.0837 -OT4 -0.0878 -0.0784 Séparé après mariage (0.0627) (0.0736) (0107) (0.0961) (0123) (0150) . -0.0834** -O137** 0.0182 -0132* -0110 -0.0866 Séparé après plaçage (0.0395) (0.0468) (0.0679) (0.0735) (0101) (0109) [page 205] Investing in People to Fight Poverty in Haïti -0107** -0112* -0.0742 -0161*** -0.0893 -0185** Veuf / Veuve (0.0382) (0.0495) (0.0596) (0.0567) (0.0865) (0.0809) -0.319* 0.0552 -0190 -031* 000535 Spouse in the household (0121) (0162) (0183) (0123) (0166) (0186) 0.00728 0.00526 0.00479 0.00763 0.00489 0.00600 Age of spouse (000482) (0.00678) (0.00711) (0.00489) (0.006971) (0.00720) -6.47e-05 -6.24e-06 -6.11e-05 -6.61e-05 -2.24e-06 -6.83e-05 Age of spouse squared Gie-05) (7.50e-05) (7.36e-05) (G5e-05) (7.57e-05) (ZHe-05) 0108** 0.0977** 0105** 0107*** 0.0939** 0106** Primary non-completed (00302) (0.0457) (0.0428) (0.03072) (0.0457) (0.0429) REV 0.231 O148°** 0175" O.225** 0153" & Sec1 non-compl (O.0346) (O.0473) (O.0543) (O.0347) (O.0476) (0.0545) sect completed 0197*** 0.229** 0.243*** 0198*** 0271** 0251** & Sec.2 non-com (0.0380) (0.0479) (0.0767) (0.0384) (0.0488) (0.0769) Sec2 completed O484*** 0.551*** O.518*** O.485*** O.540** O.526*** & university (0.0585) (O.0645) (0150) (0.0588) (0.0651) (0150) -0161*** -0113 -0168 -0159*** -010 -0156 Unemployed (0.0600) (0.0723) (0109) (0.0600) (0.0724) (0109) l -O137** -0.0796 -0176** -0128*** -0.0704 -0187** nactive (0.0488) (0.0648) (0.0732) (0.0492) (0.0652) (O.0745) Marié * Femme RE PE A NT TT Placé * Femme RS PE CE NE En union libre * Femme Célibataire * Femme Divorcé * Femme RS PE TE NT RE EE A A A CE RE AE RE TV TE 0.00149** -0.00207 0.00216** 0.00153** -0.00202 0.00234*** Land cultivated (0.000737) (0.00300) (0.000888) (0.000739) (0.00300) (0.000894) 8.27e-07 2.58e-06 -8.85e-07 6.17e-07 2.81e-06 -128e-06 Land cultivated squared (2.60e-06) (2.54e-O5) (G.01e-06) (2.60e-06) (2.54e-05) (3.02e-06) [page 206] WorldBank - ONPES | Area of residence (omitted: urban) om À À À os" | | Department (omitted: Artibonite) Grand'Anse (0.0437) (O.0777) (0.0570) (0.0437) (0.0786) (0.0570) | our |-owo | om” | ww | ow | on | (0.0327) (0.0433) (0.0500) (0.0328) (0.0435) (0.0500) Nord-Ouest (0.0374) (0.0605) (0.0509) (0.0374) (0.0606) (0.0509) (O.0360) (O.0642) (O.0475) (0.0361) (O.0642) (O.0477) Constant (0.137) (0.168) (0.273) (0.139) (0.276) Statistics Note: Standard errors are shown in parentheses. *#* p<OOI ** p<O.O5. * p<OI. [page 207] Inves n People to Fight Po iti Appendix G. Correlates of poverty and food security Table G. Correlates of poverty and food security Rural households Farm households Dependent variable EE ES) con | om | om | Cow A M TE EE D] vou | com | vous | vos | SNS NS EE D com | vw | co | vo | value NS NS NT TT D vom | com | von | com | NS NS NS ER A CS M | 00m | m0 | (00309 | (0059 | PC ES OS ME À 006) | (0535) | 058 | (0052 | A CE RES PS CS CE CS M NS ME [1 [ Gmre-os | Goo0mn) | Geo | 655605 | PC OS CE PE CS A EN [page 208] WorldBank - ONPES | Household composition Note: The table shows logit regressions with department-fixed effects. The control variable for department-fixed effects is not shown Marginal effects are reported with standard errors in brackets. ** p <O01 ** p <O005 * p<Oi [page 209] : Investing in People to Fight Poverty in Haïti Appendix H. Definition of concepts Working-age population: Population of 15 years of age or older. While Labor ques- tions are being asked to all household members above age 10, in the urban context, the age of15 is deemed more appropriate, notably to avoid capturing in employment indicators child Labor. Haiti's Labor Code (Article 335) states that the minimum emplo- yment age in all sectors is 15 years, except in the case of children working in domestic service. The Labor Code (Article 341) sets the minimum employment age for domestic work at 12 years of age. AL working children between the ages of15 and 18 must be re- gistered with the Ministry of Social Affairs and Labor. The Labor Code prohibits minors from working under dangerous conditions and prohibits children under the age of18 from working at night in industrial enterprises Employed or occupied: People in the working-age population that worked for at least an hour the week before the survey or that did not work that week but have a job that will resume in Less than a month. Unemployed: * Definition of the International Labour Organization (ILO): People in the wor- king-age population that don't have a job but are looking for one and are imme- diately available to work if they find one. + _ Extended: Contains all unemployed people under the ILO definition, plus those in- dividuals that are not actively Looking for à job either because they are discouraged of searching for a job and not finding one, are waiting for a job answer, or are in retire- ment or sick, but are available to work immediately if they were offered a job. Active population (or Labor force): People in employment or unemployment. Labor force participation rate (or economic activity rate): Percentage of the wor- king age population who are in the Labor force. Employment rate: Percentage of employed people in the working age population Unemployment rate: Percentage of unemployed people in the labor force (for both the ILO definition and extended unemployment). Underemployment: + Time-related underemployment: Employed people that work less than 35 hours per week, would like to work more hours and are willing and available to do so in the case they get a job offer. *_ Invisible underemployment: Employed individuals who earn Less than a mini- mum amount of money an employee should earn. (In this case, G 250 per day = G 7,500 monthly. This was the minimum wage before October 2012). Invisible underemployment rate: Percentage of invisible underemployed people in the employed population [page 210] WorldBank - ONPES | Time-related underemployment rate: Percentage of time-related underemplo- yed people in the employed population. Informal sector: Unincorporated enterprises (household businesses) that are not re- gistered or do not keep formal accounts and are not in the primarysector (agriculture). Informal employment: all contributing family workers, all independent workers in the informal sector, and all employees without written contracts and not bene- fiting from social protection - Not in the primary sector (agriculture). Demographic dependency ratio: Ratio of the number of demographic depen- dent people (population younger than 15 years of age or older than 70) and the number of demographic independent people (population between the ages of15 and 70). Economic dependency ratio: Ratio of the number of economically dependent people (population economically inactive between the age of 15 and 70) and the number of economically independent people (population economically active between the age of15 and 70). Childcare ratio: number of children under age 15 in a given household; variable to the used as a control variable in analysis related to female Labor force participation. Decent work: ILO defines decent work as the expression of the aspirations of peo- ple in their working lives. It involves opportunities for work that is productive and delivers a fair income, security in the workplace and social protection for families, better prospects for personal development and social integration, freedom for peo- ple to express their concerns, organize and participate in the decisions that affect their lives and equality of opportunity and treatment for all women and men. [page 211] Inves n People to Fight Po iti Appendix I. Correlates of Labor income, unemployment, underemployment, and informality in urban areas Table 11. Correlates of Labor income, unemployment, underemployment, and informality in urban areas, Haïti ni of hourly Unemployed Invisible underemployment Informal employment Independent variables RE Gender = woman -0.318°* 0175*** O493*** 0.0635** 0.0618** [| ©oxo (00253 | (00740) (00265) (©0709 | (oz | (oo Age 15-24 0.0360 0.0707 0132** 0.379** 0.0513** 0.236** [| oo (0.0782) (0206) (0.0486) (0169) (00179) (0103) Older than 55 0159 O132*** O.378*** O.0297 O.0440 —0.00823 —0.00720 fo (0123) (O.0314) (O.0970) (0.0395) (0122) (O0.0147) (0.0850) Primary completed, lower O.279*** -0.00666 -0.0344 —0.0616*** —-0172** —O.0445** —0.287*** secondary not ES @on | (00375) (©0179) (00514) (00168 | (0.0956) Lower secondary comple- O0465*** —0.0669*** —0.202*** —0137*** —0.376*** -0113*** —0.548*** ted, upper secondary not fo (0.0533) (O.0118) (0.0321) (0.0166) (O.0475) (O.0148) (0105) Upper secondary comple- 1.250*** —-0153*** —0.448*** —0.313*** —0.848*** —0.374*** —1.252*** ted, university EE) (©0159 | (O4) (00261) (Oo747) (00593) (0135) 00289 | -oo154* | -oou4** | -000771* | -00262** | -000254** | -000935 fo (0.00705) (0.003072) (O.00842) (0.000981) (0.00356) (0.007110) (0.00699) -0.000362** | 978e-05** | 0000271“ | 0000101** | 0000373** | 1808-05 | 6.35e-05 [| ©ocoio2 | 62-05 | (000109 | Guse-05) | GBte-05 | (1536-05 | (000013) on [ (0106) (O.0358) (O.0945) (O.0413) (0146) (O.05172) (0.260) Source: ECVMAS 20172; World Bank and ONPES calculations. Note: The economically active population includes people over 15 years of age only. Informal employment is defined as all contributing family workers, all independent workers in the informal sector, and all employees without written contracts and not benefiting from social protection. Invisible underemployment is defined as employees earning less than the minimum wage, which is set at G 250 a day = G 7500 a month. The proxy of labor market experience is equal to the age minus the age assumed for the last level of education completed. minus 5. Reference variables: age = people 25-55 years ofage; education = no education and incomplete primary school. Standard errors are shown in parentheses. OLS = ordinary least squares. *** p <O.O1 ** p <0.05 * p <O1 [page 212] WorldBank - ONPES | Appendix J. Mincer earnings function and Oaxaca-Blinder decomposition: a methodological clarification Mincer earnings function The Mincer earnings function is an equation named after Jacob Mincer (1958) that explains the correlation between Labor earnings and the levels of education and work experience. This equation takes the following form: Iny=c+aEDU+$EXP+OEXP?+yX+e Li Where Iny represents the natural logarithm of Labor income, in this case hour- ly Labor income, EDU EXP2 represents the level of education, is the number of years of work experience and is its squared value. We can also include other in- dividual characteristics on the right-hand side of the equation in order to obtain more precise estimates of the value of the correlations, represented by the value of their coefficients in the equation. These individual characteristics are represen- ted by and can include variables such as gender, age, industry of Labor activity, among others. We used the information of the ECVMAS 2012 in order to run the Mincer equation and find out the main correlates of Labor income in urban areas in Haïti. The re- sults, shown in Table M, confirm the existence of a gender gap in terms of hourly labor earnings, even after controlling for individual characteristics. In particular, women that share the same individual characteristics than men (such as level of education, experience, age, geographical Location, household size, number of young children in the household and industry of Labor activity) earn in average 32 percent less. Table M also shows that the youngest group of workers earn in average around 14 percent less than workers in the middle age group (i.e. workers between 25 and 54 years of age, who are the reference group), holding everything else constant. Education also plays an important role in determining Labor earnings. Al educa- tion variables are significant and the magnitude of their coefficients is as expec- ted. In particular, a higher Level of education is correlated to a higher Labor income. Holding everything else constant, someone with primary education completed is expected to earn in average 26 percent more than someone with no education at all. In line with these results, people with a first or second level of secondary completed, or a university-level of education are respectively expected to earn in average 43 percent and 119 percent more than someone without education. Labor experience affects Labor income in a positive but concave way, that is, every additional year of experience increases Labor income in a magnitude lower than the previous additional year of experience. Given that the relationship between labor income and labor experience is not expected to be linear, we include the EN [page 213] : Investing in People to Fight Poverty in Haïti squared of the level of experience in the equation and thus we have to consider its coefficient when calculating the marginal effect of Labor experience on hourly Labor income. After doing so, holding everything else constant, one extra year of Labor experience is associated to an increase of 2.65 percent in hourly Labor income."°? Table J1. Mincer equation results, urban areas, Haiti Dependent variable: Logarithm of hourly labor income Independent variable Coefficient Upper secondary completed or university T7 Source: ECVMAS 2012 Note: The control variables include number of children (younger than 15) in the household, household size, a dummy variable that indicates whether the household is or is not in Port-au-Prince, and the industry of labor activity. Reference variables: age: between 25 and 55 years; education: no education and incomplete primary school. Oaxaca-Blinder decomposition We use the Oaxaca-Blinder decomposition to study the difference between male and female hourly Labor income in urban areas in Haiti. In table 1.2, we can see that the hourly Labor income for females is around 87 percent of that of males. In total, 159 The marginal effect of Labor experience on Labor income is given by the derivative of over, which considering the coefficients is equal to 0.0268-*0.000331. When analyzing the effect of an extra year of Labor experience we replace by1, and the result is 0.0265. [page 214] WorldBank - ONPES | the difference between men's and women's earnings is about 0.46 gourdes per hour worked. This difference in wages might be, at some extend, explained by differences in individual characteristics between men and women. For instance, if men are in average better educated than women, it is expected to find a higher hourly Labor income for men than for women. However, if we control for those characteristics, the Labor earnings for women and men should be the same under no gender discrimination. The Oaxaca Blinder decomposition helps us find out which part of the gender earnings gap is explained by observable and unobserva- ble characteristics. Table J.2. Average hourly Labor income, urban areas, Haïti Men Women Difference EE PE We have calculated the Oaxaca-Blinder decomposition using three different spe- cifications. The first specification includes age and level of education as the in- dividual characteristics that could explain the gender earnings gap; the second specification includes the same observable characteristics as the first one plus the number of children in the household; while the third specification includes those included in the second one plus dummies for the industry of activity. The results are summarized in table 1.3 and figure 11. The first and second specifica- tions suggest that differences in the observable characteristics (or endowments) account for about 32 percent of the gender wage gap, while the other 68 percent remains unexplained. The industry of activity (third specification) explains a little bit more of the gender wage gap. In particular, according to the third specification, characteristics such as age, the level of education, the number of children in the household and the industry of activity explain almost 36 percent of the gender wage gap, but the other 64 percent remains unexplained. Table J.3. Gender earnings differentials, Oaxaca- Blinder decomposition, urban areas, Haiti Total 046 O46 O46 [page 215] : Investing in People to Fight Poverty in Haïti Figure J1. Blinder-Oaxaca decomposition for different specifications, urban areas, Haïti G) = (2) + Industry of Activity 35.71% (2) = (1) + Number of children in the 31.67% (1) = Age and level of Education 31.42% @ Epiained SSSSÉSÉSÉSSSSE : ON M + nn © KR © a © @ Unexplained 4 An important caveat of these results is that they might include some selectivity bias in the sense that the gender gap is calculated only for people working, thus selected into the Labor market, as well as a high probability of auto-selection into particular industries of activity. Anyhow, the magnitude of the gender earnings gap unexplained by observable characteristics suggests a worrying presence of gender discrimination in the Labor market. The fraction of the gender wage gap unexplained by observable characteristics in Urban Haiti is higher than in African and LAC countries. According to Nopo (2012), the part of the gender wage gap attributed to differences between men and wo- men that cannot be explained by observable characteristics in LAC countries is in average around 18 percent (for circa 2007). There is, however, a large variation of this result across LAC countries, for instance the highest reported is Nicaragua with 28 percent and the lowest is Colombia with 7.3 percent, but none is higher than in urban Haïti, On the other hand, Nordman et al (2013) shows that for the main cities of 7 African French-speaking countries in 2001/2002, these results range from 40 to 67 percent, which are a bit closer to the urban situation in Haiti in 2012. For exam- ple, the unexplained part of the gender gap in Lomé (Togo) is around 45 percent after controlling for sector dummies, while for Ouagadougou (Burkina Faso) it is 67 percent. [page 216] l E: Appendix K. Correlates of enrollment and progress in school Table K1. Correlates of enrollment and progress in school Marginal effect Variable PR PS MT PS MT DS PS HT DS PT PS MNT RE ES ES PT RS NT DS PS MT DS mary PS MT DS PS MT PS D DS NT ES PS HT PS HT PS MT EN [page 217] : Investing in People to Fight Poverty in Haïti Note: The regression is estimated for children aged 10-14 for overage and 6—14 for enrollment The marginal effects are evaluated at the sample means. Omitted household head education level = no schooling Omitted department = Nord-Est Robust standard errors are clustered at the household level. Significance level: * = 10 percent, ** = 5 percent, ** = 1 percent 200 [page 218] WorldBank - ONPES | Appendix L. Descriptive statistics on the shocks reported by households Table L1. Idiosyncratic economic shocks affecting households Shock Description Illness or serious accident of a household member Health Death of family member Household composition Care of a new household member Broken agricultural equipment or tools Failure of a family nonagricultural business Economic activity = Loss of salary/income of household (not due to illness/accident) Termination of aid (transfers) from family/friends Decrease in outside help Termination of aid (transfers) from government Crime Theft of goods or harvest . Food shortages in stores Economic shocks affecting the community Increase in the price of seed or fertilizer Hurricanes and floods Note: The questionnaire contained an additional question on the death ofa nonfamily household member, but responses were not reported. [page 219] : Investing in People to Fight Poverty in Haïti Table L2. Prevalence of types of shocks faced by households, by Location Type of shock Port-au-Prince Other urban Rural Idiosyncratic household shock Household composition Decrease in outside help ... shock affecting the Weather/climatic shock Number of observations Source: ECVMAS 2012; World Bank and ONPES calculations Table L.3. Impact of three main types of shocks, by household poverty status Type of shock In extreme pouerty | In pouerty, but not extreme Vulnerable, but not poor Resilient Shock1 Moe | où | où | æ | w | ER AS RS ETS EC A PS EC A RS EE AS RE NT PS ET Source: ECVMAS 2012; World Bank and ONPES calculations EN [page 220] WorldBank - ONPES | Appendix M. Coping mechanisms Table M1. Shocks: main coping mechanisms Mechanism Description Use of savings " Monetary help from friends and family Financial help Monetary help from central or local government Monetary help from religious organizations or NGOs Nutritional help from relatives or friends Nutritional hel Nutritional help from central or Local government p Nutritional help from religious organizations or NGOS Work-for-food Decreased quantity of food, number of meals consumed Decreased quality of food consumed Change in nutritional inputs Consume premature harvest Consume foods gathered in the wild Consume seeds : Active members of household engage in additional work Change in Labor output : . inactive, unemployed members engage in work he Tienneon Migration of one or more household members n p Sending children to another household Decrease in household expenditures DEGIERE® mn meneree) spending Decrease in health spending Pulling children out of school Pulling children out of school Borrowing from family or friends Debt R Borrowing from Lenders or merchants Sale of agricultural assets Sale of household durable goods (work tools, equipment) HACHES Sale of land, real estate Sale of agricultural produce, seeds Sale of cattle Sale of equipment, tools used for revenue generation Fishing more frequently Use of (common) resources Cut wood, make charcoal Increase harvest and sell natural resources Engage in spiritual activities Other mechanisms Begging Other strategy [page 221] : Investing in People to Fight Poverty in Haïti Table M.2. Coping mechanisms to address the most important shocks, by type of shock Idiosyncratic economic shock Couaria- s ALI Household Decrease tæ eco- | © me SHRCE SROCRS | Health |composi- | Agricultural Economic |inoutside Crime OMiC Mon tion help shock . Change in nutritional inputs low | 005 | oo | où | oo | on |oos) ow | où Change in household composition loco | 000 | oo | o0o | oo | oco |o0o) |oo Decrease in household expenditures loos | 005 | oo | oo | oo | oo |oo3| oo | oo | Use of (common) natural resources [oo | 000 | oo | oo | 000 | on) oo | oo | Source: ECVMAS 2012; World Bank and ONPES calculations Table M.3. Coping mechanisms for the most important shocks, households in extreme poverty Idiosyncratic economic shock Coua- ALI riate | Cli- Strategy shoc- House- Agricul- Bc- Decrease cri- eco-|matic ks Health hold com- All _ in outside Me nomic | shock position help shock Change in nutritional inputs ox | oo | où | 0% | où | ox |oo) 05 | ox. Change in household composition Dooi | oo | — | oo | oo | - | -) - |oo. Decrease in household expenditures loc | 003 | o02 | oo | ow | oos | | oo | on Use of (common) natural resources loo2 | oo | oo | ow | oo | —- | - | oo | où. Source: ECVMAS 2072; World Bank and ONPES calculations Note: — = not available. EN [page 222] WorldBank - ONPES | Table M.4. Coping mechanisms for the most important shocks, resilient households Idiosyncratic economic shock Coua- AIL shoc- riate | Cli- a ral Mic me nomic | shock position help shock Change in household com- een nn [os [om | où | où | ow | où [on] oo | 00: | expenditures Removal of child from Use of (common) natural oe Dos [oo] où | 00 | - | - |-] ce) Source: ECVMAS 2072; World Bank and ONPES calculations Note: — = not available [page 223] : Investing in People to Fight Poverty in Haïti Appendix N. Results of the multivariate analysis of shocks Table N1. Correlations of the main shocks experienced by households Des SRE DERelE (per GA Ge Only shocks Interactions, all MERS Ge Interactions, resilient penditure, In treme pouerty Main shock RE SE ES EE A PE 2 PS ES ES D 1 A D ES RS RS RS AC PE A Main shock: idiosyncratic Mocameneun "| | où | où | sw D 1 0 | 0 | w | sens [| en | or | | D 1 | 0% | 0% | em | omsrnmenmes | | où | | D A ON M ES AE EN EE Do | om | os | vs | as [| | æ | eu | D es | em | w | Ce | es | om | en | D Don | os | Main shock: covariate economic Mocameneun "| | os | om | æw | RE ER RC TC Ds | | + | où | en | DT 0m | ox | vw | RE ES ES ES D A TN M ER PE AE EE D A NT EN [page 224] WorldBank - ONPES | Besse Je Le | = | en | es | ww | as | en | ww | x | D A TN MN Main shock: covariate weather ocean |" | | os | où PE A EN EN nmmnenns [| ee | æ | er | D A NT M EC ES ES ES A NN M ae um | on | er | en Len | x | AE ES D TS EE AR EE ss [| | æ | ss | D A NN M Household characteristics D Low | en | æs | en | D [uw | 0 | ww | wn | D [of on | uw | vw | D [om | em | on | vw | D Low | om | em | on | D Jeu | ow | ww | vx D Tee | cm | ww | vw | D [om | ww | os | on | EN [page 225] : Investing in People to Fight Poverty in Haïti Note: The reference individual for the full model is a the head of a man-headed household with no formal education, but employed: the household is in Port-au-Prince and has not experienced any ofthe shocks considered. The access to some ofthe coping mechanisms is potentially correlated with income. If such a relationship exists and given that the variance in per capita expenditures is much larger among the resilient population than among the population in extreme poverty, the coefficients reflect opportunities, not merely the actual correlation with the particular strategy. To make the results more tractable, we have aggregated the shocks into three shock categories: idiosyncratic household shocks, covariate economic shocks, and covariate weather shocks. Similarly, we have aggregated the coping strategies into three categories based on frequency of use: no coping mechanism used: monetary and nutritional help; and changes in nutritional inputs, debt, sale ofassets, and other strategies utilized. The first column presents the results if only shocks are included. The second column presents the results if the shocks and coping strategies are introduced. The third and fourth columns present the results for the sample ofhouseholds in extreme poverty and for resilient households, respectively. Standard errors are in parentheses. ** p <O.01 ** p <O.05 * p <O1 208 [page 226] WorldBank - ONPES | Appendix O. Incidence maps of weather events Map O1. Flood-prone areas, Haiti Legend _ 77 Nord-Ouesth, Departments LE K€ & Tee | | Flood-prone areas CN LA By te ftibomite MM 7, | Ÿ ‘ Ent ÿ _ = se VX ne (FR Be Mérend Anse À (4 Ouest LS Source: Based on data of ‘Shakemap us2010rja6;' Earthquake Hazards Program, United States Geological Survey, Reston, VA, http://earthquakeusgs.gov/earthquakes/shakemap/global/ shake/2010a6/. Note: Flood-prone areas were identified by the United Nations Institute for Training and Research in May 2010. Map O2. Hurricanes, depressions, and tropical storms, by department, 1954-2001 Number of cases 3-4 XX, 4-6 Nord - Ouest 6-8 Nord Si Nord - Est 8-10 Artibonite À si 10-16 Le $ Centre œ Grand Anse .. sud Sud- Est n ] Sources: Based on Mathieu et al. 2003; "Shakemap us2010rja6," Earthquake Hazards Program, United States Geological Survey, Reston, VA http://earthquake.usgs.gov/earthquakes/ shakemap/global/shake/2010rja6/. [page 227] : Investing in People to Fight Poverty in Haïti Map O3. Drought-prone areas, Haïti SR Legend ce EH Drought-prone areas U Nod | Sÿ SF * Nord-Est Departments F5 à Artiboi ! A ÿ Centre CE : ET D Li = _S _ oue Sud ET = Source: Based on data of "Shakemap us2010rja6,' Earthquake Hazards Program, United States Geological Survey, Reston, VA, http://earthquakeusgs.gov/earthquakes/shakemap/global/ shake/2010ra6/. Note: Haiti's drought zones were identified through the NATHAT Project, using information from the Centre National de Météorologie of Haiti, in May 2010. Map O.4. Earthquakes, by magnitude, intensity, and economic damage, Haïti, 1701-2014 Legend — Magnitude . es Nord - Ouest Les Above $8 billion e _ Nord Between 510M to $34M Nordzest Between $1M to $10M a Less than SIM 7 centre No damage % » + 33-43 no 1860 ms © 43-57 So " (est © 57-74 ; Sus_E# © 7:85 US Sources: Based on data of "Shakemap us2010rja6,' Earthquake Hazards Program, United States Geological Survey, Reston, VA, http://earthquakeusgs.gov/earthquakes/shakemap/global/ shake/2010rja6/; Earthquake Data and Information (database). National Geophysical Data Center, Boulder CO, http://www.ngdc.noaa.gov/hazard/earthqk.shtml [page 228] WorldBank - ONPES | Map N.5. Soil Liquefaction incidents, February 2010 Legend _ = © Liquefaction incidents February 2010 = Nord ouest PE Solis liquefaction susceptibility de er 5 Ds LS LR nd ê nt À O0 1 2 3 P — SES S : Das © [a Departments _ è . \ ë centre _rand Anse h LT ue : Ouest ES Source: Based on data of ‘Shakemap us2010rja6,' Earthquake Hazards Program, United States Geological Survey, Reston, VA, http:/earthquakeusgs.gov/earthquakes/shakemap/global/ shake/2010rja6/. Note: Data on the susceptibility to the soil liquefaction hazard in Haïti and on tiquefaction (landslide) incidents in Haiti during and after the earthquake of January 12 2010, were collected through the NATHAT project in, respectively, February and May 2010. Map ©.6. Landslide incidents during and after the earthquake of January 12, 2010 Legend © Landslide incidents JS Haiti Landslide Predisposition Index 7 Nord -ouest À _ Value ge nn ( 2s NS LE 0 Nord pl # Above 2 a dire Between 1.3 and 2 sr " et ic Between O.8 and1.3 E Se > 7 dt E { - Centre Between O and O.8 « à SE te * ae Departments un Rs DMC — Se RE, “ESS ï aie > BR S = Source: Based on data of ‘Shakemap us2010ra6,' Earthquake Hazards Program, United States Geological Survey, Reston, VA http:/earthquakeusgsgov/earthquakes/shakemap/global/ shake/2010rja6/. Note: A map of landslide incidents was created through the NATHAT Project in February 2010. The landslide predisposition index built for the rainy season in the absence of an earthquake was created through the NATHAT project according to the GIPEA method in May 2010. [page 229] : Investing in People to Fight Poverty in Haïti References Acemoglu. D. and S. Johnson. 2007. “Disease and Development: The Effect of Life Expectancy on Economic Growth”. Journal of Political Economy, vol. T5, no. 6. Acosta, P, C. Calderén, P. Fajnzylber, and H. Lépez. 2006. "Do Remittances Lower Po- verty Levels in Latin America?”. In: Fajnzylber, P.; Humberto L6pez, J. (ed.) (2006). “Remittances and Development: Lessons from Latin America” The World Bank. Latin America Development Forum Series. Adams, À. M, T. Ahmed, S. E. Arifeen, T. G. Evans, T. Huda, and L. Reichenbach. 2013. “Innovation for Universal Health Coverage in Bangladesh: A Call to Action” Lancet 382: 2104-11. Adelman, M, T. Heydelk, P. Ramanantoanina, A. Latortue and M.M. Manigat. 2014. “Bac- kground paper on Education in Haiti”. Washington, DC: World Bank. Alderman, H, E. and King. 2006. “Investing in Early Childhood Development”. Research Brief, The World Bank. Archbold, Randal C. 2012. ‘Already Desperate, Haïitian Farmers Are Left Hopeless after Storm” New York Times, November 17. Arora S. 2001. “Health, Human Productivity, and Longer-Term Economic Growth" Jour- nal of Economic History 63: 699-749, Aryeetey G. C., C. Jehu-Appiah, E. Spaan, l. Agyepong, and R. Baltussen. 2012. “Costs, Equity, Efficiency, and Feasibility of Identifying the Poor in Ghana's National Health Insurance Scheme: Empirical Analysis of Various Strategies.” Trop Med Int Health 17: 43-51. Atuesta, Bernardo, Facundo Cuevas, and Aude-Sophie Rodella. 2014. “Labor Markets and Income Generation in Urban Areas.” Background paper, World Bank, Washin- gton, DC. Barrett, C. B, M. Bellemare, and J. Hou. 2010. “Reconsidering Conventional Explana- tions of the Inverse Productivity-Size Relationship," World Development, Elsevier, vol. 38(1), pages 88-97, January. Barrientos, À, and D. Hulme, 2008. Social Protection for the Poor and Poorest in De- veloping Countries: Reflections on a Quiet Revolution. Brooks World Poverty Ins- titute, Manchester, UK. Barro, Robert J. 1996. Health, Human Capital, and Economic Growth. Washington, DC: Pan American Health Organization. Barro, Robert J, and Jong-Wha Lee. 2010. “A New Data Set of Educational Attainment in the World, 1950-2010 Journal of Development Economics 104 (C): 184-98. [page 230] WorldBank - ONPES | Basu, K, and P. Van. 1998. “The Economics of Child Labor” American Economic Re- view 88 (3): 412—27. Batiston et À. 2013 “Income and Beyond: Multidimensional Poverty in Six Latin Ame- rican Countries”, Springer. Becker, G. 1964. Human Capital. New York: Columbia University Press. Beeston, Kym. 2010. “Starting from Scratch: Building a Brighter Future for Haïti's Disa- bled Children! Guardian, June 14. Bellos, À, K. Mulholland, KL. O'Brien, S. A. Qazi, M. Gayer, and F. Checchi. 2010. “The burden of acute respiratory infections in crisis-affected populations: a systema- tic review”. Conflict and Health 4:3. Bhalotra, S., and C. Heady. 2003. “Child Farm Labor: The Wealth Paradox” World Bank Economic Review 17 (2): 197—227. Bhargava, À, DT. Jamison, L. J. Lau, and C. J L. Murray. 2001. “Modeling the effects of health on economic growth," Journal of Health Economics, Elsevier, vol. 20(3), pages 423-440, May. Bloom, D, D. Canning, and J. Sevilla. 2004. “The Effect of Health on Economic Grow- th: A Production Function Approach”. World Development Vol. 32, No. 1, pp. 1-13. Bloom, D, J. D. Sachs, P. Collier, and C. Udry. 1998. “Geography, Demography, and Economic Growth in Africa”. Brookings Papers on Economic Activity, Vol. 1998, No. 2 (1998), pp. 207-29 Boesten, J. and Nana K. Poku. 2009. «Gender and HIV/Aids: Critical Perspectives from the Developing World». Surrey, UK: Ashgate Publishing Limited. 204 pp. Bongaarts, John. 2003. “Completing the Fertility Transition in the Developing World: The Role of Educational Differences and Fertility Preferences” Population Stu- dies 57 (3): 321-535. Bowser, D, and A. Mahal. 2011. “Guatemala: The Economic Burden of Illness and Health System Implications." Health Policy 100: 1159-66. Burns, J, S. Godlonton, and M. Keswell. 2010. “Social Networks, Employment, and Worker Discouragement: Evidence from South Africa” Labour Economics 17 (2): 336—44. Buvinic, M, R. Furst-Nichols, and E. Courey Pryor. 2013. ‘A Roadmap for Promoting Women's Economic Empowerment United Nations Foundation, Washington DC. Card, David. 1999. “The Causal Effect of Education on Earnings' In Handbook of La- bor Economics, vol. 3A, edited by Orley C. Ashenfelter and David Card, 1801-63. Handbooks in Economics 5. Amsterdam: Elsevier. [page 231] : Investing in People to Fight Poverty in Haïti Carletto, C, S. Savastano,and A. Zezza. 2013. “Fact or artifact: The impact of measure- ment errors on the farm size-productivity relationship, “ Journal of Development Economics, Elsevier, vol. 103(C), pages 254-261. Carpenter, S, R. Mallett, and R. Slater. 2012. “Social protection and basic services in fragile and conflict-affected situations: a global review of the evidence”. Working paper. Secure Livelihoods Research Consortium. Cavallo, E, A. Powell, and O. Becerra. 2010. “Estimating the Direct Economic Damage of the Earthquake in Haiti”, IDB Working Paper 163. CEDLAS (Center for Distributive, Labor, and Social Studies) and World Bank. 2012. ‘A Guide to the SEDLAC Socio-Economic Database for Latin America and the Cari- bbean’” March, CEDLAS, Facultad de Ciencias Econémicas, Universidad Nacional de La Plata, La Plata, Argentina; Poverty Group, Latin America and the Caribbean Region, World Bank, Washington, DC. http://sedlac.econo.unlp.edu.ar/eng/me- thodology.php. CIAT (Comité Interministerial d'Aménagement du Territoitre). 2013. « Lois et Règle- ments d'Urbanisme: Guide du Professionnel” July, CIAT, Port-au-Prince, Haïti. Cicmil, H. 2013. “Whose Education? Haiti's Girls and Haiti's Recovery”. Thinking Deve- lopment Organization. Available at: http://www.thinkingdevelopment.org/news/ whose-education-haitis-girls-and-haitis-recovery Clemens, Michael A. 2011. “Economic Impacts of H-2 Nonimmigrant Visa Eligibility for Haiti” November 8, Center for Global Development, Washington, DC. ———. 2014. “Does development reduce migration?”. Center for Global Development. Working Paper 359. Clemens, Michael À, and Timothy N. Ogden. 2013. “Migration as a Strategy for House- hold Finance: À Research Agenda on Remittances, Payments, and Development FAI Working Paper 10/2013, Financial Access Initiative, New York University, New York. Coello, Barbara, Ghbemisola Oseni, Tanya Savrimootoo, and Eli Weis. 2014. “Income Ge- nerating Activities and Barriers to Rural Development in Haiti” Background paper, World Bank, Washington, DC. Collier. 2009. “Haiti: from Natural Catastohphe to Economic Security”. Report for the Secreaty-General of the United Nations. Department of Economics, Oxford Uni- versity. Conseil Superieur des Salaires, Republic of Haïti. 2013. “Rapport relatif à la fixation du salaire minimum par secteur d'activités en Haïti” Port-au-Prince, Haïti. Cross, M, À. S. Rajkumar , E. Cavagnero and M. Sjoblom. 2014. “Background paper on Health Haiti” Washington, DC: World Bank. Cuevas,PF. F. Marzo and T. Scot. 2014. “Background paper on migration in Haiti”. Was- hington, DC: World Bank. [page 232] WorldBank - ONPES | Currie, J, and D. Thomas. 1999. “Early Test Scores, Socioeconomic Status and Future Outcomes”. NBER Working Papers 6943, National Bureau of Economic Research, INC. Davis, EP, and C.A. Sandman. 2010. “The timing of prenatal exposure to maternal cortisol and psychosocial stress is associated with human infant cognitive de- velopment”. Child Development, 81 (1), 131-138. Demombynes, G, P. Holland, and G. Leon. 2010. “Students and the Market for Schools in Haiti” Policy Research Working Paper 5503, World Bank, Washington, DC. Dercon, S. 2004. “Growth and shocks: evidence from rural Ethiopia”. Journal of Deve- lopment Economics, Elsevier, vol. 74(2), pages 309-329, August. Devereux, S. 2000. “Social Safety Nets for Poveryy Alleviation in Southern Africa”. A research report for DFID, ESCOR Report R7017 Dilley, Maxx, Robert S. Chen, Uwe Deichmann, Arthur L. Lerner-Lam, and Margaret Arnold. 2005. “Natural Disaster Hotspots: À Global Risk Analysis.” Disaster Risk Management Series 5. With Jonathan Agwe, Piet Buys, Oddvar Kjekstad, Brad- field Lyon, and Gregory Yetman. Washington, DC: World Bank. http://www.pre- ventionweb.net/files/100 _ Hotspots.pdf. Docquier, F., and H. Rapoport. 2007. “Skilled Migration: The Perspective of Develo- ping Countries.” IZA. Duflo, Esther. 2001. “Schooling and Labor Market Consequences of School Cons- truction in Indonesia: Evidence from an Unusual Policy Experiment” American Economic Review 91 (4): 795-813. Duranton, G. 2013. ‘Agglomeration and Jobs in Developing Countries.” Background paper, World Development Report 2013, World Bank, Washington, DC. Duryea, Suzanne, Sebastian Galiani, Hugo Nopo, and Claudia Piras. 2007. “The Edu- cational Gender Gap in Latin America and the Caribbean” Working Paper 5072, Inter-American Development Bank, Washington, D.C. Eastwood, R. M. Lipton and A. Newell. 2010. “Farm size” In Pingali, P. L. and R. E. Even- son, eds, Handbook of agricultural economics. North Holland: Elsevier. Eberhard, M. O, S. Baldridge, J. Marshall, W. Mooney, and G. J. Rix. 2010. “The MW 7.0 Haïti Earthquake of January 12, 2010: USGS/EERI Advance Reconnaissance Team Report” USGS Open-File Report 2010-1048, United States Geological Survey, Reston, VA. http://pubs.usgs.gov/of/2010/1048/. Échevin, Damien. 2011. “Vulnerability and Livelihoods before and after the Haïti Ear- thquake” Policy Research Working Paper 5850, World Bank, Washington DC. _______ 2013. “Characterizing Vulnerability to Poverty in Rural Haiti: À Multilevel Decomposition Approach Journal of Agricultural Economics 65 (1): 131-50. [page 233] : Investing in People to Fight Poverty in Haïti Ellis, P. 2003. Women, Gender and Development in the Caribbean: Reflections and Projections. Zed Books: London, UK. Fafchamps, M, and J. Wahba. 2006. “Child Labor, Urban Proximity, and Household Composition” Journal of Development Economic 374—97. Fafo (Fafo Institute for Applied International Studies). 2009. “Haiti Youth Survey Fafo, Oslo. Fagen, P. 2006. “Remittances in Crises: À Haiti Case Study! Humanitarian Policy Group, ODI, London. Feinstein, L. 200%. “Very Early Evidence”. CentrePiece 8 (2) Summer 2003 pages: 24-30. Ferreira, Francisco, Julian Messina, Jamele Rigolini, Luis-Felipe Lépez-Calva, Maria Ana Lugo, and Renos Vakis. 2013. Economic Mobility and the Rise of the Latin Ameri- can Middle Class. World Bank Latin American and Caribbean Studies. Washington, DC: World Bank. Filmer, Deon. 2004. “If You Build It, Will They Come? School Availability and School Enrollment in 21 Poor Countries Policy Research Working Paper 3340, Washing- ton, DC, World Bank. ________ 2008. "“Disability, Poverty, and Schooling in Developing Countries: Results from 14 Household Surveys” World Bank Economic Review 22 (1): 141-653. Foster, James, Joel Greer, and Erik Thorbecke. 1984. ‘A Class of Decomposable Poverty Measures.” Econometrica 52 (3): 761-66. Fritschel, H., 2002. “Nurturing the soil in Sub-Saharan Africa”. 2020 News & Views, July. International Food Policy Research Institute, Washington, DC Gallié, Camille, and Mario Marcellus. 2013. “Le Systeme de Protection de l'Enfant en Haïti” World Vision Haiti, Port-au-Prince, Haiti. Hanushek, Eric, and Ludger Woessmann. 2009. “Do Better Schools Lead to More Growth? Cognitive Skills, Economic Outcomes, and Causation.” NBER Working Pa- per 14633, National Bureau of Economic Research, Cambridge, MA. Harvey, Paul, Rebecca Holmes, Rachel Slater, and Ellen Martin. 2007. “Social Protec- tion in Fragile States” November, Overseas Development Institute, London. Heckman, J, and D. Masterov. 2007. “The Productivity Argument for Investing in Young Children’Review of Agricultural Economics 29(3): 446-493. Heltberg, R, A. M. Oviedo, and F. Talukdar. 2013. “What are the Sources of Risk and How do People Cope? Insights from Households Surveys in 16 Countries”. Background paper for the World Bank 2014 World Development Report. Herrera, Javier, and Sébastien Merceron. 2013. “Underemployment and Job Mismatch in Sub-Saharan Africa” In Urban Labor Markets in Sub-Saharan Africa, edited by [page 234] WorldBank - ONPES | Philippe De Vreyer and François Roubaud, 83-108. Paris: Agence Française de Développement; Washington, DC: World Bank. Hindman, H. 2009. The World of Child Labor: An Historical and Regional Survey. New York: Armonk. Hossain, À. M, and C. À. Tisdell. 2005. “Closing the Gender Gap in Bangladesh: In- equality in Education, Employment, and Earnings’ International Journal of So- cial Economics 3 (5): 439-653, IEG. 2013. “World Bank Group Assistance to Low-Income Fragile and Conflict-Affec- ted States : an independent evaluation”. Independent Evaluation Group (IEG) working paper series. Washington DC ; World Bank. IHE (Institut Haïtien de l'Enfance) and ICF International. 2014. Haïti: Évaluation de Prestation des Services de Soins de Santé, 2013. Rockville, MD: IHE and ICF In- ternational. IHSI (Haïitian Institute of Statistics and Informatics). 2007. Projections de Population Totale, Urbaine, Rurale et Economiquement Active. Port-au-Prince, Haïti: Minis- try of Economy and Finance. ______ 2010. “Enquete sur l'Emploi et l'Economie Informelle (EEE): Premiers Resultats de l'Enquete (Phase 1)” IHSI, Port-au-Prince, Haiti. _______ 2012. "Population Totale, Population de 18 Ans et Plus, Menages et Den- sités Estimés en 2012” Ministry of Economy and Finance, Port-au-Prince, Haïti. ______204 ILO. 2013. Measuring informality: A Statistical Manual on the Informal Sector and In- formal Employment. Geneva: ILO. ILO and IFC. 2013. “Better Work, Haiti: Garment Industry: 7th Biannual Synthesis Re- port under the HOPE Il Legislation IFC, Washington, DC. IPCC (Inter-Governmental Panel on Climate Change). 2013. Climate Change 2013: The Physical Science Basis. Fifth Assessment Report. New York: Cambridge Uni- versity Press. IMF (International Monetary Fund). 2011. “Investing in Public Investment: An Index of Public Investment Efficiency”. Working Paper WP/11/37. Washington, DC. Jacoby H. 1999. Access to Markets and the Benefits of Rural Roads. The World Bank, Policy Research Working Paper Series: 2028 Jadotte, Evans. 2008. “Labor Supply Response to International Migration and Re- mittances in the Republic of Haiti” Document de treball 0808, Departament d'Economia Aplicada, Universitat Autonoma de Barcelona, Barcelona. [page 235] : Investing in People to Fight Poverty in Haïti _______ 2010. “Vulnerability to Poverty: A Microeconomic Approach and Applica- tion to the Republic of Haïti” Document de Treball 10.04, Departament d'Econo- mia Aplicada, Facultat d'Economia i Empresa, Universitat Autonoma de Barcelo- na, Barcelona. _______ 2012. Brain Drain, Brain Circulation, and Diaspora Networks in Haïti. UNC- TAD. Jamison, Dean T. Lawrence H. Summers, George Alleyne, Kenneth J. Arrow, Seth Ber- kley, Agnes Binagwaho, Flavia Bustreo, et al. 2013. “Global Health 2035: À World Converging within a Generation.” Lancet 382: 1898-1955. Jann, Ben. 2008. “The Blinder-Oaxaca Decomposition for Linear Regression Models.” Stata Journal 8 (4): 453-709. Japan Government and World Bank. 2013. “Global Conference on Universal Health Coverage for Inclusive and Sustainable Growth; Lessons from 11 Country Case Stu- dies: À Global Synthesis. Krishna, A. 2007. “Poverty and Health: Defeating Poverty by Going to the Roots: Deve- lopment: Poverty 50: 63-69. Kydd, J, Dorward, A, Morrison, J., Cadisch, G.,, 2002. “Agricultural development and pro-poor economic growth in sub-Saharan Africa: potential and policy”. ADU Wor- king paper 02/04. Imperial College, Wye Lagomarsino, G. À. Garabrant, A. Adyas, R. Muga, and N. Otoo. 2012. “Moving towards Universal Health Coverage: Health Insurance Reforms in Nine Developing Coun- tries in Africa and Asia! Lancet 380: 93343. Lamaute-Brisson. 2013. “Social Protection Systems in Latin America and the Carib- bean: Haïti”. ECLAC, United Nations. Lewis, À. 1954. “Economic Development with Unlimited Supplies of Labour”. Manches- ter School of Economic and Social Studies 22:139-91 Library of Congress. 2006. “Country Profile: Haïti” May, Federal Research Division, Library of Congress, Washington, DC. http://lcweb2locgov/frd/cs/profiles/Haïti.pdf. Lipton, M. 2009. Land Reform in Developing Countries: Property Rights and property Wrongs. New York: Routledge. Lombardo, A. 2012. “Mapping of Social Protection Programmes in Haiti”. Final Report, Oxfam, UK and UNICEF. Lépez-Calva, LF, D. Battiston, G. Cruces, A.M. Lugo, M. Santos. 2013. “Income and Be- yond: Multidimensional Poverty in Six Latin American Countries,” Social Indicators Research, Springer, vol. 112(2), pages 291-314, June. Lundahl, Mats. 2011. Poverty in Haïti: Essays on Underdevelopment and Post Disaster Prospects. New York: Palgrave Macmillan. [page 236] WorldBank - ONPES | Lunde, Henriette. 2008. Youth and Education in Haïti: Disincentives, Vulnerabilities, and Constraints. Oslo: Fafo. Marshall, A. 1890. Principle of Economics. London: Macmillan. Marzo, Federica, and Mori Hideki. 2012. “Crisis Response in Social Protection” Social Protection & Labor Discussion Paper 1205, World Bank, Washington, DC. Masterson, T.. 2007. “Productivity, Technical Efficiency, and Farm Size in Paraguayan Agriculture”. Levy Economics Institute Working Paper No. 490. Mathieu, Philippe, Jean Arsène Constant, Josué Noël, and Bobby Piard. 200%. “Cartes et étude de risques, de La vulnérabilité et des capacités de réponse en Haïti” Oxfam-Haiti, Port-au-Prince. Mayer, D. et al 2000. Health, Growth, and Income Distribution in Latin American and the Caribbean: A Study of Determinants and Regional and Local Behaviour. Washington, DC: Pan American Health Organization. MENFP (Ministry of Education and Vocational Training, Haiti). 2013. Programme d’In- terventions Prioritaires en Education (PIPE): 2013-2016. Port-au-Prince, Haiti: MENFP. Montas. 2005. "La pauvreté en Haïti: situation, causes et politiques de sortie”. ECLAC, United Nations. Montenegro, Claudio E, and Harry Anthony Patrinos. 2072. “Returns to Schooling around the World’ Background paper, World Bank, Washington, DC. Moreno-Serra, R, S. Thomson, and K. Xu (2013). Measuring and comparing financial protection p.223-255. In Health system performance comparison. Edited by Irene Pa- panicolas and Peter C.Smith. MPCE (Ministry of Planning and External Cooperation). 2008. Document de Stratégie Nationale pour la Croissance et la Réduction de la Pauvreté, DSNCRP (2008- 2010): Pour Réussir Le Saut Qualitatif. Port-au-Prince, Haiti: MPCE. http://www. sdn.mefhaiti.gouv.ht/Autres/DSNCRP/DSN _TM.php. _______. 20. Rapport Final Sur La Mise En Œuvre Du Premier Document De Stratégie Nationale Pour La Croissance Et La Réduction De La Pauvreté (DSN- CRP-2008-2010) - Pour Réussir Le Saut Qualitatif. Port-au-Prince, Haiti: MPCE. http://logementquartierhaiti.files.wordpress.com/2012/03/201104-dsrn- cp-haiti-rapport-final.pdf MSPP (Ministry of Public Health and Population, Haiti). 2011. “Rapport de la Carte Sanitaire du Pays.” September, MSPP, Port-au-Prince, Haiti. ________ 2013. “Rapport des Comptes Nationaux de Santé 2010-11 June, MSPP, Port-au-Prince, Haïti [page 237] : Investing in People to Fight Poverty in Haïti Murray, C. K. XU, J. Klavus, K. Kawabata, P. Hanvoravongchai, R. Zeramdini, et al. 2003. “Assessing the Distribution of Household Financial Contributions to the Health System: Concepts and Empirical Application.” In Health Systems Performance As- sessment: Debates, Methods and Empiricism, edited by C. J. L. Murray and D. B. Evans. Geneva: World Health Organization. Nopo, Hugo. 2012. New Century, Old Disparities: Gender and Ethnic Earnings Gaps in Latin America and the Caribbean. Latin American Development Forum Series. Washington, DC: Inter-American Development Bank and World Bank. Nordman, Christophe J, and Laure Pasquier-Doumer. 2013. “Transitions in a West Afri- can Labor Market: The Role of Social Networks” DIAL Working Paper DT/2013/12, Développement, Institutions et Mondialisation, Paris. Nordman, Christophe J, Anne-Sophie Robilliard, and François Roubaud. 2013. “Decom- posing Gender and Ethnic Earnings Gaps in Seven Cities in West Africa” In Urban Labor Markets in Sub-Saharan Africa, edited by Philippe De Vreyer and François Roubaud, 271-98. Paris: Agence Française de Développement; Washington, DC: World Bank. OECD & WHO. 2003. Poverty and Health. DAC Guidelines and References Series.OECD Publishing . Olson, S. M, R. À. Green, S. Lasley, N. Martin, B. R. Cox, E. Rathje, J. Bachhuber, and J. French. 2011. “Documenting Liquefaction and Lateral Spreading Triggered by the 12 January 2010 Haiti Earthquake” Earthquake Spectra 27 (S1): S93-S116 (October). ONPES (Observatoire National Pour la Pauvreté et l'Exclusion Sociale). 2012. « Catas- trophes naturelles et accélération de la pauvreté : Le cas des cyclones Sandy et Isaac. Extract from ONPES 2012 Annual Report ». Port-au-Prince : ONPES. _____ 2014. Le travailleur pauvre en Haiti ». Port-au-Prince: ONPES. ______.Forthcoming. « Etude de pauvreté en Haiti.» ONPES, Ministry of Planning and External Cooperation, Port-au-Prince, Haïti. Orozco. 2006. “Understanding the Remittance Economy in Haiti” Paper commissio- ned by the World Bank. Institute for the Study of International Migration, George- town University. Ôzden, Ç., and M. Schiff. 2006. International Migrations, Remittances and the Brain Drain. Washington, DC: World Bank. Pedersen and Lakewood. 2001. “Determination of a poverty line for Haiti”. Fafo Institu- te of Applied International Studies Pianta, RC, and S. J. McCoy. 1997. “The first day of school: The predictive validity of early school screening”. Journal of Applied Developmental Psychology, 18, 1-22 [page 238] WorldBank - ONPES | Pierre, Y.F, G. R. Tucker, and J-FTardieu. 2009. XLost Childhoods in Haiti: Quantifying Child Trafficking, Restaveks, and Victims of Violence. Port-au-Prince: Pan American Development Foundation. PotoFi Haiti Girls Initiative (‘PotoFi”). 2012. “Gender Aftershocks: Teen Pregnancy And Sexual Violence In Haitian Girls”. Available at : http://potofifiles. wordpress. com/2012/12/summary-report-poto-fi-girls-gbv-field-survey.pdf) Psacharopoulos, George, and Harry Anthony Patrinos. 2010. “Returns to Investment in Education: A Further Update!” Education Economics 12 (2): 11-34. Raeza-Sanchez, J, A. Fuchs and M. Matera. 2014. « Background paper on Shocks and Vulnerability in Haiti ». Washington, DC: World Bank. Rajan, P. 1999. ‘An Economic Analysis of Child Labor” Economics Letters 64 (1): 99— 105. Rathje, E. M, J. Bachhuber, R. Dulberg, B.R. Cox, A. Kottke, C. Wood, R. A. Green, S. M. Olson, D. Wells, and R. Glenn. 2011. “Damage Patterns in Port-au-Prince during the 2010 Haiti Earthquake” Earthquake Spectra 27 (S1): S117—-S136 (October). Reynolds, AJ, J. A. Temple, D. L. Robertson, and E. A. Mann. “Long-term Effects of an Early Childhood intervention on Educational Achievement and Juvenile ArrestA 15-Year Follow-up of Low-Income Children in Public Schools”. Journal of the American Medical Association. 285(18):2339-2346. Ribe, Helena, David A. Robalino, and lan Walker. 2010. Achieving Social Protection for ALL in Latin America and the Caribbean: From Right to Reality. Directions in Development: Human Development 55547. Washington, DC: World Bank. RTI International. 2010. “Haïti: Early Grade Reading Assessment; Rapport pour Le Mi- nistere de l'Education et la Banque Mondiale” April, Research Triangle Institute, Research Triangle Park, NC. Sachs, J. D. 2001. “Macroeconomics and Health: Investing in Health for Economic Development”. Report of the Comission on Macroeconomics and Health, World health Organization. Scheiman, |, J. Langenbrunner, J. Kehler, C. Cashin, and J. Kutzin. 2010. “Sources of Funds and Revenue Collection: Reforms and Challenges” In Implementing Health Financing Reform: Lessons from Countries in Transition, edited by J. Kut- zin, C. Cashin, and M. Jakab, 87-118. Geneva: World Health Organization. Shefer, D. 1973. “Localization Economics in SMAS: À Production Function Analysis. Journal of Urban Economics 13 (1): 55-64. Skoufias, E, and S. Parker. 20072. “A Cost-Effectiveness Analysis of Demand and Su- pply-Side Education Interventions” FNCD Discussion Paper 227, International Food Policy Research Institute, Washington, DC. [page 239] : Investing in People to Fight Poverty in Haïti Sletten, Päl, and Willy Egset. 2004. “Poverty in Haïti” Fafo Paper 2004-31, Fafo Institu- te for Applied International Studies, Oslo. Smith, A. 1776. An Inquiry Into the Nature and Causes of the Wealth of Nations. London. Strokova, V, L. Basset, C. Clert and À. Ocampo. 2013. Background paper on Social Pro- tection in Haiti. Washington, DC: World Bank. Sveikauskas, L.1975. “The Productivity of Cities.” Quarterly Journal of Economics 89 (3): 393-113. Unal, F. G. 2008. “Small Is Beautiful: Evidence of an inverse Relationship between Farm Size and Yield in Turkey”. Working Paper No. 551, The Levy Economics Institute. UNDP (United Nations Development Programme). 2003 _______ 2004. Reducing Disaster Risk: A Challenge for Development. New York: Bureau for Crisis Prevention and Recovery, UNDP. UNECLAC (United Nations Economic Commission for Latin America and the Carib- bean). 2005. “Emploi et Pauvrete en Milieu Urbain en Haiti” United Nations, San- tiago, Chile. ______ 2009. MDG Health Chapter. Out-of-pocket health expenditure: evidence of pronounced inequalities. Chapter 5, p.77-99. ______ 2013. Social Panorama of Latin America 2012. Santiago, Chile: United Na- tions. United Nations. 2009. 2009 Global Assessment Report on Disaster Risk Reduction; Risk and Poverty in a Changing Climate: Invest Today for a Safer Tomorrow. Gene- va: United Nations. USAID (United States Agency for International Development). 2012. ‘The Early Grade Reading Assessment in Haïti” USAID, Washington, DC. Vadivelu, G. À, S.P. Wani, L. M. Bhole, P. Pathak, and A. B. Pande. 2001. ‘An empirical analysis of the relationship between land size, ownership, and soybean productivity — new evidence from the semi-arid tropical region in Madhya Pradesh, India”. Natural Resource Mana- gement Program Report no. 4. Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics. 50 pp. Verner, Dorte. 2005. “ Making the Poor Haitians Count Takes More Than Counting the Poor: A [page 240] WorldBank - ONPES | Poverty and Labor Market Assessment of Rural and Urban Haiti Based on the First Household Survey for Haiti” World Bank, April. ________ 2008. “Labor Markets in Urban and Rural Haiti: Based on the First Hou- sehold Survey for Haiti” Policy Research Working Paper 4574, World Bank, Was- hington, DC. Walz, Julie, and Vijaya Ramachandran. 2072. “Haiti: Three Years after the Quake and Not Much Has Changed’” Center for Global Development. http://www.cedev. org/blog/haiti-three-years-after-quake-and-not-much-has-changed. WHO (World Health Organization). 2001. “Macroeconomics and Health: Investing in Health for Economic Development Report of the Commission on Macroecono- mics and Health, WHO, Geneva ______ 2004. “The Impact of Health Expenditure on Households and Options for Alternative Financing’ Report EM/RC51/4, Regional Office for the Eastern Me- diterranean, WHO, Cairo. http: /www.who.int/health_ financing/documents/ cov-emrc-healthexpenditureimpact ______ 2010. Monitoring the Building Blocks of Health Systems: À Handbook of Indicators and Their Measurement Strategies. Geneva: WHO. http://www.who. int/healthinfo/systems/WHO _MBHSS_2010 _full_web.pdf. _______ “Haiti” NHA Report 2005-2006. http:/www.who.int/nha/country/hti/ en ______ 2013a. World Health Statistics 2013. Geneva: WHO. _______ 2013b. The World Health Report 2013: Research for Universal Health Co- verage. Geneva: WHO. ______ _ _ 2014a. “Haiti: National Health Expenditures” March. WHO, Geneva. _______ 2014b. “Trends in Maternal Mortality, 1990 to 2013: Estimates by WHO, UNICEF, UNFPA, the World Bank, and the United Nations Population Division” WHO, Geneva. _______ 2014c.“"Cholera, Fact Sheet 107 (February), Media Center, World Health Organization, Geneva. http://www.who.int/mediacentre/factsheets/fs107/en/. WHO (World Health Organization) and World Bank. 2013. “Monitoring Progress towards Universal Health Coverage at Country and Global Levels: A Framework Joint WHO—-World Bank Group Discussion Paper 1 (December), WHO, Geneva. Wisner, B, P. Blaikie, T. Cannon, and I. Davis. 2004. At Risk: Natural Hazards, People's Vulnerability and Disasters, 2nd ed. London: Rutledge. Wisner, B, K. Westgate, and P. O'Keaffe. 1976. Poverty and Disaster. London: New So- ciety. [page 241] : Investing in People to Fight Poverty in Haïti World Bank. 1998. “Haiti: The Challenges of Poverty Reduction’ Report 17242-HA (Au- gust), World Bank, Washington, DC. ______ 2006. “Haiti: Options and Opportunities for Inclusive Growth” Country Economic Memorandum, World Bank, Washington, DC. ______ 2007. World Development Report 2008: Agriculture for Development. Washington, DC: World Bank. ______ 20. World Development Report 201: Conflict, Security, and Develop- ment. Washington, DC: World Bank. _______ 2072. Health Equity and Financial Protection Datasheets. Latin America and Sub-Saharan Africa. www.worldbank.org/povertyandhealth. _______ 2012a. Improving Skills Development in the Informal Sector: Strategies for Sub-Saharan Africa. Washington, DC: World Bank. ______ 2012b. World Development Report 2013: Jobs. Washington, DC: World Bank. ______ 2013a. World Development Report 2014; Risk and Opportunity: Managing Risk for Development. Washington, DC: World Bank. 22 2013b. “Improving Maternal and Child Health through the Integrated So- cial Services Project.” Project appraisal document (April), World Bank, Washington, DC. _______ 2013c. Agricultural Risk Management in the Caribbean. Washington, DC: World Bank. _______ 204. "Social Gains in the Balance: A Fiscal Policy Challenge for Latin Ame- rica and the Caribbean Report 85162 rev (February), World Bank, Washington, DC, http://hdlhandlenet/10986/17198. World Bank and SEDLAC dataset. 2005/06. XU, K, D. B. Evans, K. Kawabata, R. Zeramdini, J. Klavus, and C. J. L. Murray. 2003. “Hou- sehold Catastrophic Health Expenditure: A Multicountry Analysis” Lancet 362: 111. Zapata, R. 2005. The 2004 Hurricanes in the Caribbean and the Tsunami in the Indian Ocean. Mexico City: United Nations Economic Commission for Latin America and the Caribbean. Zeng, Wu, Marion Cros, Katherine D. Wright, and Donald S. Shepard. 2012. “Impact of Performance-Based Financing on Primary Health Care Services in Haiti” Health Policy and Planning 28 (6): 596-605. [page 242] [page 243] [page 244] . À MPLES |